This proposal is to establish a cyber infrastructure to detect and visualize the brain's activities while the human body is in motion, performing various limb movements. Any disturbance in the brain-body functional coupling affects one's ability to move efficiently, causing various movement disorders such as Parkinson's disease that is the second most common neurodegenerative disorder, affecting 4 million worldwide. Acquiring deepened understanding of such movement disorders demands to study brain activity and body movements simultaneously. Due to technical constraints, traditional brain scanners such as functional magnetic resonance imaging (fMRI) require the body to remain horizontal and motionless during scanning periods. Moreover, fMRI is bulky and immobile, providing one image per second, while the brain's activity occurs at a much faster rate. Hence, there is a limited knowledge of brain dynamics that are coupled tightly to body's mobility behaviors. In recent years, "functional near infrared spectroscopy (fNIRS)" is emerging as a portable, optical neuroimaging device that provide greater benefits compared to the bulky counterparts such as fMRI. The PI takes the fNIRS technology one step further by integrating it with the maturing technology of body sensor networks (BSN) to quantify brain-body connectivity.
The project aims to establish a medical cyber-physical system (mCPS) delivering the system integration of fNIRS and BSN. Specifically, the project mainly explores the following directions: System Integration: The mCPS is a system integration aggregating two complex subsystems (20-channel fNIRS system and 17-sensor body motion suit) to produce a unique interface delivering the overarching functionality of motion-tagged neuroimaging. Challenges in this project emerge from the heterogeneity of system components and interactions among them. In this work, system integration takes place into three forms: 1) "hardware integration" for centralizing the multimodal sensor data, 2) "data integration/fusion" for conditioning and fusing multi-dimensional task-level signals, and 3) "presentation integration" for combined visualization of brain and body behaviors. Motion-Tolerant Neuroimaging: It is expected that fNIRS neuroimaging data will face challenges of motion artifacts. Brain activity data would largely be superimposed by the mechanical vibrations traversed to the scalp due to the body movements. The PI explores harmonic sum models to rectify motion-affected brain activities to extract the more accurate hemodynamic response of the brain in mobile settings. A Clinical Screening Interface for Neurologists: The PI collaborates with a neurologist specialized in treating Parkinson's disease. The end result will be a clinical tool for physicians to screen the motor exams of patient with Parkinson's disease. The clinical tool is aimed at providing quantified data of motion and brain's cortical activities displayed side-by-side for the improved diagnostic screening of Parkinson's disease.
