PREP Research Abstracts

Uncovering Key Contextual Cues Impacting Emotion Understanding

Emotion understanding is people’s ability to reason about what others might feel. Much of the literature has focused on children’s and adults’ use of faces when reasoning about emotions. Here, we consider how people incorporate multiple cues when interpreting emotions and deciding how to interact. Participants (N = 53) were shown various pieces of information and asked how confident they would be in deciding how to interact with an unknown person based solely on that information. For this study, I was most interested in a subset of the items, including external (face and body), internal (hidden emotions and motivations), and social categories (race and socioeconomic status). The differences within evidence categories were not significant (e.g., face did not differ from body). Participants, however, were more confident judging how they should interact with another person based on being able to see external features (e.g., facial movements) than on knowing their social categories (e.g., socioeconomic status).

This research was performed in the laboratory of Professor Seth Pollak.

Leveraging Natural Language Processing and Thematic Analysis to Understand Personal Sensing Acceptability in Opioid Use Disorder Patients

Opioid use disorder (OUD) is a chronic relapsing disease, affecting over 2.1 million people in the United States. Unfortunately, there are few long-term supports for people in recovery. Long-term support may be even less available for marginalized groups with known disparities in access to mental health care. Personal sensing, combined with machine learning, has the potential for scalable, long-term relapse risk monitoring. This study examined if personal sensing methods were acceptable among a national sample of people with OUD. Participants (N=200) provided sensed daily and monthly surveys, geolocation data, SMS and phone call logs, and text message content for up to one year. After using the methods for 3-4 months, participants provided open-ended responses about their experiences with each personal sensing method. Thematic analysis and topic modeling were performed to identify common themes among participants and to see if there were differences in the proportion of themes by data type and demographic group (race, gender, and income). Results showed that participants’ comments were largely positive across sensing methods. Participants reported more benefits with active methods (daily and monthly surveys) and more comments related to trust and privacy with the passive methods (geolocation, call logs, SMS text message content). There were no significant differences across themes for gender and income. However, we found significant differences in the proportion of comments related to trust, benefits, sustainability, and positive/negative affect between White participants compared to Black and Hispanic participants. A structural topic model with race/ethnicity and data type as covariates yielded 15 topics. We found significant differences between White participants and Black and Hispanic participants on six topics. This research highlights the importance of considering group differences in the early stages of treatment design and adapting these tools so that supports are equitable and accessible for everyone.

This research was performed in the laboratory of Professor John Curtin.

Hormonal Interventions and Cognition: Experimental Analysis on Lupron-Treated Rats

​Lupron is a GnRH agonist that is commercially used in treatment of precocious puberty, which when used as a puberty blocker allows transgender and nonbinary youth the option of delaying puberty related changes before deciding to take masculinizing or feminizing sex hormones. While previous work has supported the safety and utility of Lupron in gender affirming healthcare, further research is needed to clarify the relationship between Lupron mechanisms specifically and behavioral/cognitive outcomes. The present study sought to clarify the relationship between Lupron administration and cognitive outcomes of spatial learning and memory. 7 adolescent rats were randomly assigned to receive low, high, and placebo doses of Lupron and were assessed for impact on spatial learning and memory using the Barnes Maze task. Subjects completed 3 days of acquisition trials as well as three days of reversal trials one week later. Results indicated no significant differences in task acquisition across Lupron dosages, suggesting that Lupron does not impair the ability of rats to learn the task initially. We found a significant relationship between Lupron administration and reversal learning, such that subjects treated with a high dose of Lupron found the target hole much faster compared to the control and the low dose groups. The faster performance by Lupron treated rats on the reversal task suggests they might have enhanced cognitive flexibility​, providing further support and clarification on the role of Lupron on potential cognitive outcomes in gender affirming healthcare.

This research was performed in the laboratory of Anthony Auger.

The Influence of Anxiety and Irritability on Preadolescent Neural Connectivity

Anxiety and irritability are highly prevalent symptoms in youth with psychopathology, frequently appearing comorbidly. Understanding the neural basis of these symptoms may further the development of targeted, dimensionally-informed treatments. This study examined anxiety and irritability in 218 untreated preadolescent children, ages 8-12, with current anxiety disorders (n=141) or no psychopathology (n=77). An exploratory factor analysis (EFA) of parent and child reports of anxiety (SCARED scores) and irritability (ARI scores) examined the latent factors associated with these overlapping symptom domains. For a subset of participants (n=131), seed-based resting functional magnetic resonance imaging (fMRI) analyses examined associations between EFA-derived latent factors and connectivity of several brain regions associated with anxiety and irritability. The EFA uncovered three latent factors: anxiety, parent-rated irritability, and child-rated irritability. Resting fMRI analyses revealed differential correlations between EFA factors and neural network connectivity. For instance, for the posterior cingulate cortex seed, higher anxiety predicted increased connectivity with the middle occipital gyrus, whereas higher parent-rated irritability predicted decreased connectivity with the middle frontal gyrus. EFA results support the need to consider informant effects, particularly when assessing irritability in youth. Biological distinctions between children with different symptomologies imply that symptom-specific interventions may be necessary when treating anxious youth.

This research was performed in the laboratory of Ned Kalin.

Parental Support Factors of Cross-Race Friendships

Cross-race friendships impact youth in positive ways, however they occur at low rates. Parents also play an important socializing role in children’s friendships. So we investigated parents interest in having cross-race friendships for their children and their potential misperception of other parents’ interest in cross-race friendships for their children. We recruited 25 Black and White parents with 8-13 year old children to complete a survey through Prolific. The results showed that parents self-reported having a high interest in their children having cross-race friendships but underestimated in and out group parents’ interest in children having cross-race friendships. This is consequential because this misperception could cause parents to not seek contact with out group parents because they think they are not interested in these friendships for their children. Furthermore, parents may be shaping their behavior and making decisions in regards to their children based on misperceptions that can create barriers to the types of friendships their children can build. This research project is important to be able to correct these misperceptions to allow parents to be aware of other parents’ interest in children having cross-race friendships.

This research was performed in the laboratory of Professor Kristin Shutts.

Doctor-Patient Communication: What do Patients Prefer?

Effective doctor-patient communication is crucial for optimal health outcomes, significantly influencing patient satisfaction, treatment adherence, and overall health (Shen et al., 2018). Despite its critical role, evidence shows that doctors communicate differently with Black patients compared to White patients. Medical communications with Black patients are typically shorter, less patient-centered, and characterized by less positive affect. However, little is known about how patients prefer to be addressed and if there are differences in the communication preferences of Black and White patients. To address this gap, we surveyed 349 Black and White participants (Mage = 37.93, SDage = 12.44; 237 Women) on Prolific. They answered a total of 65 questions across five categories: Communication Preferences, Trust in Doctors, Information Preferences, Cultural Skills, and General Attitudes towards doctors. We found that there are no significant differences in preferences between how Black and White patients want to be treated across all categories, except for one item in the Cultural Skills section. Black participants scored higher when asked whether they wanted their doctors to consider their spiritual beliefs when suggesting treatment. There was also a significant difference in general attitudes, with Black participants showing more positive attitudes towards doctors. These results indicate that efforts should be made to minimize the gap in communication and train doctors to show the same level of respect, engagement, and communication with Black patients as they would with White patients.

This research was performed in the laboratory of Professor Markus Brauer.

Association Between Neighborhood Deprivation and ADHD in Racial-Ethnic Minority Youth

Recent studies have shown that the prevalence rate of ADHD is not equally distributed across race and ethnicity (Hoffmann et al, 2022; Coker et al., 2016; Adams et al., 2024). In addition to inconsistent reports of racial-ethnic differences in the rate of ADHD across Black, Hispanic, White and Asian youths, few studies provided clear explanations of these group differences. This poses a significant challenge in understanding and addressing racial and ethnic disparities in youth mental health. On the other hand, an increasing amount of research has focused on understanding how various aspects of the environment — also known as social determinants of health, or SDoH — contribute to mental health. Notably, many studies suggest that the impact of SDoH on mental health may vary across race-ethnicity. However, few studies have focused on ADHD, and it remains unclear whether SDoH may confer differential effects on ADHD depending on racial-ethnic status of individuals. To that end, we tested the associations  of neighborhood deprivation (as measured by the Area Deprivation Index from the US Census) with ADHD symptoms, and compared across White, Black, Hispanic, and Asian American youth from the Adolescent Brain Cognitive Development study. Using a multiple regression analysis, we found that there were small but statistically significant differences in ADHD symptoms by race-ethnicity. Additionally, racial-ethnic differences in ADHD were attenuated after accounting for neighborhood deprivation. Finally, there was a significant interaction between race-ethnicity and neighborhood deprivation; specifically, the association between neighborhood deprivation and ADHD symptoms is weaker among Black children relative to White children. Findings from this study highlighted the role of the neighborhood deprivation in explaining racial and ethnic differences in ADHD symptoms. Future research will need to also examine other prominent SDoH factors in relation to ADHD symptoms.

This research was performed in the laboratory of Professor James Li.

Sexual Assault Typologies Among College Students: A Latent Class Analysis

Sexual assault, defined as nonconsensual sexual contact through completed penetration, is an unfortunately common occurrence associated with a range of physical and mental health conditions. Historically, sexual assault has been treated as a singular construct, leading to one-size-fits all prevention programs that have not successfully reduced rates of sexual assault. The current study proposes to use latent class analysis (LCA) to identify distinct patterns in the variability of sexual assault experiences and explore relationships between differing typologies and participant disclosure and demographic variables. 322 participants with a history of an unwanted sexual experience reported on their demographics and sexual assault experiences using the Sexual Experiences Survey Revised. LCA results yielded 4 classes of distinct sexual assault patterns, with classes differing on factors such as perpetrator relationship, type of assault, assault setting, presence of others, and whether substances were involved. Chi-square analyses yielded significant associations between class membership and all disclosure types (no disclosure, informal, formal, and disclosure to the perpetrator) as well as a significant relationship between class membership and gender. No significant associations were found between class membership and Race/Ethnicity or Sexuality. Significant findings underscore the complex interplay between disclosure patterns of college students experiencing sexual assault, as well as the differential prevalence rates of sexual assault across genders. ​

This research was performed in the laboratory of Professor Kate Walsh.

Estrogen and Aggression: Rapid Effects of Estradiol Intranasal Injections on Agonistic Behavior in Monogamous, Biparental and Territorial Mice

Previous studies looking at the rapid effects of testosterone have highlighted the crucial role of aromatase activity and estrogen receptors in modulation behaviors such as aggression. Estradiol takes effects by binding to estrogen receptors with estrogen receptor alpha being implicated in increased aggression in male California mice to investigate the role of estradiol on aggressive behaviors in both males and females, we used a series of intranasal estradiol injections paired with a Resident-Intruder (RI) paradigm. Twelve unpaired California mice were subjected to a randomized concentration (1, 2, or 4 mg/ml or saline control) of estradiol and underwent the RI test four times in a period of eight days. Vaginal lavages were performed on females to determine estrus cycle. Results showed no significant changes in aggression across estradiol dosages. However, sex-specific analysis revealed higher baseline and test aggression in males compared to females, who exhibited overall decreased aggression. No significant changes were observed in stress/anxiety-like behaviors between sexes or across dosages. Future directions will explore the effect of estradiol in the vocalizations and additional testing with higher concentrations of estradiol to better understand its impact in aggressive and stress/anxiety-like behaviors.

This research was performed in the laboratory of Professor Catherine Marler.

Linking Differences in Language Use to Social Categories

Adults constantly infer social information (e.g., sex, age, place of origin, etc.) from the way someone speaks. While previous research has shown that children are sensitive to speaker identity when learning individuals’ language patterns, less is known about whether children can link speakers’ distinct grammatical patterns to different social groups. The current study addresses this gap by asking 4- to 5-year-olds to play an iPad game where they learned the different plural marking patterns of two visually-distinct alien groups: one group consistently uses po and the other ka. At test, children were tasked with listening to novel speakers and selecting which of the two alien groups these speakers belonged to based on their description of plural pictures. If children use linguistic differences to index social groups, they should be able to accurately categorize these novel speakers. Our preliminary results showed that children did not reliably select the corresponding alien group above chance. One possible account for these results is that this artificial social contrast did not capture the complexity of real-life social categories. We have thus increased the salience of the social category by adding further attributes to our social groups and are collecting data on this new version.

