Author Archives: author

Accelerating Critical Thinking to Industrial Pace and Scale Through AI: Addressing the Global Issue of Food Sustainability

DOI: 10.31038/NRFSJ.2024714

Abstract

We present a new, systematized way to teach critical thinking, using AI (artificial intelligence) incorporated into a research tool created for a newly emerging science, Mind Genomics, that is concerned with how people respond to ideas concerning everyday experiences. Mind Genomics methodology requires the researcher to develop four questions which ‘tell a story,’ and for each question to provide four alternative answers. Previous studies showed that many users experienced difficulty creating the questions. To overcome this problem, Mind Genomics incorporates AI through the mechanism of the Idea Coach. This mechanism allows the researcher to describe the problem being addressed, and then generates 15 questions the researcher evaluates and chooses for returns with 15 questions during the course of setting up the study’s story. Idea Coach provides additional analyses on the questions returned to reveal deeper structure and stimulate critical thinking by the researcher. We demonstrate the capabilities of the process by comparing the results for ‘food sustainability’ for people who are defined to be poverty stricken, first in the United States, and then in Ghana, and finally in Egypt. The effort requires approximately 10 minutes in total and is scalable for purposes of education and practical use.

Introduction: The Importance of Critical Thinking to Solve Problems

In order to address issues facing humanity, such as sustainability, it is important to be able to think clearly about the nature of the problem, and from there proceed to solutions. The importance of critical thinking cannot be underestimated, most apparently in education [1,2], but also in other areas, such as dentistry [3], not to mentioned the very obvious importance of critical thinking in areas where there are opposing parties confronting each other with the weapons of knowledge and thinking, such as the law [4]. The very idea of dealing with the United Nations’ (UN) 24 defined Global Issues (United Nations, undated) calls into play the need to understand and then deal with the problem. Critical thinking, or its absence has been recognized as a key feature in the solution of these problems. From the UN’s perspective, their 24 issues need to be addressed continually over time, strongly suggesting that the need for critical thinking is not limited in time but needs to be engaged with through time.

In today’s world, critical thinking is recognized as important for society [5]. The key question is not the recognition of critical thinking, but rather how to encourage it in a way which itself is sustainable, in a way which is cost-effective, scalable, and productive in terms of what it generates. To the degree that one can accelerate critical thinking, and even more so to focus critical thinking on a problem, one will most likely be successful . Finally, if such critical thinking can be aided by technical aids, viz., TACT (Technical Aids to Creative Thought), there is a greater chance of success. The notion of the aforementioned approach TACT was first introduced to the senior author HRM by the late professor Anthony Oettinger of Harvard University in 1965, almost 60 years ago. This paper shows how today’s AI can become a significant contributor to TACT, and especially to critical thinking about UN based problems, this one being food sustainability [6].

The topic of food sustainability is just one of many different topics of the United Nations, but one seeing insufficient progress (UN undated). From the point of view of behavioral science, how does one communicate issues regarding food sustainability? And how does one move beyond the general topic to specific topics? It may well be that with years of experience in a topic the questions become easier, but what about the issue of individuals wanting to explore the topic but individuals without deep professional experience? Is it possible to create a system using AI which can teach in a manner best called Socratic, i.e., a system which teaches by laying out different questions that a person could ask about a topic?

The Contribution of Mind Genomics to Critical Thinking about a Problem

During the past 30 years, researchers have begun to explore the way people think about the world of the everyday. The approach has been embodied in an emerging science called Mind Genomics (REF). The foundation of Mind Genomics is the belief that we are best able to understand how people think about a topic by presenting them with combinations of ideas, and instructing these people to rate the combination of ideas on a particular rating scale, such scales as relevance to them, interest to them, perceived solvability, etc. The use of combinations of ideas is what is new, these combinations created in systematic manner by an underlying structure called an experimental design. The respondent who participates does not have to consciously think about what is important, but rather do something that is done every day, namely choose or better ‘rate’ the combinations on a scale. The analysis of the relation between what is presented and what is rated, usually through statistics (e.g., regression) ends up showing what is important.

The process has been used extensively to uncover the way people think about social problems [7], legal issues [8], etc.. The process is simple, quick and easy to do, prevents guessing, and ends up coming up with answers to problems.

The important thing here is that the researcher has to ask questions, provide answers, and then the computer program matches the answers together into small groups, vignettes, presents these to the respondent, who has to rate he group or the combination.

Of interest here is the front end of the process, namely, how to ask the right question. It is asking questions which has proved to be the stumbling block for Mind Genomics, since its founding in 1993 (REF). Again, and again researchers have request help to formulate the studies. It is no exaggeration to state that the creation of questions which tell a story has become one of the stumbling blocks to the adoption of Mind Genomics.

Early efforts to ameliorate the problem involved work sessions, where a group of experts would discuss the problem. Although one might surmise that a group of experts in a room certainly could come up with questions, the opposite was true. What emerged was irritation, frustration, and the observation that the experts attending either could not agree on a question, or in fact could even suggest one. More than a handful of opportunities to do a Mind Genomics project simply evaporated at this point, with a great deal of disappointment and anger covering what might have been professional embarrassment. All would not be lot, however, as many of the researchers who had had experienced continued to soldier on, finding the process relatively straightforward. Those who continued refused to let the perfect get in the way of the good. This experience parallels what has been previously reported, namely that people can ask good questions, but they need a ‘boost’ early on [9].

The Contribution of AI in 2023

The announcement of AI by Open AI in the early months of 2023 proved to provide the technology which would cut the Gordian knot of frustration. Rather than having people have to ‘think’ through the answer to the problem with all of the issues which would ensue, it appeared to be quite easy to write a query about a topic and have the Mind Genomics process come up with questions to address that query. It was, indeed, far more enjoyable to change the ingoing query, and watch the questions come pouring out. It would be this process, a ‘box for queries’ followed by a standardized report, which would make the development fun to do.

Figure 1 shows what confronted the researcher before the advent of AI, namely an introduction page which required the researcher to name the study, followed immediately by a dauntingly empty page, requesting the research to provide our questions which tell a story. The researcher has the option to invoke AI for help by pressing the Idea Coach button.

FIG 1

Figure 1: Panel A shows the first screen, requiring the respondent to name the study. Panel B shows the second screen, presenting the four questions to be provided by the respondent.

Results Emerging Immediately and After AI Summarization

The next set of tables shows the questions submitted through the query to Idea Coach, the immediate set of 15 questions returned within 5-15 seconds. Later on, we will see the results after AI summarization has been invoked on the different set of questions.

In the typical use of Mind Genomics, the researcher often ends up submitting the squib to Idea Coach from a minimum of one time, but more typically 3-5 times, occasionally modifying the squib, but often simply piling up different questions. These different questions, 15 per page, provide a valuable resource of understanding the topic through the question. Typically, about 2/3 of the questions are different from those obtained just before, but over repeated efforts many of the questions will repeat.

Table 1 shows the first set of 15 questions for each of three countries, as submitted to Idea Coach. Note that the squib presented to Idea Coach is only slightly different for each country, that difference being only the name of the country. The result, however, ends up being 15 quite different questions for each country, questions which appear to be appropriate for the country. It is important to emphasize here that the ‘task’ of AI is to ask questions, not to provide factual information. Thus, the issue of factual information is not relevant here The goal is to drive thinking.

Table 1: Query & Questions for United States, Ghana and Egypt. These 15 questions emerged 10-15 seconds after the query was submitted to Idea Coach.

TAB 1

It is important to note that Table 1 can be replicated as many time as the researcher wishes. The questions end up allowing the researcher to look at different aspects of the problem. The results come out immediately to the researcher, as well as being stored in a file for subsequent AI ‘summarization’ described below. At the practical level, one can imagine a student interested in a topic looking at the questions for a topic again and again, as the student changes some of the text of the query (viz, the squib shown in Figure 2, Panel B). It is worth emphasizing that the Idea Coach works in real time, so that each set of 15 questions can be re-run and presented in the span of 5-15 seconds when the AI system is ‘up and running.’ Thus, the reality ends up being a self-educating system, at least one which provides the ‘picture of the topic’ through a set of related questions, 15 questions at a time. The actual benefit of this self-pacing learning by reading questioning is yet to be quantified in empirical measures, however (Figure 3).

FIG 2

Figure 2: Panel A shows the information about Idea Coach. Panel B shows the ‘box’ where the researcher creates the query for Idea Coach, in terms of ‘shaping’ the structure and information of the question.

FIG 3

Figure 3: The first six questions out of the 15 returned by Idea Coach to answer the request shown in Figure 2, Panel B. The remaining nine questions are accessed by scrolling through the screen.

It is relevant to note that AI-generated questions are beginning to be recognized as an aid to critical thinking, so that the Idea Coach strategy can be considered as part of the forefront of what might be the 21st century TACT program, Technical Aid to Creative thought (Oettinger, 1965, personal communication). Papers such as the new thesis by Danry [10] of MIT reflect this new thinking. Half-way around the world the same approaches are being pioneering in the Muslim world [11].

Once the questions are presented, it is left to the researcher to move on to completing the set-up of the Mind Genomics study, or to further request additional sets of 15 questions. When the creation of questions is complete, the researcher is instructed to provide four answer for each question. A separate paper will deal with the nature of ‘answers’ to the questions. This paper deals only with the additional analysis of the questions generated by Idea Coach.

AI Summarization and Extensions of Sets of 15 Questions

The second part of Idea Coach occurs after the researcher has competed the selection of the four questions, as well as completing the generation or selection of the four answers for each question. This paper does not deal with the creation of answers, but the process is quite similar to the creation of questions. The researcher creates the set of four questions, perhaps even editing/polishing the questions to ensure proper understanding, and tone. Once the questions are published, the Idea Coach generates four answer to each question. The entire process of summarization, for all of the set of 15 questions, takes about 15-30 minutes. The Excel file containing the ‘Answer Book’ with all summarizations is generally available 20 minutes after the questions and answers have been selected. The Answer book is available for download at the website (www.BimiLeap.com) and is emailed to the researcher as well.

We now go into each part of the summarization. The actual summarization for each set of 15 questions is presented on one tab of the Idea Book. We have broken up the summarizations into each major section, and then present the summarization by AI for the USA, followed by for Ghana, and finally for Egypt In this way the reader can see how the initial squib, the prompt to Idea Coach, differing only in the country, ends up with radically different ideas.

Key Ideas

The output from the first prompt had produced full questions. The ‘Key Ideas’ prompt strips the question format away, to show the idea or issue underlying the question. In this way, the ‘Key Ideas prompt can be considered simply as a change in format, with no new ideas emerging. Table 2 shows these ideas. It is not clear which is better to use. To the authors, it seems to be more engaging to present the ideas in the form of a question. When presenting the same material as ideas seems to be more sterile, less engaging, and without grounding.

Table 2: Key Ideas underlying the 15 questions

TAB 2

The use both of questions and of the ideas on which these questions are based have been addressed as part of an overall study of the best ways to learn. In the authors’ own words ‘Likewise, being constructive is better than being active because being constructive means that a learner is creating new inferences and new connections that go beyond the information that is presented, whereas being active means only that old knowledge is retrieved and activated.’ [12]

Before moving on to the next section, one may rightfully ask whether a student really learns by being given questions which emerge from a topic, or whether it is simply better to let the student flounder around, come up with questions, and hopefully discover other questions, either by accident, or by listening to the other students answer the same question and gleaning from those other answers new points of view [13]. The point of view taken here is that these aids to creative thought do not provide answers to questions, but rather open up the vistas, so that the questioner, research or student, can think is new but related directions. The output are additional, newly focused questions, rather than answers which put the question to rest. Quite the opposite [14]. The question opens up to reveal many more dimensions perhaps unknown to the researcher of the student when the project was first begun. In other words, perhaps the newly surfaced questions provide more of an education than one might have imagined.

Themes

With themes Idea Coach moves toward deconstructing the ideas, to identify underlying commonalities of issues, and the specific language in the questions supporting those commonalities. With ‘Themes’ the AI begins the effort to teach in a holistic manner, moving away simply from questions to themes which weave through the questions. For the current version of Idea Coach, the effort to uncover themes is done separately for each set of 15 questions, in order to make the task manageable. In that way the researcher or the student can quickly compare the themes generated from questions invoking the United States versus questions invoking Ghana, or questions invoking Egypt. Table 3 gives a sense of how the pattern of themes differ [15]. It is also important that the organization shown in Table 3, is the one provided by the Idea Coach AI, and not suggested by the researcher. Note that for Egypt, as contrasted with the USA and Ghana, Idea Coach refrained from grouping ideas into themes, but treated each idea as its own theme.

Table 3: Themes emerging from the collection of 15 questions for each country

TAB 3

Perspectives, an Elaboration of Themes

Perspectives advances the section of themes, which had appeared in Table 4. Perspectives takes the themes, and puts judgment around these themes, in terms of positive aspects, negative aspects, and interesting aspects. Perspectives are thus elaborations of themes. In other words, perspectives ends up being an elaboration of themes, useful as a way to cement the themes into one’s understanding.

Table 4: Perspectives (an elaboration of Themes)

TAB 4(1)

TAB 4(2)

TAB 4(3)

What is Missing

As the analysis moves away from the clarification of the topic, it moves towards more creative thought. The first step is to find out what is missing, or as stated by Idea Coach, ‘Some missing aspects that can complete the understanding of the topic include: ’ It is at this point that AI moves from simple providing ideas to combining ideas, and suggesting ideas which may be missing.

It is at this stage, and as the stage of ‘innovation’ that AI reaches a new level. Rather than summarizing what has been asked, AI now searches for possible ‘holes’ and a path towards greater completeness in thinking. Perhaps it is at this level of suggesting missing ideas that the user begins to move into a more creative mode, although with AI suggesting what is missing one cannot be clear whether it is the person who is also thinking in these new directions, or whether the person is simply moving with the AI, taking in the information, and enhancing their thinking (Table 5).

Table 5: What is missing

TAB 5(1)

TAB 5(2)

Alternative Viewpoints

Alternative viewpoints involve arguing for the opposite of the question. We are not accustomed to thinking about counterarguments in the world of the everyday. Of course, we recognize counterarguments such as what occurs when people disagree. Usually, however, the disagreement is about something that people think to be very important, such as the origin of climate change or the nature of what climate change is likely to do. In such cases we routinely accept alternative viewpoints.

The Idea Coach takes alternative viewpoints and counterarguments to a deeper stage, doing so for the various issues which emerge from the questions. The embedded AI takes an issue apart and looks for the counterargument. The counterargument is not put forward as fact, but simply as a possible point of view that can be subject to empirical investigation for proof or disproof (Tables 6-9).

Table 6: Alternative viewpoint, showing negative arguments countering each point uncovered previously by Idea Coach using AI.

TAB 6(1)

TAB 6(2)

Table 7: Interested audiences.
The next AI analysis deals with the interested audiences for each topic. Rather than just listing the audience for each topic, the Idea Coach goes into the reasons why the audience would be interested, once again providing a deeper analysis into the topic, along with a sense of the stakeholders, their positions, their areas of agreement and disagreement.

TAB 7(1)

TAB 7(2)

Table 8: Opposing audiences.
Once again, in the effort to promote critical thinking, the Idea Coach provides a list of groups who would oppose the topic, and for each group explain the rationale for their opposition.

TAB 8

Table 9: Innovations.
The final table selected for the Idea Coach summarization is innovations, shown in Table 9. The table suggests new ideas emerging from the consideration of the questions and the previous summarizations. Once again the ideas are maintained with the constraints of the topic and reflect a disciplined approach to new ideas.

TAB 9(1)

TAB 9(2)

TAB 9(3)

Discussion and Conclusions

The goal of the paper has been to show what is currently available to students and researchers alike. The objective of the demonstration has been to take a simple problem, one that might be part of everyday discourse, and use that problem to create a ‘book of knowledge’ from the topic, using questions and AI elaboration of the questions.

We hear again and again about the importance of critical thinking, but we are not given specific tools to enhance critical thinking. As noted in the introduction, in the 1960’s, the late professor Anthony Gervin Oettinger of Harvard University began his work on creative thought. We might not think that programming a computer to go shopping is an example of creative thought, but in the 1960’s it was (Oettinger, xxx). Now, just about six decades later, we have the opportunity to employ a computer and AI create books that help us thinking critically about a problem. We are not talking here about giving factual answers, actual ‘stuff,’ but really coaching us how to think and how to think comprehensively about the ideas within a societal milieu, a milieu of competing ideas, of proponents and opposers who may eventually agree on solutions that address or resolve issues as thorny as food sustainability. If sixty years ago teaching a computer (the EDSAC) to go shopping was considered a TACT, a technical aid to creative thought, perhaps now co-creating a book of pointed inquiry about a topic might be considered a contribution of the same type, albeit one more attuned to today. The irony is that sixty years ago the focus was on a human programming a machine ‘to think,’ whereas today it is the case of a machine coaching a human how to think. And, of course, in keeping with the aim of this new to the world journal, the coach is relevant to thinking about any of the topics germane to the journal. This same paper could be created in an hour for any topic.

References

  1. Cojocariu VM, Butnaru CE (2014) Asking questions–Critical thinking tools. Procedia-Social and Behavioral Sciences 128: 22-28.
  2. Lai ER 2011. Critical thinking: A literature review. Pearson’s Research Reports 6: 40-41.
  3. Miller SA& Forrest JL (2001) Enhancing your practice through evidence-based decision making: PICO, learning how to ask good questions. Journal of Evidence Based Dental Practice 1: 136-141.
  4. Nicholar J, Hughes C & Cappa C (2010) Conceptualising, developing and accessing critical thinking in law. Teaching in Higher Education 15: 285-297.
  5. ŽivkoviĿ S 2016. A model of critical thinking as an important attribute for success in the 21st century. Procedia-Social and Behavioral Sciences 232: 102-108.
  6. Bossert WH, Oettinger AG 1973. The Integration of Course Content, Technology and Institutional Setting. A Three-Year Report, 31 May 1973. Project TACT, Technological Aids to Creative Thought.
  7. Moskowitz H, Kover A & Papajorgji P (eds), (2022) Applying Mind Genomics to Social Sciences. IGI Global.
  8. United Nations, Undated. “Global Issues,” accessed January 27, 2024.
  9. Rothe A, Lake BM & Gureckis TM 2018. Do people ask good questions? Computational Brain & Behavior 1: 69-89.
  10. Danry VM (2023) AI Enhanced Reasoning: Augmenting Human Critical Thinking with AI Systems (Doctoral dissertation, Massachusetts Institute of Technology).
  11. Fariqh N 2023, October. Developing Literacy and Critical Thinking with AI: What Students Say. In .Proceedings Annual International Conference on Islamic Education (AICIED) 1: 16-25.
  12. Chi MTH 2009. Active-Constructive-Interactive: A Conceptual Framework for Differentiating Learning Activities. Topics in Cognitive Science 1: 73-105. [crossref]
  13. Moskowitz HR, Wren J & Papajorgji P 2020. Mind Genomics and the Law. LAP LAMBERT Academic Publishing.
  14. Niklova, N (2021) The art of asking questions: Flipping perspective. In: EDULEARN21 Proceedings Publication 2816-2825.
  15. Oettinger AG. Machine translation at Harvard 2003. In: Early Years in Machine Translation, Memoirs and Biographies of Pioneers, (ed. W.J. Hutchins), John Benjamin’s Publishing Company, Amsterdam/Philadelphia, pp. 73-86.

Understanding the Mind and Inventing the Future: The Problem of Failure to Show Up for Follow-Up Appointments with One’s Health Provider

DOI: 10.31038/ASMHS.2024813

Abstract

The paper introduces a system to deal with problems of society using SCAS, Socrates as a Service. SCAS is provided with a detailed description of a conventional problem faced by people, and in turn instructed to defined prospective mind-sets in the population who suffer with this problem. SCAS further provides information on the nature of these hypothesized mind-sets, what the mind-sets are thinking, and how the mind-sets would respond to topic-relevant slogans that would be generated to solve the problem. Finally, the paper finishes with the use of SCAS to summarize the issue, provide perspectives that people might have, and identify what next steps need to be taken, as well as innovations that should be introduced which deal with and even solve the problem. SCAS is a general approach. The paper here uses SCAS to investigate the ‘why’ patients fail to keep their doctor’s visits, and what innovations might solve the problem.

Introduction

This paper grew out of the recognition that all too often patients fail to follow the suggestions of their medical and health professionals. The topic of compliance is a large one. The focus of this paper is on the simple problem of patients not showing up at the prescribed time for their follow-up appointments. The damage which ensues can be enormous, impacting the health of the patient, the cost to the medical practice, and the disruption of a system which must accommodate the schedules of a variety of people who then must regroup and update the schedules [1].

When dealing with this problem, we are actually dealing with issues of communication interacting with motivation and habit. How does the medical establishment work with individuals to ensure that they come to scheduled appointments. The importance of this question can be easily understood when one realizes the number of reminder messages which appear on the smartphones of patients, telling them of the upcoming appointment, asking them to ‘e-check in’ and then giving them the chance to cancel and reschedule. This and other actions such as reminder phone call are the obvious effort to minimize the expensive ‘no-shows’. In recent years, the process has been automated, with AI-driven chatbots and voice interactions finding their place in the seemingly impossible to solve conundrum of getting patients to sow for their appointments [2].

The business literature recognizes the problems of ‘no-shows’. The issues underlying the no-shows are extensive, as are the suggestions for improvement. The case of medicine is particular serious for no-shows simply because one cannot necessarily move the appointment to some later time and ‘go from there.’ A person’s health is labile. Moving a scheduled appointment a month or two later, when a slot opens up, may be too late when the issue is the follow up from what can be a serious problem, and when not treated can evolve to a life-threatening one. One serious illness often comes to the fore, diabetes. The consequence of missing a follow up appoint with a doctor when the person has diabetes 2 can be severe [3-6].

The Contribution of Mind Genomics Enhanced by SCAS (Socrates as a Service)

The problem of no-shows was first brought into the world of Mind Genomics through collaboration with physicians in Chicago, IL, specifically anesthesiologist Dr. Glen Zemel. Author Moskowitz collaborated with Dr. Zemel on a variety of studies dealing with the mind of the patient in the hospital. As a practicing anesthesiologist, Zemel often recognized the issues involved in patients who fail to follow up, often even having to forego surgery on the particular scheduled date because either they ‘forgot’ (rare) or forgot to follow the requirements of avoid food for the previous 12 hours and so forth. It was these immediate issues which ended up costing the medical practice many thousands of dollars.

The problem became more acute when authors Braun and Mulvey, and later Cooper, became involved in the issue of patients who failed to follow up at specific times. These individuals suffered from a variety of metabolic disorders; the most common one being diagnosed as pre-diabetic. The failure to return at the scheduled time for a follow-up morphed from being a financial loss to a medical practice into the possibility that diabetes might develop because the pre-diabetic essentially disappeared, but presumably the condition remained with the individual.

The evolution of Mind Genomics into a much deeper use of AI opened up the questions about what SCAS might be able to contribute to an understanding of why people fail to go to follow-up appoints with their doctor after learning that they are suffering from a serious condition. Could AI provide insights, especially with the newly discovered ability to ‘prime’ AI with a detailed background of an issue, and then instruct AI to ‘flesh out’ what might be going on in the mind of a person? As we move through the topics in this paper we must keep in mind that everything presented here regarding ‘thinking’ is the result of instructing Socrates as a Service (SCAS), viz., a version of AI powered by Chat GPT 3.5 [7].

Demonstration: Priming AI to Simulate Poor Patients Living in Public Housing

The remainder of this paper presents the results of a simulation using SCAS (Socrates as a Service, a form of AI growing out of ChatGPT 3.5), and the secondary analysis, viz., AI summarization of the data generated by the SCAS simulation. The important thing to keep in mind is that there is almost no information of any substantive import presented by the user, other than the initial framing of the situation, and what the user wants to ‘discover’ by having AI simulate the answers in place of having a human being do so.

The process begins with the orientation provided to AI, shown in Table 1. The table divides into three sections.

Table 1: The input to Socrates as a Service (SCAS)

tab 1

Section 1 – Input Information to SCAS

Here, the user creates a general picture of the situation. The input positions the user as a person working in a clinic in a poor area in Brooklyn. One might this simulation with a variety of different so-called general pictures, such as stating that the area is inhabited by upper middle classes, that the person works in a concierge medical service, that the location is somewhere else. With that flexibility the user would be well on the way to parametrically exploring the different alternatives. The opportunities are endless.

Section 2 – Understanding the Mind-sets

Here the user presents SCAS with a minimum amount of information, sufficient however to allow SCAS to create mind-sets. The user does not define the concept of mind-set, nor does the user give any hint about what properties are possessed by the three mind-sets. Given only this minimal amount of information, really only one piece of information, that there are three mind-sets, the system requests AI to create names, and inner thoughts of these three mind-sets.

Section 3 – Request that SCAS Produce 12 Message and Estimate the Performance of Each Message among the Three Mind-sets

The final request generated the desired 12 messages to be evaluated by three mind-sets. It is important to emphasize that nowhere in the instructions is any information presented to SCAS program that could be considered to be a subject-relevant prompt. All of the information generated by SCAS comes from the way SCAS processes the request.

Table 2 present the first part of the output, viz., the three mind-sets, explicated in terms of what each mind-set thinks at the time of making the appointment, and then a week before the appointment. The remarkable thing emerging from Table 2 is the realistic nature of the mind-sets and their thoughts. Once could easily think that these are verbatim quotes emerging from a discussion with the patient about the issue of making and keeping medical appointments.

Table 2: The description of the three mind-sets emerging from SCAS. As noted in the text, SCAS was not given any specific material on mind-sets which to base what it returned to the user.

tab 2

Table 3 shows how each of the three mind-sets would estimate the likelihood of showing up for the follow up medical appointment if the mind-set were to be reminded through the slogan. The slogans were created by SCAS. SCAS ‘predicts’ that all 12 would be effective for Mind-Set 1 (proactive), effective for Mind-Set 3 (Anxious), but not particularly effective for Mind-Set (Carefree). Once again it should be noted that these results make sense. We expected a mind-set named Carefree not to care about any of the messages, and thus not pay attention to follow up messages with the slogans shown in Table 3.

Table 3: Estimated likelihood of showing up for the follow-up appointment, for each of 12 slogans by each of the three mind-groups. Everything was generated by SCAS, using only the input to SCAS shown in Figure 1.

tab 3

Inventing the Future Using Today’s Topics

The second part of this paper focuses on the use of SCAS to understand what to do in order to improve the compliance of patients regarding their requested follow up visit. The original use of SCAS was to allow the user to type a ‘squib’ or information about a topic and have SCAS return with a set of 15 questions. The same feature was available for SCAS to return 15 answer to a given question. These feature remain in SCAS, and led to an effort to compare the answers to the same questions when SCAS was told that the answers had to be appropriate for the 21st century (now), and then that the answers had to be appropriate for the 22nd century (75 years hence).

