A novel 4D fMRI clustering technique to examine event-related spatiotemporal dynamics of face processing in naturalistic stimuli

Cortical function is complex, nuanced, and involves information processing in a multimodal and dynamic world. However, previous functional magnetic resonance imaging (fMRI) research has generally characterized static activation differences between strictly controlled proxies of real-world stimuli that do not encapsulate the complexity of everyday multimodal experiences. Of primary importance to the field of neuroimaging is the development of techniques that distill complex spatiotemporal information into simple, behaviorally relevant representations of neural activation. Herein, we present a novel 4D spatiotemporal clustering method to examine dynamic neural activity associated with events (specifically the onset of human faces in audiovisual movies). Results from this study showed that 4D spatiotemporal clustering can extract clusters of fMRI activation over time that closely resemble the known spatiotemporal pattern of human face processing without the need to model a hemodynamic response function. Overall, this technique provides a new and exciting window into dynamic functional processing across both space and time using fMRI that has wide applications across the field of neuroscience.