This research was performed in the laboratory of Professor Jenny Saffran.

Does font size play a role in metacognition and memory?

Metamemory refers to individuals’ cognitions concerning memory processes and performances. One approach to studying this phenomenon involves the application of Judgments of Learning (JOL), which entail individuals’ predictive assessments of their future memory performance. These judgments can be influenced by various cues, with perceptual fluency, including factors such as font size manipulation, being a significant consideration. Notably, certain studies have identified a metacognitive illusion when comparing different font sizes. This illusion arises due to a substantial disconnection between individuals’ JOLs and the subsequent accuracy of recall. Specifically, individuals tend to regard larger font sizes as more memorable, despite no substantive variance in actual recall performance across font sizes. In order to explore the existence of this metacognitive illusion, our research expands the scope of evaluated font sizes (ranging from 1pt to 500pt) and introduces additional test sessions. In our preliminary findings, the pattern of predictions (i.e., JOL) closely paralleled the actual memory performance. Moreover, successive rounds of evaluation revealed adjustments in JOLs, resulting in a narrower gap between JOLs and recall accuracy, coupled with improved recall precision. These outcomes suggest the possibility that the purported metacognitive illusion may, in fact, constitute an overestimation rather than a genuine cognitive phenomenon.This research was performed in the laboratories of Professor Shawn Green and Professor Haley Vlach

Investigating children’s understanding of integer magnitudes: Does a lesson with number lines help?

Integers are whole numbers with no fractional parts that include both positives and negatives. In the United States, children typically do not learn about negative integers until the sixth grade. Few studies have examined children’s understanding of negative integers or how children’s understanding can be improved through instruction. Children often have difficulty understanding the concepts and procedures associated with negative integers. Thus, it would be valuable to know what children understand and what types of lessons are most effective. In the current study, we measured dimensions of children’s conceptual and procedural knowledge of integers. We also designed a lesson to test whether comparing magnitudes with number lines enhances integer understanding more than a lesson without number lines. We hypothesized that the fifth- and sixth-grade participants would demonstrate emerging understanding of negative number magnitudes, emerging ability to solve addition problems with negatives, and would also display understanding of the additive inverse principle (i.e., x + -x = 0). We further hypothesized that a lesson with practice comparing magnitudes would enhance children’s understanding—especially a lesson with number lines. Indeed, in a small sample (N = 12) we found that children displayed an emerging understanding of negative integer magnitudes, addition with negative integers, and good understanding of the additive inverse principle. Moreover, after practice comparing magnitudes, children displayed better understanding of negative magnitudes and better understanding of the additive inverse principle. However, children randomly assigned to a lesson with number lines did not perform better than children randomly assigned to a lesson without number lines. This research has implications for both theory and practice.This research was performed in the laboratory of Professor Martha Alibali

“Psychopathology” versus “Abnormal”: Empirically testing the differences

Many university Psychology departments have recently changed the name of their “Abnormal Psychology” courses to “Psychopathology.” Similarly, the 115-year-old Journal of Abnormal Psychology recently changed its name to Journal of Psychopathology and Clinical Science. The change has been motivated by the assumption that “abnormal” is a pejorative trope, and “psychopathology” has been selected as a viable alternative. However, there is no evidence that “psychopathology” will not also be perceived negatively. Our study, which assesses Psychology students and comprises both implicit and explicit measures, empirically investigates the terms “Abnormal Psychology,” “Psychopathology,” and “Psychology of Mental Health.” For our implicit measure, by showing participants photos of hypothetical university professors, we gain an indirect measure of participants’ preference for each term. For our explicit measures, we use both two-alternative forced-choice items and three-alternative forced-choice items to assay students’ judgments about each term. For all measures, we expect that “Abnormal” and “Psychology of Mental Health” will be preferred over “Psychopathology.” If we observe that the terms Abnormal Psychology and Psychology of Mental Health are preferred over the term Psychopathology, our study will have strong implications for those departments and journals that have recently changed their names.This research was performed in the laboratories of Professor Morton Ann Gernsbacher and Professor Gary Lupyan

Caregiver factors matter when comparing vocabulary checklists and looking-while-listening

The looking-while-listening (LWL) task (Fernald et al., 2008) and the Macarthur-Bates Communicative Development Inventories (MCDI; Fenson et. al, 1993) are both assessments used to measure infants’ word knowledge. The LWL task assesses infants’ word comprehension by measuring infants’ looking behavior following an auditory stimulus (i.e., word), while the MCDI measures infants’ vocabulary by asking caregivers to report the words their child does not understand, understands, or understands and says. As both assessments aim to capture infants’ word understanding, the words a child understands on the LWL task should align with the words their caregiver reports to be understood or said on the MCDI. However, prior studies assessing the alignment between caregiver report and the LWL task have found conflicting results; caregiver report sometimes accurately reflect infants’ LWL behavior (Styles & Plunkett, 2009), but often underestimates the number of words a child understands (Houston-Price et. al, 2007). These prior studies have not yet investigated specific caregiver factors that may influence alignment between these two measures such as caregiver reporting confidence and amount of time spent with their infant. Therefore, the current study examines whether LWL behavior and caregiver report of vocabulary aligns better when we account for caregiver factors in a sample of 13 18–20-month-olds. The current study analyzes both LWL behavior and MCDI data from 13 18–20-month-olds and their caregivers. We hypothesized that (1) caregiver report will align better with infants’ LWL behavior when caregivers are more confident about their response for a specific word and (2) caregiver report will align better with infants’ LWL behavior when caregivers spend proportionally more time with their infant. Our preliminary results suggest that caregiver report of vocabulary and the LWL task are better aligned when caregivers are more confident in their responses. Our preliminary results also suggest that the amount of time a caregiver spends with their infant influences alignment for the words that caregivers report their infant to understand, but less so for the words reported to be said or not understood. Data collection is currently ongoing. If these preliminary patterns persist with a full sample of participants, then researchers should consider adding questions about specific caregiver factors to the MCDI to improve caregiver report of vocabulary, as it is a common measure of language development used to identify early language delays.This research was performed in the laboratory of Professor Jenny Saffran

Social Norms, Referent Groups, and Support for Regulating CO2

Existing literature on social norms show that norms can influence individual behavior. These norms may be more impactful when individuals identify more closely with the referent group establishing the social norm. The focus of this research is to explore the impact of referent groups on environmental behaviors and attitudes. In study 1, we approached individuals in public with a request to sign a petition supporting the regulation of CO2 as a pollutant. Participants (n = 88) listened to one of four variations of a script which referenced whether people in a country, state, or county supported the regulation of CO2; there was also a control condition which did not have any referent group. We hypothesized that individuals in the county condition would be more likely to sign our petition, since they may identify more closely with individuals in their county, opposed to people in regions further away. Preliminary data did not show support for this hypothesis; participants in the state condition had higher signing rates. In study 2, we recruited participants (n = 1053) from CloudResearch. Similar to study 1, participants were placed into one of four conditions which corresponded to their country, state, or county of residence; there was also a control condition without a referent group. Study 2 results did not show statistical differences in support for regulating CO2 as a pollutant across the 4 conditions. If preliminary findings for study 1 persist in a greater sample size, our studies may suggest that an individual does not necessarily have to change their attitude to change their behavior.This research was performed in the laboratory of Professor Markus Brauer

Perceptual Learning: Dark-is-more Bias and Generalization

Humans tend to map darker colors to larger quantities; a phenomenon called the dark-is-more bias. While previous work has shown that information is extracted more rapidly and accurately from data presented in a way congruent with this bias, it is unclear whether this bias would also influence the learning of a new mapping between color and a response (here a horizontal distance on screen). In addition, given the dual continuous and categorical nature of the color space, we investigated to what extent the generalization to new color-distance couples was done either in a rather continuous fashion (i.e. participants interpolated between previously learned colors) or in a categorical fashion (i.e. participants tend to group new colors with previously seen colors). To investigate the first question, we had participants learn the mapping between four colors (exemplars) selected from the magma+ colorbar and the position of that color on the colorbar. These mappings were presented with the ordering of the colors either congruent (Dark-Far) or incongruent (Light-Far) with the dark-is-more bias. For the second question, once the participants had learned the position of the four exemplars, they reported the distance associated with new colors picked from the magma+ colorbar. Preliminary results indicated that (1) learning speed was greater in the Dark-Far compared to the Light-Far condition, suggesting that the Dark-is-More bias may affect the learning of new color-concept mappings, and (2) on a population level, participants generalize both in a continuous manner and in a categorical manner, although with very with strong interindividual differences.This research was performed in the laboratories of Professor Shawn Green and Professor Karen Schloss

Due to differences in experiences, do certain groups make their facial expressions more interpretable than others?

According to social-functional theory, our facial expressions signal important information to others. Smiles serve a number of purposes mostly depending on the type of smile. There are three that are currently known that have a purpose and solve some fundamental task of social living; however, this study will only be looking at two of them: the reward smile and the affiliations smile. The reward smile serves to reinforce behaviors that caused the smile, and the affiliation smile serves to create and maintain social bonds with others. For this study, we aimed to see if certain groups felt a need to make their facial expressions more clear compared to other groups. One specific group we looked at are black men due to the stigma of being dangerous or unsafe. Thus, a survey was made comparing 5 black and 5 white male face identities across 3 facial expressions for each identity (30 total images). Participants (n=18) were instructed to complete a survey of 500 total trials of triplet judgment tasks that were split into two blocks (250 trials per block). For each trial, participants were shown 3 images with one target image at the top and two choice images at the bottom, and their task was to select the choice image with the facial expression more similar to the image at the top. Once all responses were collected, an embedding was created mapping each image into clusters that participants found to be most similar to one another. In the future, we plan to analyze the results by comparing the polygon area of the black identity clusters to the white identity clusters to see which one clustered more tightly, and thus which group expressed more clearly based on participant responses. We also plan to measure the distance between each black and white identity cluster to see which group made their expressions to be more distinguished from other facial expressions. This research was performed in the laboratory of Professor Paula Niedenthal

Gene-environment interactions on trajectories of externalizing behaviors in African Americans

A recent Genome-Wide-Association Study by Karlsson-Liner and colleagues (2022) identified genetic variants that were associated with externalizing behaviors (EXT). Based on their study, a polygenic score (PGS) can be generated to estimate an individual’s genetic liability for EXT, which explains up to 10% of the variance in EXT-related traits. However, evidence is mixed regarding whether the effect of PGS on the development of EXT is moderated by psychosocial environments. Moreover, due to various complications in studying genetic effects across genetic ancestry groups, literature on PGS x Environment interaction in racial-ethnic minority groups is rare. Using five-wave longitudinal data from the National Longitudinal Study of Adolescent to Adult Health, we tested whether PGS interact with adolescent psychosocial environments (I.e., parental warmth, affiliation with delinquent peers, and school and neighborhood connectedness) in predicting later developmental trajectories of EXT in African Americans. Two dimensions of EXT, disinhibition and antagonism, were measured by substance use (e.g marijuana, tobacco, and alcohol use) and antisocial behaviors (e.g theft, aggression, property destruction), respectively. This is one of the first studies in this area working to diversify and expand knowledge on gene-environment interactions within racial-ethnic minority groups. The most significant results are between EXT and affiliation with delinquent peers. There is a strong direct correlation between increased affiliation with delinquent peers and all of the non-typical anti-social behavior trajectories. Results also suggest that an increase of affiliation with delinquent peers would directly correlate to an increase of falling within the high substance use trajectory.This research was performed in the laboratory of Professor James Li

Impairing Olfactory Bulb Adult Neurogenesis using Intranasal Recombinant Adeno-Associated Virus