The same set of 15 questions was used to compare the answers for the two centuries. The SCAS was primed to provide four separate answers to each of the 15 questions, requiring the answers to be appropriate for the 21st century (Table 4, answers A-D), and then be appropriate for the 22nd century (Tabe 4, answers E-H, italicized). Table 4 suggest that the answers for the 22nd century seem reasonable, and to be extensions of current day technology.

Table 4: Fifteen SCAS-generated topic-related questions about office visits to the medical professional. Each question shows four SCAS-generated questions assuming a year in the 21st century, and then a year in the 22nd century.

tab 4(1)

tab 4(2)

tab 4(3)

Summarization of Information Proposed – Broad Overviews Produced by SCAS

When the Mind Genomics study has been closed, SCAS creates a set of summarizations for each iteration, doing the summarizations separately. These summarizations are returned to the user in an email, usually within a half hour after the close of the study. Thus, in the not-unusual case of the user doing 10-15 iterations with different squibs, e.g., exploring different time periods with the same instructions, the user will receive one page for each effort, all of the pages becomes tabs in the one Excel workbook.

Table 5 shows one set of summarization, aptly summarized ‘Ideas’. The three summarization are key ideas, themes, and then perspectives.

Table 5: Summarization of the output from SCAS in terms of key ideas, themes emerging from the key ideas, and then a discussion of the positives and negatives of each theme.

tab 5

Key ideas simply highlights what the term suggests, namely what are the ideas presented to the user. This study generates a great number of key ideas because input to the studies comprises the basic questions and the answers pertaining only to the 21st Century both shown in Table 4.

Themes further summarize the key ideas, this time using SCAS to group together the related group of key ideas, perspectives, in turn, take these themes and provide the basis for ongoing discussion and learning, showing two alternative points of view for each theme.

The ’Human Reaction’ to These Ideas, as Envisioned by SCAS

As part of the summarization, SCAS returns with three different analyses of the sets of ideas. The analyses look at populations of people, whether these populations be defined by who they are (for both interested and opposing audiences), or by the way they think (alternative viewpoints). Table 6 shows the various groups and their reactions to the ideas uncovered by SCAS. It is important to keep in mind that these reactions are to the general ideas, not to any specific idea.

Table 6: The ‘human’ reaction to these ideas as envision by SCAS

tab 6

The final analysis deals with SCAS as an inventor. Table 7 shows two sections. The first section lists questions about what may be missing. These are typically questions which ask: How do we… ? The second section lists possible innovations, based upon the information processed by SCAS. The list of possible innovations is organized by topic.

Table 7: Using SCAS to suggest new products and services

tab 7(1)

tab 7(2)

Discussion and Conclusions

This paper emerged from recurring discussions about the real problem of ‘no shows’ in the world of medicine. The problem is a vexing one, perhaps growing because of the increasing difficulties encountered in the practice of medicine. One problem is the growing lack of affordability of medical treatments, the cost perhaps acting as a mechanism to discourage visits because of the fear of incurring expenses that are unaffordable to the patient. A second problem is the reality that doctors no longer make house calls. The patient must go to the doctor, a trip which might be difficult to schedule in view of the competing demands on the patient’s time. The third is the loss of the personal relationship between patient and doctor as the small, perhaps long-time practices are incorporated into the large medical practices. What was a personal relationship between patient and doctor (or other medical professional) now becomes a short interaction, often with the doctor’s assistant taking the necessary measurements, and the doctor meeting the patient for a few minutes debrief [8].

The importance of this paper is not in the solution in provides, but rather in the way SCAS can help focus the problem, providing a source of ideas. The speed (minutes), the extensive results in terms of the ‘human element’, and the presentation of the results in an easy-to-understand format, all suggest that those in the medical profession might avail themselves of SCAS as they enter a new subject area, if only to understand some of the issues from the part of the patient, the doctor, and the system. Scattered publication suggested only the positive, the ‘up-side’, and not the down-side of using AI and such offshoots as SCAS to solve the problem of no-shows [9].

A second aspect of the approach presented here comes from the potential of instructing SCAS to ‘imagine’ what will happen in the years to come, or even to imagine what things were like a century ago or even longer. By simply asking SCAS to assume that all the topics are to be asked from the framework of the year 2200, almost 75 years into the future, it is possible to jump-start futuristic thinking. There is no reason to assume that the answers will be ‘correct.’ On the other hand, to SCAS there is no penalty for being ‘wrong’, so that SCAS dutifully produces its best guess, once it has been properly instructed. It is that potential to focus on the future in terms of concrete questions and suggestions which make the approach attractive, especially in light of the simplicity of executing just another ‘iteration,’ albeit this time priming SCAS to guess about the future or guess about the past [10,11].

References

  1. Parsons J, Bryce C, Atherton H (2021) Which patients miss appointments with general practice and the reasons why: a systematic review. British Journal of General Practice 71(707): e406-e412. [crossref]
  2. Nadarzynski T, Miles O, Cowie A, Ridge D (2019) Acceptability of artificial intelligence (AI)-led chatbot services in healthcare: A mixed-methods study. DIGITAL HEALTH.v.5. [crossref]
  3. Kreps GL, Neuhauser L (2013) Artificial intelligence and immediacy: designing health communication to personally engage consumers and providers. Patient Education And Counseling 92: 205-10. [crossref]
  4. Lacy NL, Paulman A, Reuter MD, Lovejoy B (2004) Why we don’t come: patient perceptions on no-shows. The Annals of Family Medicine 2: 541-545. [crossref]
  5. Miller AJ, Chae E, Peterson E, Ko AB (2015) Predictors of repeated “no-showing” to clinic appointments Am J Otolaryngol 36: 411-4. [crossref]
  6. Sun CA, Taylor K, Levin S, Renda SM, Han HR (2021) Factors associated with missed appointments by adults with Type 2 Diabetes Mellitus: a systematic review. BMJ Open Diabetes Research & Care 9(1).
  7. Wu T, He S, Liu J, Sun S, Liu, K,et al.( 2023) A brief overview of ChatGPT: The history, status quo and potential future development. IEEE/CAA Journal of Automatica Sinica 10: 1122-1136.
  8. Bjerring JC, Busch, J (2021) Artificial intelligence and patient-centered decision-making. Philosophy & Technology 34: 349-371.
  9. Yun JH, Lee EJ, Kim DH (2021) Behavioral and neural evidence on consumer responses to human doctors and medical artificial intelligence. Psychology & Marketing 38: 610-625.
  10. Fogel AL, Kvedar JC (2018) Artificial intelligence powers digital medicine. npj Digital Med 1: 5. [crossref]
  11. Kuper A, Whitehead C, Hodges BD (2013) Looking back to move forward: using history, discourse and text in medical education research: AMEE Guide No. 73. Medical Teacher 35: e849-60. [crossref]

Explorations in Time Using SCAS (Socrates as a Service): Reimagining the Doctor’s Waiting Room of 1850 versus 2150 and the Evolution of that Room from 1600

DOI: 10.31038/PSYJ.2024632

Abstract

With the help of AI-based SCAS (Socrates as a Service), developed to support Mind Genomics, the study considered the nature of the doctor’s waiting room of the year 1850, followed by a paragraph about the doctor’s waiting room in 50-year intervals, from years 1600 to 2350. SCAS produced basic information about the doctor’s office as it changed over the centuries and was able to use that basic information to create even more information regarding ideas for innovation. Mind Genomics was also prompted to suggest responses of acceptors versus rejectors of the features of the 1850 doctor’s office. The paper demonstrates the simplicity, speed, and depth of information that can be obtained using AI, and the promise of the coupling of interesting reading with deeper information.

Introduction: The ‘Draw’ of the ‘What Was’ and ‘What Will Be’

A continuing theme in many aspects of life is the fascination of what was and what will be. The world of history gives people a chance to experience what happened before, and the world of ‘future studies’ for want of a better term gives people a chance to look at trends and peer into a future which might be. Indeed, the focus on the world over time, before, now, and in the future, has given the world wonderful works of history, literature, philosophy, just to name a few disciplines.The introduction of AI, artificial intelligence, has made it possible to move beyond what has been published in history and in ‘futurology.’ Through its own mechanisms of deep learning, it may be possible to get a sense of what the past may have been, not so much from reading books, but from asking AI to paint a picture of a specific issue. Even more interesting may be the attempt to do the same, not so much painting a picture of the past as a picture of the reasonably near future, a few decades from now, or perhaps a century or so. It was the development of two technologies which, when combined, opened up the focus on the past. The first was the emerging technology of Mind Genomics [1]. In simplest terms, Mind Genomics is the study of the everyday, the ordinary events, material things, and behaviors. The second is the new availability of user-friendly AI, artificial intelligence, embedded in Mind Genomics as SCAS, (Socrates as a Service), and based upon current AI systems [2]. Mind Genomics opened up the possibility of studying the everyday more deeply, looking into features, painting a picture of a situation and understanding what is important to people. The result was the realization that the ordinary events of everyday, the quotidian life, are worth studying. SCAS, embodying easy to use AI, allowed the investigation of the everyday life, not by doing experiments but rather by asking the embedded AI to assume a situation, and then report on its details.

It is important to note that this paper follows in a stream of previous work, much of it trying to digitize the historical narrative, to make history ‘come alive’ to students [3-5]. Furthermore, a great deal of interest in AI-based simulation comes from the desire to add reality and depth to history-games, which are very popular. These games try to create a realistic ‘set’ and realistic ‘behaviors.’

Exploring a Simple Topic: The Doctor’s Waiting Room Across Years

The ‘research’ presented here began with a request to SCAS to present a short description of what the doctor’s office was like:

The year is: [provided by user]. Everything that is talked about here happened in [provided by user].The doctor is a general practitioner in New York

What should the waiting room of the doctor be like, in terms of decor, in terms of people, in terms of the way people are greeted. Write you answer as five sentences in one long paragraph, simply written, in order to give the reader a complete description. Make the writing lively, and fun to read, and make the description realistic, as if the person reading the description were to be right there, at this time and this place

Table 1 shows the results for three years, 1900, 2000, and 2150., respectively. The appendix to this paper shows many more years, beginning with 1600 and going to 2300 in 50-year leaps. The first reaction to the ‘first fruits’ of this effort are summarized by the ‘astonishment.’ The paragraphs describing the mundane topic of the doctor’s waiting room seem real, as if someone were there. This led to doing the ‘experiment’ with 50-year intervals, starting in the year 1600, and proceeding to the year 2250. The Appendix those short descriptions.

Table 1: Descriptions of the doctor’s office, product for three years, 1900, 2000 and 2150

TAB 1

Exploring the Doctor’s Waiting Room in Detail – The Year is 1850

The remainder of this paper shows an AI-based exploration using SCAS. The year is 1850. The general instructions appeared above. SCAC produces the immediate output shown in Table 2. The material is similar to what appears in Table 1, as well as in the Appendix. Once again, it is important to emphasize that the paragraph is synthesized by SCAS without any information other than the year, and the directive to provide the answer as a story in five sentences.

Shortly after the completion of the session, after the Mind Genomics program finishes, SCAS produces a summarization of the results. Within the summarization appear a detailed expansion of ideas, all based upon the five sentences shown in Table 2.

Table 2: The SCAS-generated description of the doctor’s waiting room of 1850

TAB 2

Key Ideas, Themes, and Perspectives

The first set of subsequent analyses present the various ideas, this time expanded. Once again, SCAS returns with an easy-to-read analysis, all based on what SCAS had produced initially in answer to a simple question. Essentially, therefore, SCAS is producing ‘new knowledge’ based upon ‘knowledge’ it had developed simply knowing the topic and the year. Note that the perspectives are different points of view about the topics presented in the section on themes (Table 3).

Table 3: Expansion of knowledge through the key ideas, themes, and perspective regarding those themes

TAB 3

SCAS provides a sense of who would be interested in the materials, and who would be ‘opposed’ to the materials. These appear in Table 4. Once again, it is SCAS which is working on the information it first generated to provide additional information or real points of view.

Table 4: Points of view, interested versus opposed

TAB 4

Steps Towards Innovation (of Knowledge)

The final summarizations deal with questions and ideas for innovation. For this historical exploration using SCAS there is no ‘innovation’ per se. Rather, the ‘innovations’ comprise questions to answer. These are presented in the sections called ‘Alternative Viewpoint,’ and ‘What is Missing,’ both in Table 5. SCAS does return with ‘innovations,’ but this is the one section in SCAS which as yet cannot put itself into the mind of the 1850 doctor to look at the innovation of that time.

Table 5: Questions to answer, to create new knowledge about the doctor’s waiting room in 1850

TAB 5

Discussion and Conclusions

The objective of this paper is to explore how deeply one can ‘flesh out’ an otherwise modestly interesting topic, the doctor’s waiting room, although a topic which has received attention in the popular literature [6]. There is a relevant academic literature dealing with the history of doctor’s offices and their furnishings [7,8]. It is likely, however, that the material being published will interest the experts, whether these experts be those who study the history of interior design [9], or the history of medicine [10]. There is also a developing literature on the additional aspects of the doctor’s waiting room, such as design, content, etc., based upon the recognition that the waiting room is not only a place to store people, but also to make their visit pleasant [11,12], and a chance to teach them [13]. There is always a need for solid academic work the topic. It is hoped that the simulation efforts with SCAS shown here adds to the bank of knowledge and contributes to the study of the history and sociology of those in the health field and those in the field of interior design.

The real opportunity presented in this paper emerges in the world of education. The use of Mind Genomics, and especially its easy use AI embodied in SCAS can result in a great deal of relevant information being produced in minutes, with the student able to modify the requests to SCAS, and in turn get new information in virtually seconds. Afterwards, there is the major contribution to education products from the SCAS-based summarization of the information. Each iteration, the effort taking about 30 seconds per iteration, is returned with a full summarization, one Excel tab for each iteration. A student excited about the prospects, can work for 30 minutes, generating a great deal of information, with the nature of the requested information dynamically changing according to the instructions written into the squib by the user, in this case the student. One can only imagine the level of excitement as the student works with SCAS and Mind Genomics a coaches, teaching the study many things in dept, and actively interacting with the student who wants to explore the topic in different ways.

A question that can be posed is how does this AI image of a doctors waiting room across the eras, past and future, coalesce with reality? One thing that can be considered in the current and future eras is the post-COVID 19 world where telehealth and social distancing has become the norm, particularly in healthcare settings. We therefore must consider the potential future of waiting rooms with the emergence of telemedicine as less crowded [14].This is important when we consider the impact of COVID on the layout of waiting rooms, with aspects such as social distancing, spacing the time between appointments in order to prevent crowded waiting rooms, and so forth. The emergence of ancillary healthcare personnel, from the licensure of higher and higher rankings in nurse practice levels, as well as the introduction of physician assistants, have made the visit to the doctors’ office a place where there could potentially be more individuals working at the back end than patients waiting in the front.

A more casual flair is also being approached in medical offices, from patients to healthcare workers alike, with a “casualization of the workforce” occurring [15], keeping in line with recent trends in society as a whole. This casualization may likely show itself in the change of the patient waiting room, from a room psychologically separate from where the medical professionals work to simple part of a continuum of space, with far less psychological separation. This change will manifest the evolving change in power of dominance by the medical professional over the patient to one of cooperation and collaboration. One need only see the change from the formal living, dining and kitchen spaces of traditional homes to their blending in new homes, as designed by forward looking architects with their forward-looking clients.

References

  1. Moskowitz HR (2012) ‘Mind genomics’: The experimental, inductive science of the ordinary, and its application to aspects of food and feeding. Physiol Behav 107: 606-13. [crossref]
  2. Kalyan KS (2023) A survey of GPT-3 family large language models including ChatGPT and GPT-4. Natural Language Processing Journal, p. 100048.
  3. Danyun L, Jiun CY (2016) October. Historical cultural art heritage come alive: Interactive design in Taiwan palace museum as a case study. In 2016 22nd International Conference on Virtual System & Multimedia (VSMM) (pp. 1-8) IEEE.
  4. Staley DJ (2015) Computers, Visualization, and History: How New Technology Will Transform Our Understanding of the Past. Routledge.
  5. Taylor T (2003) Historical simulations and the future of the historical narrative. Ann Arbor, MI: MPublishing 6: 2.
  6. Tanner LE (2002) Bodies in waiting: representations of medical waiting rooms in contemporary American fiction. American Literary History 14: 115-130.
  7. Gainty C (2019) Why Wait? Modern American History 2: 249-255.
  8. Waltz M (2016) Patient Patients: An Ethnography of Medical Waiting Rooms. Case Western Reserve University
  9. Figueroa NI (2016) Culture, gender, and medical waiting rooms: A Kuwaiti case study. Journal of Interior Design 41: 33-46.
  10. Devlin AS (2022) Seating in doctors’ waiting rooms: Has COVID-19 changed our choices?. HERD: Health Environments Research & Design Journal 15: 41-62. [crossref]
  11. Berkhout C, Zgorska-Meynard-Moussa S, Willefert-Bouche A, Favre J. et al (2018) Audiovisual aids in primary healthcare settings’ waiting rooms. A systematic review. European Journal of General Practice 24: 202-210.
  12. Lai JCY, Amaladoss N (2022) Music in waiting rooms: A literature review. HERD: Health Environments Research & Design Journal 15: 347-354. [crossref]
  13. Maskell K, McDonald P, Paudyal P (2018) Effectiveness of health education materials in general practice waiting rooms: a cross-sectional study. British Journal of General Practice 68: e869-e876. [crossref]
  14. Swicki B (2021) “The future of the waiting room, and how telemedicine and mobile health could change it”, Healthcare IT News.
  15. Taub P (2001) The trend towards casual address and dress in the medical profession. Virtual Mentor 3: 169-171.

High XPO1 Expression can Stratify Gastric Cancer Patients with Poor Clinical Outcome

DOI: 10.31038/CST.2024921

Abstract

Objective: Local recurrence and abdominal metastasis are the main reasons for reducing the survival. It is of great clinical value to identify patients with more malignant biological features at high recurrence and metastasis risk. We want to evaluate the efficacy and sensitivity of XPO1 as a biomarker to stratify gastric cancer patients at high biological aggressive risk.

Method: We retrospectively analyzed the pathological records of 100 enrolled patients with gastric cancer who underwent gastric cancer resection in the department of surgery of our hospital from January 2017 to December 2022; all enrolled patients had complete pathological data and follow up for survival. In this study, we analyzed the immunohistochemical staining patterns of gastrectomy tissue specimens with patients with follow-up survival information and evaluated the efficacy of a novel biomarker XPO1/CRM1, also called Exportin 1.

Results: The positive IHC of XPO1 was correlated with the following factors: primary tumor volume (P value=0.05), regional lymph node invasion (P value=0.008) and TNM staging (P value=0.069). We noticed a sequential upregulation of XPO1 IHC intensity in benign lesions, borderline tumors, invasive carcinomas biological changes. Kaplan-Meier survival analysis indicated that XPO1 positivity was associated with poor survival.

Conclusions: Our results revealed XPO1 as a sensitive and useful biomarker to stratify gastric cancer patients at high biological aggressive risk. We recommend supplementing XPO1 IHC to routine pathology test to stratify individual patients for intensive therapy and stringent follow-up plans.

Highlights

  • High XPO1 can stratify tumors with more biology malignancy trend
  • High XPO1 predicts poor prognosis in gastric cancer
  • High XPO1 patients need stringent treatment and follow-up

Keywords

Biomarker, Gastric cancer, Immunohistochemistry, Prognosis, XPO1

Introduction

Gastric cancer has a high incidence and poor prognosis, particularly in China. Each year, most new cases of gastric cancer are diagnosed among Asians and Eastern Europeans [1]. In 2020, approximately 27,000 new cases will be diagnosed [2]. The survival rates for patients with gastric disease are 31% in the United States and 25% globally [3]. In addition to its high incidence, gastric cancer has a poor prognosis and survival rate. Local recurrence and abdominal metastasis substantially impaired long-term survival. Common causes of poor prognosis [3,4] include late-stage diagnosis with regional or distant metastases, intratumor heterogeneity, and chemotherapeutic resistance. The identification of novel and specific biomarkers with prognostic significance and novel targets in gastric cancer is urgently required. At present, gastric cancer remains a fatal disease with limited treatment options. In clinical practice, clinicians execute TNM staging for patients primarily based on imaging; we believe it would be more beneficial if biomarkers that can predict the intrinsic metabolic characteristics of tumor cells could be identified for clinical applications. In addition to TNM staging, for instance, more effective prognostic assessment methods for gastric cancer can be identified, and patients who are more likely to experience recurrence can be identified. Recent reports have linked elevated XPO1 expression to a poor prognosis in a variety of tumors. XPO1, also known as CRM1, is a nuclear pheherin that belongs to the importin-superfamily [5-7] and can export at least 221 NES containing proteins and several nuclear Rnas to the cytoplasm [8,9]. The presence of conserved hydrophobic NES on carrier molecules was identified by XPO [9-11]. XPO1 participates in the localization and passive transport of diverse regulatory proteins between the nucleus and cytoplasm. Presently, it is known that XPO1 regulates a number of tumor suppressor genes that play a significant role in the pathogenesis and progression of cancer. Among the cargo proteins detected to be transported by XPO1 are the tumor suppressor p53, CDK1, adenomatous colonic polyposis (APC), BRCA1 and BRCA2, survivin, etc. [12,13]. Therefore, targeting XPO1 has promising potential as a cancer treatment. Intriguingly, XPO1 inhibitors effectively discriminate between tumor and normal tissue. XPO1 inhibitors are more likely to selectively and preferentially target tumor cells. The mechanism may be that, compared to non-malignant tumors, tumor cells express more XPO1 and cancer cells have an increased rate of cell proliferation and metabolism, making them more susceptible to nuclear trafficking inhibition [14,15]. First, we selected 100 gastric cancer patients with comprehensive clinical data from the pathology center of our hospital; all of these patients underwent surgical resection of gastric cancer in our hospital. Immunohistochemical staining was used to determine the XPO1 protein expression level in paraffin-embedded specimens of gastric carcinoma. We analyzed XPO1 IHC results in various TNM stages, as well as the correlation between XPO1 positivity and patient clinical data. Second, we analyzed XPO1 positivity variations in benign lesions, ambiguous tumors, and invasive carcinomas. We observed a pathological upregulation of XPO1 in malignant transformation of tumors, indicating its role in tumorigenesis. We performed a Kaplan-Meier analysis of survival to determine the impact of XPO1 on the clinical prognosis and survival of patients with gastric cancer. High XPO1 was able to stratify high-risk patients and predict a poorer prognosis, according to the findings. We advise these patients to adhere to rigorous treatment regimens and frequent follow-up appointments. Finally, we extended our findings to additional cancer categories. By comparing pan-cancer XPO1 expression and conducting survival analyses, we identified XPO1 as a biomarker for a poor prognosis in a variety of cancer types.

Methods

Patients’ Enrollment

We selected 100 patients who underwent surgical resection for gastric cancer at Suqian Hospital Affiliated with Xuzhou Medical University between January 1, 2017 and December 31, 2022. All patients enrolled in this study were informed of the study’s purpose and procedures, and all provided written consent to participate. The included patients must have comprehensive basic and clinic pathological information. Their paraffin-embedded tissue specimens were retrieved from the pathology department archives. Bormann grade of gross morphology and WHO grade of histopathology were used as the pathological diagnostic criteria [16]. The TNM classification of the 5th edition of the International Union against Cancer (UICC) was utilized for cancer staging [17]. Patients’ clinical information was gathered, recorded, and analyzed in detail. Indicators analyzed included the patient’s gender, age, tumor size, gastric wall invasion depth, histopathological grade, regional lymph nodes, and distant metastasis. Each patient was individually contacted via telephone to inquire about their survival status and to obtain a death date from their family. None of the patients included in the study received radiotherapy, chemotherapy, or immunotherapy prior to surgery. The study protocol was approved by the Ethics Committee Board of Suqian Hospital Affiliated to Xuzhou Medical University, and all experiments were carried out in accordance with Xuzhou Medical University’s guidelines.

Immunohistochemistry Staining

Tissues embedded in paraffin were sliced into 5-mm-thick sections. The portions were deparaffinized with xylene three times for five minutes each and rehydrated with 90, 75, and 50 percent ethanol in each container for two minutes. To recover antigenicity, the sections were submerged in a 10 mmol/L citrate buffer solution (pH 6.0) and microwaved for 12 minutes. To inhibit the activation of endogenous peroxidase, the samples were treated for 12 minutes with 3% hydrogen peroxide–methanol and then rinsed with distilled water. Anti-XPO1 rabbit polyclonal antibody (sc5595; Santa Cruz Biotechnology, Santa Cruz, California; 1: 100 dilution) was applied and incubated for one hour. Following washing, sections were rinsed with TBS and incubated with horseradish peroxidase-conjugated anti-rabbit antibody (Dako Cytomation, Carpinteria, CA). Phosphate-buffered saline (PBS) was substituted for the primary antibody to create negative controls.

Interpretation and analysis of immunohistochemistry results: Two pathologists independently examined the radiographs without knowledge of the patient’s clinical history. Each slide was investigated individually using a light microscope. When the results of two pathologists’ reviews are incongruent, the conclusion of the review is reached through mutual consultation between the two pathologists. The following criteria were used to interpret the XPO1 staining results: The intensity and proportion of positive cells were used to evaluate the immunostaining for XPO1. The staining intensity scores were as follows: 0 (negative), 1 (mild positive), 2 (medium positive), and 3 (strong positive). The following four kinds of scores were calculated based on the proportion of XPO1-positive cells: 0% to 10% was 1, 11 to 50% was 2, 51 to 80% was 3, and 81 to 100% was 4. As indicated previously, the final XPO1 staining score was calculated by multiplying the intensity score by the percentage score [18]. Positive results were defined as > 10% of cells with dark brown nuclei staining, and negative results were defined as < 10% of cells with staining. We determined the cutoff point for XPO1 IHC scores using the X-tile software (Rimm Lab at Yale University, http: //www.tissuearray.org/rimmlab).

Kaplan-Meier Survival Analysis in TCGA

The Kaplan-Meier curves for overall survival (OS) have been calculated for the high/low XPO1 expression group dichotomized by the 75% quantile of XPO1 expression. The log-rank test was utilized to investigate the difference in survival between those with high and low XPO1 expression.

Statistical Analysis

We used the chi-square test to assess the relationship between XPO1 expression and various clinicopathological features of gastric cancer. Cox’s proportional hazards regression models were used to determine univariate and multivariate analyses in order to identify independent factors associated with disease-free survival and overall survival. The Kaplan-Meier method was utilized to assess the relationships between XPO1 expression and patient outcomes. *, P < 0.05, **, P < 0.01, ***, P < 0.005, and exact P values are stated in the source data for each figure panel.

