Hierarchical Bayesian Modeling of the Relationship between Task Related Hemodynamic Responses and Neuronal Excitability: a Simultaneous fNIRS/TMS Study

Abstract

AbstractBackgroundInvestigating the relationship between task-related cortical hemodynamic activity and brain excitability is challenging because it requires simultaneous measurement of brain hemodynamic activity while applying non-invasive brain stimulation. There is also considerable inter-/intra-subject variability which both brain excitability and task-related hemodynamic responses are associated with. Here we proposed hierarchical Bayesian modeling to taking into account variability in the data at the individual and group levels, aiming to provide accurate and reliable statistical inferences on this research question.MethodsWe performed a study on 16 healthy subjects with simultaneous Paired Associative Stimulation (Inhibitory PAS10, Excitatory PAS25, Sham) and functional Near-Infrared Spectroscopy (fNIRS) targeting the primary motor cortex (M1). PAS was applied to modulate the cortical function and induce plasticity. Before and after each intervention cortical excitability was measured by motor evoked potentials (MEPs), and the motor task-related hemodynamic response was measured using fNIRS. We constructed three models to encode 1) PAS effects on the M1 excitability; 2) PAS effects on the whole-time course of fNIRS hemodynamic responses to finger tapping tasks, and 3) the correlation between PAS effects on M1 excitability and PAS effects on task-related hemodynamic responses.ResultsSignificant increase of the cortical excitability was found after PAS25, whereas a small reduction of the cortical excitability was shown after PAS10 and no changes after sham. We found PAS effects on finger tapping evoked HbO/HbR within M1, around the peak of the hemodynamic time courses. Both HbO and HbR absolute amplitudes increased after PAS25 and decreased after PAS10. Cortical excitability changes and task-related HbO/HbR changes showed a high probability of being positively correlated, 0.77 and 0.79, respectively. The corresponding Pearson’s correlations were 0.58 (p<.0001, HbO with MEP) and 0.56 (p<.001, HbR with MEP), respectively.ConclusionBenefiting from this original Bayesian data analysis, our results showed that PAS modulates task-related cortical hemodynamic responses in addition to M1 excitability. The fact that PAS effects on hemodynamic response were exhibited mainly around the peak of the hemodynamic time course may indicate that the intervention only increases metabolic demanding rather than modulating hemodynamic response function per se. Moreover, the positive correlation between PAS modulations of excitability and hemodynamic brings insights to understand the fundamental properties of cortical function and cortical excitability.HighlightsHierarchical Bayesian Modeling of PAS effects on M1 excitability and hemodynamicDemonstrated PAS effects on the whole time-course of task-related hemodynamicIncreased/reduced task-related hemodynamic corresponds to higher/lower M1 excitabilityHigh probability of positive relationships between excitability and HbO/HbR changes