The process of new neuron growth in adulthood – known as adult neurogenesis – regulates cognition, anxiety, and social behavior. The lion’s share of adult neurogenesis research in mammals has focused on a key neurogenic region, the hippocampus; however, another neurogenic region in the adult mammalian brain exists, and little is known about the function of neurogenesis there. The olfactory bulb regulates social behavior in many rodent species, including that of the California mouse used in the Marler Lab, yet there is a gap in our understanding of adult olfactory bulb neurogenesis on rodent social behavior. To begin to understand this relationship, we must first identify a targeted, minimally invasive way in which to manipulate adult neurogenesis in the olfactory bulb. Combining a recently described viral method to ablate neurogenesis with intranasal delivery, we sought to ablate olfactory bulb neurogenesis in adult female and male California mice. Mice were administered a synthetic nucleoside used to label dividing cells in the olfactory bulb, and one week later were given an intranasal infusion of either viral particles or control saline. One week later, olfactory bulbs were examined for viral transfection and a reduction in recently divided cells. Our initial analysis showed that the virus significantly decreased BrdU-positive cell number in female when controlling for weight (P < 0.034). The data suggest that the virus transfects into the olfactory bulb via an intranasal pathway, and that greater and deeper transfection occur in females compared to males. Our preliminary analyses also suggest that the treatment partially ablated neurogenesis in females, but not males. Whether the sex-specific effects we observed are due to dosage, incubation time, or differences in physiology requires further analysis, however they suggest that sex may be an important factor when considering gene therapies for both research and clinical applications. This research was performed in the laboratory of Professor Catherine Marler

Examining the Relationship between childhood trauma and conduct reports in Incarcerated Individuals

Incarcerated individuals report significantly higher rates of childhood trauma exposure than the general population, with 97%of these individuals experiencing at least one adverse childhood experience. Trauma reactions have been shown to worsen behavioral outcomes, increase punishment, and further prevent release from incarceration.  Studies consistently find physical and sexual abuse to be most strongly related to institutional misconduct. The literature has established the relationship between physical abuse and later violent behavior during incarceration. Less is known about other forms of institutional misconduct that also indicate maladjustment to the prison environment. This study examines the relationship between childhood physical neglect, which is infrequently studied, and substance-related misconduct during incarceration. This study employed multiple linear regression analyses on samples of male (n= 369) and female (n= 122) incarcerated individuals to examine the relationship between childhood trauma (measured through the Childhood Trauma Questionnaire) and institutional conduct (provided by the Department of Corrections). We hypothesized that (1) greater childhood trauma severity would predict more conduct reports; that (2) greater physical neglect severity would predict more conduct reports; and that (3) greater physical neglect severity would predict more substance-related conduct reports. Finally, an exploratory analysis examined the relationship between childhood sexual abuse and conduct reports. Though the current study did not find significant relationships between childhood trauma and conduct reports, this line of research may be helpful in better informing trauma-related practices aimed at understanding how individuals with a history of childhood trauma adjust to the prison environment. The lack of a significant relationship between childhood trauma and misconduct implies that the psychosocial consequences of trauma exposure on their own were not severe enough to cause behavioral disruptions in our sample. This research was performed in the laboratory of Professor Mike Koenigs

Learning about Temporal Dependencies in an Orientation Discrimination Task

In many visual perception experiments involving a binary decision, stimulus trials are temporally independent of each other. This means that the likelihood of a current outcome is not influenced by the outcome of a previous trial (i.e., the probability of either outcome on a given trial is 50/50 as in a coin flip). However, people hold misconceptions about what random processes look like and often judge patterns of outcomes that involve frequent “switches” (more than would be expected in a truly random sequence) as being representative of a random process. This phenomenon is commonly referred to as gambler’s fallacy. In a visual perception task, this creates an effect where instead of trials being evaluated independently, the outcome on previous trials influences participants’ behavior on current and future trials. Our study looks to understand 1) whether participants can learn to make use of temporal dependencies between trials in an orientation discrimination task and 2) whether they can generalize what is learned to multiple spatial frequencies. Participants completed 900 trials of a Gabor orientation discrimination task which contained three 300-trial blocks of different spatial frequencies. The temporal dependencies between trials were such that there was an 80% chance of the current trial being a different orientation (i.e., clockwise or counterclockwise) than the previous trial. We expect participants to learn about and make optimal use of temporal dependencies such that when orientations are difficult to distinguish, they rely on the temporal statistics to inform their decision making. This study seeks to determine the extent to which participants make use of temporal statistics in decision making and whether this learned behavior generalizes to other similar conditions. This research was performed in the laboratory of Professor C. Shawn Green

Facilitations promoting social connections between children: How can we help children have successful social interactions?

Positive social interactions are important for children because they promote exploration and learning, provide protection in difficult situations, validate one’s sense of self, and provide cognitive and social benefits (Cappella, et al., 2013; Graham, et al., 2014). Although children sometimes find it easy to make friends on their own, there are situations where making friends is difficult for children (e.g., when a peer is from a different social group; Shutts, 2015). The present study aims to identify facilitations that promote social connections between children meeting for the first time, and test whether those facilitations will be especially effective for promoting social connections between children of different genders. Data collection is still ongoing, but we aim to recruit 288 5- to 10- year-olds who will all be paired with another participant of their own age; half of the pairs will be same-gender pairs and half will be mixed-gender pairs. In the study, pairs are randomly assigned to complete an interaction activity with different levels of scaffolding (i.e., structure provided by the adult) and agency (i.e., the opportunity to lead the interaction activity). Dependent measures are focused on the quality of participants’ later free interactions as well as participants’ self-reports on constructs related to feelings of connectedness to their partners. We expect that pairs assigned to conditions with high scaffolding and agency to have the most successful free interactions and report the greatest connectedness. Additionally, we predict that scaffolding will be especially effective in promoting social connections between children of different genders. Observing these predicted results will have implications for developing effective tools to foster social connections among children in contexts where children may not be able to effectively form social connections on their own. This research was performed in the laboratory of Professor Kristin Shutts

Conceptual Structure of an AI

Transformer-based large language models such as GPT-3 are trained on most of the text on the internet and are capable of generating human-like language. Given that these models are increasingly being used for making important social decisions like insurance premiums, and in customer-facing chat boxes there is a need to understand what these models’ biases are. Using methods from cognitive psychology, we prompted GPT-3 with a feature listing task and triadic judgment task using animal and tool concepts and compared its representations to human data. After establishing our methods with these more concrete concepts, we extended our approach to study people’s and GPT3’s representations of social concepts like race and gender.  Results seem to indicate similar biases in human and GPT-3’s social concepts. This research was performed in the laboratory of Professor Tim Rogers

The Brain Mechanisms That Track Context vs. Content in Working Memory

Working memory (WM), the ability to temporally maintain a small amount information in mind for guiding current and future behavior, is extensively involved in our everyday lives. Successful WM performance often requires the binding between stimulus-specific identity (content) and trial-unique context (e.g., remembering that chill pepper was on the second shelf in the kitchen cabinet). Previous studies have demonstrated that intraparietal sulcus (IPS) is sensitive to context-binding demands for non-spatial visual features, with manipulations of stimulus homogeneity (e.g., Cai et al., 2020; Gosseries, Yu, et al., 2018) and task demand (Fulvio et al., presented at CNS 2022). In this study, we further investigated the role of IPS in context-binding for spatial WM. Human subjects (n = 13) performed a dual-serial retrocueing (DSR) task during fMRI scanning. They memorized two items distinguishable by their locations and orientations, and were tested on the location (content) of the items. The item to be tested on was retro-cued by its orientation (context). Memory load for content and context was manipulated separately: load of 1 or 2. For content, increasing load from 1 to 2 led to slower and less accurate recognition performance. Delay-period activity in IPS also increased with higher content load. Critically, for context, behavioral performance showed the opposite pattern as for content: worse with a load of 1 than a load of 2. Meanwhile the activation in IPS was still higher on load-of-2 trials (easier behaviorally) than load-of-1 trials (more difficult behaviorally). Together, these results suggest that IPS is particularly important in processing context rather than representing task difficulty, and its involvement in context-binding is generalized across different domains of visual WM (spatial and non-spatial). This research was performed in the laboratory of Professor Brad Postle

Priority-Based Representational Transformations in Visual Working Memory

Working memory (WM) is critical for successful behavior in everyday situations by allowing us to briefly hold limited amounts of information for quick access according to task demands (e.g., recalling someone’s name when switching among new acquaintances in conversation). While it is known that WM can hold multiple items simultaneously, the representational scheme of prioritized memory items (PMI) and unprioritized memory items (UMI) is undetermined. Previous human neuroimaging studies suggest that UMI might be represented in an activity-silent manner via synaptic mechanisms (e.g., Rose et al., 2016), whereas other recent work has shown that UMI can be actively represented, albeit in a different format (Yu, Teng, & Postle, 2020; Wan et al., 2020; van Loon et al., 2018). To further assess the nature of the UMI representation, we used an existing EEG dataset collected while participants completed a double serial retrocuing visual working memory task (Fulvio & Postle, 2020), which involves flexible prioritization of one item among a memory set of two. Demixed principal component analysis (dPCA) was employed to project condition-averaged data into priority-specific subspaces. A scalar transform metric was used to quantify the priority-based transformations over time in these subspaces. Our results show that PMI and UMI representations are neurally separable and dynamically transformed according to their priority status. Furthermore, an inspection of individual data revealed that the PMI and UMI representations seemed to be reflected with respect to both dPC axes in the same priority-specific subspace. In combination, these results shed light on a new possibility that both PMI and UMI undergo representational transformations in WM. This strategy may be advantageous in reducing interference between memory items to serve the current behavioral goal. This research was performed in the laboratory of Professor Brad Postle

Isolating the opaque-is-more bias for colormap data visualizations

Colormap data visualizations are seen in our everyday world, including political maps during election season and humidity maps during summer. These maps use colors to represent quantities across areas. When people interpret colormaps, they are influenced by expectations of how colors map to meanings in visualizations. The dark-is-more bias leads people to infer that darker areas represent larger quantities. The opaque-is-more bias leads people to infer that more opaque areas represent larger quantities. Past work studying the opaque-is-more bias used colormaps that vary in both opacity and in lightness, meaning the dark-is-more bias was also applicable. Our experiment aimed to isolate the opaque-is-more bias and test whether it influences inferred mappings after eliminating lightness variation. We addressed this question by presenting participants with colormaps that varied in chroma while holding lightness constant (L*= 50 in CIELAB space). Participants saw colormaps on either a chromatic background or a gray background.  If the opaque-is-more bias is still present when lightness variation is held constant, then participants should choose the more chromatic side on a gray background and the gray side on a chromatic background. We found that on both backgrounds participants were more likely than chance to choose the opaque side, providing evidence for the opaque-is-more bias. However, this likelihood was reduced on the chromatic backgrounds, when the more opaque region was gray (low chroma) (Fisher’s Χ²: p < .001). This reduction on a chromatic background may be due to a new potential bias, a high chroma-is-more bias, in which people infer more chromatic colors represent more quantity. This bias would conflict with the opaque-is-more bias on chromatic backgrounds. These results contribute to our understanding of people’s expectations about colormap data visualizations, a common component to our everyday communication. Understanding these expectations will help design effective visualizations to improve visual communication. This research was performed in the laboratory of Professor Karen Schloss

PREP research projects are possible based upon support from the National Science Foundation under Grant No. 2050782

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

Motivation to be Anti-Racist: An Exploratory Study

In the wave of the Black Lives Matter and George Floyd protests in 2020, anti-racism has transitioned into the mainstream media. As a relatively new concept to the academic community, recent research has yet to provide a clear and concise definition of anti-racism. The first goal of this study is to provide a definition that addresses the nuances of racism through its three levels: personal bias, interpersonal, and systemic. Previous literature has yet to address how motivation can be tied to the three different levels of anti-racism. To address this gap, the second goal of the present study is to examine potential motivators for participation in anti-racism and the extent to which such motivators differ at the different levels (personal bias, interpersonal, and systemic). Participants will be recruited using Prolific and complete an extensive survey on several potential motivators, such as anger, closeness, values/morality, and system justifying beliefs. Results should identify the potential relationships between different motivators and how they may relate to types of anti-racism action. This exploratory study will attempt to extend the current literature on motivations and anti-racism and set the foundation for future intervention research tailored to increasing the motivation of individuals to participate in anti-racism actions. This research was performed in the laboratory of Professor Patricia Devine

