Paper Info
Title
Estimating Neural Modulation Via Adaptive Point Process Method In Brain-Machine Interface
Abstract
Brain-machine interfaces (BMIs) translate neural signals into digital commands to control external devices. During the use of BMI, neurons may change their activity corresponding to the same stimuli or movement. The changes are represented by the neural tuning parameters which may change gradually and abruptly. Adaptive algorithms were proposed to estimate the time-varying parameters in order to keep decoding performance stable. The existing methods only searched new parameters locally which failed to detect the abrupt changes. Global search helps but requires the known boundary of estimated parameter which is hard to be defined in many cases. We propose to estimate the neural modulation parameter by the global search using adaptive point process estimation. This neural modulation parameter represents the similarity between the kinematics and the neural preferred hyper tuning direction with finite range [0,1]. The preferred hyper tuning direction is then decoupled from the neural modulation parameter by gradient descent method. We apply the proposed method on real data to detect the abrupt change of the neural tuning parameter when the subject switched from manual control to brain control mode. The proposed method demonstrates better tracking on the neural hyper tuning parameters than local searching method and validated by KS statistical test.
Year
DOI
Venue
2020
10.1109/EMBC44109.2020.9175240
42ND ANNUAL INTERNATIONAL CONFERENCES OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY: ENABLING INNOVATIVE TECHNOLOGIES FOR GLOBAL HEALTHCARE EMBC'20
DocType
Volume
ISSN
Conference
2020
1557-170X
Citations 
PageRank 
References 
0
0.34
0
Authors
5
Name
Order
Citations
PageRank
Shuhang Chen113.39
Xiang Zhang210927.54
Xiang Shen313.39
Yifan Huang412.37
Yiwen Wang5559.19