PUBLICATIONS
My publications. Might not always be up to date, so find more on my Google scholar page.
2024
We used hierarchical Drift Diffusion Modeling (hDDM) to analyze proactive and reactive control deficits in individuals with schizophrenia and other psychotic disorders. Comparing people with psychosis (PwP), their first-degree relatives, and controls using the Dot Pattern Expectancy task, we found that PwP exhibited slower proactive control and perceptual processing. Machine learning analysis showed hDDM parameters effectively distinguished between these groups, highlighting specific deficits in motor/perceptual time and evidence integration, especially in proactive control.
2023
We introduce an innovative method for measuring EEG causal oscillatory connectivity, utilizing causal discovery analysis for skewed time series and spectral parameterization. We applied the Greedy Adjacencies and Non-Gaussian Orientations (GANGO) method to EEG data from a clinical trial on transcranial direct current stimulation (tDCS) for treating executive dysfunction post mild Traumatic Brain Injury (mTBI). Our findings suggest that tDCS may enhance executive functioning by increasing theta-band and decreasing alpha-band causal oscillatory connections in the prefrontal area.
We address complexities of applying causal discovery to fMRI studies, detailing nine specific challenges and the range of decisions researchers must navigate. We review a recent case study to illustrate these points and identify areas needing methodological advancements. We underscore the potential of causal discovery in fMRI analysis, suggesting its superiority over traditional methods, while also highlighting the need for further refinement of these techniques.
This study explores individual heterogeneity in Substance Use Disorders (SUDs) to improve addiction treatment. Analyzing data from 593 participants, including 173 with past SUDs, we identified three distinct neurobehavioral subtypes through latent profile analysis: a “Reward type” characterized by higher approach-related behavior, a “Cognitive type” with lower executive function, and a “Relief type” with high negative emotionality. Each subtype showed unique patterns in resting-state brain connectivity, underscoring the need for personalized treatment approaches in addiction medicine based on functional subtyping.
2022
Gregory R. Niklason, Eric Rawls, Sisi Ma, Erich Kummerfeld, Andrea Maxwell, Leyla Brucar, Gunner Drossel, Anna Zilverstand
We employ an explainable machine learning approach to examine sex and gender differences in Cannabis Use Disorder (CUD), using decision trees and SHapley’s Additive exPlanations (SHAP). We confirm that a range of factors – including environmental, personality, mental health, neurocognitive, and brain factors – significantly contribute to cannabis use levels and diagnostic status. Notably, risk factors like high openness, externalizing behaviors, and low hippocampal volume were more pronounced in men, while environmental factors like low education and instrumental support were more significant in women, highlighting distinct influences based on sex and gender in CUD.
We evaluate the impact of reduced-dimensionality Independent Components Analysis (rdICA) on EEG data quality, particularly for EEG cleaning purposes. Using 128 electrode recordings from 43 subjects, we compared standard ICA, ICA with 64 electrodes, and ICA with PCA retaining 99% and 90% data variance. The quality of ERP data was assessed focusing on the N1 and P3 components. Results showed significant changes in early sensory components for the 90% variance condition, with inconsistent effects on reliability. PCA-based rdICA can be effective for EEG cleaning if used cautiously.
Eric Rawls, Erich Kummerfeld, Bryon Mueller, Sisi Ma, Anna Zilverstand
This study presents a data-driven method to create “causal connectomes” from resting-state fMRI data, resulting in 442 sparse yet fully connected connectomes. Key findings include the identification of highly connected hubs in attentional and executive networks, each with unique connectivity profiles. The study underscores the central role and vulnerability of these networks in the human brain, paving the way for future psychiatric research applications.
Andrea M. Maxwell, Katherine Harrison, Eric Rawls, Anna Zilverstand
This study reviews gender differences in Alcohol Use Disorder (AUD), finding that psychosocial factors like lack of social support significantly impact AUD onset in adolescent girls and contribute to addiction maintenance in adult women. This gender-specific pattern suggests the need for tailored approaches in treating and understanding AUD across different genders.
This study examines the reward positivity (RewP) as a depression biomarker, revealing that while RewP varies with the salience of feedback (more positive for wins), it’s not significantly associated with depression or anhedonia. The findings also show a diminished memory for rewards in individuals with higher depression levels, highlighting the intricate relationship between RewP, memory, and depression.
Arooj Abid, Morgan Middlebrooks, Eric Rawls, Connie Lamm
This study investigates the impact of the order of emotional stimuli on cognitive control processing in an AX-Continuous Performance Task. Findings reveal that negative stimuli, especially in a random presentation order, lead to larger attention-related responses (LPP) and reduced cognitive control signals (N2), indicating an emotion-related depletion of cognitive resources. The study highlights the importance of considering stimuli presentation order in designing emotion induction tasks.
