Understanding individual differences in social perceptual decision-making
Principal Investigator – David Dodell-Feder, Ph.D.
1. PURPOSE OF STUDY
The purpose of this study is threefold: (1) Evaluate the presence of a mechanism of perceptual decision-making known as robust averaging in social perception; (2) Evaluate its association with psychotic-like experiences (PLE) in a non-clinical sample; and (3) Evaluate its association with social functioning. Our hypotheses are that individuals adaptively down-weight the influence of outliers in perception when making decisions related to social information (e.g., when determining the average facial expression of a group, ignoring extreme, outlying faces)—that is, they display evidence of robust averaging in social perception—that this mechanism is reduced in individuals experiencing PLE, and that this mechanism is associated with adaptive social behavior.
2. BACKGROUND AND RATIONALE
People confront conflicting social information all the time. Consider, for example, being accosted a group of people on a late night walk. While most of the individuals in this group seem friendly enough, smiling your way, several faces appear menacing. Is this group friendly or hostile? From a perception standpoint, how do we extract signal from this information, while adaptively ignoring the noise?
Research on human perception has revealed a mechanism by which individuals solve this very challenge. Specifically, people judge both the strength and variability of sensory information to be integrated. In the presence of variable information (i.e., noise), people adaptively down-weight extreme or outlying pieces of information in a process known as robust averaging1. Similar to the way a statistician may disregard extreme or outlying data points, our perceptual systems down-weight outlying sensory data when integrating information in the context of decision-making.
Several important questions regarding this perceptual mechanism however, remain unanswered. First, while robust averaging has been demonstrated with low-level stimulus features, such as making judgments regarding the average color of a set of different colored stimuli1,2, there’s relatively little research on whether a similar process takes place with social information. Second, altered robust averaging may be an apt way to characterize and understand pathophysiological changes associated with certain mental disorders. Specifically, psychosis has been described as a disruption to evidence-weighting whereby individuals inappropriately assign meaning to weakly supported or noisy evidence3. Researchers have even go so far as to describe the reasoning of individuals with psychosis as that of “bad statisticians,”4 characterized by reduced ability to make use of adaptive perceptual strategies, such as robust averaging. Recent research suggests that in the case of low-level perception, again, identifying the average color of a display, robust averaging is less likely to occur in individuals experiencing PLE2. Given the social cognitive and functioning disturbances often observed in individuals experiencing PLE and psychotic disorders5, it stands to reason that reduced robust averaging with social information may be contributing to those alterations. Third, while the face value of robust averaging social information is clear, we seek to demonstrate empirically for the first time, that this process is linked with real-world social outcomes.
These issues in mind, we will conduct a study to evaluate the presence of robust averaging when processing social information, and the association of this process with PLE and social functioning. Findings have the potential to shed light on critical processes that support social information processing, and how those processes are altered in psychopathology risk states, which in turn, may inform the development of new treatments.
3. ADMINISTRATIVE ORGANIZATION
The participating UR department is the Department of Psychology on River Campus.
4. STUDY DESIGN
The study design is correlational. As such, subjects will not be randomly assigned to activities. All subjects will complete the same measures, which include one measure to assess robust averaging and several self-report measures to assess PLE, social functioning, demographic characteristics, and study experience. All measures will be completed in-person in a single session in Dr. Dodell-Feder laboratory in Meliora Hall on River Campus.
5. SUBJECT POPULATION
300 participants will be enrolled in the study. If a subject withdraws from the study, we will replace them in order to meet our enrollment goal. Enrollment will be open to students in the Department of Psychology’s study pool (SONA). As such, subjects will include students, who are considered a vulnerable population. Students are considered in a subordinate position relative to the experimenters. Additionally, given that study staff includes undergraduate or graduate-level psychology students at the University of Rochester, student participants may be familiar with some members of the study staff outside of a research context. However, student participants are unlikely to have a student-teacher relationship with any members of the study staff. Additionally, students are able to drop out of the study with no penalty. Participation is voluntary. All psychology courses offer alternative routes for extra credit besides study participation.
6. INCLUSION AND EXCLUSION CRITERIA
Inclusion criteria: Native English speaking or native-like fluency; 18 years of age or older; and normal or corrected-to-normal vision.
