Visualizing Outliers in High Dimensional Functional Data for Task fMRI data exploration

TL;DR

This paper introduces visualization techniques for high-dimensional task fMRI data using depth-based methods, enabling efficient outlier detection and exploration of individual variability before neuroscientific analysis.

Contribution

The paper presents novel depth-based visualization methods tailored for high-dimensional task fMRI data, facilitating outlier detection and sample understanding.

Findings

Effective 2D representations of fMRI data achieved

Outliers and variability identified visually

Demonstrated on motor and language tasks

Abstract

Task-based functional magnetic resonance imaging (task fMRI) is a non-invasive technique that allows identifying brain regions whose activity changes when individuals are asked to perform a given task. This contributes to the understanding of how the human brain is organized in functionally distinct subdivisions. Task fMRI experiments from high-resolution scans provide hundred of thousands of longitudinal signals for each individual, corresponding to measurements of brain activity over each voxel of the brain along the duration of the experiment. In this context, we propose some visualization techniques for high dimensional functional data relying on depth-based notions that allow for computationally efficient 2-dim representations of tfMRI data and that shed light on sample composition, outlier presence and individual variability. We believe that this step is crucial previously to any inferential approach willing to identify neuroscientific patterns across individuals, tasks and brain regions. We illustrate the proposed technique through a simulation study and demonstrate its application on a motor and language task fMRI experiment.