基于数据驱动的闭环脑机接口设计

基于神经元峰电位的植入式脑机接口开展相关研究，通过搭建大脑皮层仿真模型，并在控制理论分析的基础上进行自发单关节运动任务。使用自适应维纳滤波器完成神经元放电活动的线性解码器设计。通过分析发现在视觉反馈信息缺失时，解码器性能严重下降。针对此问题，使用基于数据驱动算法的紧格式无模型控制算法产生刺激信号来刺激仿真大脑皮层感觉区神经元，使其跟踪存在感官反馈时感觉区神经元的放电活动。由于感觉区神经元放电信息的恢复，解码器在感觉反馈信息缺失时的性能也得到了恢复。最后，通过仿真验证了基于数据驱动算法的人工感官反馈有效性，并与整定的PID控制算法对比，结果验证了本文设计的闭环系统的有效性。

Design of Closed Loop Brain Computer Interface Based on Data Driven

In this paper,we consider the simulation model of cortical neural network and the analysis of the single joint movement tasks based on the control theory.The linear decoder design of neuron discharge activity is designed using adaptive Wiener filter.The performance of the decoder is severely reduced when the visual feedback information is missing.Hence,this paper further utilizes a data driven tightly formatted model-free control algorithm to generate stimulus signals such that the neurons in the sensory area of the cerebral cortex can be stimulated to track the neuronal firing activity in sensory areas.Due to the recovery of neuron discharge information in the sensory area,the performance of the decoder in the absence of sensory feedback information can be also restored.Finally,the simulation results illustrate the effectiveness of the artificial sensory feedback based on data driven algorithm,which also verify the performance of the closed loop system in this work,compared with PID algorithm.