Optimization of Task Allocation for Collaborative Brain-Computer Interface Based on Motor Imagery

Abstract

Objective: Collaborative brain-computer interfaces (cBCIs) can make the BCI output more credible by jointly decoding concurrent brain signals from multiple collaborators. Current cBCI systems usually require all collaborators to execute the same mental tasks (common-work strategy). However, it is still unclear whether the system performance will be improved by assigning different tasks to collaborators (division-of-work strategy) while keeping the total tasks unchanged. Therefore, we studied a task allocation scheme of division-of-work and compared the corresponding classification accuracies with common-work strategy's.

Approach: This study developed an electroencephalograph (EEG)-based cBCI which had six instructions related to six different motor imagery tasks (MI-cBCI), respectively. For the common-work strategy, all five subjects as a group had the same whole instruction set and they were required to conduct the same instruction at a time. For the division-of-work strategy, every subject's instruction set was a subset of the whole one and different from each other. However, their union set was equal to the whole set. Based on the number of instructions in a subset, we divided the division-of-work strategy into four types, called "2 Tasks" … "5 Tasks." To verify the effectiveness of these strategies, we employed EEG data collected from 19 subjects who independently performed six types of MI tasks to conduct the pseudo-online classification of MI-cBCI.

Main results: Taking the number of tasks performed by one collaborator as the horizontal axis (two to six), the classification accuracy curve of MI-cBCI was mountain-like. The curve reached its peak at "4 Tasks," which means each subset contained four instructions. It outperformed the common-work strategy ("6 Tasks") in classification accuracy (72.29 ± 4.43 vs. 58.53 ± 4.36%).

Significance: The results demonstrate that our proposed task allocation strategy effectively enhanced the cBCI classification performance and reduced the individual workload.

Keywords: collaborative brain-computer interfaces; common-work; division-of-work; motor imagery; task allocation.

FIGURE 1

Experimental paradigm of a motor imagery task. At the beginning of each trial, a red fixation cross was presented at the center of the screen to remind subjects to prepare for the following task. At the first second, a symbol of instruction appeared on the screen for 4 s, subjects were instructed to perform the indicated motor imagery (MI) task up to the fifth second. This time period of 4 s was defined as a MI epoch. Then, “Rest” was displayed for 2 s to remind participants to have a rest.

FIGURE 2

Workflow of the division-of-work strategy for the proposed MI-cBCI system. Arrows indicate different types of motor imagery instructions as shown in Table 1. [+]/[−] in “panel C” means taking the following instruction as a positive/negative class. +/− in “panel D” means a positive/negative decision label.

FIGURE 3

The data processing procedure of a single user for offline modeling. X_A represents the training dataset of subject A. x means a certain class of data. [+]/[−] means taking the following data as a positive/negative class. CSP and SVM indicate CSP filters and SVM classifiers, respectively. We use the symbol F to represent the feature matrix. acc is the abbreviation of accuracy.

FIGURE 4

The data processing procedure for the feature fusion method. means that m is processed by component k (a filter or a classifier) to obtain data n. Mutual Info and Dv are the abbreviations of mutual information and decision value, respectively.

FIGURE 5

The data processing procedure for the decision fusion method.

FIGURE 6

Averaged time–frequency maps across 19 subjects for six types of MI tasks at the location of C3 and C4 electrodes. Blue indicates ERD; red indicates ERS. Black dashed line indicates the onset and offset of motor imagery.

FIGURE 7

Averaged topographical distribution for six types of MI tasks at α (8–13 Hz) and β (14–28 Hz) bands. Blue regions indicate the involved areas where ERD occurs during the MI period.

FIGURE 8

Classification accuracy curves of the feature and decision fusion methods for cBCI and single-user BCI.