Date of Award
Master of Science (MS)
Resting state functional MRI (rsfMRI) has been proven to be a valuable tool in clinical applications such as pre-surgical mapping, but there is not yet a functional and usable algorithm that can be used by physicians in a clinical setting to evaluate an individual patient for diseases and aberrant brain connectivity. If a physician wants to evaluate a patient in this way, the rsfMRI data must be looked at “by hand,” i.e. the physician must manually evaluate the data and identify the functional ICN’s and whether they are normal or aberrant. An algorithm that would automate this process and supplement the physician’s evaluation would be very valuable and would decrease the time needed while increasing accuracy of the data analysis. The algorithm could be used in clinical applications as discussed, or academic and research applications to explore the neural basis of neurological disorders and deficits (epilepsy, etc). rsfMRI data is significant for the proposed solution as it provides maps of functional brain connectivity within functionally specific neural networks, and those connectivity maps can help identify normal and abnormal brain conditions. Whether an ICA approach based on standard networks or an ROI seed based approach which utilizes temporal correlation is used, the end goal of this research is to develop and refine an imaging biomarker for aberrant brain connectivity. The biomarker algorithm should be able to detect the two main types of aberrant connectivity: increased (when abnormal brain connections are present) and decreased (when normal brain connections are missing). The algorithm should then correlate the connectivity patterns to a normative reference data set and create prioritized classification matches to that reference data set. This will allow identification of the aberrant connectivity patterns. Data from the Human Connectome Project (HCP) will be used to create the normative reference data set. The algorithm will finally be verified using simulated test data and test statistics.