Connectome Fingerprinting
“Connectome Fingerprinting” is an applied machine learning technique that estimates the unique functional topography of an individuals brain by mapping their unique inter-areal network-level interactions. See my recent publication for more infomation.
Sensory-selectivity Cortical Mapping
Sensory selectivity mapping is a novel approach to mapping the precise layout of distinct region in frontal cortex at the individual level. The lab’s first work in this area [1] mapped 4 bilateral auditory- and visual-selective regions and my subsequent work [2] has shown that a significant portion of frontal cortex can be characterized based upon mapping sensory-selectivity vs. sensory-independence.
Corpus Callosum Classification
Corpus callosum atrophy is a strong predictor of increasing deasease burden and cognitive decline in multiple sclerosis. We aim to produce a machine learning method to automatically segment the corpus callosum using multi-modal neuroimaging input and release an integrated set of tools for image processing, tissue segmentation and modal (re)training. While most image segmentation tools make use of a single modality for simplicity, the inclusion of predictors derived from several image types leads to a modal that is robust to poor image quality.