Capturing Task-evoked Brain Dynamics Based on Eigenmode Reconstruction

Motivation: Traditional method treats fMRI signals as linear and use predefined hemodynamic response functions to convolve with the fMRI signal to detect neural activity, which may lead to inaccurate activation maps. Goal(s): To identify task-evoked brain dynamics at the individual level with high spatiotemporal resolution. Approach: We used eigenmodes to reconstruct each time point of task fMRI signals and perform statistical inference to characterize of task-evoked activation dynamics. Results: Our method captured individual, task-evoked activations in motor tasks, highlighting personalized, temporally resolved neural dynamics. Impact: Our approach enables the capture of individualized, task-evoked brain activation patterns, offering potential advancements in personalized diagnostics and interventions for mental health conditions by identifying unique neural activity signatures and temporal dynamics in task engagement.