To evaluate fatigue real time and the laparoscopic training effect,and to find the relationship between them.Data acqui-sition module and fatigue analysis module were designed based on MATLAB.During the training,EEG data were collected by Brainlink and sent to data acquisition module wirelessly by bluetooth,the raw data′s artifacts were eliminated by the way of blind source separa-tion(BSS).And then,4 kinds of EEG rhythm were detected using the method of wavelet packet analysis.Fatigue values were calcu-lated using Semi supervised regression model of Mahalanobis distance.This module's threshold value was evaluated using receiver op-erating characteristic(ROC)curve.When fatigue values exceed the threshold, it indicated that fatigue occured and warning tips for termination of training appeared.Eye movement signals were collected by Tobii X 1 Light Eye Tracker,the effect of laparoscopic training was evaluated using the parameters of completion time,fixation time,error accounts,fixation accounts,the percentage of fixation time and the percentage of fixation/saccade.In 1~6 times training,fatigue values rised without fatigue occurrence,and the training effects became better;in 7~15 times, fatigue values changed relatively flat, and training effect tended to a stable plateau; in the 16~20 times,fatigue values rised significantly with fatigue occurring and the training effects declined.The fatigue level has effects on training effect.When the fatigue value reaches the fatigue threshold of 10,namely fatigue occurrence,the train should be terminated.%实时评价模拟腹腔镜训练者训练时的疲劳值,并分析疲劳与训练效果的关系.基于MATLAB平台设计了数据采集模块和疲劳分析模块.在训练过程中,Brainlink采集脑电数据,通过蓝牙将数据发送到数据采集模块,然后通过盲源分离的方法去除伪迹,应用小波变换提取4种节律波,采用马氏距离的半监督回归模型得出疲劳值.采用ROC曲线确定模型的阈值,当超过阈值时表示疲劳事件发生,报警提示终止训练;利用Tobii眼动仪采集眼动信号,计算完成时间,错误数,注视点个数,注视时间百分比以及注视/眼跳时间百分比来评价训练效果.在1~6次训练中,疲劳值略微上升但疲劳还未发生时,训练效果逐渐变好;在7~15次中,疲劳值略微上升,训练效果趋于稳定;在16~20次中,疲劳值呈上升趋势,疲劳发生时,训练效果下降.训练者的疲劳程度对训练效果有一定的影响.当疲劳发生时,训练效果下降,当疲劳值达到疲劳阈值10时,应终止训练.
展开▼