The effect of direct color-concept associations on interpretations of colormap information visualizations

Colormap information visualizations are used to communicate patterns of data in a wide variety of domains, such as showing storm patterns in weather maps and showing brain activity in neuroimaging brain maps. Observers have expectations about how visual properties of colormaps should map onto quantities in the data, and colormaps that violate those expectations are harder to interpret. One expectation is that darker colors map to larger quantities, known as the dark-is-more bias. However, it is unclear how strong this bias will be when data in the colormap represent concepts that are directly associated with lighter colors, such as sunshine. In this project, we investigated how the dark-is-more bias influences people’s inferred mappings in the presence of direct color-concept associations. We first asked participants to rate their associations between each of 71 colors and each of two concepts: shade and sunshine. We selected eight pairs of light and dark colors. In four pairs, both colors were similarly, weakly associated with both concepts (low association difference). In four other pairs, the lighter color was more associated with sunshine and the darker color was more associated with shade (high association difference). We interpolated between the colors in each pair to create color scales, which we used to generate colormaps. A new set of participants were presented with 10 colormaps from a given color scale and were asked to choose the side that represented data indicating more sunshine or shade. Color scales and concepts varied between-subjects-. For shade, the darker side was chosen more often for both the low and high association difference conditions. This is not surprising given both the dark-is-more bias and direct color-concept associations suggest the darker side has more shade. For sunshine, there was a significant effect of association difference. For low association differences, participants overall chose the darker side as having more sunshine, whereas for high association differences, they chose the lighter side more often. Thus, direct associations overrode the dark-is-more bias when they conflicted. These results indicated that it is important to consider both direct and relational associations to anticipate people’s expectations about the meanings of colors in colormap data visualizations. This research was performed in the laboratory of Professor Karen Schloss

Polygenic Scores for ADHD and the Role of Parenting for Black and White Youth

Attention-deficit/hyperactivity disorder (ADHD) has a lifetime prevalence rate of 7.2% and is characterized by symptoms like inattention, impulsivity, and hyperactivity that are present in multiple settings. Despite negative life outcomes such as increased rates of incarceration, substance use, and physical health complications, little is known about the genetic and environmental underpinnings of ADHD in White vs. Black youth. Though most genetic studies are limited to White populations, the first aim of this study addressed whether polygenic scores (PGS) for ADHD would predict ADHD equally well for Black (n = 1,503) and White (n = 4,489) youth from the National Longitudinal Study of Adolescent to Adult Health. The second aim tested whether supportive parenting, a prominent environmental factor involved in ADHD that includes parental warmth, care, and involvement, might mediate PGS and ADHD symptoms for each racial/ethnic group. We found that supportive parenting mediated this relationship for White youth (p = .05), but not Black youth (p = .48). Further, there was a significant negative relationship between supportive parenting and ADHD symptoms for both White (p < .01) and Black (p < .01) youth. These results indicate a strong need for ethnically diverse genetic studies to create more inclusive PGS and suggest that supportive parenting affects ADHD symptoms for both Black and White youth. More research using culturally sensitive parenting measures is needed to elucidate how parenting effects may differ across these two groups. This research was performed in the laboratory of Professor James Li

Sexual Assault Resistance Strategies, Blame and Mental Health Outcomes

Sexual assault has been associated with poor mental health outcomes, especially post traumatic stress disorder (PTSD) and depressive symptoms. The resistance strategies a survivor may or may not employ (forceful, non-forceful or passive)as well as the blame one attributes (directed towards self or rapist) during an assault may affect these outcomes. To our knowledge, however, these factors have not been studied in tandem. In this study, we hypothesized that greater use of passive resistance strategies and higher levels of self-blame would both be positively associated with post traumatic stress and depressive symptomatology.We used online survey data from 62 male and female students from a public university in the Northeast who identified as having an unwanted experience and reported on their resistance strategies, rape attributions, PTSD and depressive symptoms. As predicted using bivariate correlations, greater use of passive resistance strategies was significantly positively correlated with PTSD (r= .51,p> .001) and depression (r= .27,p> .05), and higher levels of self-blame were also positively associated with PTSD (r=.40,p> .01) and depression (r= .31,p> .05). These findings could be important targets for secondary prevention work as they suggest that forceful sexual assault resistance strategies are protective against negative mental health outcomes in university students. Clinical mental healthcare providers may also use these findings to inform treatment of patients with sexual victimization histories by targeting the role of self-blame in the development and maintenance of PTSD and depression. This research was performed in the laboratory of Professor Kate Walsh

Parents’ Perceptions of Children’s Gender Bias

Children begin to display gender biases very early in development (Martin et al., 2010), reflecting beliefs about what is acceptable for boys and girls in domains of activities and abilities (Blakemore & Hill, 2008; Golombok & Rust, 1993; Liben et al., 2002). Such biases are concerning because they can limit opportunities based on gender (Bian, 2017). This has caused many to question the source of these biases. Parents have been implicated as socialization agents who may contribute to the development of children’s social biases (Scott et al., 2020), yet little is known about how parents think about different manifestations of children’s gender biases. The present research seeks to assess the extent to which mothers and fathers believe gender biases are acceptable in daughters and sons and how concerned they would be if their daughters or sons expressed gender biases about activities and abilities. Mothers and fathers of 3–10-year-olds (N=399) were recruited via Prolific and completed an online survey. Participants indicated the extent to which they approved of children’s gender bias and their concern about their children’s bias across a variety of domains. Parents were more concerned about ability-related biases (e.g., intelligence) than activity-related biases. When considering biases related to ability, fathers who were considering the behavior of sons were less concerned than fathers considering the behavior of their daughters and also less concerned than mothers. The discussion will focus on how these results illuminate the need to increase parents’ concern about the consequential nature of children’s gender biases. This research was performed in the laboratory of Professor Kristin Shutts

Manipulations of animacy affect item and order memory in sentence-like lists

Sentence-like lists of words are recalled better than random lists of words. This phenomenon is naturally accounted for in emergent theories of verbal working memory, which argue that memory abilities are derived from language skill. However, different emergent theories of verbal working memory make different predictions about the kinds of linguistic representations that matter. Limited emergent theories propose word representations impact only memory for what words have been encountered. Fully emergent theories propose word representations impact both memory for what words have been encountered, as well as the order in which those words appeared. We presented 100 native English speakers with sentence-like memory lists that differ in the subject noun’s animacy (e.g. important – LOBSTER/PUDDING – assigned – tall – orthodontist – pool). Results show that participants were more likely to recall the list with the animate subject noun (lobster) than the inanimate subject noun (pudding). This supports the fully emergent perspective, suggesting that lexico-syntactic constraints constrain both item and order memory. This research was performed in the laboratory of Professor Maryellen MacDonald

Learning Mechanisms and Emotion Concept Development

Emotions are complex social phenomena, expressed in variable ways across different situations. Despite this variability, humans perceive emotions in term of conceptual categories. There is considerable evidence that labeling supports infants’ object categorization. The acquisition of emotion concepts is essential for successful social interactions, and little is known about how children learn these concepts. This research examines how a total of 60 3- and 4-year-old children learn novel emotion categories in one of the two experimental conditions supervise and unsupervised learning task. Through unsupervised learning, children might form emotion concepts by observing the distribution of naturally occurring emotions in their environment. In contrast, through supervised learning children might form emotion concepts based around the verbal labels and explicit scripts. This study provides insights to ongoing debates about the role of labels in emotion concept development and it determines how labels influence children’s ability to learn novel emotion categories. In a previous study we provided feedback to children and children were at chance at baseline, and they only learned in the supervised condition. In this study we removed feedback, and children were above chance at baseline and test phase in both conditions, but participants in the supervised condition did not performed greater than the unsupervised condition. This research was performed in the laboratory of Professor Seth Pollak

Examining Sex Differences in Childhood Trauma and Recidivism

Exposure to childhood trauma is overwhelmingly prevalent among incarcerated populations with over 90% of this population reporting at least one traumatic event during their lifetime. Such exposure has been found to increase recidivism, defined as committing an offense that results in a new sentence, among this population. However, possible sex differences in this relationship remain unclear, specifically when examining physical and sexual abuse. The current study employed logistic regression with moderation in a sample of male (n = 536) and female (n = 91) incarcerated individuals to determine whether there are sex differences in the relationship between childhood trauma and recidivism. Measures included the Childhood Trauma Questionnaire (total score, physical, and sexual abuse subscales) as well as DOC recidivism data. We hypothesized that (1) childhood trauma would increase the probability of recidivism, with this relationship being stronger in males; and (2) physical abuse would increase the probability of recidivism, with this relationship being stronger in males. An exploratory aim examined the relationship between sexual abuse and recidivism, with no specific hypothesis regarding the nature of this relationship. Contrary to hypotheses, there were no significant sex differences in the effect of overall childhood trauma, nor physical or sexual abuse, on recidivism. While the current study failed to find a link between childhood trauma and the probability for recidivism between sexes, childhood trauma is still a relevant concern for the DOC and the mental health services they provide. This research was performed in the laboratory of Professor Mike Koenigs

Childhood unpredictability and the development of impulsivity 

Harshness has been a well-studied area in child development research, whereas unpredictability is a newer, emerging dimension. In addition to environmental harshness and other well-established risk factors, exposure to heightened unpredictability may alter children’s reward processing, leading to increased impulsivity. Impulsivity describes issues with emotional or behavioral self-control and is a risk factor for maladaptive outcomes such as eating disorders and substance abuse. In this study, we explored the effects of environmental unpredictability on the development of impulsivity using the demographically diverse sample in the Adolescent Brain Cognitive Development (ABCD) study (n = 3,773). The Urgency, Premeditation (lack of), Perseverance (lack of), Sensation Seeking, and Positive Urgency (UPPS-P) Impulsive Behavior Scale (ABCD version) at 2-year follow-up was used to assess impulsivity. As there are no direct measures for unpredictability in the ABCD study, we approximated an unpredictability composite score using items of parental monitoring, parental psychological and behavioral problems, marital status, residential history, and neighborhood crime. Results indicated that, controlling for harshness, unpredictability was associated with increased impulsivity (β = 0.32, p < 0.0001) at 2-year follow-up. In conclusion, unpredictable environments were prospectively associated with the development of impulsivity. Understanding the environmental factors that contribute to the development of impulsivity helps guide interventions to provide stable and consistent environments for the benefit of child development. This research was performed in the laboratory of Professor Seth Pollak

PREP research projects are possible based upon support from the National Science Foundation under Grant No. 2050782

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

The Gender (Im)Balance in Citations in the Cognitive Neuroscience Literature

There are many possible factors that contribute to the gender inequities present in academia, the prevalence of which can impact career potential and trajectory. Understanding of the extent of, and possible reasons for, this prevalence is important in the campaign to reduce and then eliminate this inequity. In this study, we focus on citation practices. A recent study of gender balance in citation patterns in five top neuroscience journals has revealed undercitation of  women-led papers (Dworkin et al., 2020). Here we ask whether this imbalance generalizes to the domain of cognitive neuroscience. To address this, we applied the methods of the Dworkin et al. study to the Journal of Cognitive Neuroscience. First, we determined the gender proportions in the authorship of the Journal of Cognitive Neuroscience publications during 2009-2020. Next, we calculated the extent to which the citation patterns in the Journal of Cognitive Neuroscience reflected the journal’s authorship. We found that papers with men as both the first and last author were cited more often than expected whereas papers with a woman as the first or last author were cited less often than expected. We also found that men-led and women-led papers both have qualitatively the same citation patterns. These results indicate that there is a systemic issue that influences citation practices. We will then work to find effective and efficient ways for authors to be aware of the issue and take steps to mitigate it. This research was performed in the laboratory of Professor Brad Postle

The relationship between emotion-modulated eyeblink startle magnitude, anxiety, and emotional memory across the lifespan

