A0410
Title: Spatiotemporal random field framework for signal detection: Simulation studies and stop-signal task fMRI application
Authors: Theophilus Acquah - University of northern Colorado (United States) [presenting]
Abstract: The aim is to develop and apply a spatiotemporal random field framework for signal detection in task-based functional magnetic resonance imaging (fMRI). The approach integrates random field theory with functional time series modeling through a voxel-wise two-way repeated measures ANOVA, targeting the dynamic detection of task-related neural activation in a stop-signal task. Using preprocessed fMRI data from 20 participants, activation was assessed within a middle axial brain slice with within-subject factors of task (manual, vocal, pseudoword) and run (Run 1, Run 2). The global test statistic Xmax, defined as the maximum Z-statistic across all voxels and time points, enables robust detection of peak activation while controlling for multiple comparisons. F-statistics were transformed into Z-statistics and thresholded conservatively to identify significant regions. Spatial heatmaps and time series plots illustrate how activation profiles vary by task, run, and their interaction. Distinct activation emerged in motor and prefrontal regions, with task, run interactions showing temporally specific effects in frontal areas, supporting models of inhibitory control and adaptive cognitive processing. This framework demonstrates the value of combining random field methods with spatiotemporal inference to uncover complex patterns in fMRI data.
