Localizing brain-behavior relationships remains a key goal of cognitive neuroscience and clinical neurology. The relationship between lesion site(s) and behavior may be assessed using either functional or structural imaging. In functional neuroimaging
(e.g., functional magnetic resonance imaging (fMRI)), brain activation is often presumed to indicate a causal role of the observed region in some cognitive process; however, the data generally admit a variety of other explanations, often including the
possibility that the activity is epiphenomenal to the process of interest. Structural imaging-based lesion-symptom mapping (LSM) has long been used to study brain-behavior relationships, and complements what can be learned from functional
neuroimaging by providing high-quality evidence that the integrity of a brain region is necessary for the normal performance of the measured function.
The project involves implementation of 3-D and 2-D sliced segmentational VLSM using various modified and custom architectures. This would be compared to simple SVR-LSM techniques. This is project is proposed to be the first of such kind of work.
The current version contains the extraction, EDA and visualization of fMRI based lession data, in EDF data-fomrat. The entire analysis was done in python.