ECE 532 : Using fMRI derived functional network features for disease and stimuli classification

Abstract

Functional magnetic resonance imaging (fMRI) is an imaging technology which is used clinically and in research to investigate the spatiotemporal dynamics of neural activity in the brain. The measured signals represent a Blood Oxygen Level Dependent (BOLD) response that corresponds to functional neural activations. In resting state fMRI, subjects relax and steady state functional activations can be observed in the brain. Analysis of these functional activations allows us to model the interactions of individual brain networks, whose activity patterns can vary significantly as a result of a neurological or psychiatric disease. Another type of experiment is the so called block design. In these experiments, patients are presented with a series of stimuli, and their neurological response to these stimuli are measured. Each individual stimulus and corresponding neural activation signal comprise a block. Our goal is to apply Machine Learning tools learned in this class to implement a classifier that can make accurate predictions about a patient based on functional network features derived from fMRI images. In this lab, we will go through two prediction exercises. In the first, we will attempt to diagnose if a patient has Schizophrenia or not, using functional network features derived from fMRI images. In the second exercise, we will classify the type of stimuli a patient received as visual or auditory based on their fMRI data; for example, we will try to predict whether a patient has seen an image of a frog or heard its ’ribbit’ instead.

2 Figures and Tables

Cite this paper

@inproceedings{Rusk2014ECE5, title={ECE 532 : Using fMRI derived functional network features for disease and stimuli classification}, author={Sam Rusk and Aritra Biswas and Chris Fernandez}, year={2014} }