The socioemotional selectivity theory predicts that older adults will show enhanced emotional responses to positive stimuli because time is perceived limited as people age, motivating shifts in priorities to optimize meaningfulness and emotional functioning, resulting in better emotion regulation. Nevertheless, this theory does not explicitly consider older adults with dysregulated emotions, such as anxiety. We performed preregistered secondary analyses of MIDUS Refresher Neuroscience Project data (n = 115; age range 26-76, M = 47.91, SD =11.75) to examine how emotional responses and memory in older adults with high levels of trait anxiety  may differ from the predictions of the socioemotional selectivity theory. We hypothesized that older adults with higher trait anxiety will have larger startle eyeblink magnitudes to negative images than older adults with lower trait anxiety, and that older individuals with higher trait anxiety or larger startle eyeblink magnitudes during negative images will show poorer emotional memory than older adults with low trait anxiety or smaller startle eyeblink magnitudes during negative images. As predicted, the interaction between age and anxiety on negative startle eyeblink magnitudes was significant (p = 0.043), such that for older adults, as anxiety increased so did the negative startle magnitude. The interaction between age and negative EBR magnitudes on the total number of negative images recalled (p = 0.410) was not significant, nor was the interactions between age and levels of anxiety on negative images recalled (p = 0.243). These findings suggest the need for further research on emotional responses and disorders across the lifespan, including examinating if lingering affect after negative picture offset may be the way anxiety contributes to negative memory biases. This research was performed within the Center for Healthy Minds under the supervision of Associate Scientist, Stacey Schaefer

Explaining Misinformation Through the Lens of Bayesian Cue Combination

There is no question that our memory is not perfectly accurate. Research on the misinformation effect has repeatedly shown that humans have a strong tendency to incorporate outside information. While this is considered a failure of memory, a Bayesian perspective posits that this phenomenon is a result of sensible combinations of memory and outside information rather than a failure of memory. A focus of the study is to determine if the two sources of information in a misinformation paradigm were weighted properly and to identify the situations in which they were not. This experimental design aimed to create a new paradigm that could measure the magnitude of the misinformation effect and explain the misinformation effect through Bayesian Cue Combination. The study additionally focused on factors that influenced the misinformation effect. Participants (n=15) viewed a set of 5 letters, numbers or symbols, saw a set of 3 math questions, saw the same set of letters with 2 letters in a new location at varying distances (i.e. misinformation event), and then were asked to locate where the letters were in  the first viewing. Additionally, the participants viewed trials where only the event and math filler were given, which was used to measure general memory ability.  The results gathered confirm that the paradigm created was successful in measuring the magnitude of the misinformation effect. The results also revealed that the distance of misinformation and the participants memory ability influence the magnitude of the misinformation effect, which are in line with Bayesian Cue Combination predictions. Closer distances between the event and misinformation were combined appropriately, whereas further distances were not. Additionally, those who were more uncertain of their own memory (as measured by higher error in the letters with no misinformation) were more susceptible to misinformation than those with lower error. Future researchers could look into using this paradigm to study the effects of co-witness expertise, the participant’s confidence on the misinformation, and their interaction on susceptibility to the misinformation effect. This research was performed in the laboratory of Professor C. Shawn Green

Investigating multiple representations of simple and ratio magnitudes

Humans can process different forms of magnitude information. This experiment is interested in two types of magnitudes, simple and ratio. Simple magnitudes are defined as “the raw amount of a substance being measured”, for example comparing the length of a line to another line. Ratio magnitudes are different and emerge from the relationship between two simple magnitudes. An example of a ratio magnitude can be how much of one initial line can fit into a second but separate line. This experiment investigated two competing hypotheses that focus on whether simple and ratio magnitudes are processed jointly or independently. Participants consisted of preschoolers (n = 42), 2nd graders (n = 31), 5th graders (n = 29), and undergraduates (n =32). Participants completed a series of nonsymbolic simple and ratio magnitude discrimination tasks, choosing the larger of two magnitudes.  Both types of magnitudes were presented in four different nonsymbolic formats: dots, lines, circles, and irregular blobs. Paired sample t-tests indicated a format effect, such that participants performed better with continuous formats than with discrete formats on all age groups for both comparison tasks (line acuity > circle acuity > blob acuity > dot acuity). Moreover, for both ratio and simple magnitude comparison tasks, accuracy improved with age. This research was performed in the laboratory of Professor Percival Matthews

Incorporating Interventions in Intelligent Tutoring Systems to Enhance Conceptual Knowledge of Mathematics

Intelligent Tutoring Systems are software-based tools for teaching. Past work has shown that such systems are generally effective for helping students learn procedures, but ineffective for helping students learn fundamental concepts. The purpose of this study was to test interventions to increase conceptual knowledge in an Intelligent Tutoring System for students learning about linear equations. We incorporated diagrams and warm-up activities into an ITS. 30 middle-school students participated in a study that used a pretest-intervention-posttest design. For the interventions, students were randomly assigned to the original version of the ITS (control) or the version enhanced with diagrams and warm-up activities (experimental). Overall, students gained conceptual knowledge from pretest to posttest; however, gains were comparable in both conditions. Unexpectedly, students who correctly completed more problem-solving steps within the ITS did not make greater gains in conceptual knowledge, and this also did not vary by condition. Students who correctly completed more steps made gains specifically in their understanding of the concept of “doing the same thing to both sides of the equation”, but these gains also did not vary by condition. Among students who correctly completed many steps, those who did not like the diagrams showed less success at posttest than students who liked the diagrams. These findings suggest that the interventions were not successful at enhancing students’ conceptual knowledge. Future studies will examine other approaches to integrating diagrams into Intelligent Tutoring Systems.  This work will inform the design of future Intelligent Tutoring Systems. This research was performed in the laboratory of Professor Martha Alibali

Predicting Intergroup Attitudes and Behaviors from Peer Descriptive Norms

The Focus Theory of Normative Conduct states that descriptive social norms (e.g., how most of one’s peers behave) influence behavior when concentration on them is initiated. Also, previous research found peer groups to be influential depending on proximity (the stronger one identifies with a peer group the stronger the influence of the norm). In addition, researchers have focused on ways to improve diversity and inclusivity among individuals, but literature has lacked in showing which intergroup attitudes and behaviors are predicted from descriptive norms. We ran a correlational study examining the questions of which intergroup attitudes and behaviors are associated with descriptive norms regarding peer inclusiveness. Data collection was completed using Prolific (N = 401). The survey covered multiple attitudes and behaviors (e.g. intergroup anxiety, allophilia, value of diversity in workplace and workplace policy) regarding diversity and inclusion and perceptions of norms from four different peer groups. We hypothesized that attitudes and behaviors regarding diversity and inclusion would be positively correlated with perceptions of peers’ descriptive norms. We also hypothesized that friends as a peer group would be most influential regarding diversity and inclusion descriptive norms. The data supports our hypotheses, as attitudes and behaviors were positively correlated with perceptions of peers’ descriptive norms, and friend norms were most influential in being a predictor of several intergroup attitudes and behaviors. Implications include the need to examine the causal relationship between perceptions of descriptive norms and intergroup attitudes and behaviors, and whether strategies to change norm perceptions will result in more positive attitudes and behaviors. This research was performed in the laboratory of Professor Markus Brauer

Predicting transphobia: A replication and extension study

Transgender individuals, whose gender identity does not align with their sex assigned at birth, regularly face transphobia, which is characterized by negative emotions, beliefs, and behaviors. Recent research suggests certain constructs, such as right-wing authoritarianism and social dominance orientation, serve as direct predictors of negative attitudes and beliefs toward transgender individuals (Makwana et al., 2018; McCullough et al., 2019). However, little research has examined predictors of transphobic behavior alongside transphobic attitudes and beliefs. To address this gap, the present study will examine potential predictors of transphobic behaviors, attitudes, and beliefs. Participants will be recruited using Prolific and complete an extensive survey on a number of psychological constructs, such as need for closure, intergroup disgust sensitivity, endorsement of traditional gender roles, and false beliefs about transgender individuals. Results should identify new predictors of transphobia and give a more nuanced understanding of anti-trans prejudice. Being novel and exploratory, the present study will attempt to replicate and extend the current literature outlining predictors of transphobia and set the foundation for future intervention research tailored to decrease bias toward transgender individuals. This research was performed in the laboratory of Professor Patricia Devine

How White People Engage in Racial Allyship with the Black Community

In response to the recent deaths of unarmed Black people at the hands of police officers, there has been a widespread call for White people to become more effective allies and better support Black people. Recent research suggests that although White people understand the principles of allyship in the abstract, they often fail to effectively act as an ally when Black people share a specific instance of subtle bias. Specifically, White people tend to challenge, rather than accept, their Black friend’s interpretation that bias occurred, a response Black people reported was unhelpful. The present research examines whether White participants are more likely to challenge (rather than accept) a Black person’s experience of potential bias if the incident is more ambiguous (vs. blatant). Participants were randomly assigned to a scenario in which their Black friend disclosed an incident of potential bias that was either ambiguous or blatant. Following the scenario, participants were asked to respond to their friend’s disclosure and answer closed-ended questions about what motivated their response. We measured challenging vs. accepting with both open- ended and closed- ended questions. We found that participants who responded to the blatant (vs. ambiguous) bias scenarios were more likely to acknowledge that their friend had experienced bias and less likely to suggest other explanations for the incident. This research provides insight into how White people respond to subtle instances of bias and when White people may fail to act as an effective ally. This research was performed in the laboratory of Professor Patricia Devine

PREP research projects are possible based upon support from the National Science Foundation under Grant No. 1757785

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

Children’s multicultural literature: A content analysis

Children begin to show racial biases at an early age. Researchers and laypeople, alike, have suggested multicultural children’s literature as a tool to reduce these biases. However, little empirical research has investigated this claim. The current project takes the first step in examining the content of children’s multicultural literature. Two independent coders analyzed the content of the top 100 children’s books featuring African Americans as main characters. Our findings suggest that there are four typologies of multicultural books. The most popular books were about daily life and historical events, followed by books that celebrated differences, and then those that taught moral lessons. We found that the majority of the books used language that appreciated, rather than ignored racial differences. Moreover, White authors were more likely to ignore race in their books, compared to Black and biracial authors. Future directions will investigate the effects that these books have on children from different backgrounds and will determine who is using these books as a method for racial socialization. This research was performed in the laboratory of Professor Kristin Shutts

What It Is and What We Can Be: Using Social Norms to Influence Intergroup Attitudes and Behaviors

The Focus Theory of Normative Conduct states that social norms, when made salient to individuals, results in norm-conforming behavior. The focus of this study is on the causal link between descriptive peer norms and intergroup attitudes and intentions. For this project, we compared a negative descriptive norm about a high prevalence of discriminatory behavior among peers to a positive descriptive norm of pro-diversity attitudes among peers (a control condition is also included). The study recruited participants (n = 89) from Amazon Mechanical Turk where each participant was paid $0.40 for their participation. Participants were asked to complete two separate surveys. The first survey made salient the descriptive social norms for either the negative descriptive norms condition, positive descriptive norms condition, or the control condition using a data interpretation activity. The second survey consisted of four self-report Likert scale items that measured intergroup anxiety, concern about discrimination, tolerance of racism, and behavioral intentions. This survey also included a manipulation check that measured perceptions of peer norms regarding discrimination and inclusivity. To assess whether participants were engaged in the data interpretation activity, content analyses on the responses to the interpretation questions were conducted. Linear models were constructed in R to assess the relationship between norm condition and each of the outcome measures. The results show that while most individuals do engage in the data interpretation task and demonstrate an understanding of the normative information that the data imply, there was no influence of the descriptive peer norms information on norm-conforming intergroup attitudes and behavioral intentions. This research was performed in the laboratory of Professor Markus Brauer

Individual differences in perceptual adaptation, perceptuo-motor adaptation, and perceptual learning.