Results

Gastric Cancer Exhibits Higher XPO1 with Immunohistochemistry

The clinical characteristics of the patients were summarized and exhibited (Table 1). The ages of the patients ranged from 30 to 85 years. High XPO1 expression was specifically correlated with TNM stage (p=0.003), tumor stage (p=0.05), and lymph node metastasis positivity (p=0.007). In contrast, no significant correlation was found between XPO1 expression and other clinical factors, including gender, tumor diameter, age, and status of distant metastasis. To determine if there are any differences in XPO1 expression between gastric patient samples and normal gastric tissues, we compared XPO1 expression between gastric cancer tumors and normal gastric epithelial tissues. Each clinic sample contains comprehensive information regarding the pathology cell type and tumor stage. Two pathologists independently evaluated the results of XPO1 immunohistochemistry staining, with no knowledge of the patient’s clinical history. When there was disagreement, a conclusion was reached via consensus. Evaluation of the immunostaining was based on the intensity and percentage of XPO1-positive cells. The stain’s intensity was measured as follows: 0 (negative), 1 (weakly positive), 2 (moderately positive), and 3 (strongly positive). In addition, the percentage of XPO1-positive cells was scored based on four categories: 1 for 0 to 10%, 2 for 11 to 50%, 3 for 51 to 80%, and 4 for 81 to 100%. Multiplying the intensity and percentage scores produced the final XPO1 staining score. The emblematic images of IHC were displayed. We observed a substantial difference in XPO1 expression between tumor and normal tissue samples. Strong XPO1 positivity was observed in gastric tumors, and XPO1 intensity increased with TNM stages II, III, and IV (Figure 1).

Table 1: Demographic characteristics of the 100 gastric cancer patients. High XPO1 expression was associated with TNM stage (p=0.003), tumor stage (p=0.05) and positive lymph node metastasis (p=0.007). No significant correlation was discovered between XPO1 expression and other clinical parameters, such as gender, age, tumor diameter, and distant metastasis status.

 

n

Negative (%) (n = 55)   Positive (%) (n = 45)

P-value

Gender (M: F)

Age (years)

Longest diameter (cm)

T stage

T1

T2

T3

T4

Nodal stage

N0

N1

N2

N3

Distant metastasis

M0

M1

TNM stage

I

II

III

IV

 
 
 
 

33

24

38

5

48

34

14

4

97

3

44

17

33

6

 
 
 
38: 17(69%: 31%)

56.38

3.65

27(82%)

11(46%)

16(42%)

1(20%)

37(39%)

20(59%)

4(29%)

1(25%)

62(64%)

0(0%)

34(77%)

11(65%)

15(45%)

2(33%)

 
 
 
 
26: 19(58%: 42%)

57.65

4.76

6(18%)

13(54%)

22(58%)

4(80%)

11(61%)

14(51%)

10(71%)

3(75%)

35(36%)

3(100%)

10(23%)

6(35%)

18(55%)

4(67%)

 
 
 
 
 
 
 

0.531

0.723

0.125

0.05

0.007

0.072

0.03

 

fig 1

Figure 1: Gastric tumors in TNM II, III, IV stages exhibit increased XPO1 intensity with immunohistochemistry staining. Representative immunohistochemistry results for anti-human XPO1 staining were presented. Gastric tumors were strongly positive for XPO1, and XPO1 intensity increased with TNM II, III, IV stages. Compared with gastric cancer samples, the expression of XPO1 in normal tissues was limited or absent.

Compared with gastric cancer samples, the expression of XPO1 in normal tissues was limited or absent. There was a statistically significant difference between adjacent non-tumor tissues and tumor-infiltrated areas in XPO1 expression, P = 0.001. The rate of positivity in normal tissue was 6%, whereas the rate of positivity in tumor areas was significantly higher (45%). High XPO1 expression was detected in 45 of 100 (45%) gastric cancer tissues, while only 6 of 100 (6%) normal gastric tissues displayed XPO1 expression (Table 2). These results indicated that XPO1 signaling was strongly activated in gastric cancer.

Table 2: Overall XPO1 expression in tumor and surrounding normal tissues. IHC was employed to investigate the expression of XPO1 in gastric cancer. There was a statistical difference in the XPO1 expression between tissues adjacent non-tumor tissues and tumor-infiltrated areas (p=0.0001).

XPO1

Normal tissue

Cancer

negative

1–10%

11–50%

51–100%

P-value

94 (94%)

4(4%)

2(2%)

0(0%)

0.0001

25(55%)

13(23%)

18(18%)

44(4%)

XPO1/CRM1, also called Exportin 1, Cancer-related genes, FDA approved drug targets.

XPO1 Plays a Role in Tumor Initiation and Progression

Previous research indicates that XPO1 exports tumor suppressor genes from the nucleus and promotes tumorigenesis. We hypothesized that XPO1 facilitated tumor initiation, i.e., that XPO1 levels would increase during the carcinogenesis process. We compared the variance in XPO1 IHC intensity among benign lesions, ambiguous tumors, and invasive carcinoma groups. Although there was no XPO1 positivity in benign lesions, there was an increase in borderline tumors. Strong XPO1 positivity was observed in invasive carcinomas (Figure 2).

fig 2

Figure 2: Gastric tumors exhibit higher XPO1 expression, which predicts shorter disease-free survival and overall survival. (A) Presentative XPO1 IHC staining in benign lesions, borderline tumors, and invasive carcinoma groups. There was no XPO1 positive staining in benign lesions, however, invasive carcinoma showed a very strong XPO1 positive staining. (B) STRING analysis showed the top genes interacting with XPO1 in gastric cancer, of which the top correlated genes were TP53, CDKN1B, RANBP2, NUP98, NUP214. (C) Gastric cancer had higher XPO1 expression than the normal tissues, results calculated from TCGA gastric cancer cohort. (D) Higher XPO1 expression predicts shorter disease-free survival in gastric cancer, p=0.031. (E) Higher XPO1 expression predicts shorter overall survival in gastric cancer, p=0.039.

To gain a more detailed understanding of how XPO1 may interact with other genes. We analyzed the top genes in TCGA gastric cancer cohorts that correlate with XPO1. According to gene STRING analysis, the XPO1 gene is closely related to a number of genes that promote malignancy. STRING analysis showed the top interacting genes with XPO1 in gastric cancer, of which the top genes were TP53 (responds to diverse cellular stresses, induce cell cycle arrest, apoptosis, senescence, DNA repair, or changes in metabolism), CDKN1B (cyclin-dependent kinase inhibitor, which shares a limited similarity with CDK inhibitor CDKN1A/p21), RANBP2 (RAN binding protein 2, enables SUMO ligase activity), NUP98 (the 96 kDa nucleoporin is a scaffold component of the nuclear pore complexes), NUP214 (the protein encoded by this gene is localized to the cytoplasmic face of the nuclear pore complex). Gastric cancer had higher XPO1 expression than the normal tissues. Higher XPO1 expression was related with shorter disease-free survival, p=0.031 and overall survival, p=0.039 in gastric cancer. To gain a more detailed understanding of how XPO1 may interact with other genes. We analyzed the top genes in TCGA gastric cancer cohorts that correlate with XPO1. According to gene STRING analysis, the XPO1 gene is closely related to a number of genes that promote malignancy. Results revealed a considerable increase in XPO1 during the progression of gastric cancer from benign lesions to borderline tumors and then to the terminal invasive carcinoma (Table 3).

Table 3: Increased expression of exportin 1/XPO1 located both in nuclear and cytoplasm. The specific number of XPO1 IHC stain location in benign lesions, borderline tumors, and invasive carcinoma groups were summarized and shown.

Number. of patients (%)

XPO1
Expression

Invasive carcinomas, N=70 Borderline tumors, n=20 Benign lesions, n=10

P*

Nuclear and Cytoplasmic Negative 20 (28.5)

Positive 50 (71.4)

16(80)

4 (20)

 10 (100)

0 (0)

0.001

Nuclear Negative 41 (58.6)

Positive 29 (41.4)

18 (90)

2 (10)

10 (100)

0 (0)

0.002

Cytoplasmic Negative 49 (70)

Positive 21 (30)

18 (90)

2 (10)

10 (100)

0 (0)

0.000

XPO1/CRM1, also called Exportin 1, Cancer-related genes, FDA approved drug targets.

Nuc and Cyt, Nuclear and cytoplasmic.

*Chi-square test.

Since XPO1 is present in both the nucleus and cytoplasm, both patterns were evaluated separately. Immunohistochemistry for XPO1 was negative in benign lesions. In borderline tumors, XPO1 positivity was more prominent than in benign lesions. Nuclear (2 of 20) and cytoplasmic (2 of 20) expression was moderate in 4 of 20 borderline tumors. XPO1 nuclear expression was detected in 29 of 70 invasive carcinomas (41.4%), whereas XPO1 cytoplasmic expression was detected in 21 of 74 tumors (30%). In the majority of tumors, both expression patterns were found concurrently, albeit with differing intensities. XPO1 facilitates the transport of tumor suppressor genes outside of the nucleus and may facilitate and accelerate tumorigenesis, as suggested by these findings.

TP53 Mutant Gastric Cancer had Higher XPO1 Expression

To confirm the localization of XPO1, we examined the human protein atlas and cell atlas. We discovered that XPO1 was predominantly localized in the nucleus and cytoplasm of cancer cells, which is consistent with our results. The specific intracellular XPO1 localization were examined and analyzed (Figure 3).

fig 3

Figure 3: XPO1 has many subcellular locations, with cytosol and nucleus as the two most frequent sites. (A) Representative confocal images stained with anti-XPO1 (CAB010184) antibody. In addition to localized at the cytosol & vesicles, XPO1 mainly localize to the nucleoplasm & nuclear membrane. XPO1, also called exportin 1, was cancer-related genes, a transporter which localized to the nucleoplasm (enhanced), and nuclear membrane (enhanced). (B) The specific subcellular XPO1 location were examined and analyzed from COMPARTMENTS(C) TP53 mutant gastric cancer exhibited higher XPO1 expression. (D) Subcellular locations of XPO1 from the Human Protein Atlas (HPA) COMPARTMENTS, cytosol (5), nuclear membrane (5), nucleoplasm (2), vesicles (2).

The XPO1 localization intensity was calculated. We observed that XPO1 can be found in numerous locations within the cell, with the cytosol and nucleus being the most common. XPO1 facilitates tumorigenesis and confers drug resistance by transporting the tumor suppressor TP53. Given that p53 was a cargo protein for XPO1, it was hypothesized that inhibiting XPO1 could activate TP53. We observed that TP53 mutant gastric cancer has increased XPO1 expression. The significance of TP53 mutational and functional status on XPO1 inhibitor sensitivity in gastric cancer cell lines and the functional role of apoptosis signaling mediated by TP53 were correlated with nuclear accumulation of TP53.

Pan-Cancer XPO1 Expression and Survival Analysis

Finally, we wanted to extend our discovery to other cancer types. We investigated The Cancer Genome Atlas (TCGA) database for pan-cancer XPO1 expression analysis in cancer and normal tissues. Box graphs were used to illustrate the differential gene expressions (Figure 4).

fig 4

Figure 4: Pan cancer XPO1 expression and Kaplan-Meier survival analysis. (A) We studied the differential expression between tumor and adjacent normal tissues for XPO1 in all TCGA tumors. Distributions of gene expression levels were displayed using box plots. The statistical significance computed by the Wilcoxon test was annotated by the number of stars (*, p-value < 0.05). (B) Kaplan-Meier survival analysis between XPO1 expression and clinical outcome in multiple cancer types. In the kidney renal papillary (KRP) carcinoma, bladder tumor, cervical squamous (CS), liver hepatocellular cancer (LHC), and esophageal adenocarcinoma (EA) cohort, high XPO1 predicts shorter overall survival. The p value was labeled on each graph. (C) In lung adenocarcinoma (LA), pancreatic ductal adenocarcinoma (PDA), pheochromocytoma and paraganglioma (PP), sarcoma, and uterine corpus endometrial carcinoma (UCE) cohort, high XPO1 predicts shorter overall survival. The p value was labeled on each graph.

The Wilcoxon test’s statistical significance was indicated by the number of stars (*: p-value 0.05). We were able to determine whether XPO1 was up- or down-regulated in tumors relative to their normal counterparts for each cancer type. The 33 malignancies analyzed by the TCGA/Pan Cancer Initiative were represented schematically according to their tissue of origin. XPO1 was expressed substantially more (red) in cancerous tissues than in normal tissues, with the exception of ovarian, prostate, thyroid, and uterine cancers (black). We then performed a Kaplan-Meier analysis of survival between XPO1 expression and clinical prognosis in multiple types of cancer. In the cohort of patients with kidney renal papillary (KRP) carcinoma, bladder tumor, cervical squamous (CS), liver hepatocellular cancer (LHC), and esophageal adenocarcinoma (EA), a high XPO1 level predicts a shortened overall survival. In lung adenocarcinoma (LA), pancreatic ductal adenocarcinoma (PDA), pheochromocytoma and paraganglioma (PP), sarcoma, and uterine corpus endometrial carcinoma (UCE) cohorts, XPO1 expression was associated with shorter overall survival at a 75% quantile threshold. Collectively, these findings demonstrated that XPO1 is a potential broad-spectrum biomarker for cancer prognosis and could be a therapeutic target for treatment.

Discussion

In this retrospective study, we evaluated the immunohistochemical staining of XPO1 in gastric tumor samples and investigated the correlation between XPO1 level and multiple clinicopathology factors in predicting its clinical significance in patients with gastric cancer. XPO1 levels were substantially elevated in cancer samples compared to normal counterparts. Statistically, the degree of XPO1 positivity did not correlate with tumor size; however, higher expressions were found in patients with higher T values, more regional lymph node invasion, and advanced TNM staging, which could predict a substantially lower survival rate. We examined the differential expression of XPO1 in various phases of gastric cancer and the correlation between XPO1 immunohistochemical staining and patient clinical characteristics. Our results demonstrated that XPO1 is a valuable biomarker for stratifying gastric cancer patients based on their biologically malignant nature. Chemotherapy and surgery for gastric cancer have improved over the past few decades [19,20]. Nonetheless, patients with gastric cancer continue to have a poor prognosis due to therapeutic failure and disease progression [21]. Identification of novel and validated prognostic biomarkers in practice has clinically significance for gastric cancer. In this study, we discovered that XPO1 was a useful marker in gastric cancer that had the potential to be used as a candidate for targeted therapy. The regulation of material transport across the nuclear membrane was essential for maintaining homeostasis, which required the correct nuclear-cytoplasm positioning of large molecules; nevertheless, this process was typically dysregulated in cancer cells [22]. XPO1, an export receptor responsible for the nuclear-cytoplasm transport of multiple proteins and RNA species, was frequently overexpressed or mutated in human malignancies and served as a potential oncogenic driver [23]. Unlike small molecules, which can passively diffuse through the nuclear pore complex (NPC), larger cargo molecules (>40 kDa) require active transport via transport receptors [24-26], which belong to the karyopherin beta family and are classified as importins (nuclear import), exportins (nuclear export), and transportins (for both import and export) [25]. Studies show that exportins are potential targets in tumorigenesis [27,28], of which XPO1 was the most important and well-studied target. XPO1 was initially identified as a chromosomal mutation in the yeast Schizosaccharomyces pombe [29]. XPO1, also known as CRM1, transported over 200 proteins, the majority of which were tumor suppressors and oncoproteins [29-31]. CRM1-mediated cargos include p27, p53, FOXOs, nucleophosmin, PI3K/AKT, Wnt/-catenin, BCR-ABL, p21, NF-kB, APC, and Rb; these cargos all play important roles in tumorigenesis [28,32]. For the first time, we investigated the clinical and prognosis value of XPO1 in gastric cancer. IHC analysis revealed a higher XPO1 concentration in gastric cancer tissues compared to normal gastric tissues. Consistent with previous research, our findings indicated that a variety of malignancies exhibited a higher level of XPO1 expression than their normal counterparts [33,34]. In addition, elevated XPO1 levels in gastric cancer were associated with certain clinical-pathologic factors, including AJCC stage, positive lymph node metastasis, and tumor grade. The Kaplan-Meier analysis demonstrated that the disease-free survival and overall survival of patients with increased XPO1 expression were shorter than those of patients with negative or decreased expression. A univariate analysis revealed that XPO1 expression, AJCC stage, and lymph node metastasis were correlated with gastric cancer patients’ survival (both disease-free survival and overall survival). High levels of XPO1 and advanced AJCC staging independently predicted unfavorable disease-free survival and overall survival outcomes for patients with gastric cancer, as determined by multivariate analysis. XPO1 overexpression was identified in solid tumors and hematologic malignancies and was reported as an indicator of poor prognosis and potential drug resistance in cancers [35]. One potential mechanism for XPO1 overexpression was associated with altered transport, which promoted cancer-promoting outcomes [36]. XPO1 facilitated the import of growth regulatory proteins, such as c-myc or BCR-ABL, into the cytoplasm and consequently activated downstream signaling, resulting in sustained cell proliferation. Similarly, tumor suppressor proteins (TSPs), such as p53, p21, Rb, and p27, were rendered inactive by exportin and lost their ability to inhibit uncontrolled cell proliferation. Collectively, these findings support the notion that XPO1 inhibition is an attractive therapeutic target for its ability to target a variety of hallmarks of oncogenesis signaling. In addition, the combination of SINE compounds with existing standard regimens in multiple cancer types was feasible and well tolerated in clinical trials. Common inhibitors of nuclear export (SINE) XPO1 antagonists included KPT-185, KPT-276, KPT-251, and KPT-330, which were reported to inhibit the proliferation of triple-negative breast cancer (TNBC) cell lines and also demonstrated efficacy in human breast cancer xenograft models. Mechanically, SINE compounds inhibit XPO1 and suppress STAT3 trans-activation, thereby inhibiting the oncogenic potential of TNBC and their clinical application [36]. Priming cancer cells with XPO1 inhibitors followed by doxorubicin, melphalan, bortezomib, or carfiltiamob may sensitize de novo and adaptive cancer cell lines to drug resistance [37]. Inhibiting the activation of the XPO1 pathway would accelerate the apoptosis of tumor cells and induce cell cycle arrest [38,39]. In summary, XPO1 expression or upregulation may replicate the natural process of gastric cancer bio-evolution, and XPO1 may therefore predict and stratify patients with a poor prognosis. In another sense, we may consider the XPO1 level as a molecular staging biomarker for oncologists employing intensive surgical intervention or chemotherapy. High XPO1 expression in gastric cancer was a reliable molecular biomarker for staging and prognostic prediction during both the diagnostic and treatment phases. High XPO1 expression in gastric cancer is indicative of an aggressive phenotype requiring intensive treatment and careful monitoring. Our findings supported XPO1 as a novel prognostic biomarker for patients with gastric cancer, and targeting XPO1 may provide a beneficial strategy for gastric cancer patients with positive XPO1 expression, which is typically accompanied by TP53 mutation. As mentioned previously, inhibiting XPO1 signaling with SINE may restore the functions of common tumor suppressors. Thus, targeting XPO1 in gastric cancer may provide new treatment options for gastric cancer patients, particularly those with advanced disease and a high recurrence risk. In addition, our pan-cancer analysis of the TCGA dataset revealed that XPO1 was commonly elevated in all cancer types. Consequently, our findings illuminated the potential universal application of XPO1 inhibitors in multiple types of cancer. Future clinical studies are required to evaluate the therapeutic effects of KPT-SINE compounds (small molecules for XPO1) alone and in combination with XPO1-targeted therapy. Our research had several limitations. We detected XPO1 positivity solely through immunohistochemistry, so there is a possibility for diagnostic error. Several other techniques, such as immune blotting and qRT-PCR for mRNA expression, have been considered in an effort to achieve more precise diagnostic results. Second, all clinical data, including recurrence and survival rates, were retrospectively collected. Thirdly, the relatively small number of patients enrolled in our study may result in a lack of statistical power; therefore, a larger prospective study is needed in the future.

Availability of Data and Materials

The data generated in the present study may be requested from the corresponding author through 1822991734@qq.com

Authors’ Contributions

Conception and design were performed by Rui Wang and Yanli Cheng. Data analysis and interpretation were performed by Ruimin Wang. Manuscript writing was performed by Rui Wang and revised by Yanli Cheng. Final approval of manuscript was performed by all authors who read and approved the final manuscript.

Grant Support

Rui Wang is founded by China Scholarship Council (202206920039). This research was supported by funds from Natural Science Foundation of Suqian Science and Technology Bureau (K201903, Z2018076, Z2018213 and Z2022065). Jiangsu Association for Science and Technology (JSTJ-2022-004).

Ethics Approval and Consent to Participate

The patient reported in this study was informed for the purpose and process of this study and had written informed consent according to the guidelines of the hospital’s human associated research.

Patient Consent for Publication

Not applicable.

Competing Interests

The authors indicated no potential conflicts of interest.

Declaration of Interest

The authors declare that there is no competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Financial Support

Rui Wang is founded by China Scholarship Council (202206920039). This research was supported by funds from Natural Science Foundation of Suqian Science and Technology Bureau (K201903, Z2018076, Z2018213 and Z2022065). Jiangsu Association for Science and Technology (JSTJ-2022-004).

Acknowledgements

The authors would like to thank Dr. Xiaohong Shi for comments and discussion on the manuscript. We also would like to thank Dr. Quanquan Guo for data analysis.

References

  1. Rawla P, A Barsouk (2019) Epidemiology of gastric cancer: global trends, risk factors and prevention. Prz Gastroenterol 14(1): 26-38. [crossref]
  2. Gu E (2020) . SCDb: an integrated database of stomach cancer. BMC Cancer 20(1): 490. [crossref]
  3. Rawla P, A Barsouk (2019) Epidemiology of gastric cancer: global trends, risk factors and prevention. Prz Gastroenterol 14(1): 26-38. [crossref]
  4. Stahl P (2015) Heterogeneity of amplification of HER2, EGFR, CCND1 and MYC in gastric cancer. BMC Gastroenterol 15(1): 70-78. [crossref]
  5. Fornerod M (1997) CRM1 is an export receptor for leucine-rich nuclear export signals. Cell 90(6): 1051-60. [crossref]
  6. Fukuda M (1997) CRM1 is responsible for intracellular transport mediated by the nuclear export signal. Nature 390(6657): 308-11. [crossref]
  7. Ossareh-Nazari B F. Bachelerie, C. Dargemont (1997) Evidence for a role of CRM1 in signal-mediated nuclear protein export. Science 278(5335): 141-4. [crossref]
  8. Fu SC (2013) ValidNESs: a database of validated leucine-rich nuclear export signals. Nucleic Acids Res 41(Database issue): D338-43. [crossref]
  9. Okamura M H. Inose and S. Masuda (2015) RNA Export through the NPC in Eukaryotes. Genes (Basel) 6(1): 124-49. [crossref]
  10. Monecke T A. Dickmanns, R. Ficner (2014) Allosteric control of the exportin CRM1 unraveled by crystal structure analysis. FEBS J 281(18): 4179-94. [crossref]
  11. Cautain B (2015) Components and regulation of nuclear transport processes. FEBS J 282(3): 445-62. [crossref]
  12. Wang A Y, H Liu (2019) The past, present, and future of CRM1/XPO1 inhibitors. Stem Cell Investig 6(3): 69-78. [crossref]
  13. Schmidt J (2013) Genome-wide studies in multiple myeloma identify XPO1/CRM1 as a critical target validated using the selective nuclear export inhibitor KPT-276. Leukemia 27(12): 2357-65. [crossref]
  14. Angus L, PJ. van der Watt, VD Leaner (2014) Inhibition of the nuclear transporter, Kpnbeta1, results in prolonged mitotic arrest and activation of the intrinsic apoptotic pathway in cervical cancer cells. Carcinogenesis 35(5): 1121-31. [crossref]
  15. Kuusisto HV, DA Jans (2015) Hyper-dependence of breast cancer cell types on the nuclear transporter Importin beta1. Biochim Biophys Acta 1853(8): 1870-8. [crossref]
  16. Schlemper RJ ( 2000) The Vienna classification of gastrointestinal epithelial neoplasia. Gut 47(2): 251-5. [crossref]
  17. Hermanek P (1987) [TNM classification of malignant tumors: the new 1987 edition]. Rontgenblatter. 40(6): 200. [crossref]
  18. Gong W (2005) Expression of activated signal transducer and activator of transcription 3 predicts expression of vascular endothelial growth factor in and angiogenic phenotype of human gastric cancer. Clin Cancer Res 11(4): 1386-93. [crossref]
  19. Barchi LC (2016) MINIMALLY INVASIVE SURGERY FOR GASTRIC CANCER: TIME TO CHANGE THE PARADIGM. Arq Bras Cir Dig 29(2): 117-20. [crossref]
  20. Kinoshita J (2021) Current status of conversion surgery for stage IV gastric cancer. Surg Today 11(4): 1386-93. [crossref]
  21. Digklia A, AD Wagner (2016) Advanced gastric cancer: Current treatment landscape and future perspectives. World J Gastroenterol 22(8): 2403-14. [crossref]
  22. Gravina GL (2014) Nucleo-cytoplasmic transport as a therapeutic target of cancer. J Hematol Oncol (7): 85-96. [crossref]
  23. Azizian NG, Y Li (2020) XPO1-dependent nuclear export as a target for cancer therapy. J Hematol Oncol 13(1): 61. [crossref]
  24. Schmidt HB, D Gorlich (2016) Transport Selectivity of Nuclear Pores, Phase Separation, and Membraneless Organelles. Trends Biochem Sci 41(1): 46-61. [crossref]
  25. Ullman KS, MA Powers, DJ Forbes (1997) Nuclear export receptors: from importin to exportin. Cell 90(6): 967-70.
  26. Jamali T (2011) Nuclear pore complex: biochemistry and biophysics of nucleocytoplasmic transport in health and disease. Int Rev Cell Mol Biol 287(3): 233-86. [crossref]
  27. Sun Q et al. (2016) Inhibiting cancer cell hallmark features through nuclear export inhibition. Signal Transduct Target Ther (1): 160-175. [crossref]
  28. Das A et al. (2015) Selective inhibitors of nuclear export (SINE) in hematological malignancies. Exp Hematol Oncol (4): 79-86. [crossref]
  29. Stade K et al. (1997) Exportin 1 (Crm1p) is an essential nuclear export factor. Cell 90(6): 1041-50. [crossref]
  30. Senapedis WT, E Baloglu, Y Landesman (2014) Clinical translation of nuclear export inhibitors in cancer. Semin Cancer Biol (27): 74-86. [crossref]
  31. Xu D, NV Grishin, YM.Chook (2012) NESdb: a database of NES-containing CRM1 cargoes. Mol Biol Cell 23(18): 3673-6. [crossref]
  32. Senapedis WT, E Baloglu, Y Landesman (2014) Clinical translation of nuclear export inhibitors in cancer. Semin Cancer Biol (27): 74-86. [crossref]
  33. Zhang S (2012) ROR1 is expressed in human breast cancer and associated with enhanced tumor-cell growth. PLoS One 7(3): e31127. [crossref]
  34. Zhang S (2012) The onco-embryonic antigen ROR1 is expressed by a variety of human cancers. Am J Pathol 181(6): 1903-10. [crossref]
  35. Sun Q et al. (2016) Inhibiting cancer cell hallmark features through nuclear export inhibition. Signal Transduct Target Ther (1): 160-179. [crossref]
  36. [36]. Cheng Y et al. (2014) XPO1 (CRM1) inhibition represses STAT3 activation to drive a survivin-dependent oncogenic switch in triple-negative breast cancer. Mol Cancer Ther 13(3): 675-86. [crossref]
  37. Turner JG (2014) Inhibition of CRM1-dependent nuclear export sensitizes malignant cells to cytotoxic and targeted agents. Semin Cancer Biol (27): 62-73. [crossref]
  38. Huang WY (2009) Prognostic value of CRM1 in pancreas cancer. Clin Invest Med 32(6): E315. [crossref]
  39. Kim J (2016) XPO1-dependent nuclear export is a druggable vulnerability in KRAS-mutant lung cancer. Nature 538(7623): 114-117. [crossref]

Exploring Thought Patterns in Those with a History of Adverse Childhood Events and Trauma: An AIEmpowered Mind Genomics Cartography

DOI: 10.31038/ASMHS.2024812

Abstract

Respondents evaluated vignettes comprising a combination of simple phrases, designed to describe ordinary, common experiences that may be upsetting for individuals who have undergone adverse childhood events (ACE) and childhood trauma. The vignettes were systematic combinations of 2-4 stand-alone answers to four questions, each question generating four answers (aka messages, elements). The four questions generated phrases describing reactions to different types of childhood trauma, namely history of childhood sexual abuse, exposure to a caretaker or other adult with substance abuse, living in a lower socioeconomic status, and exposure to crime and gun violence, respectively. Each respondent rated a unique set of 24 vignettes constructed according to an underlying experimental design, with the 24 vignettes comprising an experimental design ready for regression modeling. The respondents rated each vignette on a five-point scale, assessing their immediate emotional reaction if they were to experience the events in the vignette and the likelihood that experiencing the events would evoke negative memories from the past. In addition to the 24 vignettes, respondents answered 8 yes/no questions regarding experience with Adverse Childhood Events (ACE). Regression analysis linked the elements in the vignettes to ratings. Three mind-sets emerged, defined by the pattern of coefficients: a strong response to emergency precautions, safety barriers, and depictions of substance abuse; a strong response to financial hardship or perceived hardship; and a strong response to startling sensory input and sexual content, respectively, The regression coefficients showed varied by individuals who answer “yes” to different ACE questions.. By employing AI within the framework of Mind Genomics, this study reveals relations between the what a person suffers as a child and how the person relates to emotionally-sensitive messages, evaluated years later. The Mind Genomics approach coupled with AI creates a new tool to understand the nuances of emotional responses associated with distinct types of traumas.

