Navigating cognition: A comparison of resting-state and task-based fMRI localization

Poster Session C - Sunday, April 14, 2024, 5:00 – 7:00 pm EDT, Sheraton Hall ABC

Caroline Nettekoven¹ (cr.nettekoven@gmail.com), Da Zhi¹, Ladan Shahshahani^1,2, Ana Luisa Pinho¹, Joern Diedrichsen¹; ¹Western University, ²Brown University

A major challenge to understanding human cognition from neuroimaging studies is that regions involved in higher-order cognition vary highly between individuals. Numerous groups have therefore pursued a precision mapping approach, using localizing data to define functional regions at the individual level. However, which type of localizing data yields the highest precision is unclear. In this study, we compared the ability of resting-state and task-based (3T) fMRI data acquired in 19 subjects to localize individual functional regions in the cerebellum. The cerebellum’s tightly packed and highly variable functional regions present a challenge for localization, offering an ideal test case. From the task, data, we constructed task batteries either focused on motor function, language, or working memory (task-specific localizer) or several domains (task-general localizer). We used a hierarchical Bayesian model to adapt a cerebellar group atlas according to the localizer data. We then used these individualized regions to predict functional boundaries in held-out data of the same individual using the Distance-Controlled Boundary Coefficient. Performance was evaluated across the whole cerebellum, and separately for cerebellar motor, action, multiple demand and social regions. In cerebellar motor regions, rest performed worse than task-based localization, whereas for cognitive regions rest and task performed similarly. Across the whole cerebellum, task-general localizers outperformed task-specific localizers, whereas a combination of language and working memory tasks was advantageous for mapping cerebellar social-linguistic regions and multiple demand regions. We believe these findings will be useful in guiding the localization of cerebellar functional regions in future studies of cerebellar cognition.

Topic Area: METHODS: Neuroimaging