Authors:

Aanya Krishnan

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Functional connectivity studies have demonstrated that patterns of brain connectivity can predict individual differences in behavior. However, most prediction approaches rely on whole-brain connectivity, which may mask relationships that are specific to particular functional networks. This project tested the hypothesis that individual brain networks exhibit selective predictive relationships with distinct mental health dimensions. Resting-state fMRI data from 75 participants were analyzed using 13 functional networks defined through individual-specific parcellations. For each network, connectivity patterns were used to predict five clinical self-report measures: reward sensitivity (BIS/BAS), ADHD symptoms (SRADHD), depression (BDI), worry (PSWQ), and emotion regulation (ERQ). Kernel tidge regression with 50 iterations of cross-validation was implemented to generate stable prediction estimates while controlling for age. Distinct patterns emerged across networks. The Reward network most strongly predicted ADHD symptoms (r = 0.40) and worry (r= 0.31). SomatomotorLateral connectivity showed the strongest association with depression (r = 0.25), while the Dorsal Attention network demonstrated a moderate negative relationship with depression (r = -0.27). Single cross-validation iterations showed high variability (SD = 0.05-0.08), demonstrating that multiple iterations were necessary for stable estimates. These findings demonstrate that functional networks exhibit construct-specific predictive profiles rather than uniform predictive strength across behaviors. These network-specific profiles provide a foundation for developing targeted neurobiological markers, potentially enabling more precise identification and monitoring of specific mental health dimensions.

Source:

University of Illinois Urbana-Champaign

Topics:

No topics listed

Co-authors:

Aanya Krishnan