Task-driven Functional Connectivity Modulation in the Human Cortex

Event Date:
January 27th 9:00 AM - 10:00 AM

Meeting ID: 928 0482 8495 Passcode: 185518

Speaker: Crispin Foli

Research Advisor: Prof. Ajiboye


Title: Task-driven Functional Connectivity Modulation in the Human Cortex

Abstract: In addition to the primary motor cortex (M1), the Anterior-intraparietal area (AIP) and the Inferior Frontal Gyrus (IFG) are two very attractive sites for Brain-Machine Interfaces (BMIs) targeting hand movement restoration, given their demonstrated activation during movement planning. Together, these three areas constitute the visuomotor grasp network, which underlies the execution of a large variety of grasps and hand movements in both humans and Non-Human Primates (NHPs). Even though task-driven functional interactions between the different components of the grasp network have been vastly explored in NHPs, a high-resolution temporal profile of these interactions at the spatial detail level of single neuronal assemblies have not yet been demonstrated in humans. An improved understanding of the dynamics of information flow within this network could prove essential to the design and implementation of more efficacious BMIs.

The current study in the Ajiboye Lab records neural activities intracortically from all these key hand areas (AIP, IFG and M1), thus providing an opportunity for a high-resolution characterization of task-driven spatiotemporal information flow within the grasp network homologue in humans. The use of multiple high-density microelectrode arrays in our recording setup, however, presents some challenges in using well established statistical techniques for inferring functional connectivities between different brain areas, including high computation. The first goal of my project is to design an optimal data analysis framework for characterizing information flow suited for neural data recorded from multiple high-density microelectrode arrays. Preliminary results, based on computational modeling, show that the “true” connectivities (information flow patterns) between different nodes within a complex brain network can be accurately computed by first performing multivariate decomposition of the high dimensional signals from the recording microelectrode arrays, followed by Multivariate Autoregression of the resulting low dimensional signals. When tested on empirical data from closed-loop aperture control experiments, this framework produced AIP-IFG interaction dynamics consistent with the current view of information processing within the grasp network in NHPs.
The second goal of my project then is to use this tool to characterize the specificity and task-dependent modulation of the grasp network homologue in the human cortex by studying the relative levels of spatiotemporal information flow within this network while a human subject performs different tasks such as speech, closed-loop grasp control and mental computation.