7. RECRUITMENT METHODS
Potential subjects will be recruited through SONA. An posting for the study will be included in the SONA database, which will describe the study aims, eligibility criteria, duration, and credit compensation (see “SONA Advertisement”). Interested potential subjects will have the opportunity to read about the study and contact the PI before signing up to participate. Those who choose to participate will select a date and time. Potential subjects who sign up to participate may receive a reminder email before their appointment.
8. CONSENT PROCESS
The consent document will be created using a REDCap-based electronic consent form, consistent with Office of Research IT REDCap e-Consent Workflow. Potential subjects will participate in the consent in-person. Specifically, upon arrival at the study session, study staff explains the research study again in detail both verbally and in written form with the consent form. This is done in a private, comfortable room at the Social Cognition & Psychopathology space in Meliora Hall. It is communicated in the consent and via verbal explanation of the study procedures that the potential subject’s participation is voluntary, and the subject can stop at any point without penalty. To ensure comprehension, after describing all procedures, study staff uses open-ended questions to ask the potential subject to describe the study, the procedures, and the risks in their own words. Study staff then answers any questions regarding the research before subjects sign the consent. There is no time limit to this procedure. Formal consent is obtained when the subject provides their written signature (via cursor) on the eConsent via University-owned computer. Once the consent form is signed and submitted, subjects will be able to receive a printout of the paper copy, download a PDF, and/or receive an email with a PDF attachment of the signed consent form.
9. STUDY PROCEDURES
After providing informed consent, subjects will complete the following measures on a computer. We anticipate the entire study taking approximately 30-60 minutes to complete.
Robust Averaging Task: Following prior studies1,2,6, in this task, subjects will complete trials while seated in front of a computer during which they view a fixation cross, see an 8-array display of faces expressing emotions that vary from very negative to very positive, which is masked after several seconds, and then make a judgment regarding the average emotion of the group of faces. Faces are taken from a validated database of face stimuli7 that depict men and women of varying ages displaying happy, angry, or neutral faces of varying intensities with a closed mouth. Example individual faces are depicted below. The experimental task is preceded by a set of instructions read by the experimenter and several practice trials to ensure adequate comprehension of the task.
Figure. Example face stimuli from the robust averaging task. (A) depicts angry to happy morph, (B) depict neutral to happy morph, and (C) depicts neutral to angry morph. Each face is morphed on a continuum from 1 (e.g., least intense display) to 50 (e.g., most intense display). Image adapted from prior study using the same stimuli6.
After completing the robust averaging task, subjects will complete the following self-report measures in a Qualtrics survey (see “Qualtrics Survey Measures”). Embedded within the questionnaire are several control questions to ensure adequate attention (e.g., “Please select option 4 to indicate you’re paying attention.”).
1. Revised Green Paranoid Thoughts Scale8. In this 18-item questionnaire of delusion-proneness, subjects answer questions assessing ideas of reference (e.g., “I often heard people referring to me”) and ideas of persecution (e.g., “Certain individuals have had it in for me”) using a 0-4 scale.
2. Peters Delusion Inventory9: In this 21-item questionnaire, subjects are asked whether they’ve ever experienced a variety of beliefs (e.g., “Have your thoughts ever been so vivid that you were worried other people would hear them?”), and if so, the degree of distress they cause, the subject’s preoccupation with the belief, and the subject’s conviction in the belief.
3. Cardiff Anomalous Perceptions Scale10: In this 32-item questionnaire, subjects are asked about the presence of a variety of perceptual aberrations (e.g., “Do you ever hear noises or sounds when there is nothing about them to explain?”). For any endorsed experience, subjects are then asked about the distress it causes, how distracting the experience is, and how often the experience happens.
4. UCLA Loneliness Questionnaire11: In this 20-item questionnaire, subjects answer questions regarding feelings of loneliness (e.g., “How often do you feel alone?).
5. Multidimensional Scale of Perceived Social Support12: In this 12-item questionnaire, subjects answer questions regarding their perceived level of social support from a variety of sources (e.g., “I have a special person who is a real source of comfort to me.”).
6. Friendship Network Satisfaction Scale13: In this 14-item questionnaire, subjects answer questions regarding the intensity and quality of their friendships (e.g., “I spend free time with my friends”).