How Mind Genomics works

Mind Genomics works by presenting respondents with vignettes. A vignette is a combination of messages (elements, answers to questions), with the elements being ‘stand-alone’ phrases. An underlying experimental design prescribes the elements appearing in each vignette. The design used here is a so-called ‘permuted design.’ Each respondent evaluates the same design of 24 vignettes, but the actual combinations differ from respondent to respondent. Mind Genomics provides a way to obtain information from people without the people ‘gaming the system,’ and indeed without any prior knowledge being necessary.

There are two features of Mind Genomics which make it valid and reliable:

  1. Mind Genomics works at the granular, concrete level, the everyday, to generate both the test elements: The elements paint ‘word pictures.’ The test stimuli become short word pictures, albeit not connected but rather seemingly put together at random. The combination is more realistic, more similar to the world of the everyday, with the different aspects of the daily word present together in no clear pattern. Yet people are able to navigate through the ordinary world. In the same fashion, the vignettes comprise these stand-alone phrases. After a moment of shock, the respondent usually relaxes, and simply ‘grazes; through the information in what seems to be a relaxed fashion. Thus, Mind Genomics measures responses to ‘real information, viz., particulars such as descriptions of events as opposed to general ideas. Starting the evaluations from a granular experience differs from conventional scales of today in that it does not only restrict the responses to an intellectualized version of a response. Most researchers find themselves restricted to working with and considering only generalized ideas rather than specifics, but with Mind Genomics, the survey respondent is the one who more easily “abstracts” the general ideas as they are based on their answers to the particulars. by this method.
  2. Mind Genomics makes it impossible for people to ‘game’ the system. People do not necessarily tell the truth, either by deliberately lying, or unconsciously misrepresenting themselves to be more agreeable or more correct as they interact with the interviewer. It remains the nature of today’s science to rely on scales that can be easily ‘gamed [1]. Due to human nature and the inherent drive to be positively perceived, the validity of a scale cannot be easily assured using conventional methods.

The history of Mind Genomics traces back to three sources, the combination of which ended up synthesizing this new discipline. The history traces back to experimental psychology, particularly in the realm of psychophysics; statistical specifically in experimental design; and consumer research, with a focus on discerning and delineating patterns in decision-making processes. The ‘experiment’ in Mind Genomics presents respondents with vignettes, combinations of succinct, stand-alone phrases which create a scenario. The respondent reads the vignettes, one at a time, and then rates the vignettes. Afterwards, the underlying experimental design enables the deconstruction of the responses into the contribution of the individual elements. In order to make the analysis simple and the results straightforward for a manager to understand, the Mind Genomics program transforms the ratings into new binary variables, taking on the value of 0 or 100. The transformations are explained below. Mind Genomics endeavors to uncover interpretable patterns within the construct, such as trauma, doing so with simple stimuli, but ‘cognitively meaningful’ ones. In the Mind Genomics study, respondents ultimately generate cohesive patterns of data, as illustrated in the results below. It is crucial to underscore that throughout the study, respondents are intentionally guided to abstain from delving into the messages within the various combinations in an attentive way, pondering the message. Instead, the subconscious assumes control, steering respondents towards responses that may initially seem arbitrary. However, it is important to recognize that these responses are far from random.

Introduction to the Study

The science behind Mind Genomics allows researchers to explore the human experience in the context of a wide array of topics. Among these factors, childhood trauma stands out as a potent force shaping an individual’s perception and interaction with the world [2]. Childhood trauma, defined as adverse experiences occurring before the age of 18, encompasses a range of events such as abuse, neglect, and household dysfunction. Its profound impact on psychological development and overall well-being has been extensively documented in psychological and medical literature [3,4]. One of the most significant consequences of childhood trauma is the development of complex post-traumatic stress disorder (cPTSD) and hypervigilance [5]. These conditions manifest as heightened sensitivity to potential threats, pervasive feelings of fear and anxiety, and difficulty regulating emotions. Individuals who have experienced childhood trauma often navigate daily experiences through a lens colored by these psychological scars, which significantly alters their perceptions and reactions compared to neurotypical individuals. Whereas the neurotypical individual may take for granted the ability to engage with daily experiences without undue distress, those who have experienced childhood trauma face unique challenges. Simple tasks such as witnessing conflict, engaging in social activities, and managing emotions may turn into formidable obstacles. The omnipresent threat of triggering memories or emotions associated with past trauma can cast a shadow over even the most mundane activities, leading to avoidance behaviors and social isolation. The power behind the use of AI in the context of Mind Genomics allows researchers to expand this existing knowledge about psychology and medicine.. By applying Mind Genomics techniques to the study of reactions to daily experiences among individuals who experienced childhood trauma, researchers can identify differences in cognitive processing and information integration. This approach allows for a more comprehensive understanding of how past trauma influences the interpretation and response to everyday stimuli.

Setting up the Mind Genomics study on trauma

The raw material for vignettes – questions and answers: The first step registers the study, gives it a name, and proceeds to the heart of the approach, namely creating four questions, and then for each question creating four answers. The answers themselves must be simple, stand-alone phrases which paint a word picture. These answers, called ‘elements’ will end up being combined with each other in vignettes. Figure 1, Panel A shows the schematic request for the four questions. It is in this first encounter with the requirement to think of four questions than many researchers encounter difficulties. While discussing the etiology of trauma is manageable, crystallizing these etiologies proves challenging when the task is to create standard questions with answers which reflect mundane life experiences triggering trauma. Producing the questions separates the diagnosticians/therapists from statistically oriented researchers. Diagnosticians and therapists aim for a comprehensive understanding of the patient, whereas statisticians prefer straightforward numerical data for tallying questions. Consequently, the task of creating a list of questions that tells a coherent story becomes arduous. The answer to the dilemma is AI, in the form of Idea Coach, which embodies SCAS (Socrates as a Service). SCAS was created to make the process less arduous, and in some cases make the process a learning experience. Figure 1, Panel B introduces the Idea Coach, embodying SCAS. In this stage, the user articulates the issue by writing the ‘squib,’ a colloquial term for the text typed into the box. The squib may undergo multiple edits to refine the type of question desired. Panel C showcases some of the AI’s output. Finally, Panel D displays the four questions selected after user editing, preparing them for answers. Users can also rerun the Idea Coach, doing so in an iterative fashion, editing the squib when desired to see what emerges. Within the program (www.bimileap) the SCAS system embedded in Idea Coach requires about 10-15 seconds per iteration. Even with squib editing, users can generate 10 pages of 15 questions in about 8-10 minutes. The resulting ‘question book,’ combined with AI summarization, proves to be a valuable resource, as further illustrated below.

FIG 1

Figure 1: The first part of the Mind Genomics study, showing the four panels to help the user select questions.

Figure 2 the second phrase, the selection of answers to each of the four questions. Panel A shows the request for four answers. The user can simply press the Idea Coach button. The ‘squib’ is the question selected by the user. Panel B shows 8 of the 15 answers emerging with 10-15 seconds. Once again the user can iterate to educate themself by looking at the different answers to the question, or the user can actually edit a particular question.

 

FIG 2

Figure 2: The Mind Genomics Template showcasing the question and the AI-generated possible answers to the question that the user can select. Panel A shows the template for the four answers for Question #2. The question comprises the full set of modifications to the prompts, to create answers which are in the proper form. Panel B shows eight out of 15 answers for this iteration.

In summary, the Mind Genomics study entails a structured process to generate questions and answers, incorporating AI assistance (SCAS) and iterative refinement. The outcome is for a comprehensive exploration of complex topics, such as the effects of trauma on daily responses to various stimuli. This phase often becomes a critical juncture where researchers may encounter challenges, finding it easier to discuss cases of trauma but challenging to crystallize the discussion into standard questions. Those focused on diagnostics and therapy seek a closer, deeper understanding of the patient, whereas those concentrating on statistical analysis prefer simple numbers to tally on questions. Both will end up being satisfied with the combination of AI, creative thinking, and quantitative analysis using qualitative inputs.The final set of questions and answers appear in Table 1. These crafted inquiries and responses embody the collaborative effort to mold the content and structure of the answers. The overarching goal is to create a collection of meaningful standalone phrases (elements) which paint word pictures about the topic, and which can be combined together in small vignettes comprising 2-4 elements.

Table 1: The final set of questions and answers emerging from the collaboration of the user and SCAS (the AI embedded in the Idea Coach).

TAB 1

The self-profiling classification question: The setup process advances by formulating classification questions. The classification questions are simple questions which provide additional information about who the respondent IS, what the respondent has EXPERIENCED, etc. In this study, eight questions based on the original ten Adverse Childhood Events questions were asked as part of the study introduction to each respondent. Adverse Childhood Experiences (ACEs) and the ACE survey are integral components of research aimed at understanding the long-term impacts of childhood trauma on health and well-being. The concept of ACEs originated from a groundbreaking study conducted by the Centers for Disease Control and Prevention (CDC) in collaboration with Kaiser Permanente’s Health Appraisal Clinic in San Diego, California, in 1997.The ACE study surveyed over 17,000 adult respondents, collecting information about their childhood experiences of abuse, neglect, and household dysfunction. The study identified ten specific types of adverse experiences [3]:

  1. Physical abuse
  2. Emotional abuse
  3. Sexual abuse
  4. Physical neglect
  5. Emotional neglect
  6. Household substance abuse
  7. Household mental illness
  8. Parental separation or divorce
  9. Domestic violence
  10. Incarcerated household member

Respondents were asked to indicate whether they experienced any of these events during their childhood and adolescence. The study revealed a significant association between ACEs and a myriad of negative outcomes across the lifespan, including physical and mental health issues, substance abuse, risky behaviors, and socioeconomic challenges. For the purposes of Mind Genomics, the original ten questions were reduced to eight (Table 2) questions as seen in Table 2. Figure 3 shows the templated format in the Mind Genomics set-up for the ACE question.

Table 2: The eight ACE questions presented to respondents in the self-profiling classification part of the Mind Genomics interview, BEFORE the evaluation of the 24 test vignettes.

TAB 2

 

FIG 3

Figure 3: How the ACE question is inserted into the Mind Genomics study at the time of set-up.

The final phase of the setup process involves crafting the introduction to the study and then establishing the scale to be utilized. Table 3 shows the introduction and the rating scale in table form. Figure 4, Panel A displays the introduction presented to the respondent. Typically, brief and direct, this introduction focuses the respondent’s attention on the task. Recognizing the increased shortening of attention spans, the introduction is written in an abbreviated way. In some cases, however, such as legal cases involving childhood trauma, the introduction to a Mind Genomics might warrant a more detailed introduction to ensure that respondents are informed of the case’s background. Panel B illustrates the five-point rating scale, customizable by the user to align with the project objectives. In this study, the rating scale captures two dimensions. The first dimension of the scale assesses the immediate feeling of being upset if one or more of the events in the vignette were to occur. The second dimension of the scale gauges the perceived ability of the vignette to evoke a distressing memory.

Table 3: The introduction to the respondent and the text of the five answers to the question

TAB 3

 

FIG 4

Figure 4: Respondent orientation. Panel A shows the set-up screen for the orientation that the respondent reads. Panel B shows the set-up screen for the rating scale that the respondent will use to rate each vignette.

The Respondent Experience

The respondent experience begins with the completion of a questionnaire which asks them demographic questions followed by the 8 questions about Adverse Childhood Event (Figure 5). The respondent’s experience was simplified by a pull-down questionnaire, with each tab had to be individually pulled down.

FIG 5

Figure 5: The pull-down menu for the self-profiling classification questionnaire

After the self-profiling classification, the Mind Genomics program initiates the presentation of vignettes to respondents. Figure 6 shows an example, the top presenting a concise introduction to the study at the top, middle showing the rating, and the bottom showing both the vignette presented as a set of lines (two to four elements shown in three successive lines), followed by the rating scale. This format enables respondents to ‘graze’ through the information, avoiding the fatigue that could result from reading 24 dense paragraphs, each consisting of two to four sentences with connectives.

FIG 6

Figure 6: Example of a vignette

The vignettes were developed with the following properties.

  1. Each respondent assessed a total of 24 vignettes, the 24 vignettes comprising a complete experimental design.
  2. The 24 vignettes for each respondent each comprised a minimum of two elements and a maximum of four.
  3. Each vignette contained at most one answer (referred to as ‘element’) from a specific category, ensuring that no vignette presented conflicting information of the same type.
  4. Every element appeared five times among the 24 vignettes and was absent 19 times,.
  5. Each question or category contributed to 20 out of the 24 vignettes.
  6. Each respondent evaluated unique vignettes, a distinctive feature of Mind Genomics studies. The creation of the different sets of vignettes means that the research covered a great deal of the design pace, freeing the user from having to select the ‘most promising’ elements ahead of the research. This approach, called the permutated design [6] allows the researcher to use the approach at any stage of the research.
  7. Every respondent assessed a precisely crafted set of 24 vignettes, with all 16 elements being statistically independent of each other, and capable of independent analysis, particularly OLS (ordinary least squares) regression analysis, also known as curve fitting. This approach enhances the statistical significance of each respondent’s contribution to the study.

Field Specifics and Data Preparation

As demonstrated in Figures 1-6, the system has transitioned into a do-it-yourself (DIY) framework. Within this DIY paradigm, user engagement extends to recruiting respondents through an online panel aggregator equipped with a built-in API. The Mind Genomics platform empowers users to define the target population based on criteria such as country and age, which is then incorporated into the API. Users incur a nominal recruitment fee, gaining access to a pool of online volunteers facilitated by the provider, Luc.id Inc., boasting access to hundreds of millions of volunteer panelists worldwide. In the context of this study, test respondents were invited based on their prior agreement to participate in studies, accruing points towards rewards for their participation. The user’s role involved specifying the details, completing the payment through a credit card, and triggering email invitations to the target respondents. This streamlined process demonstrated efficiency, with the study involving 101 respondents, requiring less than four hours for completion. From the perspective of the study respondent, the actual duration was approximately 3-4 minutes. Respondents initiated the study by clicking on the embedded link, progressing through a brief ‘hello’ page, a self-profiling classification (Figure 5), and an introduction to the study itself. Subsequently, respondents evaluated the set of 24 unique vignettes organized according to the aforementioned experimental design (Figure 6). Figure 7 shows the median response time by test order, indicating that once respondents grasped the task, the median response time dropped to less than two seconds. The ability to swiftly inspect a vignette and assign a rating meant that the effort to read the vignettes amounted to 2-3 minutes. As noted above, in this type of study respondents adopt a ‘grazing’ approach, akin to superficially inspecting their surroundings, aligning with the Mind Genomics objective to capture data from individuals engaged in casual observation and rating. This approach aims to derive patterns not from deliberate, conscious efforts but from the more typical automatic and almost instinctive responses. Psychologist Daniel Kahneman’s conceptualization of rapid evaluation as ‘System 1,’ distinct from the considered and slower ‘System 2,’ aligns with the behavioral outcomes shaped by Mind Genomics [7].

FIG 7

Figure 7: How the median response time to the vignette changes across the 24 test positions, from start (test order 1) to end (test order 24).

The respondent provided a rating for each vignette on a five-point scale, constructed to encompass two dimensions: the immediate negative feeling upon imaging one or more elements in the vignette were to happen to them and/or the sense of the vignette potentially evoking a memory of a negative feeling. To extract meaningful information from the scale for statistical analysis, it is necessary to create new ‘binary dependent variables’ through simple transformations. These transformations yield variables suitable for OLS (ordinary least-squares) regression, allowing the researcher to uncover the relation between the presence/absence of the element and the respondent’s ratings. Table 4 shows the transformations. After the transformations, a prophylactic step was taken, adding a vanishingly small random number (<10-5) to every newly created binary scale value.. This addition ensured a slight variability in each binary dependent variable, a requisite for the regression analysis to function effectively and prevent potential ‘crashes’ due to the lack of variability in the dependent variable.

Table 4: The transformations

TAB 4

 

Initial Analyses – Do the Averages of the Binary Variables Different Across Key Subgroups?

The initial analysis examines the average of each newly created binary variable, first for the total panel, and then for the key groups defined by the self-profiling classifications. Three new subgroups were created, based upon the speed of rating the vignette (quick, intermediate, slow). The Mind Genomics system produce a great deal of data, especially with the many newly created binary variables, such as R1x to R5x, etc. To make the analysis more tractable, we consider only two newly created binary variables, R5x (most negative response to the vignette) and R1x (least negative response to the vignette), along with response time. Table 5 presents the average ratings. Particularly noteworthy is the inverted U pattern for R5x concerning respondent age, with the highest R5x occurring in the age group of 35-44. Additionally, respondents who answered quickly exhibited a notably high value for R5x. This finding means quite simply that when the respondent finds the vignette stressful, the respondent is likely to react more quickly to the vignette.

Table 5: Average ratings of R5x, R1X and response time by key groups

TAB 5

Relating the Presence/Absence of the Elements to R5x (Strongest Negative Response to a Vignette)

The next level of analysis is to link together the assignment of the most strongly negative response (R5x) to the 16 elements appearing in the vignettes. OLS regression will uncover the pattern. The equation estimated by OLS regression uses all the relevant data from the respondents in the defined group. The equation is expressed as R5x = k1A1 + k2A2.. k16D4. These coefficients appear in Table 6 as 16 rows, with each column corresponding to a key subgroup. The coefficient shows the link between the presence of the element as the average value of R5x. We look for high linkages, viz., a coefficient of 15 or higher. Table 6 shows elements with strong linkages of coefficients 15 or higher. These are all shaded. Table 6 It also shows coefficients greater than 10-14, meaning relevant elements, but not strong drivers of R5x. Empty cells correspond to coefficients less than 10. With large numbers of coefficients to be displayed it is often more productive to eliminate low values for the coefficients, making it easier to discover the patterns, and see elements with strong linkages. Table 6 shows some strong elements, but the pattern is difficult to discern. Most groups show very few coefficients above 10 (viz., Total Panel, males, females), which can make those elements which show a coefficient of 15 or higher stand out more amongst the total elements. What is also interesting is how certain traumatic experiences had more significant elements. Individuals who reported economic disadvantages had five strong performing elements, whereas Childhood Sexual Abuse showed only two strong performing elements. Respondents who had a caretaker or exposure to mental illness or suicidality and subjects who had a caretaker or exposure to someone who was in jail or involved in criminal activity showed only one strongly performing element. Only one element,, witnessing a bar fight, turned out to be a strong performer for more than one group of traumatic events. This element was a strong performer for three groups, those reporting childhood sexual abuse, those reporting poverty, and those reporting having an incarcerated household member.

Table 6: How the elements ‘drive’ R5x by Total panel and by key subgroups. Note: F. Female, M. Male, Q1x. Sexual Abuse, Q2x. Emotional Neglect, Q3x. Poverty, Q4x. Emotional Abuse, Q5x. Physical Abuse, Q6x. Household Substance Abuse, Q7x. Household Mental Illness, Q8x. Incarcerated Household Member.

TAB 6

Mind-Sets Emerging from Clustering the Respondents based on R5x

A hallmark of Mind Genomics is the search for underlying groups of people based upon their ways of perceiving the presented elements. Rather than looking for general groups of people defined by how they behave or how they look at the world in general, Mind Genomics focuses on the ‘granular,’ where experience is real and immediate. The approach to clustering combines mathematics and interpretation. The mathematics uses the coefficients emerging from the OLS regression, those coefficients showing how each of the elements ‘drives’ the response. Recall that each binary dependent variable could be expressed as a linear combination of ‘weights,’ expressed by the equation: BDV (Binary Dependent Variable) = k1A1 + k2A2… k16D4. The mathematics creates an individual equation for each of the 101 respondents, relating the presence/absence of the elements to that individual’s ratings of R5x, the key dependent variable. The data returning from this first step comprises 16 columns, each column corresponding to one of the 16 elements, and 101 rows, each row corresponding to one of the 101 respondents. The numbers inside the cells are the coefficients. The second step clusters the 101 respondents, dividing the respondents first into two groups, and then into three groups, called clusters. The clusters are defined as comprising ‘similar patterns’ of coefficients, similarity in turn defined as high Pearson correlation between the 16 coefficients of pairs of respondents. The actual mechanics of computation do not consider the meaning of the elements. Rather, each cluster comprises individuals who show highly correlated patterns of coefficients. The centroids of the different clusters are quite different from each other [8]. The process of clustering ends up generating different groups of respondents, the patterns with a group or cluster being similar, and the average pattern of each cluster differing from the average pattern of the other clusters. Once each respondent is assigned to a cluster, the clusters become new subgroups, allowing the user to estimate the coefficients for each cluster. When the 101 respondents are clustered by the pattern of their coefficients, viz., by their responses to test elements, three radically different mind-sets emerge, each mind-set comprising elements with far higher coefficients. It is in the nature of clustering by mind-sets that the user isolates groups of people with clearly different and clashing responses to the same elements. The conflicting answers frequently obscure the underlying patterns as they negate each other, creating the misleading notion that there is nothing to examine. However, a clearer picture can emerge by reconciling these contradictions. Table 7 shows these three emergent mind-sets, and the substantially higher coefficients which emerge once the mutually canceling effects are diminished by separating the mind-sets from each other. We are left with several R5x coefficients of 20 and greater. AI was able to find a general pattern between each mind-set- MS1- “strong response to safety issues”, MS2- “strong responses to financial realities of the everyday”, and MS3- “strong response to everyday events that could be misinterpreted.”

