Defense: Spatio-Temporal Connectome Data Analysis using Machine Learning and Visualization

Ran Xu
Technology & Information Management PhD Candidate
Location
Virtual Event
Advisor
Angus Forbes

Join us on Zoom: https://ucsc.zoom.us/j/97208498399?pwd=TXlUR1BKUUFPa2hpdUpxaWliU0pIUT09 / Passcode: 756931

Description: Analysis and comparison of multiple time-varying connectomes is crucial for studying brain diseases such as Depression, Schizophrenia, and Alzheimer’s Disease (AD). Recently, many visualization techniques have been proposed to assist clinical psychiatrists in exploring complex connectome datasets but majority of those visualizations are focused on static connectome features.

We develop TempoCave, a novel visualization application for analyzing dynamic brain networks, or connectomes. TempoCave provides a range of functionality to explore metrics related to the activity patterns and modular affiliations of different regions in the brain. We present a real-world use case that analyzes pre- and post-treatment connectome datasets investigating the use of rumination related cognitive behavior therapy (R-CBT) to treat major depressive disorder (MDD), indicating that TempoCave can provide new insight into the dynamic behavior of the human brain. To further facilitate connectome trends over time, we developed ConnectVis, an extension of TempoCave, that explores multiple dynamic connectomes. A temporal view shows the modular dynamics while a spatial view shows topological features with regions as nodes and functional connectivities as edges. We demonstrate the effectiveness of ConnectVis through two case studies involving real-world fMRI data.

Analysis of temporal connectome is limited due to high dimensionality of the data. To address this problem, we introduce a new data analysis technique that is specifically designed for dynamic functional connectomes by combining a novel temporal graph neural network and a differentiable mask mechanism. Our technique can classify depression group vs control group with high accuracy and the mask is used to identify the significant brain regions that contribute to depression. Along with our visualization tool, we have a spatio-temporal connectome analysis pipeline. We demonstrate the effectiveness of our pipeline through two use cases.