7. Demographics Questionnaire: This questionnaire ask subjects about their demographic characteristics including age, sex, gender, race, ethnicity, education, and comprehension of English, among other characteristics. We will also ask questions regarding the subject’s study experience.
10. AUDIO/VIDEO RECORDINGS
Not applicable.
11. RISKS TO SUBJECTS
Potential risks of participation related to the procedures include the following:
• Risk of boredom and fatigue (low risk level).
• Coercion and undue influence (low risk level).
• Breach of confidentiality (low risk level).
• Discomfort or distress when completing self-report questionnaires (low risk level).
In addition to training study staff on study protocols and ongoing discussion and monitoring of subject experience during regular study meetings (see “Data and Safety Monitoring Plan”), the above risks will be minimized in the following ways:
• Risk of boredom and fatigue. Subjects will be offered breaks throughout the testing session.
• Coercion and undue influence. It will be made clear to subjects at first contact and prior to all subsequent testing sessions, that participation is voluntary, they can stop at any time without penalty, and opting out will not affect their treatment in any way.
• Breach of confidentiality and risk of reputation. All data obtained from subjects enrolled in the study will be kept confidential. Subjects will be assigned a de-identified code under which their data will be stored. The database containing identifying information connecting the subject’s name to their de-identified subject ID will be kept in a secure REDCap database. Only IRB-approved study staff will have access to these data. All data collected will only be labeled with the de-identified study ID.
• Discomfort or distress when completing the questionnaires: Subjects can skip questions and there’s no evidence to suggest that answer questionnaires of the type included in this survey leads to persistent discomfort after the survey concludes.
There are alternatives to participation. Specifically, psychology professors may offer an alternative assignment for extra credit in place of participating in the current study. This alternative assignment is determined at the discretion of the faculty member, but involves about the same investment in time per credit point as participating in this study.
12. POTENTIAL BENEFITS TO SUBJECTS
We do not anticipate any direct benefits from participation.
13. COSTS FOR PARTICIPATION
There are no costs for participation.
14. PAYMENT FOR PARTICIPATION
Subjects will be provided 1.0 SONA credit after the end of the testing session.
15. SUBJECT WITHDRAWALS
Subjects will be withdrawn for non-compliance (e.g., refusing to perform tasks in the manner instructed). Any data collected from a withdrawn subject will be kept for analysis. Withdrawn subjects from the study will be replaced. Conditions for subject withdrawals are described in the consent form and verbally explained to subjects during the consent process.
16. PRIVACY AND CONFIDENTIALITY OF SUBJECTS AND RESEARCH DATA
To protect privacy, all study procedures take place in a quiet, comfortable, and private room in the PI’s laboratory with minimal study staff. To protect confidentiality, all data collected in this study is recorded using a subject ID, which is a de-identified string of letters and numbers (e.g., “robavg001”) that does not allow for identification. The key to this coding system, linking subject ID to personal information (name and email) will reside in a secure REDCap database accessible only to study staff. Data will be collected and stored on password protected platforms, such as Qualtrics and Box, that only study staff will have access to. All data will be kept for at least 7 years to abide by guidelines provided by the American Psychological Association (APA). Consistent with American Psychological Association’s ethical principles (https://www.apa.org/ethics/code) and scientific journal requirements, after completion of the study, anonymous study data may be shared on an open science data repository (e.g., Open Science Framework). See Electronic Data Security Assessment Form for additional information.
15. DATA / SAMPLE STORAGE FOR FUTURE USE
As described above, data will be stored on UR Box and the anonymous data will also be available for future researchers on a data repository (e.g., Open Science Framework).
16. DATA AND SAFETY MONITORING PLAN
Data and safety monitoring will be the responsibility of the PI and the study coordinator, with the principal investigator assuming primary responsibility. Monitoring of study activities will occur on a weekly basis. During these meetings, study staff will meet to address issues of recruitment and screening, data collection, human subjects testing issues, and data management and analysis, and to perform ongoing training to ensure diagnostic reliability and fidelity to study protocols. As the current study does not involve procedures that are greater than minimal risk, we do not anticipate adverse events or serious adverse events related to the study procedures. If an event requires reporting, the RSRB will be notified in a timeline and manner consistent with the “Guideline for Reporting Research Events” (http://www.rochester.edu/ohsp/documents/ohsp/pdf/policiesAndGuidance/801a_GDL_Reporting_Research_Events.pdf).