2021
Eric Rawls, Rebecca White, Stephanie Kane, Carl E. Stevens Jr., Darya L. Zabelina
This study explores preferences for fractal patterns, using Jackson Pollock’s paintings and matched random Cantor sets. EEG data revealed that participants generally preferred more complex (higher-dimensional) fractals. Brain activity analysis showed that parietal alpha and beta power tracked the complexity of these fractals, with parietal alpha power closely linked to aesthetic preference. This suggests that the brain’s appreciation of fractal complexity and preference is reflected in specific patterns of neural activity.
Jarrod Eisma, Eric Rawls, Stephanie Long, Russell Mach, Connie Lamm
This study examines frontal midline theta activity across different cognitive control strategies, including response inhibition, proactive/reactive control, and conflict monitoring. EEG data from 176 participants showed that higher theta power is consistently linked to greater cognitive control demands. Notably, the intensity of theta activity varies with each control strategy, suggesting that while it’s a general marker for cognitive control, its role differs depending on the specific strategy employed.
We investigated brain activity related to prediction errors (PEs) in reinforcement learning, using EEG to analyze mediofrontal event-related potentials (ERPs). Negative reinforcement leads to specific ERP patterns, with mediofrontal ERPs signaling unsigned PEs during the P2 potential and signed PEs during the FRN/RewP and frontal P3. This aversion positivity correlates with increased central delta power for more aversive outcomes, indicating its role in negative reinforcement learning. In contrast, positive reinforcement PEs did not significantly alter ERP patterns, despite influencing behavior.
We employed Causal Discovery Analysis (CDA) with Human Connectome Project data to uncover the complex neurobehavioral drivers of Alcohol Use Disorder (AUD). Analyzing 926 participants, including 22% with AUD, researchers distilled 100 measures into 18 domains and examined 12 brain networks. The resulting causal model revealed a hierarchy: brain connectivity influences cognitive abilities, which affect social, affective, and psychiatric functions, ultimately impacting AUD severity. This approach underscores the significant roles of cognitive, social, and affective factors in understanding and addressing AUD.
2020
This study explores the neural mechanisms of conflict adaptation, where response to conflicting stimuli quickens over time. Using EEG during a flanker task, we found that conflict levels and adaptation modulate the P3 component of the event-related potential (ERP), associated with the locus coeruleus norepinephrine system. Key findings include delta oscillation phase predicting reaction times and affecting P3 amplitude, supporting the theory that phasic NE release resets cortical activity, contributing to ERP generation and environmental monitoring.
Eric Rawls, Vladimir Miskovic, Shannin N. Moody, Yoojin Lee, Elizabeth A. Shirtcliff, Connie Lamm
This EEG study investigates prediction errors (PEs) in reinforcement processing, focusing on feedback-related negativity (FRN) and frontal midline theta (FMΘ) during a monetary incentive delay task. The results indicate distinct cognitive processes for these measures: FRN responds to errors in both positive and negative reinforcement, sensitive to feedback level, while FMΘ is responsive to outcomes in positive reinforcement and control conditions, but not in negative reinforcement. This suggests FRN reflects unsigned PEs (salience signal), and FMΘ is associated with negative cues and cognitive control needs.
2018
In a study with 75 undergraduates, we used event-related potentials (ERPs) to examine how effortful control relates to aggression. We focused on three ERP components: P2, N2, and P3. Results revealed that N2 activation, particularly in negative contexts and high-conflict trials, is crucial in moderating the relationship between effortful control and aggression, suggesting that individual differences in neural processing efficiency play a key role in managing aggression.
Mejdy Jabr, Greg Denke, Eric Rawls, Connie Lamm
We investigated the connection between violent video gameplay, aggression, and neurocognitive processes using event-related potentials (ERPs) N1 and P3. Results indicate that video game players show different N1 and P3 responses to emotionally-charged imagery compared to non-players. Notably, smaller P3 amplitudes, associated with heightened aggression, were observed in players. The findings suggest that selective attention to violent content and desensitization are crucial in understanding the link between video gameplay and aggression.
Greg Denke, Eric Rawls, Connie Lamm
We examined the relationship between attentional conflict, anxiety, and emotional eating using dense-array EEG. Participants played the attentional blink game, highlighting how excessive focus on negative stimuli can affect subsequent action processing. Results indicated that N2 activation, an event-related potential linked to conflict processing, moderates the association between anxiety and emotional eating. This suggests that higher anxiety and more negative N2 activation contribute to emotional eating, pointing to ineffective conflict processing as a factor in poor emotion regulation.