Schedule of Events | Search Abstracts | Invited Symposia | Symposia | Poster Sessions | Data Blitz

Psychopathology and Dimensional Anxiety Shifts Brain State Landscapes Away From Affective-Control Attractor Dynamics

Poster Session C - Sunday, March 8, 2026, 5:00 – 7:00 pm PDT, Fairview/Kitsilano Ballrooms

Chichi Chang¹ (cc4870@columbia.edu), Stone Su¹, Hengda He¹, Paul Sajda¹; ¹Columbia University

Neuropsychiatric disorders are increasingly recognized as dynamic network dysfunctions, yet research remains focused on static connectivity frameworks (e.g. static functional connectivity) and categorical diagnostic labels (e.g. DSM). These traditional approaches are often limited in capturing the fluid, dimensional nature of psychopathology. To bridge this gap, we employ Energy Landscape Analysis (ELA) to map brain activity as attractor states. Unlike static models, ELA characterizes the stability and occupancy of these states, linking neural dynamics directly to transdiagnostic traits such as behavioral inhibition and anxiety. Here, we applied ELA to resting-state fMRI from 149 transdiagnostic patients and 96 healthy controls from the Transdiagnostic Connectome Project. Data were parcellated into Yeo seven networks, and pairwise maximum entropy modeling was used to identify stable brain state attractors. We compared the frequency of occupancy of these states between groups and correlated state frequencies with Behavioral Inhibition System (BIS) and State-Trait Anxiety Inventory (STAI-Trait) scores. Compared to controls, patients showed increased occupancy of a DMN-dominant attractor (t=4.190; p<0.001) and decreased occupancy of an affective-cognitive control attractor engaging DMN, salience, control, dorsal attention, and limbic networks (t=-4.079; p<0.001). Across participants, higher BIS and STAI-Trait scores correlated positively with DMN-dominant occupancy (BIS: r=0.22; p<0.01; STAI: r=0.19; p<0.01) and negatively with affective-cognitive control occupancy (BIS: r=-0.19; p<0.01; STAI: r=-0.18; p<0.01). Psychopathology and anxiety converge on a shared bias toward internally focused, DMN-dominant states, with reduced engagement of affective-cognitive control networks. This statistically significant but modest effect suggests that lower occupancy of affective-control states may underlie transdiagnostic vulnerability to anxiety.

Topic Area: EMOTION & SOCIAL: Emotion-cognition interactions