Table 7: The three emergent mind-sets after clustering on the basis of R5x

TAB 7

By dissecting the specific elements within each mind-set, we gain insight into the commonalities among them and uncover potential narratives about the individuals they represent. For instance, in MS1, characterized by a “strong response to safety issues,” individuals exhibit a pronounced negative reaction to seven specific elements, as Table 7 shows. These elements collectively reflect situations involving safety drills, emergency scenarios, and the witnessing of potential crisis situations. Individuals within this mind-set possess a heightened sensitivity to safety and emergency preparedness. Without considering additional factors such as Adverse Childhood Experiences (ACE), a psychologist might infer that these individuals have likely encountered various forms of heightened emergency situations, perhaps including military veterans or individuals with backgrounds in high-stress environments. Specifically, those who endured childhood sexual abuse exhibited a strong negative reaction to mishandling a firearm. Whereas it may be challenging to directly infer the correlation between childhood sexual abuse and firearm mishandling, it underscores the complex psychological impact of trauma and its potential influence on perceptions of safety and threat. In contrast, individuals who grew up in poverty, often residing in high-crime areas in the USA, displayed heightened reactions to elements such as sirens, active shooter drills, high-security measures, and witnessing excessive alcohol and drug consumption. This connection is more readily understandable, as individuals from socioeconomically disadvantaged backgrounds may have been exposed to environments characterized by increased risk and danger, thus developing heightened sensitivities to these specific stimuli. These findings highlight the intricate interplay between childhood experiences, trauma, and individual reactions to various elements. By elucidating these connections, we gain valuable insights into the nuanced ways in which past experiences shape perceptions and responses, guiding more targeted and effective interventions for those impacted by trauma. Mind-Set 2 is dominated by individuals with obvious financial insecurities whether that be real- struggling to pay bills or make ends meet and being unable to afford necessary medical treatment- or imagined such as feeling pressured to keep up with fashion trends or brands. An intriguing observation is that none of the elements within Mind-Set 2 corresponded to any Adverse Childhood Experiences (ACE). This absence of ACE alignment in Mind-Set 2 aligns with the nature of financial insecurity, which may predominantly stem from present-day circumstances rather than from childhood traumas. Financial pressures, both real and perceived, can exert substantial influence on individuals’ daily lives and psychological well-being, often overshadowing the impact of past adverse experiences. Furthermore, the unique composition of elements within Mind-Set 2 underscores the diverse manifestations of financial insecurity, ranging from tangible economic hardships to more intangible pressures of societal expectations. This multifaceted nature highlights the complexity of individuals’ experiences and the various factors contributing to their mind-set and behavior. For example, experiencing childhood poverty did not correlate to a mind-set of financial insecurities as an adult, which may seem perplexing at first glance. However, a deeper analysis reveals that individuals who endure childhood poverty might develop coping mechanisms to navigate persistent financial pressures, thereby mitigating the fear of financial strain in adulthood. Alternatively, they could be driven by an increased motivation to break free from the cycle of poverty, thereby reducing the impact of financial insecurities on their mind-set as adults. Finally, Mind-Set 3, characterized by a “strong response to everyday events that may be misinterpreted,” shares some similarities with Mind-Set 1. The difference, however, comes from the focus on situations commonly encountered by everyone, albeit with an elevated response. These elements encompass everyday occurrences which might typically evoke minor reactions but can trigger heightened distress for individuals belonging to this mind-set. For instance, encountering sudden or expected loud noises, such as a car backfiring, may prompt annoyance or momentary surprise for many individuals. However, for someone with a history of childhood trauma and hypervigilance, this noise could induce a more profound and prolonged state of distress. Moreover, the only specific adverse childhood event associated with Mind-Set 3 is having a household member with mental illness or suicide. This association underscores the profound impact of familial dynamics and mental health struggles within the household on an individual’s psychological well-being. The absence of other ACEs in this mind-set suggests that the elevated responses to everyday events may be influenced primarily by experiences within the immediate family environment, rather than broader childhood traumas. However, this is a hypothesis that would benefit from further exploration given the nature of the elements strongly responded to by MS3, specifically crowded places, loud noises, and sexually explicit content, respectively. A final point to make is that only out of all of the elements that were significant within a specific ACE category as seen in Table 6, each element had an intersection with one of the three Mind-Sets in Table 7. The only except was witnessing a bar fight, which elicited a strong negative response for individuals with a history of poverty, and those who had a household member who involved in criminal activity or incarcerated. This unique response suggests that the impact of witnessing a bar fight may transcend the typical associations with specific ACE categories and instead be influenced by broader environmental factors, such as socioeconomic status and exposure to violence in the community. This insight underscores the complexity of trauma and its intersectionality with various life experiences. By recognizing the specific elements that trigger strong responses within each ACE category and their corresponding mind-sets, we gain a deeper understanding of the interconnected factors contributing to individuals’ psychological well-being.

Meta-analysis: Contribution of Gender, Adverse Childhood Events, and Elements to Responses

Table 6 shows how the elements were associated with the eight ACE questions. Table 7 shows which elements strongly drive negative responses among respondents in the three mind-sets. The final analysis reveals patterns emerging when we reduce the stringency of our criterion importance by reducing the cut-off level of a coefficient of +8 or higher. using a coefficient of 8 or higher to represent strong coefficient. In this analysis, ACE, elements, and the three mind-sets were analyzed in a meta-analysis. The three objectives of this meta-analysis were: 1) Uncover the specific ACE questions and elements associated with the three mind-sets, 2) Explore gender association with each mind-set, 3) Explore response time (RT) as a function of gender and ACE experience. Table 8 shows the grand models, incorporating all predictors, for Total and for mind-sets. The grand model was created by OLS regression, using the presence/absence of the 16 elements for each vignette, but incorporating two new sets of predictor variables. One group was gender. Since there were only two genders considered, male and female, respectively, there was only one predictor. This was ‘Female.’ A respondent could either be a female or not a female. Thus, in Table 8 there is no coefficient for males. The second group was the ACE experience. For each of the eight ACE questions, the respondent was coded ‘1’ when the respondent reported having that ACE experience, and coded ‘0’ when the respondent reported not having that ACE experience. The result was a new set of 25 predictors for OLS, comprising 16 elements, one gender, and eight ACE experiences. Table 8 shows the strong performing elements shaded, as well as long response times shaded. For the binary dependent variable, R5x, a great deal has to do with who the respondent is, as show by the gender and by the ACE variable, with little additional effect due to the element. In contrast, response time is affected more by mind-set and by specific message. In this meta-analysis, few elements emerge as drivers of negative feelings. In fact, only one element “hearing someone make an inappropriate sexual joke or comment” was associated with a specific mind-set, in this case Mind-Set 1. Mind-Set 1, characterized by heightened concerns about safety issues, is associated with emotional abuse, household mental illness, sexual abuse and household substance abuse. These experiences likely contribute to a heightened sensitivity to safety, and a tendency to perceive threats more acutely. In contrast, Mind-Set 2, marked by strong responses to real or perceived financial insecurities, exhibits correlations with household mental illness, sexual abuse, incarcerated household member and emotional neglect. Noteworthy, childhood poverty, which might intuitively be thought to align with concerns about financial stability in adulthood, does not fit into this mind-set. Additionally, Mind-Set 2 stands out as the only mind-set to display a gender bias toward female respondents, indicating potential gender-specific vulnerabilities related to financial insecurities. Within this mind-set, one of the elements was “feeling pressured to keep up with expensive fashion trends or brands,” which tends to weigh more heavily on individuals identifying as women. Mind-Set 3, characterized by strong responses to everyday events that could be misinterpreted, shows correlations with emotional neglect, poverty, and household substance abuse. This suggests that individuals in this mind-set may have experienced childhood environments where misinterpretation of everyday events was prevalent, potentially leading to heightened sensitivity to ambiguous stimuli in adulthood.The overlap of household mental illness, sexual abuse, emotional neglect, and household substance abuse across two out of the three identified mind-sets in this meta-analysis highlights the profound and interconnected nature of these adverse childhood experiences (ACEs) as shaping individuals’ psychological profiles and responses to their environment. Furthermore, it is noteworthy that physical abuse is the only ACE that does not align with any of the three identified mind-sets. These discrepancies underscore the complexity of trauma and its varied manifestations, indicating that some ACEs may elicit responses which transcend the thematic boundaries of the identified mind-sets. Overall, the patterns observed in the correlation between ACEs and mind-sets highlight the diverse pathways through which childhood traumas influence adult psychological profiles.

Table 8: Meta analysis, relating the rating of 5x (negative feelings) and RT (response times) to gender, to the eight ACE experiences, and to the presence/absence of the 16 elements.

TAB 8

Discussion and Conclusions

Traditionally, studying human cognition and emotion has been a complex and labor-intensive process, all too often relying on subjective assessments and limited sample sizes. The particular questions to ask, the elements, have always been a stumbling block to the researcher seeking to understand the ‘unwritten rules for appropriate stimuli’, relevant when working with patients in particular, with people in general. The AI contribution here through SCAS, focusing as it does on questions to ask and answers to use, provides a new tool to explore how people think. SCAS and its underlying AI empower the user the ability to iterate again and again in real time, understanding the topic more deeply by reading, thinking, and revising questions and answers, all in real time. The research ‘teaches’ as the user sets up the study, doing so in a way which engages because the material, ranging from the squib to the questions to the answers, is relevant, and the AI is hyper-focused. One should view AI as a tool that to enhance and to augment the capabilities of human researchers rather than something which replaces them.. In the context of Mind Genomics research, AI is a powerful ally, extending the reach and scope of human imagination in the creation of the test stimuli. As shown here with the study of trauma, it is the human researcher who can set up the study, and who can guide the analysis. Mind Genomics, with its focus on vignettes as experimental stimuli, offers a platform to delve into the complexities of trauma. These vignettes serve as windows into the subconscious, allowing individuals to explore thoughts and emotions, thinking and feeling. With AI’s assistance, the researcher can generate the elements, the raw material of the vignettes, ensuring that these elements resonate with diverse audiences to capture the essence of trauma across various contexts. The elements generated in this study by SCAS in the Idea Coach are witness to the ease with which Mind Genomics coupled with AI create some of the raw material needed to understand the mind. By gaining a deeper understanding of how trauma affects emotional processing, researchers can develop more effective interventions and treatments for individuals who have experienced childhood adversity. Overall, the integration of AI into the study of emotional responses to childhood trauma represents a convergence of cutting-edge technology and compassionate inquiry. AI-driven algorithms could potentially help identify personalized treatment plans tailored to an individual’s specific needs and circumstances, leading to better outcomes and improved quality of life. By harnessing the power of AI, researchers can unlock new insights into the complex interplay between personal experiences and psychological outcomes, ultimately paving the way for more effective interventions and support systems for those affected by trauma. Ultimately, the fusion of Mind Genomics, AI, vignettes, and statistics represents a paradigm shift in the study of trauma. It offers a holistic framework for unraveling the complexities of human experience, empowering us to explore, understand, and address trauma with unprecedented depth and precision. Through this interdisciplinary approach, we pave the way for innovative solutions and interventions which resonate with the intricacies of the human mind.

References

  1. McCambridge J, Witton J, Elbourne DR (2014) Systematic review of the Hawthorne effect: New concepts are needed to study research participation effects. Journal of Clinical Epidemiology 67: 267-2 [crossref]
  2. Schauer M, Elbert T. (2010) Dissociation following traumatic stress: Etiology and treatment. Zeitschrift für Psychologie/Journal of Psychology 218: 109127.
  3. Felitti VJ, Anda RF, Nordenberg D, Williamson DF, Spitz AM, Edwards V, Marks JS, et al. (1998) Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: The Adverse Childhood Experiences (ACE) Study. American Journal of Preventive Medicine 14: 245-258. [crossref]
  4. Austin A. (2018) Association of Adverse Childhood Experiences with Life Course Health and Development. North Carolina Medical Journal 79: 99-103. [crossref]
  5. Cloitre M, Courtois CA, Charuvastra A, Carapezza R, Stolbach BC, Green BL, et al. (2011) Treatment of complex PTSD: Results of the ISTSS expert clinician survey on best practices. Journal of Traumatic Stress 24: 615-627. [crossref]
  6. Gofman A, Moskowitz H [. Isomorphic permuted experimental designs and their application in conjoint analysis. Journal of Sensory Studies 25: 127-145.
  7. Kahneman D, (2011) Thinking, Fast and Slow. Macmillan.
  8. Aristidis Likas, Vlassis NA, Verbeek JJ (2003) The global k-means clustering algorithm. Pattern Recognition 36: 451-461.

Mind Genomics and Today’s Realpolitik: The Conflict About Teaching Young Students Ideas that Could Be Propaganda

DOI: 10.31038/JIAD.2024121

Abstract

AI embedded in the platform of Mind Genomics was used to synthesize mind-sets regarding the attitudes of parents towards what is being taught to their children. The AI program emerged with the three mind-sets being Traditionalists, Concerned Parents, and Progressives, respectively. The AI program clearly summarized the values of these three mind-sets when instructed to define what the mind-sets believed to be wrong with today’s education, to be ok with today’s education, and finally to be excellent with today’s education. The AI program, Idea Coach, was instructed to create 20 statements, and to predict how each of the three mind-sets would score each statement in terms of whether or not it bothered them, and whether or not this statement described what was happening in their schools. The answers made intuitive sense. The process shows the power of AI as an aid to critical thinking and understanding, as well as a novel way to deal with a complicated topic even before doing any reading or research.

Introduction

During the past several years a variety of news stories have appeared regarding the discovery that students in classrooms around the United States may be receiving propaganda in their daily lessons. One need only read stories about propaganda to realize that something may be going on, although we do not know. This issue has come to the fore recently with the spread of news about the influence of the Chinese Communist Part (CCP), such as this headline from the Oklahoma Council of Public Affair in August 2023 Tulsa Schools Linked to Chinese Communist Entity [1]. We are accustomed to hearing this in other countries and situations, such as the recently revealed but long-known fact that the ‘teachers’ in Gaza, supported by the United Nations (UNWRA) are teaching anti-Israel and anti-Semitic material, lesson material that is found in their textbooks [2]. The issue is here in the United States as well, as expressed in a November 22, 2023, story by Zachary Faria in the Washington Examiner: The Democratic Party’s panic over losing control of narratives in schools has led it to pursue a new path to propagandize to children: mandatory “media literacy” classes to teach students about “fake news.”. Media literacy lessons are now mandatory in California in English, science, history, and even math classes throughout every grade level, thanks to a law signed by Gov. Gavin Newsom (D-CA) last month. Delaware, Illinois, and New Jersey are also among the states requiring these lessons. Among other things, the California law worries about the effects of “online misinformation” that has “threatened public health.” [3]

The issue of teachers teaching what the parent’s believe the student should not be learning appears to be a systemic issue, perhaps one plaguing all of society. Increasingly, teacher values clash with parent’s values, the conflict played out in the arena of education. ‘School boards have a central position in educational governance. They have to guarantee quality, monitor results and intervene if needed’ [4]. The issue is explained even more elegantly by [5]’…while school administrators are challenged to turn schools around with limited time and resources quickly, their efforts are not a silver bullet. Engaging community requires committed partnerships that support schools to advance quality learning. Community school councils, an organizing strategy, focus on addressing potential threats and enhancing strengths for student success’.

In the United State, irate parents have verbally attacked the local school boards, often the attack making news. The complaining parents are called ‘terrorists’. The local education establishment gather around to defend the teacher and castigate the irate parents. . “New whistleblower information has revealed that the FBI targeted parents who spoke out against their local school boards’ COVID policies, after prodding from education officials. It all started in September, when the National School Boards Association (NSBA) sent a letter to the Department of Justice (DOJ) requesting federal intervention into the alleged “domestic terrorism” that is citizens disagreeing with school board officials” [6]

The conflict between parents and the educational system has created the opportunity for academic investigation on a broad front, ranging from who is teaching to whom, and most important, what is taught, with what emphasis, and with what objective [7-9].

The opportunity to explore these issues using AI embedded in the Mind Genomics platform, www.bimileap, allows us to look at the topic in a new way, and with some new tools. The tool, Idea Coach, allows us to specify a problem, and after specifying the problem, put AI to work to generate the data.

The approach presented here originates in the effort by the authors to change the way we discover how people think. The traditional methods began with qualitative discussions, in which people could surface their concerns, and in which a trained ‘listener’ or ‘moderator’ could elicit information from people leading to insights. There is an entire discipline of qualitative research, popular, growing, requiring training to understand what is really being communicated in an interview or in a group discussion. It is from these qualitative interviews with parents as well as observing what is being reported in the media that the importance of the curriculum being used with the students has emerged.

Beyond qualitative work is the effort by researchers to measure the minds of peoples, or if not the minds, then measure the attitudes of people. These measurements are done by surveys, in which the researcher creates a list of questions about topics, and instructs the respondent, the survey taker, to rate the different topics or questions on one or another scale. Most of the readers by now should be familiar with surveys which seem to follow every transaction of a business nature, with the survey attempting to quantify the different aspects of the experience. Typically the surveys are either done for general attitudes or for specific attitudes, but in either case the surveys fail to get to the granularity of the experience. Nonetheless, the researcher executing the survey ends up with a measure of performance or importance of an experience relevant to the group commissioning the survey

It was against this background of surveys and the failure to deal with the granularity of the data that the notion of presenting respondents with combinations of messages emerged. The research leading to this effort had been developed by mathematical psychologists under the name conjoint measurement [10], and popularized by Wharton professors [11]. The underlying idea was to present the respondent with combinations of messages about a situation, obtain their reaction to the messages and then deconstruct the messages into the contribution of the components.

The trajectory of science would lead to some developments in this effort to understand the mind of the person. The first effort was the creation of a DIY, do-it-yourself system, to run these experiments on topics, with the user providing questions which told a story, and for each question provide four answers, and then mix and match the answers (also called elements), to create a set of combinations (called vignettes). The respondent or survey taker would be presented with these combinations and instructed to rate the combination. Each respondent evaluated a full set of combinations, these combinations constructed by an underlying experimental design which prescribed the component elements of the vignette. The later analysis generated an estimate of the contribution of each element.

The requirement to develop the four questions which tell a story proved to be a stumbling block. It was to address this issue that the approach, now called Mind Genomics, first attempt to teach users, but then with increased effort changed direction. Rather than teaching students, it became easier to create an AI-powered system called Idea Coach which, upon receiving a background of the study, would come up with the offering of 15 questions, and then for each question 15 answers. The task facing the researcher was now to write a coherent ‘squib’ requesting the 15 questions. Once the questions were chosen it was easy to create 15 answers for each question. Figure 1 shows the process.

fig 1

Figure 1: The basic Mind Genomics set-up. Panel A shows the request for four questions which tell a story. Panel B shows the AI-powered Idea Coach. The user can instruct the embedded AI to provide information.

Introducing the Problem to AI

The remainder of this paper shows the process of creating an orientation to AI, instructing AI to provide the relevant information, and then generating the necessary materials from the internal processing of AI. The approach can be done on the current Mind Genomics platform (www.bimileap.com). The one caveat is that with the same orientation to AI the output from Idea Coach varies, even when run several times. The reason for the variation is not known. The AI does try to follow instructions. Occasionally the AI returns with a message saying that it cannot fulfill the task, providing one or another reason for not be able to do so. Yet, despite the lack of perfect reproducibility, it is the very richness and instructional value from the AI output which motivates this paper.

Table 1 shows the introductory ‘squib’ or description provided to the Idea Coach program embedded in the BimiLeap program. The structure of Table 1 is important to elucidate because with AI small deviations from the pattern can end up causing the AI to deliver the wrong material., or to be incomplete. The parts of the introductory squib play different roles.

Table 1: The input ‘squib’ to Idea Coach. The input describes generalities of the mind-sets, requesting that AI provide specifics for each mind-set.

tab 1

Topic (Sentences 1-3)

In the ‘topic’ the user must provide the AI with some sort of background. The AI output from Idea Coach is sensitive to the phrases. In order to allow a systematic exploration of the impact of the set-up on the results, the set-up material is structured into three short sentences, each sentence spatially separated from the others by a blank line. In this way it becomes possible to change the orientation quickly, either help the AI provide the necessary information, or even to help explore what happens when the orientation is changed by adding something about the year in history for which the information is desired, or the country and year for which the information is desired

Posited Mind-Sets (Sentence 4)

The orientation states that there are three mind-sets, but does not define what a mind-set is, nor give any information about the mind-sets, other than there are three. In other uses of Mind Genomics and specifically Idea Coach, the authors have occasionally defined the mind-sets, whether these definitions be very tight and specific, or whether these definitions be ‘broad stroke.’ As will be shown in the next few paragraphs, simply defining that there are three mind-sets suffices for the AI to provide three radically different groups. From iteration to iteration there may be some changes in the nature of the three mind-sets, but each iteration makes sense.

Request for AI to Generate 20 Statements (Sentence 5)

These statements pertain to what the education process should do to ensure that the children are not subject to propaganda in the form of education. The instruction to AI is to provide a reasonably short phrase (15 words or less), that statement is realistic, and that each statement is a stand-alone sentence. AI has no problem following these directions.

Instruct AI to ‘Rate’ Each Statement on a Defined Four-point Scale (Sentences 6 and 7)

The AI is to assume that the statement was read by each of the three mind-sets, respectively. The scale is defined completely. The ingoing assumption is that the AI understands the meaning of the question, understands the meaning of the rating scale, and can assume the role of the mind-set.

Provide an Addition Eight Pieces of Information about Each Mind-set (Sentence 8)

Once presented with the request in the squib, it takes the Idea Coach approximately 10-15 seconds to provide answers. Each the user requests a new ‘run’ the Idea Coach begins anew. Occasionally Idea Coach cannot immediately answer the question immediately, returning with an apology. Later, however, after the effort has finished, Idea Coach will return with each of the iterations summarized. Those iterations that could not be addressed immediately and about which the Idea Coach apologized end up having been answered, however. The only problem seems to be the ability to provide the answers immediately and then move on to the next iteration.

We move on with the section of the squib requesting information about the three mind-sets. The information appears in Table 2. The columns show the three mind-sets, the rows show the answers to the eight questions. If were to trace back the mind-sets to the complaint about propaganda in education, it would be Mind-Set 2, the so-called ‘Concerned Parents’ who would be the ones most likely to fear the propaganda.

Table 2: Specifics for the three mind-sets created by AI

tab 2

It is important to recognize that the user provided no information to the Idea Coach other than the ‘operating hypothesis’ that there exist three different mind-sets in the population. Despite the paucity of direction given to the AI in Mind Genomics, Table 2 shows a quite reasonable division of points-of-view across the mind-sets. Table 2 also shows hypothesized demographic distributions which make intuitive sense.

Perhaps the strongest evidence for the usefulness of AI comes from ratings assigned to the mind-sets. Part of the request to Idea Coach was to suggest the likely rating to be assigned to each of 20 statements. The rating scale was two-sided, one side of the scale talking about ‘bothers me’, and the other side of the scale talking about ‘occurs in my school’. The mind-sets emerging from AI focus on topics of opinion. The AI did not request any guidance about the mind-sets, but rather simply presented them. The three mind-sets differ in their concerns and values, as Table 3 shows.

Table 3: AI-generated ratings for the 20 statements according to the three hypothesized mind-sets, as well as the deconstruction of the ratings into what statements ‘bother me’, and what statements describe ‘what occurs in my school’.

tab 3

The pattern of scaled responses appears to be more intuitively correct than might have been expected. The middle and right sets of columns show two letters, N corresponding to ‘NO’ for that rating category, Y corresponding to ‘YES’ for that rating category. The pattern of Y’s make sense. The ‘Traditionalists’ are not bothered by any of the statements. The ‘Concerned’ are bothered by the content of what is being taught. The ‘Progressives’ are concerned about fairness of what is being taught, and the ability of students to become ‘critical thinkers’.

Table 3 does not show dramatic ideological differences among the hypothesized mind-sets, but rather gives a sense of modest, nuanced differences. Furthermore, the pattern breaks down when we consider the second part of the scale dealing with ‘occurs in my school’. There seems to be no clear pattern here for any of the hypothesized mind-sets.

At the end of the iteration, once the user has either gone to the next iteration to obtain new ideas from AI in the Idea Coach routine or has proceeded to select questions and answers, the material created by Idea Coach is sent to a summarizer. The summarizer comprises a series of prompts which end up deconstructing the information and reconstructing the material into new perspectives. The first summarization comprises the analyses shown in Table 4. This summarization shows 15 new questions, 20 key ideas, and 20 themes. This pattern of summarizations is a legacy from the original summarizer, done when the focus of the Idea Coach was to present sets of 15 question to a simple squib, and set of 15 answers to a question. Despite being a legacy summarization, the three sets of statements provide additional topics for consideration, as well as different ways of stating the key issues.

Table 4: AI-generated summarization of the of the key issues, presented to the user in the Idea Book. The analysis is a legacy summarization used in simpler forms of the Idea Coach.

tab 4(1)

tab 4(2)

The Idea Coach further analyzes the material, presenting ideas, and for each idea three aspects. These aspects are called Plus (positive aspect), Minus (negative aspect or difficulty), and Interesting (long term benefit). The perspectives appear in Table 5.

Table 5: Perspective on the different ideas, showing short term benefits (Plus), short term problems (Minus), and long-term benefits (Interesting).

tab 5(1)

tab 5(2)

The next set of analyses provided in the Idea Book by the Summarizer deal with the different receptions that the ideas will receive. The first is the Alternative Viewpoints, or different ways of dealing with the topic. The second is the Interested Audiences, those who will accept the ideas. The third is the Opposing Audiences, those who will reject the ideas. Once again the AI embedded in Idea Coach provides a fairly thorough analysis of these viewpoints and responses to the material, an analysis which dramatically augments the understanding of the topic. Table 6 shows these groups of analyses.

Table 6: How the ideas are received, by three different groups suggested by Idea Coach

tab 6

The final set of analyses appears in Table 7. These analyses consider what is missing, and innovation. Once again, the analyses are done completely by AI, working only the information generated from the input squib given to Idea Coach. That information, in turn, comprised only the suggestion of three mind-sets, as well as some modest background of a few lines provided to Idea Coach at the start of the process.

Table 7: Suggestions for the future, based upon an analysis of ‘what is missing’, and suggested ‘innovations’.

tab 7

Discussion and Conclusions

This paper was generated in response to issues about possible propaganda in schools, a response to news articles and broadcasts from the media. The issue of propaganda is school emerged out of an interest in the actions of the Federal Government versus local school boards, where protesters were considered to be part of a rebellious criminal element. It is issues such as these, issues which inflame the emotions and which call into question the basic rights and liberties of people, which become interesting topics for Mind Genomics.

The original approach of Mind Genomics would have been to introduce the topic by a small squib, the writeup given to the Idea Coach, that writeup simply describing the situation as reported by the media and specifying questions to ask. The next step in the original Mind Genomics would have been to generate sets of 15 questions, select four questions which ‘told a story’, and for each question generate four answers. The respondent would then have been exposed to small vignettes; combinations of these messages would have rated the vignettes with each respondent rating a different set of 24 vignettes. The analysis of the ratings would be by accepted statistics (Ordinary Least Squares to relate elements to ratings; Clustering to define new to the world groups or mind-sets). The outcome of this straightforward approach, done in the space of an hour or two using human respondents would have generated the type of data shown in Table 2 (ratings of each statement by each group).

The approach presented here takes Mind Genomics into an entirely new direction, one fully directed by the artificial intelligence built into the Idea Coach. What emerges as most remarkable is the depth of information from a few lines of request. The AI builds upon itself, providing information at the basic level, viz., the statements, and then building on that information and nothing else to generate the wealth of information presented. What is even more interesting is that the paper deals with the results of one iteration taking about 15-30 seconds for immediate results, and about 30 minutes wait for the summarizer in the Idea Book. Not reported here are the results of 20 of the iterations, each done in 15-30 seconds, occasionally with a small change to the squib, for example to specify that the analysis is to reflect what would have happened say in 1900 vs 2000, or what would have happened had the user specified two mind-sets, or four or five or even many more mind-sets. That parametric investigation awaits the attention of a graduate student for their thesis work.

References

  1. Carter R (2023) Tulsa school linked to Chinese Communist entity. August 3, 2023 Oklahoma Council of Public Affairs. 2023.
  2. UN Watch (2023) UN Teachers Call To Murder Jews, Reveals New Report
  3. Faria Z (2023) The newest Democratic propaganda in schools: ‘Fake news’ classes. Washington Examiner, November 22. 2023.
  4. Honignh M, Honingh MR, van Thiel S (2020) Are school boards and educational quality related? Results of an international literature review, Educational Review 157-172.
  5. Medina MA, Grim J, Cosby G, Brodnax R (2020) The power of community school councils in urban schools. Peabody Journal of Education 95: 73-89.
  6. Institute for Free Speech (2020) FBI Targets Outspoken Parents, School Boards Silence Them.
  7. Rubin JS, Good RM, Fine M (2020) Parental action and neoliberal education reform: Crafting a research agenda. Journal of Urban Affairs 42: 492-510.
  8. Sampson C, Bertrand M (2022) “This is civil disobedience. I’ll continue.”: The racialization of school board meeting rules. Journal of Education Policy 37: 226-246.
  9. Shuffelton A (2020) What parents know: risk and responsibility in United States education policy and parents’ responses. Comparative Education 56: 365-378.
  10. Luce RD, Tukey JW (1964) Simultaneous conjoint measurement: A new type of fundamental measurement. Journal of Mathematical Psychology 1: 1-27.
  11. Green PE, Krieger AM, Wind Y (2001) Thirty years of conjoint analysis: Reflections and prospects. Interfaces 31(3_supplement), S56-S73.