We all have an inherent ability to adapt to various environments that we encounter. Day-to-day, we make subtle adjustments that take a few seconds, or minutes that allow us to better interact with our environment. For example, when we walk out of a dark environment, our photoreceptors located in our retina adjust their sensitivity to the light allowing us to locate various objects and people. In other instances, we may require much longer time scales, such as months or years, to learn important structures that drive our behavior. Learning and adaptation have been extensively, but independently studied, and as such, the relationships between adaptation and learning have hardly been explored. Namely, the literature on adaptation focuses on measures that extrapolate magnitude, or the strength of adaptation. To help bridge the gap between these research domains, our study addressed the question: Is there a common process that links our ability to learn and adapt across different time scales? Here, we focused on the rate of adaptation, and learning, specifically, the rate of change across time. This method allowed us to directly compare changes across tasks that vary by time scale, and also provided insight into individual differences as rates vary across individuals. We collected rates of learning and adaptation from three tasks: a dot stimulus that induced a motion aftereffect, a motor adaptation task, and a dot-motion learning task. Within all three of the relationships, we found positive correlations ranging between .14 to .55. These findings were additionally supported by a principal component analysis, where the first component accounts for 56 percent of the variance. With sixteen participants, these findings are very promising for future replication with a larger sample size. This research was performed in the laboratory of Professor C. Shawn Green

A Novel Program to Improve Children’s Rating of Facial Emotion Intensity

Children use facial cues to predict others’ feelings, desires, and behaviors. Understanding the intensity of a facial cue may provide meaningful information beyond assigning an emotion label. Intensity inferences can guide how to respond to a situation (e.g. a person might react differently if someone is mildly annoyed versus deeply enraged). Given that using facial cues is critical for social competence, researchers are developing training programs to improve people’s ability to identify and use facial cues. This study examines whether a training program – which has been shown to improve adults’ ability to rate facial intensity (Leitzke, Plate, & Pollak, Under Review) – can improve children’s recognition of facial cue intensity. Twenty-five children ages 7 to 11 were recruited to complete a five-day training program (data collection ongoing). Participants viewed morphed facial stimuli ranging from 100% angry to neutral to 100% happy in 5% increments and rated the intensity of the facial cues. They completed three types of blocks (baseline, training, assessment) and, during the training block only, received feedback indicating how their response compared to the objective morph percentage (participant error). Contrary to the results with adults, there was not a statistically significant reduction in participant error after training on day 1 or after the entire length of the training program. However, it is possible that children need more exposure than adults to benefit from training, given that the pattern of results suggests some reduction in participant error. Continuing data collection, assessing the generalizability of these training effects to different emotions and stimulus models, and conducting this study with clinical populations are appropriate next steps. This research was performed in the laboratory of Professor Seth Pollak

Machine-assisted prediction of alcohol use disorder severity from Facebook.

Early diagnosis of alcohol use disorder (AUD) is crucial for a successful prognosis. Unfortunately, people typically do not seek out treatment until alcohol-related problems have become severe. College populations are especially affected by high rates of alcohol use, and it is estimated that 25% of college students meet the criteria for AUD. Due to the ineffective nature of current AUD screening methods available on college campuses, we proposed social media as a passive, scalable alternative for collegiate AUD screening. We collected complete Facebook downloads from 831 college students at a research university in Madison, WI. All participants were interviewed with Module E of the Structured Clinical Interview for the DSM-V (SCID) to determine an AUD severity score based on symptom count. To predict individual symptom count, we extracted features from linguistic content within Facebook private messages, comments, liked pages and demographics data using the Linguistic Inventory and Word Count (LIWC). A regularized Poisson regression model was trained using repeated 10-fold cross validation to predict AUD symptom count from social media linguistic features. While the model suggests Facebook data contain relevant information for forming an AUD diagnosis (AUC = .61), current model performance is yet to be sufficient for clinical implementation. Future avenues for increasing model performance include utilizing additional data sources within Facebook, implementing additional language processing techniques (e.g. bag of words, word embedding), and consideration of alternative classification models. Ultimately, more research is needed to determine the optimal input variables for predicting AUD. This research was performed in the laboratory of Professor John Curtin

Social support, strain and the aging brain.

Previous research suggests that having healthy, supportive social relationships is associated with improved executive functioning across the lifespan, but a growing body of evidence suggests that the strenuous components of our social relationships may actually play a larger role in shaping executive functioning in later life. This study investigated whether social support could act as a protective factor against social strain’s negative impact, by focusing on the interaction between social support and social strain in relation to executive function. A sample of 138 adults between the ages of 25 to 75 completed a questionnaire assessing the levels of support received from or strain caused by their familial relationships. Each participant also underwent a structural magnetic resonance imaging (MRI) scan to collect T1-weighted structural image data. As the dorsolateral prefrontal cortex (DLPFC) is often attributed to be a key player in executive function, we examined cortical thickness of the DLPFC in relation to executive functioning, and familial support and strain. DLPFC thickness was indexed as FreeSurfer-estimated cortical thickness in the caudal middle frontal region of the Desikan-Killiany atlas. Executive functioning was indexed as switch cost performance (related to inhibitory control) on the CANTAB Stop and Go Switch Task. Surprisingly, we did not identify a statistically significant association between the DLPFC thickness and switch cost, contrary to existing literature. No significant relationships were found between executive functioning and the interaction between familial support and strain, or between cortical thickness in the DLPFC and the interaction between familial support and strain. Further research may be necessary to determine if varying sources of support and strain, such as spousal or from friendships, may be associated with executive function outcomes. This research was performed in the laboratory of Professor Richard Davidson

How children and adults update vocal emotion categories.

Previous research has shown the relevance of statistical distribution in the formation of perceptual category boundaries. A 2018 study from the Child Emotion Research Lab (Plate., Wood., Woodard., & Pollak., 2018) specifically tested and validated the salience of statistical distribution in an emotion recognition paradigm, with facial expressions ranging from neutral to angry. In this follow-up study we examined whether children and adults displayed the same tendency to shift perceptual boundaries. We exposed 8-10-year-old children and college undergraduates to different distributions of “calm” and “angry” vocal expressions. In our preliminary results, we found strong evidence of both children and adults being able to update their category boundaries of “anger” or “calm” depending on the frequency of the auditory expression. These results may potentially inform our broader understanding of how children dynamically interpret and react to emotions according to the frequency with which they are exposed to them. This research was performed in the laboratory of Professor Seth Pollak

PREP research projects are possible based upon support from the National Science Foundation under Grant No. 1757785

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

Hold up! How expected vs. surprising actions affect toddlers’ novel word learning

In this study we tested to see if children, would learn the names of novel objects better when an adult picked up and labeled an expected versus a surprising object. Children watched videos of an adult seated at a table with a novel object to their left and right. The adult repeatedly picked up and looked at the same novel object. The locations of the two novel objects were then switched and children watched as the adult picked up and labeled the same novel object from before (expected condition) or picked up and labeled the other novel object (surprising condition). Immediately afterwards, we tested children’s novel word learning using the looking-while-listening paradigm. On each trial, children saw pictures of both novel objects and heard a sentence labeling one. We tracked children’s eye movements to measure their accuracy in looking to the correct object both before and after it was labeled. Preliminary evidence demonstrates that children successfully learned new words in the expected condition – their accuracy in looking to the correct object significantly increased after it was labeled. Children did not learn words in the surprising condition – their accuracy in looking to the correct object did not increase after it was labeled. It is difficult to interpret the results in the surprising condition, however, because children’s accuracy was above chance before the object was even labeled. A follow up experiment will make methodological changes to correct the differences in accuracy before labeling. Data collection is still underway and planned future analyses will examine children’s looking behavior during the teaching videos and whether individual differences in children’s attention to the correct object during teaching predicts their success in word learning at test. This research was performed in the laboratory of Professor Jenny Saffran

Cultural differences in parent-child discussions of germ and cold weather theories of the common cold

Previous research has shown that children display intuitive thinking when it comes to establishing causal relationships related to certain illnesses. As they get older, instead of growing out of these intuitive theories, they seem to coexist with new and scientific theories, since adults also seem to display these intuitive beliefs. Cold weather theory is an example of intuitive thinking where cold weather and dampness are claimed to cause the common cold. Germ theory, on the other hand, is the scientifically supported explanation for how illnesses such as the common cold are contracted, where the cold virus enters the body after a person comes into contact with germs and bacteria. We also know that individuals from different cultures support these beliefs at different rates. This study examined cold weather and germ theory beliefs in Mexican and European-American parents and children (ages 4-6). The parent-child pairs first read a picture story book and then later individually answered questions regarding cold weather and germ theory beliefs. Results from the questionnaire data showed that coexistence of theories was present in both cultural groups. We found that Mexican parents and children endorsed more cold weather theory beliefs than European-American parents and children. Parents and children of both cultures endorsed germ theory beliefs at similar rates. These results are important because they demonstrate the role culture plays in the use of these theories in parents and children. This research was performed in the laboratory of Professor Karl Rosengren

Adding cognition to GitHub

GitHub, a social coding platform, has become a rich source of information. Researchers seek to understand the structure, growth, and influence of the social environment it has become. As part of a DARPA Challenge to model the GitHub environment, this project aims to model the users of GitHub. We were able to create a descriptive cognitive model classifying three types of users on GitHub, core developers, active users, and passive users, affirming previous research. From this cognitive user model we were able to create a predictive model of user behavior. Specifically, we created a model that, when given a repository, the user that owns the repository (the owner), and the user that performed an action on the repository (the actor), could predict the type of action performed on the repository by the actor. We also determined that the actor’s information was more predictive of the action taken than the owner’s information. This research was performed in the laboratory of Professor Joe Austerweil with Nicole Beckag

False beliefs and Mr. Spock

This research was performed in the laboratory of Professor Tim Rogers

Grocery shopping habits as a predictor of memory decline

Dementia of the Alzheimer’s Type (DAT) is a devastating illness that is challenging to diagnose due to co-morbidity with other neurodegenerative illnesses and difficulty discerning it from other causes of dementia. Diagnostic tests, such as PET scans and CSF collection, are expensive, and non-definitive. Additionally, patients are typically screened once they exhibit cognitive difficulties, which is well into the biological development of the disease. One common cognitive test of memory is the semantic fluency task, which taps only one component of memory (semantic memory) and is only administered once or twice yearly. We propose a new variation of this task, the Grocery Fluency Task, which tests participants memory of grocery items they bought the previous week. The task tests episodic memory of items bought and can be used to test theories of semantic decline. Additionally, the task holds the potential to provide information about other lifestyle changes that occur over the course of the disease, such as changes in habits and food preferences that may make it easier to discern DAT from other causes of dementia. The goal of this task is to maintain the ease of use and feasibility of cognitive memory testing, while allowing for the detection of smaller changes in specific memory components used to accomplish everyday tasks. We believe that the task will be useful for basic research, diagnostic testing, and will produce better support for those living with DAT or caring for DAT patients. This research was performed in the laboratory of Professor Joe Austerweil

The role of interference vs. load in visual short-term memory capacity

Contralateral Delay Activity (CDA) has been used as a neural measure for the capacity of visual short term memory (VSTM). However, whether the CDA reflects other aspects of VSTM besides the number of items is still not clear. In our current study, using colors, orientations and contrasts, we estimated the memory capacity (Cowan K as an index) in different set sizes and different heterogeneity conditions for VSTM tasks. Our results showed the memory capacity in heterogeneous condition was significantly larger than that for homogeneous conditions. We also recorded EEG (electroencephalogram) data when participants performed the behavioral tasks. Results demonstrated that even in same set size conditions, the CDA amplitude was different in heterogeneous versus homogenous conditions. The implications of this study show how the CDA may not only reflect the number of items, but the interference between items. Heterogeneous items may produce less interference in memory, leading to a greater VSTM capacity. This research extended our understanding of the CDA by being able to measure characteristics of VSTM beyond the number of items stored. This research was performed in the laboratory of Professor Bradley Postle

Mom, What Happened to Mufasa in The Lion King? The Prevalence of Death in Children’s Animated Films

Many studies have focused on how children learn about death. However, there has been little research done on whether animated films can provide children with death-related information. We analyzed 50 full-length children’s animated films from the past five decades to explore how they portrayed death. We found that most of these films contained one or more death scenes and the majority of these deaths were biologically accurate. We suspected that these scenes may spark death-related conversations between parents and their children. However, most of these deaths were of minor characters. Parents may not talk about them because minor character deaths are generally not seen as important as main character deaths. This research was performed in the laboratory of Professor Karl Rosengren

