Student Speaker: Camille Gomez-Laberge Neuroscience Program, Ottawa Health Research Institute and Department of Systems and Computer Engineering, Ottawa, ON, Canada Exploring cerebrovascular dynamics in the human brain: steps toward corroborating cellular and molecular findings using in vivo neuroimaging data Background: Understanding the regulatory mechanisms of cerebral blood flow and its dysfunction remains a central topic in the neurological scientific literature. Ample evidence now indicates that certain factors related to age, disease, and medication can dramatically alter cerebrovascular regulation. Studying the effects of these factors in the living human brain may be possible by exploratory analysis of the dynamics measured by whole brain functional neuroimaging modalities such as MRI and PET. Theory: We propose a unified exploratory method for the analysis of the cerebrovascular response to event-related stimuli as measured by the blood oxygen level-dependent MRI signal. A key problem in employing exploratory analysis methods is that after the voxels have been clustered into groups having similar dynamics, the statistical significance of each cluster remains unknown. Classical hypothesis testing is not appropriate, being designed to operate on balanced groups of data. Instead, we apply a modern Bayesian technique to calculate each cluster’s significance in the overall hierarchical structure of the data. Methods: Voxels having correlated time series are clustered using the fuzzy k-means algorithm. The Bayesian hierarchical model is fit to the data using a temporal feature based on the cross-correlation of each cluster with the stimulus time series. Data are acquired from simulation and from healthy volunteers during an event-related motor task. Results: Results on simulated and real data are compared with classical hypothesis testing. Results show that the proposed method is capable of identifying territories having distinct responses to stimulus. Performance estimates from simulated data are also given. Conclusion: The study of stimulus response with neuroimaging modalities may further elucidate how neurovascular disease impacts the human brain as a whole.