Neural Prostheses
A. Developing an innovative precise neuromodulation system for treating Parkinson’s disease
Introduction :Parkinson’s disease (PD) is a neurodegenerative disease that affects more than 10 million people in the world. Although medication such as levodopa or dopaminergic agonist has been widely used to suppress the motor deficits, medication is just a symptomatic therapy which becomes less or even adversely effective after long-term usage. Deep brain stimulation (DBS) has been a promising alternative for treating the PD by suppressing abnormal neural synchrony in the cortico-basal ganglia circuit. However, contemporary clinical treatment turns on DBS continuously after its surgery implant, and such “open-loop” DBS is found to induce both motor and psychiatric sideeffects. Two major obstacles for DBS are (1) a limited understanding of the mechanism underlying DBS effects; (2) the lack of a high-density neuromodulation device able to deliver high spatiotemporal stimulation protocols and record neural activity across the entire cortico-basal-ganlia circuit. Trough interdisciplinary collaboration with neuroscientists, medical doctors in neurology, biomedicine experts, our goal is to develop more effective and less invasive neuromodulation procedures for treating PD, as well as to provide in-depth understanding of the functional dynamics during neuromodulation in the cortico-basal-ganlia circuit.
Characteristics : This research relies on close collaboration with neuroscientists, medical doctors, and biomedicine experts. The main objective of our lab is to develop a multifunctional microchip able to monitor and modulate neural activities by electricity, light, etc.
Pre-requisites:Knowledge in electronics, VLSI design, basic neuroscience, and basic electrochemistry, signal processing, machine learning algorithms
Professional skills obtained:Knowledge in neuro-electrophysiology, Methodology in modeling and designing bio-electronic interfaces, Design of mixed-mode integrated circuit for neural recording and stimulation, as well as for detecting neurotransmitters
B. Suppressing pathological spindles with high efficacy by Closed-loop DBS
Introduction :The high-voltage spindle (HVS) is a spike-and-wave, neural oscillation that has been identified as a pathological signature related to the dopamine depletion in Parkinsonian rats. It is of great interests to investigate whether suppressing HVSs helps to improve the motor deficits associated with the HVS, or even to delay the progression of the Parkinson’s disease. As each HVS episode occurs randomly and lasts for only a few seconds, an adaptive neuromodulator is demanded for predicting the onset of HVSs and suppressing the HVSs by deep-brain stimulation at precise timing. This study uses an adaptive neuromodulator, called NeuLive (https://www.bioproweb.com/series/neulive/#product=neulive) consisting of a lightweight(~4g) neuro-interfacing headstage and a wireless data hub. The proposed algorithm for detecting the HVS is based on the autoregressive modelling at interval (ARt), whose parameters are adaptable on-line by a Kalman-filter framework. In our pilot study, the algorithm is proved able to predict the onset of HVS with high precision (96%) and short latency (61ms). This study further implements the algorithm in the wireless neuromodulator for experiments with freely-moving Parkinsonian rats. The stimulation duration as short as 0.2s is found sufficient for suppressing the HVS effectively, even though the mean duration of HVS episodes is 4.3s. This intriguing finding indicates that precise control on the timing of brain stimulation is crucial for increasing stimulation efficacy while reducing stimulation dosage and thus its side-effects.
Watch our presentation at SfN2021: https://youtu.be/vlGa8imVqf8
Characteristics : This research relies on close collaboration with neuroscientists to collect physiological data from animal experiments, to design algorithms for detecting pathological signatures automatically, and finally to verify the efficacy of closed-loop neuromodulation in animal study.
Pre-requisites:Knowledge in signal processing, embedded system, FPGA, and machine learning.
Professional skills obtained:Knowledge in neuro-electrophysiology and neural diseases, Modeling neural activities, Machine learning algorithms for predicting and detecting pathological neural activities reliably.
