This proposal describes a five-year career development program to prepare the candidate, Dr. Nan Xu, for a career as independent investigator at a major academic research institute, with the expertise of modeling dynamics of brain causal system to provide novel insights into the basic pathophysiology of neurologic disorders. This proposal develops upon Dr. Nan Xu’s expertise in mathematical modeling and algorithm development to translate model inferences in neurologic disorders and therapeutic practices; while training her to ask scientific questions relevant to clinical practice and neurophysiological pathology. The PI will be mentored at Biomedical Engineering at Georgia Tech and Emory University by Drs. Shella Keilholz (blood oxygenation level dependent (BOLD) dynamics and their neurophysiological origins), Vince Calhoun (translational image analysis and biomarkers), and Jason Allen (clinical training in neurologic disorders). Their complementary expertise will help PI to bridge the gap between her analytical expertise and problems in clinical neurology and cognitive neuroscience that need to be addressed. In the brain causal system, evidence has shown that the directed information transfer may comprise not only a strength, but also a duration and a capacity. The latter two metrics imply important neuroscientific aspects that have not been well studied in the past. They may play key roles in pathological dysfunctions such as the vestibular syndrome which occurs in 80% of patients following concussion. In PI’s thesis work, an innovative measure was developed to predict the first two metrics. Building upon this work, novel theoretical and computational approaches will be developed in this study to further evaluate the temporal variability in the strength and duration of information transfer (K99) , and then to characterize the information capacity as well as to evaluate the temporal variability of all three causal metrics (R00) . The causal estimates of resting BOLD data will be validated against the estimates of the concurrently recorded local field potential (LFP) data, and task- evoked BOLD data ( for both K99 and R00 ). Finally, findings ( in both K99 and R00 ) will be translated into clinical studies of patients with different severity of vestibular syndromes. The specific aims are to: (1) model and compute the time-varying spatiotemporal functional causal interactions among functional brain regions; (2) evaluate the reliability and sensitivity of the estimated time-varying causal metrics using multimodal brain imaging data of rodents and human, and (3) access the time-varying causal patterns in patient brains with different severity of post-concussive visual motion sensitivity, which is one type of vestibular syndromes that were commonly occurred in patients after concussion. Successful project completion would potentially transform the rapidly evolving field of dynamics modeling in brain causal system, facilitate basic neuroscience discovery, enable comprehensive identification of neurologic disorders, and inspiring new animal experiments for studying neurological diseases.

Project Narrative This project aims to develop a novel computational paradigm for comprehensively characterizing and interpreting the causal system and its variability in functional brain networks. The findings will be translated in clinical studies of patients with different severity of neurological disorders to guide therapeutic practices, and particularly, post- concussive vestibular impairment syndrome will be used as one illustrative example in this study. Upon conclusion, we will understand the information transfer in resting state brain networks as well as its role in differentiating the subtypes of vestibular impairment.