Evolving the Doctor’s Waiting Room: Applying AI to Visioning the Future, a Cartographic Approach

DOI: 10.31038/IMROJ.2024911

Abstract

SCAS (Socrates as a Service), incorporating ChatGPT 3.5, was instructed to create ten visions for the future waiting room of the doctor, based upon positioning the waiting room as part of the continuum of healthcare, rather than just a room housing people before they are admitted to see the medical professionals. SCAS ‘fleshed out’ each of the 10 visions in paragraph form, and then generated 15 questions about aspects of implementing these visions. In the subsequent iteration, SCAS was given the 15 questions it had previously created, and instructed to provide answers to each question, as well as estimate the difficulty involved in achieving what the answer specified. At the end of the process, about 15 minutes later, the program returns with a detailed Microsoft Excel file, the Idea Book, showing each iteration, and providing additional insights generated by the AI-based ‘Summarizer.’ The process presented here shows the power of AI to help create the future by allowing easy-to-create, quick-to-run queries, and providing detailed answers and additional subsequent analysis. With the turn-around time in seconds, and with a system which is iterative, the user can explore a topic in depth, generating a strong educational experience for any topic that one can imagine. The strength of the approach is the ability to help one think and envision in a way which is absorbing, user-driven, and often filled with surprises.

Introduction

What the future will bring is always a topic of conversation. It becomes especially interesting when the focus is on something relevant to the daily lives of people, and when the evolution may improve the quality of life. This paper is an attempt to envision the doctor’s waiting room as part of a smooth transition from greeting the patient to treating the patient. The strategy uses AI to help envision the future, doing so in a way which requests specificity from AI. Furthermore, the objective is to integrate OpenAI’s freely available AI ChatGPT 3.5, a large language model (LLM), into an easy-to-use system with a limited number of queries, automatically built-in. All the work is done behind the scenes by ChatGPT 3.5 [1]. The objective is to create a system which lets the user focus on the work, not on the effort to code. As a result, the user spends time learning through iterations, the immediate feedback allowing a quick change of query, with the iterations eventually generating the information desired. The notion here is that AI becomes the collaborator in the up-front development, and in some cases a participant who estimates the difficulty of the task [2,3].

The approach used in this paper is called SCAS, an abbreviation for Socrates as a Service. The name comes from the belief that the approach presented here is a coach to creative thinking. Rather than Socrates ‘pulling out’ ideas from the mind of slaves through questioning, the idea is for people to pull out ideas from the mind of AI through questioning. The questions are not meant to be factual, but conceptual. The user is challenged to let Socrates give expansive answers, not constrained ones, while at the same time ensuring that these expansive answers are feasible and produce something tangible. [4] called this approach ‘Homo Silicus.’

Systematizing New Idea Creation

Creating new ideas is a well-accepted business objective. Business always needs to create ‘new’ for the simple reasons that business must keep up with changes, whether the changes be responses to truly new situations and their requirements, or activities to maintain the interest of fickle customers. The world changes, people get tired of what they have, the two of these factors combining to force the creation of the new as a standard part of one’s operations, one’s business.

With all the focus on ‘new,’ it should come as no surprise that there are myriad publications with their pronouncements of how to create the new, hundreds of thousands of experts who are ready to advise about the ‘new’ at a moment’s notice and for an agreed-upon fee, not to mention the multitudinous best practices which dot the horizon of business.

The introduction of AI has changed the world of new product development. With user-friendly technology such as that embedded in Open AI’s ChatGPT 3.5, it has become possible to provide AI with a specification, a request, and let AI return with the suggestions. It is from experience with AI in this format that the study reported here arose. The paper focuses on the application of AI to envision the patient’s waiting room, an integral part of the medical facility, as a room where the healing process begins.

The process presented here grew out of the effort to understand how people respond to the world. The typical process of understanding the mind of a person with respect to something like the patients’ waiting room is to ask questions about the waiting room, to identify the relevant features, and their importance. The typical approach is to use some type of scale and instruct the respondent to rate the importance of different features that the waiting room might feature. It is from this type of scaling that the designer gets a sense of what is important. It is up to the designer to then take the information and create the prototype rooms to be evaluated for acceptance, utility, and so forth. Sometimes the survey is put aside or occurs after a discussion with participants about the new patient’s waiting room. In this case a trained interviewer discusses the nature of the waiting room, either with one or two people in a so-called ‘depth interview’ or convenes a group of people called a ‘focus group’. In both cases the researcher ends up conducting a structured interview in the form of a conversation. The foregoing process is a standard part of most research approaches to understand what is needed in a new product.

One emerging problem embedded deeply in the foregoing conventional process is the failure to deal with the details. Surveys, discussions, and so forth can provide direction, but it is the execution of these ideas which is important. In research this is known as the gap between ‘strategy,’ the big idea, and ‘execution,’ making the particulars a reality and bringing that into the world. The emerging science of Mind Genomics was developed to incorporate the specifics, the executions, into the research process [5]. Rather than instructing the respondent to rate general ideas, hoping that these ideas would end being well executed, the Mind Genomics approach took its cue from nature. The approach was to create small vignettes, combinations of concrete ideas, specific descriptions, instruct respondents to rate these vignettes, and then through statistics (regression, clustering) deconstruct the response to the part-worth contribution of the individual concrete elements, the granular information shown to the respondent . The outcome was the recognition that valuable data could be obtained by focusing on the specifics of a situation, rather than focusing on large, featureless, general ideas. A good strategy was to generalize from the pattern of particulars, easier to do than particularizing from a featureless generality. It was this realization which pushed the focus of Mind Genomics to focus on the concrete, the granular.

It was the pattern of questions and answers, begun with Mind Genomics, which was to evolve to the approach presented here. In its current form, Mind Genomics requests the user to create four questions regarding a topic (see Figure 1, Panel A), and then for each of the four questions create four answers (see Figure 1, Panel B). The process thus creates 16 elements, the four sets, each comprising four answers. It was the ensuing difficulty with having a user, develop four questions which eventuated in introducing AI was used as a coach, called, not surprisingly, the Idea Coach. The user would type background to a topic in the form of an AI prompt, which Mind Genomics calls a ‘squib’ (Figure 1, Panel C). In turn, the Idea Coach would then use AI to develop the four questions. (Figure 1, Panel D). The same approach could be used to create four answers to each question. The process of using AI a system to generate questions for a specific topic, or answers for a specific question soon moved AI to a central role in Mind Genomics.

fig 1

Figure 1: The original process in the set-up of a Mind Genomics study. Panel A shows the request for four questions which ‘tell a story’. Panel B shows one question, with the request to provide four answers. Panel C shows the background to the project and the desired results, the so-called ‘squib’, typed into the Idea Coach. Panel D shows 5 of the 15 questions provided by SCAS operating within Idea Coach.

It soon became apparent that the embedded AI in Mind Genomics was being used far more to create ‘books of knowledge’. The user would create a set of questions to ‘brief’ the AI, and then run the AI several times, occasionally just iterating, but often changing the input slightly and then iterating. The emerging focus was using the embedded AI to provide a range of alternative answers to the same question. Furthermore, the AI had been programmed to provide additional summarization, viz., post-emergent analysis. The system ended up being renamed SCAS, Socrates as a Service, in recognition of this evolution from a system which recommended questions and answers into a system to educate people on a topic, using AI to help people think critically about a topic, imagine alternative realities, and create by typing descriptions of those realities.

Using AI to Drive the Future

The immediate stimulus driving this set of experiments resulted from a fortunate accident. than simply asking a question and getting the SCAS to provide 15 answers, something that was done a dozen times in the set up of a Mind Genomics study, the user this time created a more detailed briefing, by writing down both the relevant input for Idea Coach, but then additional, background information that was usually part of the thinking, but never submitted to the AI in Idea Coach. The AI was ‘briefed’ in a far more detailed way, with thoughts, requests, hypotheses, and so forth, again by accident, but an accident which proved remarkably productive as shown below.

Table 1 show the entire briefing. Table 1 is divided into two parts, shaded and unshaded, respectively. The shaded parts provide modest details, specifically the nature of the room (part of the flow, first room in a sequence of room), followed by a statement of ‘fact’, viz., that there are ten mind-sets. There is no supporting evidence given for the statement about ten mind-sets, however. The user could just have easily specified three mind-sets or 13 mind-sets. The second part of Table 1 shows the general instructions to the AI. These are to focus on the room itself, and afterwards on the reaction of people to this room. Again, Table 1 is far more extensive. We can contrast this extensive brief with the ordinary brief of that time, which would be something like: ‘What are questions to ask about the nature of the patient’s waiting room, if we think of the next decade, starting in 2030?’

Table 1: The briefing given to AI about the waiting room outside the doctor’s office

tab 1

The actual creation of the AI briefing takes a few minutes. Once the AI is provided with the briefing, the AI is invoked through SCAS, the Socrates as a Service procedure, embedded in the Mind Genomics program. Within 10-20 seconds the AI returns with preliminary answers along with an additional set of 15 questions for consideration..

Table 2 presents the 10 visions of the patient’ waiting room, attempting to follow the request to discuss the room itself as an integral part of the visit, and of the continuum from waiting room to examination room. Each of the 10 visions is labelled, followed by a description of the room itself, and then followed by the response of the patient in that room.

Table 2: The 10 visions for the patient’s waiting room as part of the continuum for health and healing. Each vision comprises a name, a description of the room, and then a description of the feelings of people towards that ‘now-evolved’ space.

tab 2

The important thing to keep in mind is that the AI was not given any information at all, other than this was to be a patient’s waiting room but was also going to be part of the spatial continuum, and thus an integral part of the ‘path’ for health and healing.

As noted above, each iteration was accompanied by 15 questions to answer. These questions are based on the topic. The questions emerge automatically. The next step was to copy the 15 questions into a new run of Idea Coach. Table 3 shows the 15 questions (underlined), two answers to each question provided by SCAS and for each question an estimate of the difficulty of achieving that answer (Easy vs Difficult).

Table 3: The 15 questions generated by SCAS when it provided information about the vision, along with two answers to each question, and the estimated difficulty of implementing the answer. The questions were ‘answered’ by a separate run of SCAS, where the questions were inserted as part of the briefing to AI.

tab 3

Summarizing and Expanding the Ideas; Themes and Perspectives

SCAS generates a great deal of information when it creates visions (Table 2) and generates questions (Table 3). One of the benefits of AI is the ability to summarize the information in a succinct way. Table 4 shows two summarizations. The top of Table 4 shows the key ideas. The bottom of Table 4 shows the four themes identified by SCAS, followed by the good (plus), bad (minus) and novel (interesting) aspects of each theme. It is in Table 4 that we begin to see a secondary analysis by AI of the information generated by AI.

Table 4: Key ideas, themes and perspectives

tab 4

Table 5 moves the analysis to who would be interested in these ideas, and who would be opposed to them. Once again we see a deeper secondary analysis of the information provided by AI.

Table 5: Interested versus opposing audience to the topic ideas

tab 5

Driving Towards Innovation

Table 6 shows the final analysis provided by AI, comprising three sets of analyses/questions which drive towards innovation. The first comprises a set of 10 questions about what is missing. This first set is meant to challenge the user. The second comprises a set of 16 alternative viewpoints, contrasting answers to a question. The third focuses on innovations, comprising a set of four general categories or topics, each with several suggestions.

Table 6: Steps toward innovation, comprising ‘what is missing,’ alternative viewpoints, and then categories or topics with several suggestions each.

tab 6(1)

tab 6(2)

tab 6(3)

Discussion and Conclusions

The focus of this paper is to present an approach to solving specific problems, once these problems are specified and provided to AI in a useful manner. The AI approach used here, SCAS, Socrates as a Service, was developed to help critical thinking, and to focus attention on potential answers. Rather than focusing the time on elaborating the question, SCAS builds on the richness of AI, especially in the world of design. The effort shown here revolves around the desire to structure a development effort, or perhaps even a research effort.

As we consider what has been presented here, it should be kept in mind that the work presented here belongs to the world of philosophy of science, specifically the topic of ‘hard to define’ problems [6]. [7] presented the issue which summarizes the thrust of this paper. In their words: “‘The formulation of a problem is often more essential than its solution’….To raise new questions, new possibilities, to regard old problems from a new angle, requires imagination’ Yet, how the creative process unfolds in framing ill-defined problems remains an open question. Indeed, untangling the interrelation between problem framing and the creative process in its lower-level aspects can inform scholars.. Problem framing entails building mental representations that simplify the problem..”

At this point, once the results of the effort have been presented, it may well be productive to contrast what has gone on in this paper versus the standard scientific or brainstorming effort which characterizes current research. Conventional research begins with observations, a hypothesis, and then establishes the validity or the falseness of that hypothesis through one or more experiments. The objective of science is to amass many hypotheses that have been shown to be correct, or more precisely, have not been falsified from experiments. The current research follows a different approach, one which ‘explores’ rather than confirms. The term ‘cartography’ is appropriate here. The approach here is ‘mapping’ the topic, almost as an explorer maps an area that is unknown. Knowledge becomes the accretion of these cartographies, and human advancement may end up through the application of these cartographies, the applications showing up in changes in the worlds of the material and of experience, respectively.

References

  1. Binz M, Schulz E (2023) Using cognitive psychology to understand GPT-3. Proceedings of the National Academy of Sciences 120(6)
  2. Kitadai A, Tsurusaki Y, Fukasawa Y, Nishino N (2023) Toward a novel methodology in economic experiments: Simulation of the Ultimatum Game with large language models. 2023 IEEE International Conference on Big Data 3168-3175.
  3. Gray K, Dillon D, Tandon N, Yuling G (2023) Can AI language models replace human participants? Science & Society. Trends Cogn Sci 27: 597-600. [crossref]
  4. Horton J.J (2023) Models as Simulated Economic Agents: What Can We Learn from Homo Silicus? National Bureau of Economic Research Working Paper 31122.
  5. Moskowitz HR (2012) ‘Mind Genomics’: The experimental, inductive science of the ordinary, and its application to aspects of food and feeding. Physiology & Behavior 606-613. [crossref]
  6. Garbuio M, Lin N (2021) Innovative idea generation in problem finding: Abductive reasoning, cognitive impediments, and the promise of artificial intelligence. Journal of Product Innovation Management 38: 701-725.
  7. Pham CTA, Magistretti S, Dell’Era C (2023) How do you frame ill-defined problems? A study on creative logics in action. Creativity and Innovation Management 32: 493-516.

Synthesis and Biological Evaluation of Chalcone Analogues as HIV Latency Reversing Agents

DOI: 10.31038/JPPR.2024713

Abstract

The primary barrier of eradicating HIV virus is the existence of latent reservoirs. To overcome this problem, it is urgent to identify specific and effective agents that can activate HIV latent reservoirs and bind with effective anti-retroviral therapy to eventually eliminate HIV virus. In previous study, we reported a chalcone analogue called Amt-87 that can significantly reactivate the transcription of latent HIV. However, it is of limited use due to high toxicity. Here, we have synthesized a series of structurally modified analogues of Amt-87 and evaluation for their reactivation of HIV-1 latency revealed one excellent active compound 4k, which can be taken up for further studies.

Introduction

It is more than 30 years passed since HIV was first reported in 1981 [1]. There are many drugs designed by targeting enzymes which are important for replication event during different stages of HIV-1 life cycle, preventing HIV-1 virus replication, therefore decrease the virus in the host blood to an undetectable level [2,3]. However, HAART is unable to completely eliminate the virus due to the existence of latent viral reservoirs, which are formed early after host infection and undetectable by common clinical tests [4], thus the host immune system can be invaded again upon the treatment interruption. HIV latency reservoir in the host CD4 T cell is the main obstacle to cure HIV-1 as reported recent years. Effective curative strategies aiming at eradication of HIV-1 virus are being developed [5-7]. One significant therapy which is called “shock and kill” has been proposed to achieve HIV eradication, there are two steps in this approach, using latency-reversing agents (LRAs) to reactivate the latent proviruses in the “shock” phase and then combining with the HAART to make the reactivated cells be sensitive to host immune system and cytopathogenicity response in the “kill” phase [8,9]. In this strategy, finding drugs to activate HIV latency catches much attention. Currently there are several activators have been proved, for example HMAB, SAHA, prostration JQ1 and so on, but most of these drugs are toxic to host cell and have some side effects [9]. Thus, highly efficient and specific drugs for reactivating HIV latency is important for HIV-1 eradication. Chalconoids, natural phenol compounds, have an array of biological activities including anti-inflammatory, antioxidant, antitumor and antibacterial activities and inhibit angiogenesis in vivo and in vitro [10-14]. Chalcone analogues have attracted a great deal of interest due to their synthetic and biological importance in medicinal chemistry. In our previous study, we reported the synthesis and biological evaluation of a chalcone analogue Amt-87, which can significantly reactivate the transcription of latent HIV proviruses and act synergistically with known LRAs such as prostratin and JQ1 to reverse HIV latency [15]. Unfortunately, Amt-87 is less active and requires high concentrations to reactivate latent HIV. Herein, we describe the optimization and characterization of the structure−activity relationship (SAR) of chalcone scaffold to generate a series of compounds with enhanced HIV latency reactivation activity.

Results

Synthesis of Chalcone Derivatives

Chalcone derivatives 4a-4k were prepared according to a two-step procedure (Scheme 1) previously reported. Condensation of 1-(2-hydroxy-5-methylphenyl)ethanone with 2- or 3-carboxybenzaldehyde in the presence of alkali gave carboxylic acid derivatives 3a and 3b. Treatment of 3a or 3b in CH2Cl2 with various amines in the presence of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDCI) and N-hydroxybenzotriazole (HOBt) afforded various amides 4a-4k with yields ranged from 30% to 50% (Scheme 1).

scheme 1

Scheme 1: Synthesis of chalcone amide derivatives 4a-4k. Reagents and conditions: (a) NaOH, EtOH, r.t., 24 h. (b) various amines, anhydrous HOBt, EDCI, DCM, 12 h.

Reactivation of latent HIV-1 in J-Lat A2 cells and structure–activity relationship. In order to identify compounds that can efficiently reactivate latent HIV-1, the J-LatA2 latency model was used in a screen that involved flow cytometry. This cell line was generated by infecting Jurkat T cells with viral particles containing an HIV-1 vector expressing Tat and GFP under the control of the viral 5’ LTR, with an IRES (internal ribosome entry site) separating the two open reading frames [15]. All synthesized chalcone compounds were evaluated for their activation of latent HIV-1 in J-Lat A2 cells and Amt-87 was used as the positive control. The results are presented in Table 1. The structure–activity relationship (SAR) was further explored as described below. Among these structurally related chalcone derivatives, we found that 4b, 4d, 4g were more active than Amt-87 in terms of HIV latency-reversing activity. Specifically, 4g showed the most optimal properties in the latent HIV-1 reactivation activity, inducing 14.6 ± 2.1% and 20.5 ± 2.8% GFP production in J-Lat A2 cells at 25 μg/mL and 50 μg/mL, respectively. What’s more, compound 4g has no significant cytotoxicity at a concentration of 50 μg/mL. These observations may demonstrate the potential of chalcone derivatives as LRA leads.

Table 1: Summary of HIV reactivation of Chalconoids in J-Lat cells

tab 1

Levels of HIV-1 activation and cell viability were evaluated by FACS analysis. Data are presented as mean ± SD of at least three independent experiments performed in triplicate.

Discussion

In this study, we synthesized and biologically evaluated chalcone derivatives for their HIV latency-reversing activity using HIV-LTR based cell model. Among these new derivatives, compound 4g effectively induced GFP production in J-Lat A2 cells at 25 and 50 μg/mL, while displaying no significant cytotoxicity at these concentrations. This study provides valuable data for the future development of chalcone as effective LRAs.

Materials and Methods

Chemistry

General procedure for the synthesis of Chalcones [15]. Claisene Schmidt condensation. To a solution of the corresponding aldehyde (1 equiv) and corresponding acetophenone (1 equiv) in EtOH (3 mL for 1 mmol of acetophenone) was added NaOH (5 equiv). The reaction mixture was stirred at room temperature for 24 h and neutralized with 10% HCl solution to form yellow precipitate. The yellow precipitate was filtered and washed with appropriate amount of water. The crude product was purified by chromatography using hexane/EtOAc and recrystallized by MeOH to give the desired compounds.

(E)-2-(3-(2-Hydroxy-5-Methylphenyl)-3-Oxoprop-1-en-1-yl)Benzoic Acid (3a)

Following the general procedure for the Claisen-Schmidt condensation, 1-(2-hydroxy-5-methylphenyl)ethanone 1 (1.1 mmol, 165.2 mg) and 2-carboxybenzaldehyde (1.0 mmol, 150.1 mg) were used to give 3a as a yellow solid. Yield 45.7%. 1H NMR (600 MHz, C5D5N): δ 8.41 (d, J=8. 2 Hz, 2H), 8.19 (d, J=15.5 Hz, 1H) , 8.12 (d, J=17.2 Hz, 1H), 8.07 (d, J=1.0 Hz, 1H), 7.88 (d, J=8.2 Hz, 2H), 7.31 (dd, J=1.8, 8.4 Hz, 1H), 7.1l (d, J=8.4 Hz, 1H), 2.18 (s, 3H). 13C-NMR (150 MHz, C5D5N): δ 193.9, 168.3, 161.6, 143.9, 138.7, 137.7, 137.2, 130.5 (2C), 129.0 (2C), 128.3, 126.4 (2C), 120.5, 118.2, 20.0. ESI-HRMS (-): m/z [M-H]- calcd for C17H13O4, 281.0819, found, 281.0816.

(E)-3-(3-(2-Hydroxy-5-Methylphenyl)-3-Oxoprop-1-en-1-yl)Benzoic Acid (3b)

Following the general procedure for the Claisen-Schmidt condensation, 1-(2-hydroxy-5-methylphenyl)ethanone 1 (1.1 mmol, 165.2 mg) and 3-carboxybenzaldehyde (1.0 mmol, 150.1 mg) were used to give 3b as a yellow solid. Yield 50.5%. 1H NMR (600 MHz, DMSO-d6): δ 12.3 (s, 1H), 8.40 (s, 1H), 8.18 (d, J=7.9 Hz, 1H), 8.08-8.14 (m, 2H), 8.03 (d, J=7.7 Hz, 1H), 7.89 (d, J=15.6 Hz, 1H), 7.62 (t, J=7.7 Hz, 1H), 7.40 (dd, J=1.7, 8.3 Hz, 1H), 6.92 (d, J=8.3 Hz, 1H), 2.33 (s, 3H); 13C NMR (150 MHz, DMSO-d6): δ 193.9, 167.4, 160.3, 144.0, 137.8, 135.4, 133.5, 132.2, 131.8, 131.0, 130.2, 129.7, 128.5, 123.5, 120.8, 118.0, 20.4; ESI-HRMS (-): m/z [M-H]- calcd for C17H13O4, 281.0819, found, 281.0814.

General Procedure for Synthesis of Amide Compounds 1a-1p. A Mixture of 3 (3a or 3b)

(1 equiv), HOBt (1.2 equiv) and EDCI (1.2 equiv) was dissolved in CH2Cl2, and stirred for 30 min. The mixture was then added with appropriate amine (2.0 equiv), and stirred at the room temperature for 12 h. After completion of the reaction, the mixture was concentrated in vacuum to give the crude product. The crude product was purified by column chromatography with hexane/EtOAc, and recrystallized with EtOAc to afford pure products.

(E)-N-(2-(1H-Indol-2-yl)Ethyl)-2-(3-(2-Hydroxy-5-Methylphenyl)-3-Oxoprop-1-en-1-yl)Benzamide (4a)

The title compound 4a was obtained by the reaction of compound 3a with 2-(1H-indol-2-yl)ethan-1-amine following the general procedure. Yellow solid, yield 57.5%, mp 182-183°C. 1H-NMR (600 MHz, CHLOROFORM-d): δ 12.49 (s, 1H), 8.08 (d, J=15.6 Hz, 1H), 7.57-7.59 (m, 1H), 7.52 (d, J=7.9 Hz, 1H), 7.45 (td, J=3.2, 8.9 Hz, 4H), 7.38 (dt, J=1.6, 7.5 Hz, 1H), 7.30-7.35 (m, 2H), 7.26 (dd, J=2.0, 8.4 Hz, 1H), 7.23 (d, J=8.1 Hz, 1H), 7.09 (t, J=7.6 Hz, 1H), 7.06 (d, J=2.2 Hz, 1H), 6.99-7.03 (m, 1H), 6.88 (d, J=8.4 Hz, 1H), 3.80 (d, J=6.2 Hz, 2H), 3.05 (t, J=6.5 Hz, 2H), 2.27 (s, 3H); 13C-NMR (151 MHz, CHLOROFORM-d): 192.6, 167.6, 166.7, 160.5, 141.7, 136.9, 136.6, 132.0, 131.3, 129.9, 129.2, 128.6, 127.8, 127.0, 126.7, 126.4, 121.9, 121.3, 121.2, 118.6, 118.5, 117.6, 117.4, 110.3, 64.6, 39.0, 29.5; HRMS calcd for C27H24N2O3 [M-H]- 423.1714, found 423.1718.

(E)-2-(3-(2-Hydroxy-5-Methylphenyl)-3-Oxoprop-1-en-1-yl)-N-(2-(Thiophen-2-yl)Ethyl)Benzamide (4b)

The title compound 4b was obtained by the reaction of compound 3a with 2-thiopheneethanamine following the general procedure. Yellow solid, yield 57.5%, mp 182-183°C. 1H-NMR (600 MHz, CHLOROFORM-d): δ 12.47 (s, 1 H), 8.08 (d, J=15.41 Hz, 1H), 7.67 (d, J=7.70 Hz, 1H), 7.58 (s, 1H), 7.46 (d, J=15.41 Hz, 1H), 7.37-7.42 (m, 2H), 7.34 (t, J=7.4 Hz, 1H), 7.22-7.25 (m, 1H), 7.04 (d, J=5.0 Hz, 1H), 6.85 (d, J=8.4 Hz, 1H), 6.81-6.83 (m, 1H), 6.79 (d, J=3.1 Hz, 1H), 5.98 (br. s., 1H), 3.68 (q, J=6.4 Hz, 2H), 3.10 (t, J=6.6 Hz, 2H), 2.26 (s, 3H); 13C-NMR (151 MHz, CHLOROFORM-d): 193.5, 168.6, 161.5, 142.6, 141.0, 137.6, 137.6, 133.3, 130.4, 130.2, 129.6, 128.0, 127.7, 127.5, 127.1, 125.6, 124.1, 123.0, 119.6, 118.4, 41.5, 29.7, 20.6; HRMS calcd for C23H21NO3S [M-H]- 390.1169, found 390.1171.