Learning to see in virtual reality

Initial perception of a virtual environment involves systematic errors that do not occur when we perceive the real world. These errors can be reduced by providing performance feedback during training. Our research sought to understand how skills learned during training in a 3D motion-perception task transfer beyond the context of training. We trained participants on a “3D pong” task using a virtual reality device and tested whether the training transferred to performance on a different virtual reality device. Performance, as measured by the percent of targets intercepted, on the trained device improved after training. However, this increase of performance did not transfer to the other device. These results suggest that learning in 3D is context-specific and that for the human visual system to make minimal errors, experience is essential to become familiar with each new context, its unique cues, and knowledge of how to use those cues. This research was performed in the laboratory of Professor Bas Rokers

The Role of Socialization in Recalling Death Experiences and its Relation to Coping and Bereavement

Researchers have investigated socialization extensively through the lens of gender, race and culture. There is, however, a gap of research exploring how children are socialized to understand death. Death socialization is important because the death of a loved one is still one of the most traumatic experiences an individual can experience (Dowrenwend & Dowrenwend, 1974). This study’s purpose is to analyze the role of children’s socialization of death and its influence on the coping and bereavement process. Using a survey, we analyzed 319 participants using Amazon MTurk and asked them to recall their childhood death socialization and death experiences of a close friend, family member, or pet. A structural equation model suggested socialization and gender as predictors of how people cope with death experiences. Future researchers should look at the role of pets in socialization, among other factors that facilitate the coping and bereavement process. This research was performed in the laboratory of Professor Karl Rosengren

Comparing the Effects of Prejudice Interventions in Reducing Attitude Resistance

Prejudiced attitudes are tied to individuals’ social identities, which lead to the favoring of their in-group over out-groups. Several prejudice interventions have yet to overcome attitude resistance outside of laboratory settings, because such interventions rely on systematic processing. The present study sought to overcome attitude resistance and reduce prejudice by comparing the effects of entertainment media, which relies on heuristic processing, with those of other established prejudice interventions. In this field experiment, 105 participants were recruited throughout the U.S. and randomly assigned to one of four conditions: entertaining narrative, imagined contact, counter-stereotypic imaging, or control. After completing a short activity, participants in the interventions completed resistance and prejudice measures, while those in the control condition completed the prejudice measures only. Contrary to our hypothesis, preliminary findings reveal that the entertaining narrative was not the most effective in overcoming resistance and reducing prejudice. Implications and future directions are discussed. This research was performed in the laboratory of Professor Markus Brauer

Long-Term Training on the Approximate Number System: A Psychophysical Approach

Humans have a non-symbolic number sense that complements symbolic number sense training in formal education. Many researchers have asked whether this non-symbolic number sense, also known as approximate number system (ANS), relates to and affects mathematics ability. However, all ANS studies present stimuli centrally and investigate effects between some ANS task and some mathematics task. This study investigates psychophysical questions related to ANS training:  1) Is ANS trainable?, and 2)  Does ANS training transfer to different quadrants,  to a lower level task: enumeration,  to a higher level task: ratio comparison, and to arithmetic ability?, and 3) Are improvements explained by attentional learning? Participants took part in long term training (28 sessions) on the standard ANS task parafoveally. Participants’ ANS acuity (did/not) improve throughout training (and/but not) between pretest and posttest. A psychophysical approach can guide parameters of ANS interventions for those with dyscalculia. This research was performed in the laboratory of Professor Shawn Green

Examining the Long-term Effects of Infant Rearing Experiences on Delay of Gratification Tasks by Adult Rhesus Macaques (Macaca mulatta)

Delay of gratification has been measured across various species (primates, rodents, birds). For Rhesus macaques, impulsive-aggressive behaviors are more prominent in animals that had disrupted social relationships during infancy (i.e., nursery-rearing). No previous study has examined the impact of differential infant rearing experiences on impulsivity in terms of cognitive control in rhesus monkeys. Further, little is known about their performance in delay maintenance tasks. Twelve male rhesus macaques (6 nursery-reared; 6 mother-reared) participated in a delay of gratification task in order to test the hypothesis that nursery-reared monkeys would show less ability to delay gratification compared to mother-reared counterparts. Five phases of delay maintenance tasks were conducted. Nursery-reared monkeys delayed longer than mother-reared monkeys, but only in Phase 1, (p = .03). The present study demonstrates that rearing experience may impact cognitive control, but more studies are needed to confirm this hypothesis. This research was performed in the laboratory of Professor Allyson Bennett

The Effect of Explicit Knowledge in Problem Solving and Learning

Previous research has indicated that non-verbalized knowledge is not thoroughly evaluated, allowing room for inconsistencies and gaps. When people do verbalize what they know via explanation, they are more likely to notice the subtler—but also more generalizable—features of a problem. The present study investigated the role of explicit knowledge in problem solving and learning. We utilized three versions of Bongard problems: a Source problem, where a feature was introduced; a Transfer problem, where the learned feature was useful in solving the problem; and an Irrelevant problem, where the feature was salient but not relevant to the solution. We hypothesized that participants would solve Bongard problems more accurately when they verbalize their solutions. Our data suggest that, while there was no effect on Transfer problems, participants who verbalized their solution where better able to inhibit the learned irrelevant feature in Irrelevant problems. This research was performed in the laboratory of Professor Gary Lupyan    

Scientific Thinking and Data Interpretation

Data interpretation is an important concept to teach children from a young age, as it is imperative to being able to properly understand academic results and apply them to the real world. A number of other general cognitive abilities have been shown to correlate with mathematical analyses, which are, in turn, related to data interpretation, but scientific thinking seems to be the most likely to have an effect on how people interpret data. Ten adult undergraduate students had their data interpretation skills, scientific thinking, inhibition, arithmetic ability, proportional reasoning, general intelligence, working memory, and reading comprehension assessed independently. Scientific thinking was not associated with data interpretation and none of the other general cognitive abilities were correlated with interpretation of data. Future studies will include a more representative sample size, a more comprehensive measure of scientific thinking, and a multiple regression analysis of the relation between scientific thinking and data interpretation. This research was performed in the laboratory of Professor Martha Alibali        

Learning of Gender versus Number Regularities by Native Speakers of English

Previous research has shown that people have trouble learning dependencies that their own language does not have. However, this research wasn’t well controlled for properties of the languages and manner of instruction. Using an artificial language paradigm, we investigated whether native English speakers would be better at learning gender or number dependencies. Participants learned an artificial language accompanied by a visual world made up of monsters. We assessed learning of gender and number dependencies by an error monitoring test, which contained gender and number agreement error sentences. Results showed that participants were better at learning gender rather than number dependencies, which was surprising given that English does have number but does not have gender. Further, we explored the role of experience in learning a new language through a language background questionnaire. We found that experience with another language was a significant predictor of how well participants learned these dependencies. This research was performed in the laboratory of Professor Maryellen MacDonald

You Test What You Can Name: The Role of Dimensional Nameability in Testing Hypotheses During Category Learning

Historically, language has had a tremendous amount of theoretical and empirical speculation about its role in thought. In the cognitive sciences, the mainstream view is that language is separated from thought. Recently, language has seen a range of assertions that demonstrate its role in cognition. These claims illustrate the importance of language in determining cognitive outcomes. In particular, language’s role in category learning has come under investigation. In tandem with literature on active and passive hypothesis testing in category learning, we investigated how language impacts learning performance via dimensional nameability (i.e., how likely a set of words is available that succinctly highlight what is important). In a series of studies, dimensional nameability predicted performance on learning physical Bongard problems. Dimensional nameability did not improve or hinder performance in either active or passive learning. Limitations on the design of the categorization task may have reduced a main effect. This research was performed in the laboratory of Professor Gary Lupyan

Effects of Collaboration on Learning How to Solve Equivalence Problems

We investigated how collaboration influences learning in second and third grade students solving equivalence problems. Children worked in three types of pairs to solve problems during the collaboration session. Some pairs consisted of two children who demonstrated prior knowledge of equivalence, some pairs consisted of two children who did not demonstrate prior knowledge, and other pairs consisted of one child who did and one child who did not. We hypothesized that strategy disagreement between children would lead to higher scores at posttest. Results indicated that disagreement does not predict learning. This suggests that there may be other pressing factors that contribute to children’s learning, such as noticing problem structure. This research was performed in the laboratory of Professor Martha Alibali

Statistical Learning of Paired Associates in Children

Two studies investigated whether children could learn patterns in stories. In both experiments, children listened to a story and completed a pair-matching task on an iPad. We presented children with pairs listed in random order to explore pattern learning. In the first study, this task asked children to identify predictive pairs (pairs with 100% transitional probability), and the results showed that children were able to learn the pattern. In the second experiment, we designed a more difficult test of learning by asking about both the predictive and non-predictive pairs (pairs with 30% transitional probability on average). In the second experiment, children were unable to learn the predictive pairs. Further research should examine why there was this shift in learning and define an efficient testing phase before continuing to manipulate attention. This research was performed in the laboratory of Professor Jenny Saffran

Mexican and U.S. Children’s Understanding of Death: A Cultural Perspective

Past research examining children’s understanding of death has focused mainly on finding a universal developmental trend. We investigated the development of a biological concept of death in context by looking at how culture, parental beliefs, and past experience with death might reveal earlier competencies in children’s understanding of death. We compared data from 62 children from Mexico (ages 4-6) and 102 children from the U.S. (Rosengren,  Miller, Gutiérrez, Chow, Schein, and Anderson, 2014). We found that children in the U.S. had a more biological understanding of death than children in Mexico. We suggest that the way in which children are socialized into cultural practices surrounding death might impact the way they conceptualize death. This research was performed in the laboratory of Professor Karl Rosengren

Interactions Between CRF and Noradrenergic Alpha-1 Receptors in the Rat Prefrontal Cortex

The prefrontal cortex (PFC) modulates higher cognitive processes including working memory.  Recent studies in our laboratory indicate that the stress-related neuropeptide, corticotropin-releasing factor (CRF), acts within the PFC to impair working memory, similarly to noradrenergic alpha-1 receptor activity.  Therefore, we investigated whether the CRF and noradrenergic systems interact in the PFC. Fluorescent double-label immunohistochemistry was used to detect alpha-1 receptors and CRF in the rat PFC. Over 100 CRF- and alpha-1 receptor-immunoreactive cells within the PFC were counted.  Our results indicate that CRF and alpha-1 noradrenergic receptors are highly co-localized in the PFC, which has implications for drug development in the treatment of prefrontal dysfunction. This research was performed in the laboratory of Professor Craig Berridge    

Let’s Play Pong: Errors in 3D Target Estimation Depend on Eccentricity

When directly observing moving objects, humans often surprisingly confuse motion in depth reporting approaching objects as receding and vice versa. However, lateral motion is accurately perceived. As objects move in depth, their corresponding motion on the retinas is small compared to retinal motions corresponding to laterally moving objects. We hypothesize that small motions impose greater uncertainty in visual processing giving rise to greater confusion in depth. This hypothesis predicts that the nature of confusion will change in the periphery where laterally moving objects have smaller corresponding motion on the retinas relative to objects moving in depth. Using a virtual reality environment, we measured participants’ accuracy in estimating the direction of objects moving at fixation and in the periphery. When fixating objects, participants exhibit significantly more misreports of motion in depth than of lateral motion, thus replicating previous findings. We find a significant increase in misreports of lateral motion for estimation of the direction of objects moving in the periphery. No impact is observed in estimation of motion in depth when compared to performance at fixation. These results support our hypothesis and suggest that the nature of retinal motion impacts the reliability of the information used by visual processes responsible for 3D object motion direction estimation. This research was performed in the laboratory of Professor Bas Rokers        

Working Memory Precision: Slots versus Resources

Two models aim to explain the capacity limits of visual working memory (VWM): the slots model and the distributed resources model (DRM). The slots model states each visual item is held in an individual slot, where the capacity limit is up to four objects regardless of the number of features. The DRM states that memory resources are allocated among the various items kept in VWM, where the precision of memory is dependent on the number of features regardless of the number of objects. To test the effects of varying feature load independently of object load, participants completed a memory task involving varying featural loads of color and direction. We found increasing featural load of direction resulted in a drop off of precision, while increasing featural load of color did not result in a drop off of precision. This suggests that memory mechanisms for direction operate under the control of the DRM, while color operates under the control of the slots model. This research was performed in the laboratory of Professor Brad Postle