17. DATA ANALYSIS PLAN
Following other studies that have evaluated robust averaging in similar paradigms1,2, we will estimate the weight each participant gives each element in the 8-face array when making decisions by conducting logistic regressions with the element rank as the predictor (rank 1-8 in terms of emotion intensity) and decision (“positive” or “negative”) as the outcome, on a trial-by-trial basis. Element weights (i.e., standardized beta values) will be averaged across trials for each participant. We will conduct two sets of analyses with these weights. First, we will test the quadratic effect of element rank on beta weights with a regression analysis. If participants are robust averaging, we should observe a u-shaped curve in the association between rank and weight, such that outlying ranks are given less weight. Second, we will average the weight of inlying (elements 3-6) and outlying (elements 1, 2, 7, and 8) ranks. If individuals are robust averaging, inlying ranks should receive comparatively higher weights. To evaluate the association between robust averaging, PLE, and social functioning, scores on the self-report measure (e.g., Revised Green Paranoid Thoughts Scale) will be included in the models described above as an additional predictor.
18. REFERENCES
1. De Gardelle V, Summerfield C. Robust averaging during perceptual judgment. Proc Natl Acad Sci USA. 2011 Aug 9;108(32):13341–13346.
2. Larsen EM, Jin J, Zhang X, Donaldson KR, Liew M, Horga G, Luhmann C, Mohanty A. Hallucination-Proneness is Associated With a Decrease in Robust Averaging of Perceptual Evidence. Schizophrenia Bulletin. 2024 Jan 1;50(1):59–68.
3. Kapur S. Psychosis as a State of Aberrant Salience: A Framework Linking Biology, Phenomenology, and Pharmacology in Schizophrenia. AJP. 2003 Jan 1;160(1):13–23.
4. Moritz S, Pfuhl G, Lüdtke T, Menon M, Balzan RP, Andreou C. A two-stage cognitive theory of the positive symptoms of psychosis. Highlighting the role of lowered decision thresholds. Journal of Behavior Therapy and Experimental Psychiatry. 2017 Sep;56:12–20.
5. Green MF, Horan WP, Lee J. Social cognition in schizophrenia. Nat Rev Neurosci. 2015 Oct;16(10):620–631.
6. Ji L, Pourtois G. Capacity limitations to extract the mean emotion from multiple facial expressions depend on emotion variance. Vision Research. 2018 Apr;145:39–48.
7. Tottenham N, Tanaka JW, Leon AC, McCarry T, Nurse M, Hare TA, Marcus DJ, Westerlund A, Casey B, Nelson C. The NimStim set of facial expressions: Judgments from untrained research participants. Psychiatry Research. 2009 Aug;168(3):242–249.
8. Freeman D, Loe BS, Kingdon D, Startup H, Molodynski A, Rosebrock L, Brown P, Sheaves B, Waite F, Bird JC. The revised Green et al., Paranoid Thoughts Scale (R-GPTS): psychometric properties, severity ranges, and clinical cut-offs. Psychol Med. 2021 Jan;51(2):244–253.
9. Peters E, Joseph S, Day S, Garety P. Measuring Delusional Ideation: The 21-Item Peters et al. Delusions Inventory (PDI). Schizophrenia Bulletin. 2004 Jan 1;30(4):1005–1022.
10. Bell V, Halligan PW, Ellis HD. The Cardiff Anomalous Perceptions Scale (CAPS): A New Validated Measure of Anomalous Perceptual Experience. Schizophrenia Bulletin. 2006 Apr;32(2):366–377.
11. Russell DW. UCLA Loneliness Scale (Version 3): Reliability, Validity, and Factor Structure. Journal of Personality Assessment. 1996 Feb;66(1):20–40.
12. Zimet GD, Dahlem NW, Zimet SG, Farley GK. The Multidimensional Scale of Perceived Social Support. Journal of Personality Assessment. 1988 Mar;52(1):30–41.
13. Kaufman VA, Perez JC, Reise SP, Bradbury TN, Karney BR. Friendship Network Satisfaction: A multifaceted construct scored as a unidimensional scale. Journal of Social and Personal Relationships. 2022 Feb;39(2):325–346.