(E)-2-(3-(2-Hydroxy-5-Methylphenyl)-3-Oxoprop-1-en-1-yl)-N-Octylbenzamide (4c)

The title compound 4c was obtained by the reaction of compound 3a with octan-1-amine following the general procedure. Yellow solid, yield 57.5%, mp 182-183°C. 1H-NMR (600 MHz, CHLOROFORM-d): 12.47 (s, 1H), 8.09 (s, 1H), 7.69 (d, J=7.7 Hz, 1H), 7.59-7.61 (m, 1H), 7.50 (d, J=15.4 Hz, 1H), 7.45 (dd, J=1.3, 7.5 Hz, 1H), 7.42 (dt, J=1.2, 7.6 Hz, 1H), 7.36-7.39 (m, 1H), 7.24 (dd, J=2.0, 8.4 Hz, 1H), 7.19 (s, 1H), 6.86 (d, J=8.4 Hz, 1H), 3.38-3.42 (m, 2H), 2.27 (s, 3H), 1.14-1.33 (m, 12H), 0.80 (t, J=7.1 Hz, 3H); 13C-NMR (151 MHz, CHLOROFORM-d): 192.4, 167.6, 160.5, 141.6, 137.0, 136.6, 132.1, 129.2, 129.2, 128.5, 126.9, 126.8, 126.6, 122.0, 118.5, 117.3, 39.3, 30.8, 28.6, 28.2, 28.2, 26.0, 21.6, 19.6, 13.1; HRMS calcd for C25H31NO3 [M-H]- 392.2231, found 392.2233.

(E)-N-Benzyl-2-(3-(2-Hydroxy-5-Methylphenyl)-3-Oxoprop-1-en-1-yl)Benzamide (4d)

The title compound 4d was obtained by the reaction of compound 3a with benzylamine following the general procedure. Yellow solid, yield 57.5%, mp 182-183°C. 1H-NMR (600 MHz, CHLOROFORM-d): 12.48 (s, 1H), 8.11 (d, J=15.6 Hz, 1H), 7.65 (d, J=7.7 Hz, 1H), 7.55 (s, 1H), 7.42-7.45 (m, 2H), 7.39 (t, J=7.4 Hz, 1H), 7.31-7.34 (m, 1H), 7.21-7.25 (m, 3H), 7.16-7.19 (m, 1H), 6.83 (d, J=8.4 Hz, 1H), 6.21 (br. s., 1H), 4.54 (d, J=5.5 Hz, 2H), 2.24 (s, 3H), 1.18 (br. s., 1H); 13C-NMR (151 MHz, CHLOROFORM-d): 193.5, 168.5, 161.5, 142.6, 137.7, 137.6, 137.5, 133.2, 130.4, 130.2, 129.6, 128.9, 128.0, 128.0, 127.8, 127.7, 127.6, 123.0, 119.5, 118.3, 44.3, 20.6; HRMS calcd for C24H21NO3 [M-H]- 370.1449, found 370.1448.

(E)-N-(Tert-Butyl)-2-(3-(2-Hydroxy-5-Methylphenyl)-3-Oxoprop-1-en-1-yl)Benzamide (4e)

The title compound 4e was obtained by the reaction of compound 3a with 2-methylpropan-2-amine following the general procedure. Yellow solid, yield 57.5%, mp 182-183°C. 1H-NMR (600 MHz, CHLOROFORM-d): 12.48 (s, 1H), 8.06 (d, J=15.4 Hz, 1H), 7.64 (d, J=7.5 Hz, 1H), 7.58-7.60 (m, 1H), 7.47 (d, J=15.5 Hz, 1H), 7.32-7.41 (m, 3H), 7.22 (dd, J=2.0, 8.4 Hz, 1H), 6.83 (d, J=8.6 Hz, 1H), 5.63 (br. s., 1H), 2.26 (s, 3H), 1.41 (s, 9H); 13C-NMR (151 MHz, CHLOROFORM-d): 192.5, 167.2, 160.5, 141.7, 137.9, 136.5, 131.7, 129.2, 128.9, 128.5, 126.9, 126.7, 126.6, 121.8, 18.5, 117.3, 51.3, 28.7, 27.8, 19.5, 13.1; HRMS calcd for C21H23NO3 [M-H]- 336.1605, found 336.1607.

(E)-2-(3-(2-Hydroxy-5-Methylphenyl)-3-Oxoprop-1-en-1-yl)-N-Phenethylbenzamide (4f)

The title compound 4f was obtained by the reaction of compound 3a with 2-phenylethan-1-amine following the general procedure. Yellow solid, yield 57.5%, mp 182-183°C. 1H-NMR (600 MHz, CHLOROFORM-d) 12.48 (s, 1H), 8.09 (d, J=15.41 Hz, 1H), 7.68 (d, J=7.70 Hz, 1H), 7.59 (s, 1H), 7.47 (d, J=15.59 Hz, 1H), 7.41 (br. s., 1H), 7.33-7.36 (m, 2H), 7.25 (d, J=8.25 Hz, 1H), 7.18-7.22 (m, 3H), 7.15-7.17 (m, 2H), 7.12 (d, J=7.15 Hz, 1H), 6.87 (d, J=8.44 Hz, 1H), 3.67-3.72 (m, 2H), 2.89 (t, J=6.79 Hz, 2H), 2.27 (s, 3H);13C NMR (151 MHz, CHLOROFORM-d) 192.5, 167.6, 160.5, 141.5, 137.5, 136.7, 136.6, 132.2, 129.3, 129.2, 128.5, 127.7, 127.7, 126.9, 126.7, 126.5, 125.6, 122.0, 118.5, 117.4, 40.1, 34.5, 19.6;HRMS calcd for C25H23NO3 [M-H]- 384.1605, found 384.1603.

(E)-3-(3-(2-Hydroxy-5-Methylphenyl)-3-Oxoprop-1-en-1-yl)-N-(2-Methoxyphenyl)Benzamide (4g)

The title compound 4g was obtained by the reaction of compound 3b with 2-methoxyaniline following the general procedure. Yellow solid, yield 57.5%, mp 182-183°C. 1H-NMR (600 MHz, CHLOROFORM-d): 12.49 (br. s., 1H), 8.49 (br. s., 1H), 8.42 (d, J=7.0 Hz, 1H), 8.12 (br. s., 1H), 7.75-7.83 (m, 2H), 7.68 (d, J=7.3 Hz, 1H), 7.61-7.65 (m, 1H), 7.59 (br. s., 1H), 7.42-7.45 (m, 1H), 7.18-7.22 (m, 1H), 7.00 (dt, J=1.7, 7.8 Hz, 1H), 6.90-6.94 (m, 1H), 6.81-6.84 (m, 2H), 3.83 (s, 3H), 2.25 (s, 3H); 13C-NMR (151 MHz, CHLOROFORM-d): 192.3, 163.4, 160.5, 147.2, 142.8, 136.7, 135.0, 134.3, 130.9, 128.4, 127.5, 127.1, 126.4, 126.1, 123.2, 120.4, 120.1, 118.9, 118.5, 117.3, 109.0, 54.8, 19.5; HRMS calcd for C24H21NO4 [M+Na]+ 410.1363, found 410.1358.

(E)-3-(3-(2-Hydroxy-5-Methylphenyl)-3-Oxoprop-1-en-1-yl)-N-Octylbenzamide (4h)

The title compound 4h was obtained by the reaction of compound 3b with octan-1-amine following the general procedure. Yellow solid, yield 57.5%, mp 182-183°C. 1H-NMR (600 MHz, CHLOROFORM-d): 12.51 (s, 1H), 8.05 (s, 1H), 7.77 (d, J=15.4 Hz, 1H), 7.68 (d, J=7.7 Hz, 1H), 7.61-7.65 (m, 2H), 7.59 (s, 1H), 7.38 (t, J=7.6 Hz, 1H), 7.22 (d, J=8.4 Hz, 1H), 6.83 (d, J=8.4 Hz, 1H), 6.41-6.49 (m, 1H), 3.39 (q, J=6.2 Hz, 2H), 2.26 (s, 3H), 1.24-1.31 (m, 3H), 1.17-1.24 (m, 9H), 0.78-0.81 (m, 3H); 13C-NMR (151 MHz, CHLOROFORM-d): 193.4, 166.9, 161.5, 144.0, 137.7, 135.7, 135.1, 131.7, 129.4, 129.2, 128.6, 128.1, 126.9, 121.3, 119.6, 118.3, 40.3, 31.8, 29.7, 29.3, 29.2, 27.1, 22.6, 20.6, 14.1; HRMS calcd for C25H31NO3 [M+Na]+ 416.2196, found 416.2193.

(E)-3-(3-(2-Hydroxyphenyl)-3-Oxoprop-1-en-1-yl)-N-(2-(Thiophen-2-yl)Ethyl)Benzamide (4i)

The title compound 4i was obtained by the reaction of compound 3b with 3-thiopheneethanamine following the general procedure. Yellow solid, yield 57.5%, mp 182-183°C. 1H-NMR (600 MHz, CHLOROFORM-d): 12.49 (br. s., 1H), 7.99 (br. s., 1H), 7.70-7.76 (m, 1H), 7.62 (t, J=8.5 Hz, 2H), 7.55-7.59 (m, 2H), 7.33-7.37 (m, 1H), 7.21 (dd, J=1.9, 8.3 Hz, 1H), 7.07-7.09 (m, 1H), 6.87 (dd, J=3.3, 5.0 Hz, 1H), 6.81-6.83 (m, 1H), 6.79 (d, J=2.8 Hz, 1H), 6.63 (br. s., 1H), 6.56 (br. s., 1H), 3.64-3.68 (m, 2H), 3.07-3.10 (m, 2H), 2.25 (s, 3H); 13C NMR (151 MHz, CHLOROFORM-d): 193.4, 167.0, 161.5, 143.9, 141.1, 137.8, 135.4, 135.1, 131.9, 129.4, 129.3, 128.7, 128.1, 127.2, 126.8, 125.5, 124.1, 121.3, 119.6, 118.3, 41.5, 29.8, 20.6; HRMS calcd for C23H21NO3S [M+Na]+ 414.1134, found 414.1129.

(E)-3-(3-(2-Hydroxyphenyl)-3-Oxoprop-1-en-1-yl)-N-Phenethylbenzamide (4j)

The title compound 4j was obtained by the reaction of compound 3b with 2-phenylethan-1-amine following the general procedure. Yellow solid, yield 57.5%, mp 182-183°C. 1H-NMR (600 MHz, CHLOROFORM-d): 12.49 (br. s., 1H), 7.96 (s, 1H), 7.69-7.75 (m, 1H), 7.57-7.60 (m, 2H), 7.53-7.57 (m, 2H), 7.31-7.35 (m, 1H), 7.22-7.25 (m, 1H), 7.19-7.22 (m, 2H), 7.14 (d, J=7.3 Hz, 3H), 6.80-6.82 (m, 1H), 6.49 (br. s., 1H), 6.44 (br. s., 1H), 3.64 (q, J=6.8 Hz, 2H), 2.86 (t, J=6.9 Hz, 2H), 2.25 (s, 3H); 13C-NMR (151 MHz, CHLOROFORM-d): 192.3, 165.9, 160.5, 142.9, 137.8, 136.7, 134.4, 134.0, 130.8, 128.4, 128.2, 127.8, 127.7, 127.6, 127.1, 125.7, 125.6, 120.2, 118.5, 117.3, 40.3, 34.6, 19.5; HRMS calcd for C25H23NO3 [M+Na]+ 408.1570, found 408.1565.

(E)-N-(2-(1H-Indol-3-yl)Ethyl)-3-(3-(2-Hydroxy-5-Methylphenyl)-3-Oxoprop-1-en-1-yl)Benzamide (4k)

The title compound 4k was obtained by the reaction of compound 3b with 2-(1H-indol-3-yl)ethan-1-amine following the general procedure. Yellow solid, yield 57.5%, mp 182-183°C. 1H-NMR (600 MHz, CHLOROFORM-d): 12.49 (s, 1H), 8.08 (d, J=15.6 Hz, 1H), 7.57-7.59 (m, 1H), 7.52 (d, J=7.9 Hz, 1H), 7.45 (td, J=3.2, 8.9 Hz, 4H), 7.38 (dt, J=1.6, 7.5 Hz, 1H), 7.30-7.35 (m, 2H), 7.26 (dd, J=2.0, 8.4 Hz, 1H), 7.23(d, J=8.1 Hz, 1H), 7.09 (t, J=7.6 Hz, 1H), 7.06 (d, J=2.2 Hz, 1H), 6.99-7.03 (m, 1H), 6.88 (d, J=8.4 Hz, 1H), 3.80 (d, J=6.2 Hz, 2H), 3.05 (t, J=6.5 Hz, 2H), 2.27 (s, 3H); 13C NMR (151 MHz, CHLOROFORM-d): 192.6, 167.6, 166.7, 160.5, 141.7, 136.9, 136.6, 132.0, 131.3, 129.9, 129.2, 128.6, 127.8, 127.0, 126.7, 126.4, 121.9, 121.3, 121.2, 118.6, 118.5, 117.6, 117.4, 110.3, 64.6, 39.0, 29.5; HRMS calcd for C27H24N2O3 [M+Na]+ 447.1679, found 447.1674.

Biochemistry

Reagents and Cell Lines

All the cell line used in this study were either purchased from American Type Culture Collection (ATCC, Manassas, VA) or described as previously [16].

Flow Cytometry Screening

The Flow cytometry screening assay was performed in Jurkat-based cells line(J-LatA2), containing LTR-Tat-IRES-GFP-LTR [15]. After incubation with Amt-87 as indicated or 0.1% dimeththyl sulfoxide (DMSO) as negative control, J-Lat A2 were harvested and washed twice with 1× phosphate buffered saline (PBS) and analyzed the expression level of GFP by flow cytometer (Epics Altra, BECKMAN COULTER). The changes of GFP expression level can indirectly reflect the transcription level of HIV.

Notes

The authors declare no competing financial interest.

References

  1. Siegal FP, Lopez c, Hammer GS, Brown AE, Kornfeld SJ, et al. (1981) Severe Acquired Immunodeficiency in Male-Homosexuals, Manifested by Chronic Perianal Ulcerative Herpes-Simplex Lesions. New England Journal of Medicine 305: 1439-1444. [crossref]
  2. DD HO, Neumann AU, Perelson AS, Chen W, Leonard JM, et al. (1995) Rapid Turnover of Plasma Virions and Cd4 Lymphocytes in Hiv-1 Infection. Nature, 373: 123-126. [crossref]
  3. Wei XP, Ghosh SK, Taylor ME, Johnson VA, Emini EA, et al. (1995) Viral Dynamics in Human-Immunodeficiency-Virus Type-1 Infection. Nature 373: 117-122. [crossref]
  4. Eisele E, Siliciano RF (2012) Redefining the viral reservoirs that prevent HIV-1 eradication. Immunity 37: 377-388. [crossref]
  5. Deeks SG, Brigitte A, Ben B, Monsef B, Scott C, Nicolas C, et al. (2012) Towards an HIV cure: a global scientific strategy. Nature Reviews Immunology 12: 607-614. [crossref]
  6. Mohammadi P, Julia dI, Miguel M, Raquel M, István B, et al. (2014) Dynamics of HIV latency and reactivation in a primary CD4+ T cell model. PLoS Pathog. 10. [crossref]
  7. Archin NM, Margolis DM (2014) Emerging strategies to deplete the HIV reservoir. Curr Opin Infect Dis 27: 29-35. [crossref]
  8. Karn J (2011) The molecular biology of HIV latency: breaking and restoring the Tat-dependent transcriptional circuit. Current Opinion in Hiv and Aids 6: 4-11. [crossref]
  9. Barton KM, Burch BD, Soriano-SN, Margolis MD et al. (2013) Prospects for Treatment of Latent HIV. Clinical Pharmacology & Therapeutics 93: 46-56. [crossref]
  10. Nowakowska Z (2007) A review of anti-infective and anti-inflammatory chalcones. Eur J Med Chem 42: . 125-37.
  11. Boumendjel, A, Julien B, Pierre-AC, Edwige N, Madeleine B, et al. (2008) Antimitotic and antiproliferative activities of chalcones: forward structure-activity relationship. J Med Chem 51: 2307-2310. [crossref]
  12. Ouyang, Y. Juanjuan L, Xinyue C, Xiaoyu F, Sun S, et al. (2021) Chalcone Derivatives: Role in Anticancer Therapy. Biomolecules 11. [crossref]
  13. 1Sharma R, Rakesh K, Rishi K, Sukriti K, Anukriti K, et al. (2015) A Review on Mechanisms of Anti Tumor Activity of Chalcones. Anticancer Agents Med Chem. 16: 200-211. [crossref]
  14. Elkanzi NAA, Hajer H, Ruba AA, Wassila D, Fatin MZ, et al. (2022) Synthesis of Chalcones Derivatives and Their Biological Activities: A Review. ACS Omega 7: 27769-27786. [crossref]
  15. Wu J, Ming TA, Rui S, Hui-ru W, Diao Y, et al. (2017) A chalcone derivative reactivates latent HIV-1 transcription through activating P-TEFb and promoting Tat-SEC interaction on viral promoter. Sci Rep 7.
  16. Li ZC, Jia G, Yuntao W, Qiang Z (2013) The BET bromodomain inhibitor JQ1 activates HIV latency through antagonizing Brd4 inhibition of Tat-transactivation. Nucleic Acids Research 41: 277-287. [crossref]

The Use of Pine Apple Juice in the Elimination of Egg Stickiness in Catfish (Clarias gariepinus), Reared in a Hatchery, Sebeta, Ethiopia

DOI: 10.31038/CST.2021115

Abstract

This study was conducted from May to June 2018 at the National Fishery Hatchery in Ethiopia and aimed to assess the effectiveness of pineapple juice in the elimination of egg stickiness and improved hatchability rate in Clarias gariepinus. There were four different concentrations of pineapple juice tested: 0% (without juice), 1%, 3%, and 5%. The result revealed that egg stickiness was reduced dose-dependently (p < 0.05), with the 5% concentration showing the greatest reduction (59%) compared to the control group (77%). However, it was noted that higher concentrations of pineapple juice have adversely affected egg hatching and fertilization rates. The minimum observed fertilization rate (45%) was observed at the highest concentration tested (5%), while the maximum (77%) was observed at lower concentrations (1%). Similarly, the minimum hatching rate (51.5%) was observed at higher concentrations (5%), with the maximum (71.3%) at 3% pineapple concentrations. Despite these variations, no significant difference (p > 0.05) in the survival rate of juvenile catfish was observed across different concentrations of pineapple juice. This finding highlights the potential of pineapple juice as a practical and accessible approach for improving catfish hatching operations. This study helps to optimize catfish breeding procedures and demonstrates the efficacy of natural treatments in aquaculture practices.

Keyword

Clarias gariepinus, Pineapple juice, Egg stickiness, Hatchability, Fertilization

Introduction

Aquaculture is crucial for meeting the world’s expanding food demand. Sustainable aquaculture output is strongly dependent on good reproduction and efficient hatching methods. Fish seed mass propagation, whether artificial or semi-artificial, is essential to modern fish farming. Due to seasoned breeding habits in most culturable species and the fact that fish must be caught at a time that may not coincide with ideal production conditions, the natural supply of fingerlings is erratic, unstable, and insufficient. Of these species, the African catfish (C. gariepinus) reproduces according to its seasonal breeding behavior at particular periods of the year [1-3].

The African catfish is a valuable tropical aquaculture fish in Africa, Asia, and Europe. This is because the species is suitable for aquaculture due to its fast growth rate, tolerance of high stocking density and poor water quality, acceptance of both artificial and natural feeds, and high market demand, and high production in the pond, 2.5 times higher than tilapia. It is also the second most important commercial fish species in Ethiopia’s capture fishery next to the Nile tilapia. Despite this, the lack of good quality and quantity of seeds has been a major constraint in aquaculture development. Thus, at larvae and fry stages, low survival rates and poor growth performance are the major challenges in catfish production in Ethiopia. This is most probably related to egg stickiness, which needs to be reduced before eggs can be incubated successfully [3-7].

Chemical treatments like tannic acid, urea, and acylase enzyme are frequently used in traditional ways for reducing egg stickiness. However, the extensive use of such chemicals raises concerns about potential detrimental implications on water quality, fish health, and the general ecological balance of the aquaculture system. This procedure also takes a minimum of one hour and a great deal of labor and experience. As a result, this circumstance forces people to look for more eco-friendly alternatives, like pineapple juice [8-12].

In catfish hatcheries, recent studies on the use of pineapple juice as an alternate egg-sticking reduction agent have produced encouraging results. According to preliminary research, bromelain in pineapple juice significantly decreased the stickiness of catfish eggs, offering a healthy and natural alternative. Pineapple is widely available in many tropical nations, including Ethiopia. It includes several proteolytic enzymes, primarily bromalinases, which can breakdown proteins. The skin and hump of pineapple contain tannin chemicals, which belong to the phenol group. Tannins can eliminate egg adhesion because they can bind and precipitate protein compounds due to the presence of functional bond groups that interact strongly with protein molecules, resulting in large and complex cross-links, known as tannin-protein [13-17].

With sustainability becoming a top priority in the aquaculture sector, this study attempts to thoroughly explore green ways to enhance the hatchability of catfish eggs. Therefore, in order to overcome the drawbacks of chemical agents, this study aimed to evaluate pineapple juice’s potential as a natural substance to remove egg stickiness and increase the hatchability rate in catfish. This is in line with the necessity of ethical and ecologically friendly aquaculture methods, guaranteeing the sustainability of Ethiopia’s catfish aquaculture sector.

Materials and Methods

Broodstock Selection and Hormone Administration

The experiment was conducted from May to June 2018 at the National Fisheries and Aquatic Life Research Center, NFALRC, hatchery. Three males and three females of C. gariepinus broodstock, weighing between 958 and 1000 grams, were taken from the NFALRC ponds. Priority was given to individuals presenting an elongated papilla and a slightly swollen reddish urogenital organ for male brooder selection. While, criteria including a well-rounded and soft abdomen, extending anteriorly beyond the pectoral fins to the genital opening were used for female brood stock selection. The broodstocks were acclimatized in separated tanks in hatchery for 24 hours.

The female brooders were artificially stimulated using 0.3 ml of common carp pituitary gland per kilogram of body weight, as described by Olanrewaju et al. The injection was done intramuscularly above the lateral line just below the dorsal fin. The female brooder was given light abdominal pressure twelve hours following the injection, at which point the ovulated eggs were gathered and weighed to the closest 0.01 g in a dry, clean plastic dish. The milt was collected by sacrificing the male broodstock. The milt and eggs were mixed and two drops of saline water were added to activate the milt and stirred gently with a feather to fertilize the eggs. The eggs were left for one minute to provide sufficient contact with the milt [18,19].

Pineapple Juice Preparation and Application

Pineapple juice was prepared by squeezing freshly peeled fruit from a local market in Sebeta, Ethiopia. Three pineapple juice concentrations, including 1% (1 ml of juice to 99 ml of clean water), 3% (3 ml of juice to 97 ml of clean water), and 5% (5 ml of juice to 95 ml of clean water), as well as a control group without pineapple juice, were prepared to investigate dose-dependent effects. In order to represent each concentration of pineapple juice, the fertilized eggs were divided into four equal parts weighing 800 g each. These portions were then divided in triplicate. After pouring the juice solution over the eggs, the mixture was constantly swirled for approximately three minutes using a feather.

Experimental Design and Water Management

For egg hatching, a total of twelve 200-liter plastic bowls were utilized, one for each of the four treatments in triplicate. A mesh was placed above the water surface within the bowl to provide a platform for placing fertilized catfish eggs treated with pineapple juice. The eggs were gently poured over the mesh, ensuring uniform distribution above the water surface. In order to guarantee sufficient oxygenation, a constant water flow was maintained during the trial. The ideal conditions for embryonic development were preserved by controlling the flow rate. In order to facilitate the incubation of catfish eggs, the water temperature was carefully monitored and changed to 24 to 27°C. Water quality indicators such as temperature, pH, and dissolved oxygen levels were carefully monitored at regular intervals throughout the 36-hour experiment. The hatching success criterion was defined as the emergence of larvae from egg capsules. Dead eggs, which are white and transparent, are removed immediately.

Data Collection

Egg adhesiveness was assessed by determining the percentage of eggs that formed clumps. After 15 minutes of incubation, the eggs that had formed clumps within each container were counted. The percentage of adhesion, fertilization, hatching, and survival rate were determined using the standard formula (Eq1 to Eq4) [20,21].

FOR

Statistical Analysis

The collected data on egg stickiness and hatchability rates for each concentration was statistically analyzed using R statistical software, with a one-way ANOVA used determine the significant differences between the treatment and different variables. Pearson’s correlation was used to assess the relationship between pineapple juice concentrations and fertilization and hatchability parameters [22].

Results and Discussion

Physicochemical Parameters

The result of major physicochemical parameters is presented in Table 1. Our study indicated that water temperature (24-27°C), dissolved oxygen (>5 mgL-1), pH (7.12-7.16), and ammonia levels (< 0.1 mgL-1) remained within optimal ranges for effective hatching in catfis. Effective water quality management strategies are critical for providing an optimal environment for catfish embryogenesis [23,24].

Table 1: Physicochemical water quality parameters with mean ± SD

Parameters

Range

Mean ± SD

DO (mgL-1)

5.8 – 6.6

6.2 ± 0.4

pH

7.12 – 7.16

7.1 ± 0.2

Temperature (°C)

24 – 27.5

25 ± 0.3

Ammonia (mgL-1)

0.1 – 0.15

0.1 ± 0.05

Effect of Pineapple Juice Concentrations on Cat Fish Egg Stickiness

Supplementing catfish eggs with pineapple juice resulted in a significant reduction in stickiness (ANOVA, p < 0.05) at varied doses (Figure 1). It was found that all of the treatments reduced the adhesiveness of C. gariepinus eggs to varying degrees. The control group, without pineapple juice, had the highest adhesion value (78.6%) showing the least reduction in egg stickiness. Interestingly, the greatest concentration, 5%, showed the highest reduction in egg adhesion (61%). The 3% pineapple juice concentration resulted in a medium reduction (67.3%). Subsequent post-hoc Tukey’s tests indicated a significant decrease in egg stickiness for the 3% and 5% concentrations compared to the control. This could be explained by the different concentrations of the enzyme bromelain in pineapple juice, which breaks down protein in a sticky fish egg layer. Fish eggs are sticky when they come into touch with water because of their high protein cortical envelope. The more easily the bromelain enzyme can break protein molecules into smaller forms, the more effective it is at eroding the protein layer found in catfish eggs. Our findings corroborate this theory because we found that the amount of egg stickiness decreased more noticeably at greater pineapple juice concentrations [25-28].