Top-Down Control of the Phase of Alpha-Band Oscillations as a Mechanism of Temporal Attention

Much of the research on temporal expectation has focused on stimulus-driven modulation of perception, rather than top-down control of temporal attention. Here we examined the effects of a cue containing information as to when a stimulus would appear on behavioral performance and on the progress of alpha-oscillations (9-13 Hz) in visual cortex. We collected electroencephalography (EEG) recordings while participants completed a target discrimination task in which some targets appeared at a predictively-cued latency, while others appeared at an unpredictable latency. We found that top-down temporal attention improved perceptual processing, demonstrated by an increase in accuracy following a predictive cue at the short delay. The concurrent electrophysiological data revealed accompanying modulation of the phase of alpha such that phase differences were seen prior to target onset between predictively and unpredictively cued trials at both latencies. Furthermore, the phase during these attended windows was biased towards that optimal for stimulus detection for each individual participant.  These data demonstrate a functional consequence of the phase of alpha, suggesting that these oscillations are not merely a noisy byproduct of an idling system, but rather a feature under top-down control that may serve as a mechanism for directed attention. This research was performed in the laboratory of Professor Brad Postle

The Impact of Task Difficulty and Individual Differences on Learning

Learning is optimized when the difficulty level of a task is “just hard enough” (Lui, 2012). However, it is unclear whether participants learn best when they are presented with a task in which the difficulty level adaptively increases/decreases in response to the participants’ performance (i.e., the staircase method), or when difficult and easy trials are randomly intermixed (i.e., the constants method). As such, the purpose of the current study was to examine whether perceptual learning differs depending on whether participants receive training with a staircase, or a constants, method. Thirty-two participants were placed into either a staircase or a constants training group, and completed a perceptual learning task. All participants began with a constants pre-training block, after which they completed 3 training blocks using either the staircase or the constants method. All participants then completed a constants post-test. A variety of individual difference factors were also examined to determine whether they could predict the amount of learning that resulted from both staircase and constants training. While both groups of participants showed significant improvements after training, there was no influence of task on learning performance. Additionally, there were no significant correlations between individual difference factors and pre-post scores for either group, although many of the relationships approached significance. This study demonstrates that, regardless of their subjective differences, training with the staircase and constants methods lead to similar improvements in perceptual learning. However, further research is needed to better understand the influence of individual differences on perceptual learning. This research was performed in the laboratory of Professor C. Shawn Green

Developing a Flicker Paradigm to Assess Visual Working Memory in Early Childhood

The current study focuses on the development of visual working memory in early childhood. We tested predictions of a dynamic neural field model. This model suggests that adults are more likely to correctly identify the correct stimulus in a change detection task involving colors when high-similarity colors are presented, as compared to when low-similarity colors are presented. An adaptation of this model with parameters scaled to simulate children’s performance suggests that the opposite phenomenon should occur early in development, with children being more accurate in their responses when low-similarity colors are presented. In this study, we developed a new “flicker” paradigm to test children’s ability to detect changes in color on high- versus low-similarity trials. Twenty-seven 3-year-old and 16 5-year-old children participated in this study, in addition to 10 adults. Results indicate that in the current task, all age groups showed the same general pattern of better performance on high-similarity trials, contrary to the predicted developmental difference.  We consider how these results can be used to understand how memory may influence behavior differently depending on the structure of the task. This research was performed in the laboratory of Professor Vanessa Simmering

Individual Differences in Learning Relational Categories: A Look at the Importance of Labeling in Same/Different Tasks

Relational categories, or categories of entities that share a common abstract relation but not any perceptual commonalities, are thought to be critical for abstract thought. They are generally difficult for children and nonhuman animals to learn. Although most adults easily learn relational categories, in experiments exploring a particular type of relational category—the same/different relation—a minority of humans and all non-human animals (specifically pigeons) categorize same/different arrays that have levels of variance based on how varied they are rather than based on the discrete categories of “same” and “different.” Because previous research suggests that labels aid in categorization (specifically in relational categories), we ask whether ability to label relational categories in general predicted individual differences in categorizing instances of same and different. We conducted a correlational study to examine the relationship between ability to categorize and name same/different arrays, and ability to complete and name relational analogies. Results indicate a trending, though not significant, relationship between relational analogy ability and same/different categorization ability, such that those who can name relational analogies are slightly better at categorizing and naming same/different arrays. Future research will be needed to better understand the nature of this relationship. This research was performed in the laboratory of Professor Gary Lupyan

Testosterone May Reveal Individual Differences in Future Winning Ability in Sexually Naïve Male California Mice

Recent work in humans has demonstrated that there are individual differences in pleasure-motivated or “hedonic” aggression among men linked to their salivary testosterone response to violent or aggressive stimuli. Corresponding work in the monogamous California mouse (P. californicus) related to the “winner effect” and conditioned place preference (CPP) plasticity in sexually naïve (SN) vs. pair-bonded (PB) males hints that a testosterone-revealed preference of SN male mice to approach male conspecifics may predict higher levels of “hedonic” aggression in these males. We test this hypothesis using repeated “male-female preference tests,” a new paradigm derived from partition tests and CPP studies, followed by an aggressive encounter for 30 SN male P. californicus. Although male-female preference appears to be plastic and subject to test-retest reactivity and related more to the area stimuli inhabit than the animals themselves, and although our aggression index was insufficient to characterize the aggressive behaviors of SN males, our results revealed that males that initially prefer to inhabit the same space as male conspecifics injected with T had a significantly higher chance of winning their aggressive encounter compared to any of the other three groups. This difference in winning ability is not due to the steroid effect of T, as female-preferring males never won an aggressive encounter, even if injected with T. These findings suggest that individual differences in male-female preference may reflect individual differences in aggression circuitry that influence fighting ability. This research was performed in the laboratory of Professor Cathy Marler

Active Collaborative Learning and Verbal Coordination in Children

In this paper we investigate the effects of role-relevance and task-relevant communication as mechanisms of collaborative learning in young children. Five dyads of 5-6 year olds and eleven dyads of 7-8 year olds participated in an active category-learning task in order to learn a one dimensional categorization rule. The task required pairs to collaborate in order to learn a rule, where each child controlled either the relevant dimension or irrelevant dimension. Analyses of post-collaboration categorization accuracy and strategies found that role-relevance had an effect on learning. Analyses of verbalizations during the collaborative task revealed that children spontaneously engage in complex verbal coordination, which significantly aids children’s learning. This research was performed in the laboratory of Professor Martha Alibali

The Effects of State and Trait Brooding on Attentional Biases After an Acute Stressor

Rumination is an important risk factor for psychopathology, most notably depression. Previous research has linked rumination to cognitive biases for negative information, but the mechanism behind why this occurs is poorly understood. Our current study aims to link two important theories of rumination, the Control Theory account (Watkins, 2008) and the Attentional Scope model (Whitmer & Gotlib, 2012), by showing that both context-specific and negative biases can be induced as a result of brooding, the maladaptive subtype of rumination. Specifically, we aimed to determine whether induced brooding induces biases toward context-specific, goal-related information, which translates into attentional biases toward negative information. We induced brooding by presenting participants with a false intelligence test (Remote Associates Task or RAT) on which they either performed well (control condition) or poorly (experimental condition). We then measured attentional biases using an emotional Stroop Task. We found that the hard RAT induced negative affect and state anxiety, but both versions of the RAT induced state brooding. Neither version of the task induced changes in response times to goal-related or negative information on the Stroop task, but RT to both word types demonstrated trend-level correlations with trait brooding. Finally, trait brooding moderated the relationship between RT to negative and context-specific words such that individuals higher in trait brooding had a strong, positive correlation between reaction times to these two types of information. Our results provide new insights into the relationship of rumination and cognitive processing after a negative event, which could be relevant for understanding development of disorders such as depression. This research was performed in the laboratory of Professor Richard Davidson

Gender Differences in Semantic Representation of Food

Differences in experience are reflected in semantic structure, which can be measured using multidimensional modeling of participants’ similarity judgment tasks. Is it possible to detect group-level differences and individual differences utilizing the similarity spaces created from multidimensional modeling? Participants were recruited using Mechanical Turk, given a survey, and separated into gender to answer 80 triads in a similarity judgment task. 349 men provided 26695 responses and 331 women provided 25510 responses to create separate similarity spaces for each gender. Differences between the spaces were analyzed by looking at clusters that formed in each space and by identifying words that occupied dramatically different locations in each similarity space. Women answered more consistently than men. The similarity space derived from their answers had more tightly bundled and delineated clusters than the men’s. This research was performed in the laboratory of Professor Tim Rogers

3D Motion Perception in Virtual Reality Environments

Accurate motion estimation in our 3D environment is crucial for survival. However, surprising and systematic errors have been reported in the laboratory setting that, if made in the real world, could have disastrous consequences. We hypothesize that visual processing in the real world takes advantage of multiple sources of sensory information, some of which may be degraded or non-existent in traditional psychophysical tasks. Using a carefully controlled virtual reality environment, we manipulated two cues that are important for 3D motion estimation: target contrast and the head motion relative to the surrounding visual environment. Our goal was to identify the association between the degradation of sensory information and participants’ accuracy in 3D motion estimation. We found that performance was negatively impacted by increased sensory uncertainty due to reduced target contrast. We also found evidence that different participants employed different response strategies. These strategies were differentially impacted by the perceptual stability of the surrounding visual environment. Those participants who appeared to adopt a more heuristic-based strategy were less impacted by this cue than participants who attempted to base their responses on all of the available sensory information. These results demonstrate the importance of individual differences in sensitivity to 3D motion information and in the strategies adopted to respond to 3D motion. Moreover, these results extend our understanding of 3D motion perception under sensory uncertainty. Further investigation of the relationship between sensory information acquisition and performance, as well as individual differences in task strategies utilizing that information will provide additional insight into the broader implications of our results. This research was performed in the laboratory of Professor Bas Rokers

Democratizing small group discussion

Viewing emotional facial expressions influences sensory judgments

Abilities in updating intrinsic reference frames in 4-year-olds

Foreign language discrimination by 3- to 6-year-old monolingual children

Exploring the motivation to express prejudice: a scale validation

Selective attention to motivationally significant drug stimuli: Differential effects of psychopathic and antisocial personality disorder symptoms

Using text analysis to evaluate utility value

Asymmetries in stereoscopic motion sensitivity across simple 3D environments

Processing positive emotional stimuli consciously and non-consciously: a study of behavior and peripheral physiology

(Note: PREP research projects are typically part of larger studies that will be published in the peer-reviewed literature. The PREP Journal will present abstracts from student student research papers until the corresponding complete study has been published.)

Alcohol’s Effect on Uncertain Reward

Gender and relational aggression in adolescence

Does stereotype threat underlie gender differences in self-reported empathy?

Investigating the mechanisms of values affirmation-mediated achievement gap reductions

The effects of cognitive training on alpha-band power underlying visual short-term memory

Are obstetric complications related to autism? A twin study

A preliminary investigation of the relationship between schizotypal and autistic traits

Psychostimulant action within the prefrontal cortex on sustained attention

Organization of vasopressin within the amygdala

Reading comprehension: Effects of SES and dialect on healthy literacy comprehension

(Note: PREP research projects are typically part of larger studies that will be published in the peer-reviewed literature. The PREP Journal will present abstracts from student student research papers until the corresponding complete study has been published.)

The effect of detailed instruction on autistic performance on the Wechsler Comprehension Subtest

Effects of testosterone on DNA methylation in the adult rat male brain

Children’s accuracy in assessing power from nonverbal behavior

Understanding of equations: The effects of priming on undergraduates’ encoding and solving of Equivalence Problems

Discriminating facial emotion: Neural pathways in Autism Spectrum Disorders

Odor event related potentials (oERPs): A pilot study to capture detection of neutral, negative, and no odors

The effect of timing of maternal and paternal psychopathology on children’s internalizing and externalizing behaviors

Ultrasonic vocalizations in pair bonded Peromyscus californicus

(Note: PREP research projects are typically part of larger studies that will be published in the peer-reviewed literature. The PREP Journal will present abstracts from student student research papers until the corresponding complete study has been published.)