Cognitive Process Models Reveal Meaningful Brain-Behavior Associations in the Adolescent Brain Cognitive Development...

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Psychiatry

Background:

• Understanding individual differences in brain activity and behavior is crucial in neuroimaging.
• Computational models offer mechanistic insights into cognitive task performance.

Purpose of the Study:

• To apply a formal computational model to the Adolescent Brain Cognitive Development Study stop-signal task (SST).
• To quantify relationships between computational model parameters and brain activation using machine learning.
• To assess the construct validity of the computational model using phenotypic measures.

Main Methods:

• Utilized a large dataset (n=6469) of 9- to 11-year-olds' behavioral and neuroimaging data from the SST.
• Employed a machine-learning-based predictive modeling approach to link model parameters with functional magnetic resonance imaging (fMRI) data.
• Examined associations between model parameters and phenotypic measures of inhibition, impulsivity, and cognition.

Main Results:

• Formal SST model parameters explained significantly more variance in fMRI activation (R2 up to 26.86%) compared to empirical performance measures (R2 up to 20.89%).
• Neuroimaging data were strongly associated with evidence accumulation for 'go' choices and attentional lapses ('trigger failure') in stopping.
• Salience network activity correlated with evidence accumulation, while visual cortex activity correlated with attentional lapses.

Conclusions:

• Computational cognitive modeling provides a more sensitive approach to linking brain function and behavior.
• The study validates the computational model's construct validity through associations with behavioral phenotypes.
• Findings highlight the utility of computational models for uncovering nuanced brain-behavior relationships in development.