FIG 1

Figure 1: Egg adhesion across different pineapple juice concentrations. Letters a and b indicate significant differences between treatments (P < 0.05).

Our result aligns with previous studies indicating the potential of pineapple juice to affect adhesive properties with dose-dependent stickiness reduction in catfish eggs. The egg of C. gariepinus has a sticky layer, which is composed of glycoprotein. Because of its proteolytic action, bromelain may interact with the glycoproteins on the egg surface, changing their sticky characteristics and changing the way the egg clumps.

Furthermore, the slight decrease in stickiness that was seen at lower pineapple juice concentrations (1%), which points to a possible threshold effect wherein enzyme activity might not be enough to cause appreciable changes in sticky characteristics. This suggests a dose-dependent effect of pineapple juice on egg adhesiveness, with lower concentrations potentially leading to greater egg aggregation. Excessive egg clumping may result in restricted water flow around individual eggs and poor oxygenation. Furthermore, it reduces oxygen levels and increases fungal infection, leading to higher mortality rates.

Fertilization and Hatchability Rate

The result of fertilization and hatching rates are presented in Figure 2 with significant variation among treatments (ANOVA, p < 0.05). In comparison to the control group, treatment with 1% concentration resulted in a higher improvement in fertilization rates (77%), which was less than the values reported by Nwerich and Igill-Iboi (88%), Sukendi et al. (89.5%), Egwenomhe et al. (98.6%) in C. gariepinus, and Thal and Ngo (89.5%) in C. carpio. Surprisingly, the 5% concentration, despite reducing egg stickiness significantly, showed a decline in fertilization rates (45%). The highest fertilization value obtained in this study was not related to the smallest adhesion value obtained, which suggests that additional influencing elements as well as the egg adhesion value affect the fertilization value. This is in agreement with the previous studies which reported the highest fertilization rate in the lowest concentrations of pineapple juice solution. This research implies that pineapple juice improves fertilization rates at specific concentrations, while greater concentrations may have a negative impact on sperm-egg interactions [13-16].

FIG 2

Figure 2: The fertilization and hatching rate across treatments. Letters indicate a significant difference between treatments (p < 0.05).

The egg’s adhesiveness makes it more difficult for the embryo to hatch successfully in artificial breeding. This is because the eggs have become agglutinated, covering the micropyles and making it more difficult for sperm to fertilize them. As a result, there are less opportunities for sperm to come into touch with the eggs, which lowers the likelihood that the eggs will be fertilized. Other substances, such as pineapple juice, are used to crack the egg layer, which accelerates fertilization by allowing sperm to make direct contact with the eggs without being impeded by the mucus coating on the egg surface [14-16].

There was significant variation in hatching rate among treatments (p < 0.05). The highest hatching rate was obtained in the 3% pineapple juice concentrations (71.3) and showed a decrease (51.5%) in the 5% concentrations. Similar results were found in the previous studies confirming the dose-dependent trend in egg hatchability of 86.6 % at 1% concentration and decreasing to 25.3 % at 5% concentration. Similarly reported 78.1 % hatching rate at 0.75% pineapple juice concentration and decreased to 75.6% at 1% concentration [18].

This emphasizes the critical role of optimal pineapple juice concentration in influencing not only fertilization but also hatchability in catfish. This shows the effect of pineapple juice in reducing the adhesive layer of catfish eggs. In normal cases, when the eggs come in contact with water for a while, and clump reducing the prospect of the eggs hatching. In general, destocking of eggs using pineapple is quick and simple and requires three minutes only instead of the one hour for conventional desticking techniques. This technique is easy to adopt and advantageous by dramatically reducing the egg handling period.

Survival Rates of Hatched Fish

The survival rates of fish in treatment groups were extensively observed after hatching (Table 2). The larvae survival rate varied from 84% in 1% pineapple concentration to 90% in 3% pineapple concentration, with no significant difference across treatments (p > 0.05). The general survivability of the growing larvae may not be directly impacted by the enzymatic activity of bromelain, which mainly modifies protein structures and sticky characteristics. Similar results were observed in the previous studies [16,17].

Table 2: Survival rate of catfish fries at different pineapple juice concentrations

Concentrations

Number of hatchlings

Survival rate (%)

0% (control)

261

87

1%

259

84

3%

315

90

5%

291

89

The Response of Different Variables to Pineapple Concentrations

The relationship between different parameters and pineapple juice concentrations is indicated using Pearson’s correlation (Figure 3). The pineapple juice concentrations exhibited a strong negative correlation with egg adhesion (ANOVA, p < 0.001, r2=-0.93), indicating that higher concentrations of pineapple juice led to reduced egg adhesiveness. This is associated with the presence of proteolytic enzymes in the pineapple juice that can inhibit the activities of glycoprotein which could have the potential of increasing the sticky layer of fish eggs. Similar reports were found on the negative correlation between pineapple concentrations and fish egg adhesiveness. Although not statistically significant, there was a negative correlation found between the concentrations of pineapple juice and the rates of fertilization and hatching. Similarly, report state a decrease in fertilization and hatching rate with a corresponding increase in the concentration of the pineapple juice. On the other hand, hatching rates and fertilization rates exhibited a positive correlation (ANOVA, p < 0.001, r2=0.87), indicating a comparable response to the concentrations of pineapple [16,17].

FIG 3

Figure 3: Person’s correlation showed the relationship between pineapple juice concentrations and different parameters.

The findings underscore the potential of pineapple juice as a natural agent for manipulating reproductive parameters in aquatic organisms, offering insights into its application in aquaculture practices for improving breeding success and egg quality. However, pineapple juice concentration matters at a certain concentration and might be detrimental over the optimal dosage.

Conclusion and Recommendation

In conclusion, the application of pineapple juice solution has demonstrated efficacy in reducing the stickiness of Clarias gariepinus eggs, offering a practical solution for hatcheries in Ethiopia. Our findings suggest that this method not only effectively mitigates the adhesive properties of the eggs but also presents a user-friendly approach that does not necessitate specialized expertise for implementation. Moreover, while higher concentrations of pineapple juice may have adversely affected egg hatching and fertilization rates, it’s noteworthy that the positive impact on egg-destocking adhesiveness was observed. Therefore, this approach holds promise as a simple and accessible solution for hatchery operations, allowing for improved handling and management of catfish eggs without requiring extensive training or technical skills. To optimize hatchery operations and improve the management of Clarias gariepinus eggs in Ethiopia, the following recommendations are proposed:

  • Incorporate pineapple juice solution application to reduce egg stickiness effectively.
  • Utilize lower concentrations of pineapple juice to maintain optimal hatching and fertilization rates while mitigating egg adhesiveness.

Acknowledgements

We would like to acknowledge the National Fisheries and Aquatic Life Research Center (NFALRC) for financing the research work.

References

  1. Ojutiku RO (2008) Comparative Survival and Growth Rate of Clarias gariepinus and Heteroclarias Hathclings Fed Live and Frozen Daphnia. Pakistan Journal of Nutrition 7(4).
  2. Adebayo OT (2006) Reproduction performance of African catfish Clarias gariepinus broodstocks on varying maternal stress. Journal of Fish Biology 1(2).
  3. Dadebo D, Gebre-Mariam Z, Mengistou S (2011) Breeding season, maturation, Fecundity and condition factor of the African catfish Clarias gariepinus Burchell 1822 (Pisces: Clariidae) in Lake Chamo, Ethiopia. Ethiopian Journal of Biological Science 10(1).
  4. FAO (2020) The State of World Fisheries and Aquaculture. Sustainability in action. Rome.
  5. Amachree D, Akinrotimi OA, Dabor PO (2018) Effects of Different Physiological Saline Concentrations on the Reproductive Performance of Clarias gariepinus (Burchell, 1822). Journal of FisheriesSciences 12(4).
  6. Ponzoni RW, Nguyen NH (2008) Proceedings of a Workshop on the Development of a Genetic Improvement Program for African Catfish, Clarias gariepinus. Accra, Ghana, pp 131.
  7. Abdelhamid AM, Radwan IA Mehrim A, Abdelhamid AF (2010) Improving the survival rate of the African catfish, Clarias gariepinus. Journal of Animal and Poultry Production, Mansoura University 1: 409-414.
  8. Zarski D, Krejszeff S, Kucharczyk D, Palinska-Zarska K, Targonska K (2015) The application of tannic acid to the elimination of egg stickiness at varied moments of the egg swelling process in pikeperch, Sander lucioperca (L.). Aquaculture Research 46 (2).
  9. Asraf A, Muchlisin ZA, Siti-Azizah MN (2013) Removal of eggs adhesiveness of African catfish (Clarias gariepinus) at different concentrations of urea solution. Aceh International Journal of Science and Technology 2(3): 94-97.
  10. Linhart O, Rodina M, Gela D, Kocour M, Rodriguez M (2003) Improvement of common carp artificial reproduction using enzyme for elimination of egg stickiness. Aquatic Living Resources 16: 450-456.
  11. Linhart O, Rodina M, Gela D, Flajšhans M, Kocour M (2004) Enzyme treatment for elimination of egg stickiness in tench (Tinca tinca), European catfish (Silurus glanis L.) and common carp (Cyprinus carpio L.). Fish Physiology and Biochemistry 28: 507-508.
  12. Horvath L, Tamas G, Seagrave C (2002) Carp and Pond Fish Culture. Blackwell Science, pp. 23-56.
  13. Nwachi OF, Igill-Iboi M (2014) Experimental Trials of the use of Pineapple juice in disticking Clarias gariepinus egg in indoor hatchery at Niger Delta Area. Nigeria. Journal of Natural Sciences Research 4(18).
  14. Sule OD, Adikwu I (2004) Embryonic development in clarias gariepinus (buchell, 1822) under laboratory conditions. Animal Research International 1(2).
  15. Egwenomhe M, Yusuf A, Ihenyen E, Sadiq HO, Nwabuma EI (2022) The Effects of Pineapple and Orange Juice as De-Adhesion Agents on Clarias gariepinus (African catfish) Eggs. Dutse Journal of Pure and Applied Sciences 8: 70-76.
  16. Sukendi T, Putra RM Nuraini AD (2022) The use of pineapple extract to remove egg adhesiveness in the production of Asian redtail catfish fry (hemibagrus nemurus cv). OP Conf. Ser: Earth Environ.
  17. Payne DE, Martin NR, Parzych KR, Rickard AH, Underwood A, Boles BR (2013) Tannic acid inhibits Staphylococcus aureus surface colonization in an IsaA-Dependent Manner. Infection and Immunity, American Societr for Microbiology 81: 2-13.
  18. Zakes KD, Zdzisław Z, Jakub R (2005) The use of tannic acid to remove adhesiveness from pikeperch, Sander lucioperca, and eggs. Aquaculture Research 36: 1458-1464.
  19. Olanrewaju AN, Agbelege OO, Saidu AK (2009) Growth and production of fingerlings of Dutch Clarias gariepinus in indoor tanks in arid zone, (FISON) 25-28. pp. 8-10.
  20. Legendre M, Linhart O, Billard R (1996) Spawning and management of gametes, fertilized eggs and embryos in Siluroidei. Aquat Living Resour 9: 59-80.
  21. R Core Team (2018) R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing, Vienna, Austria.
  22. Islam MM, Amin AA, (2016) The induced breeding of common carps (Cyprinus carpio) in Bangladesh. Indian journal of science and technology 23(84).
  23. Liu Y, Hu J, Zhou S, Yang R, Qin J, Ma Z, Yang Q (2018) Effect of Acute Ammonia Stress on Antioxidant Enzymes and Digestive Enzymes in Barramundi Lates calcarifer Larvae. Indian Journal of Anaesthesia.
  24. Ataguba GA, Okomoda VT, Ukpong P (2022) Optimization of induced breeding of Clarias gariepinus using different sources of water. Livestock Research for Rural Development 34: 80-88.
  25. Kudo S (1982) Ultra Structural Modifications of fertilization Envelope in the Carp Egg. Growth and Differ 24 (4).
  26. Utami DP (2010) The Effect of Addition of Pineapple Extract and Different Cooking Times on the Quality of Duck Flesh [Thesis]. Surakarta: Eleven March University.
  27. Ljubobratović U, Csengeri I, Kucska B, Balogh E Lengyel S (2017) Comparison of the Procedures for Adhesiveness Removal in Pikeperch Eggs with Special Emphasis on the Effect of Tannic Acid. Turkish Journal of Fisheries and Aquatic Sciences 469: 461-469.
  28. Thai BT, Ngo TG (2004) Use of Pineapple Juice for Elimination of Egg Stickiness of Common Carp (Cyprinus carpio L.). Asian Fisheries Science 17: 159-162.

Verbal Judo and Prospective Crime – How AI + Mind Genomics Can Synthesize Mind Sets for Training

DOI: 10.31038/MGSPE.2024434

Abstract

This paper shows how the emerging technology of AI (large language models), coupled with the world-view of the importance of mind-sets contributed by Mind Genomics can help negotiators to learn how to deal with situations where violence may take place unless the hostage situation is defused. The approach technique uses artificial intelligence embedded in the Mind Genomics platform to specify the situation, and then to hypothesize the thinking and appropriate actions for the negotiator in the presence of different mind-sets. The paper shows how to interact with AI to synthesize these mind-sets, then how to recognize them, what these mind-sets might be thinking, and suggested communications. The paper focuses on educating police negotiators through the use of easy-to-access technology, and inexpensive, rapid learning by iterating the questions asked to AI. Rather than waiting for actual data, the paper uses artificial intelligence based in large language models to provide information that can be adjusted to a specific situation, providing the opportunity to look at the same issue but from different perspectives provided by the nature of the situation and the mind-set of the person who must be disarmed without causing damage.

Keywords

Artificial intelligence, Mind genomics, Mind-set, Synthesize, Training

Introduction

Look at the daily news to see how many times persons or groups turn hostile to police officers or conduct harm to property or people and must be bargained with by police negotiators. The problem is what police and other negotiators should say in specific instances, when they are immersed in the situation. Negotiations, such as family violence and criminal hostage-taking, requires talent, sensitivity to others, and understanding what to say, knowing what works, knowing what doesn’t. After years of experience, expert negotiators learn what to say and its consequences. How do we teach new hires about new situations? Can negotiation opportunities be synthesized? Artificial bodies designed to offer feedback are used in medicine. Is it possible with artificial minds to teach negotiation skills?

There are many different situations where individuals or groups become hostile to police officers, or where these groups become hostile and do hostile acts to property or to people and must be negotiated with by police negotiators. The issue is what to do in particular situations. The knowledge base of negotiating with criminals is not new, spanning centuries, a fixed part of the human comedy. The characters may change, the causes may evolve, society may change what it considered to be “out of bounds” behavior, but the problem remains [1-3]. In the end where the crime or prospective crime is under the “control” of a person, one may have the opportunity to “talk through” that situation with the person, reaching a positive outcome. What does one say, how does one say it, and when? This ability to talk through requires a skill that must be developed, a sensitivity to individuals, a sense of knowing what to say, what will work and what will not work.

Over many years of experience, professional negotiators begin to understand what they can say, what they should say, the repercussions of what they say. The question becomes, how do we transmit this information for new situations to people who are just coming on the job? Is there a way to synthesize negotiation opportunities?

Advances in artificial intelligence based on LLMs (large language models) have made AI more attractive [4,5]. Data analysis using statistics used to be challenging. Once the data were analyzed, the user faced the onerous task of interpreting the results, and then putting that interpretation into text. The result was often tortured prose rather than felicitous and well expressed. Today’s LLMs now enable an almost-human teaching interface, both at the level of the input, and at the level of the output. More important, however, is the emerging reality that the LLMs need not produce factual results, requiring statistical and factual accuracy. We can instruct the LLM to suggest how reality is structured, and then investigate the different nuances within that constructed reality. Everything offered today is assumed to be generalities presented for learning purposes.

The emerging science of Mind Genomics — which has been developing for almost 40 years — helps when it becomes the precursor to LLM simulation. Mind Genomics is empirical, working with combinations of phrases, having people read the combinations, and rate each one. Through statistical modeling (regression and clustering), Mind Genomics discovers the language which “convinces,” which makes the person answer “yes.” The process uncovers mind-sets, groups of individuals who respond differently to the same set of phrases or messages. These mind-sets are uncovered through empirical evaluations with real people [6,7]. Subjects such as law, medicine, as well as the minds of children and adults have been topics of Mind Genomics studies [8-11].

The combination of Mind Genomics thinking about mind-sets and LLMs to provide “content” generates a new world of opportunities to understand the way people think, doing so quickly by simulation. The approach taken in this paper uses LLMs to provide deep, albeit simulated content, for these mind-sets. In a sense, we use Mind Genomics “thinking” to inform and guide the start of the LLM on a topic area, with the result that we generate a system to learn quickly.

The actual system is embedded in the Mind Genomics platform (BimiLeap.com). The access to the LLM is through the embedded Idea Coach. Idea Coach, in turn, uses an entire structure of inquiry called SCAS, Socrates as a Service. With Idea Coach, the user can type in the topic, and a request to the LLM. The SCAS translates this request, giving it to the LLM. The process is straightforward, requiring only an account, and the proper information provided to Idea Coach.

Phase 1 – Setting up the scenario for the LLM

After creating a Bimileap account and studying, the user is instructed to use Idea Coach to enter ideas. The format is a “squib” similar to that shown in Table 1. One has to be certain that the syntax is correct, and that everything is closed-ended. The LLMs are powerful, but in this case not particularly forgiving. After tries, however, with feedback given in 15 seconds, one eventually understands the syntax and what to do for the specific case.

Table 1: The “squib” or query (background, specific request) given to SCAS (Socrates as a Service), the AI embedded in the Mind Genomics platform, BimiLeap.com.

tab 1

Table 1 shows the request, the aforementioned squib. Table 1 supplies very little information to the AI, Socrates as a Service, but requests a fair amount of testable, actionable information that can be validated because of their concrete nature.

There is no underlying “reason” for selecting six mind-sets, other than the desire to see what the system returns. In fact, across the different iterations, seven iterations emerged, not six. They can be discussed and even tested in real-life.

As a point of information, the brief or squib in Table 1 is the toughest aspect of the procedure and sometimes requires many iterations, viz., attempts. It takes the large language model about 15 seconds per iteration. An inexperienced user might require 20 iterations to get the “syntax” of the request “just right.” Those 20 iterations might require 5-10 minutes.

The artificial intelligence SCAS generates seven mind-sets, shown in Tables 2A-2G, one more mind-set than was specified in Table 1. This over-delivery, as well as corresponding under-delivery, often happens because the mind-sets are developed in separate iterations. Generally, each iteration produces one or two mind-sets. The user ends up requesting more iterations to complete the full number. The under-delivery of one, two, occasionally three, mind-sets in an iteration ends up being not problematic because, as noted above, each iteration requires about 10-15 seconds. The speed and ease to iterate often ends up with some mind-sets repeated in different iterations, and the happy coincidence of more mind-sets than were planned for. Finally, when a mind-set repeats in several iterations, the language surrounding the mind-set may change, and some of the finer points may change from iteration to iteration. Most of the “knowledge” will be the same, but some may be fresh ideas across iterations.

Table 2A: Results for the MACHO Mind-Set

tab 2A

Table 2B: Results for the RADICALIZED Mind-Set

tab 2B

Table 2C: Results for the MENTALLY ILL Mind-Set

tab 2C

Table 2D: Results for the EXTREMISTS Mind-Set

tab 2D

Table 2E: Results for the ANGRY INDIVIDUALS Mind-Set

tab 2E

Table 2F: Results for the CRIMINAL Mind-Set

tab 2F

Table 2G: Results for the DESPERATE INDIVIDUALS Mind-Set

tab 2G

From Tables 2A-2G, it becomes apparent that there is an education to be had. Whether any of these results represent actual mind-sets or not, whether these are the slogans or not, the results are interesting and educational. One even gets a sense of the effectiveness of the rating by artificial intelligence, which is fascinating in and of itself. We might even have put in things like, provide slogans that are not effective, and we could do that for each of these as well. In the interest of space and brevity, we have limited our exploration of the mind-set.

Deeper Analyses through SCAS

Early empirical investigations in Mind Genomics in the late 1990s and onward revealed these mind-sets in topic after topic, study after study. It was the task of the user to understand these mind-sets. Fortunately, the mind-sets were clearly demarcated, and easy to differentiate and to discuss.

In order to discover patterns, SCAS in the BimiLeap program has been set up with a set of fixed analyses in order to extract more information from each iteration. Table 3 shows a summary table about what was learned. The results were analyzed in eight ways, such as key concepts, topics, views, interested audiences, opposing audiences, and three stages toward innovation (alternate viewpoints, what is lacking, and potential innovations). For this paper, this deeper AI analysis was reduced to one summary table.

Table 3: Summarized “deeper analysis” of the results, combining the results from all seven mind-sets, although ordinarily done on a mind-set basis.

tab 3(1)

tab 3(2)

The actual deeper analysis does not occur at the time that the user goes through the iterations. Rather, for each iteration, the material is stored in an Excel workbook as a separate worksheet. Only afterwards, when the study is closed, does the second level of analysis take place. This deeper analysis is done separately for each worksheet, going through all eight steps, returning with recommendations, analyses, themes, and so forth. The seven deeper analyses were put into the “Excel Idea Book.” The Idea Book was returned by e-mail approximately 15 minutes after the project was closed.

It is important to keep in mind that despite the “hand-craft” work to combine information from seven deeper analyses, one per mind-set, the actual sentences and information shown in Table 3 are taken directly from the AI-generated output. Thus, the reality is that we have AI providing second level, deeper analysis of AI-generated ideas, those ideas originally emerging from a simple squib and a paragraph of very little detailed information, viz., Table 1.

Discussion and Conclusions

Artificial intelligence-generated results are only now becoming of interest to scientists in the world of behavior, such as psychologists. It is only now that there is a focus on the nature of ideas generated by AI. The reason may be the increasing use of AI to act as an aid-to-thought, perhaps a “technical aid to creative thought” in the words of the late Harvard professor of computer science, Anthony Gervin Oettinger [12].

In the words of the late Harvard professor may appear off-putting for a scientific publication but they should not be. We are not offering factual knowledge, but frameworks in which one may incorporate certain bits of information to obtain insight. We are not interested in factual information that can be obtained by digging into the data, such as how do the mind-sets distribute in the population, or really anything better understood empirically. This whole text has been a teaching tool to comprehend a possibly extreme situation in an hour or two. The approach presented here empowers novices and professionals alike to describe a scenario and understand its aspects, ranging from attitudes, behaviors, motives, public reactions, and far more.

Much more should be said about this realistic approach. Pragmatism will succeed. In fact, the late psychologist George Miller says that most of one’s knowledge gets “chunked,” with just the general structure maintained [13]. We might consider this approach to be a “cartography,” a mapping of the relevant features of a topic. As a cartographer explores an area, looking at its geographies, lay of the land, and flora and fauna, the professional might explore the mind of a criminal for the best way to communicate.

In the end, the world-view espoused by this paper is pragmatics. The age of enhanced learning is coming. This paper may provide in a short time a new approach, a new way of extracting insights, of synthesizing the real world. The approach demonstrated here emerges not so much by slow learning from books and the formation of concepts, but rather human thinking augmented by the “creative powers” embedded in artificial intelligence, ready to be released through the appropriate prompts. With the power of AI embedded in accessible LLMs, people can now ask more focused questions using a simple set of queries, the power of easy use through SCAS. The final output is made user-friendly, the outcome of the AI efforts returned in the form of easy-to-read material.

In the end, the goal of this project is to suggest ways to think about the topic of communicating in a situation which is potentially dangerous. To reiterate the caveat, we’re not looking for an accurate description of each mind-set, but rather interested in providing a general framework. It would be interesting to see whether or not the untutored person, inexperienced in criminology and verbal negotiations with criminals, would use this information, or perhaps even just specific parts of the information. Which parts are useful? Which parts seem useful, but are really irrelevant to practice? The approach is well worth the try, simply because the format here presented by Socrates as a Service, SCAS, using LLMs, generates the key information easily with virtually instant analyses of iteration after iteration, each iteration returned in 15 seconds, with a deeper analysis an hour later returned by email, and done for each iteration.

Acknowledgement

The authors wish to thank Vanessa Marie B. Arcenas for her assistance in preparing this manuscript.

Abbreviations

AI: Artificial Intelligence; LLM: Large Language Model; SCAS: Socrates as a Service

Competing Interests

The authors have no conflict of interest to disclose.

References

  1. Thomson JB (1870) The psychology of criminals. Journal of Mental Science 16(75): 321-350.
  2. Abrahamsen D (1944) Crime and the human mind. Columbia University Press.
  3. Samenow S (2014) Inside the Criminal Mind (Newly Revised Edition). Crown.
  4. Kasneci E, Seßler K, Küchemann S, Bannert M, et al. (2023) ChatGPT for good? On opportunities and challenges of large language models for education. Learning and Individual Differences 103: 102274.
  5. Chang Y, Wang X, Wang J, Wu Y, et al. (2023) A survey on evaluation of large language models. ACM Transactions on Intelligent Systems and Technology.
  6. Sen A and Srivastava M (2012) Regression analysis: theory, methods, and applications. Springer Science & Business Media.
  7. Duran BS and Odell PL (2013) Cluster analysis: a survey (Vol. 100). Springer Science & Business Media.
  8. Moskowitz HR, Wren J and Papajorgji P (2020) Mind genomics and the law. LAP LAMBERT Academic Publishing.
  9. Gabay G and Moskowitz HR (2019) “Are We There Yet?” Mind-Genomics and Data-Driven Personalized Health Plans. The Cross-Disciplinary Perspectives of Management: Challenges and Opportunities, 7-28.
  10. Moskowitz HR, Gofman A, Beckley J and Ashman H (2006) Founding a new science: Mind genomics. Journal of Sensory Studies 21(3): 266-307.
  11. Mendoza C, Mendoza C, Rappaport S, Deitel Y, et al. (2023) Empowering Young Students to Become Researchers: Thinking of Today’s Gasoline Prices. Mind Genomics Studies in Psychology and Experience Volume 2(2): 1-14.
  12. Bossert WH and Oettinger AG (1973) The Integration of Course Content, Technology and Institutional Setting. A Three Year Report, 31 May 1973. Project TACT, Technological Aids to Creative Thought.
  13. Saaty T (2016) Seven is the magic number in nature. Proceedings of the American Philosophical Society 160(4): 335-360.