The subject wears a head cap with Ag-AgCl active electrodes that connect to the amplifier (Active Two; Biosemi Inc., the Netherlands) with which electroencephalogic (EEG) signals are measured from the brain scalp. The task was to use motor imagery of the right ankle by 30 deg. The EEG signals produced during the task and during a resting period were detected. The EEG signals were sent to the amplifier where A/D conversion was done. The signal was sent through a USB receiver (via optic fiber) to a PC, where data were processed using software (LABVIEW2015; National Instruments Corp., USA). The sampling frequency was 2048 Hz. A fuzzy template matching (FTM) algorithm was used. The optimal value was obtained (FTM learning). When the FTM-processed EEG signal surpassed the threshold, a triggering signal was generated and sent to data acquisition devices (DAQ: USB-6000; National Instruments Corp., USA), and was sent by the trigger sending device (sampling frequency 2048 Hz, MARQ; Kissei Comtec Co. Ltd., Japan) to the trigger receiver. This receiver was connected to another PC where software (Matlab2016; The Mathworks Inc., USA) processes controlled triggering signals. Here, the triggering signal sends commands where the rehabilitation device operates the pullup by 30 deg for a duration of 20 s. The thresholds were set by the naked eye, initially to lower to detect motor imagery signals, and later refined to exclude weaker signals of FTM signals of resting EEG. The rehabilitation device is activated by air pressure generated by compressed air supply through an air compressor (YC-4; Yaezaki-Kuatsu Co. Ltd., Japan) to the electropneumatic regulator (RTR-200-1; Koganei Corp., Japan). Control signals were sent through an I/O board (MF634; Humusoft, Janap). The rotation angle was measured by a potentiometer (SVO1; Murata Manufacturing Co. Ltd., Japan). Data from the potentiometer were sent through the I/O board to the PC. The system continues to be activated for a certain time (20 s) once it obtains a triggering signal. Only after activation is finished does the next triggering signal become effective.
4.2 Experimental procedure
Five volunteers (age: 22–24 years old, all healthy male students) participated in the experiment. The experiment followed the Kwansei Gakuin University regulation of ethics for the Protection of Human Subjects of Medical Research, which was approved by the campus committee. Informed consent was acquired from each subject. This study protocol conforms to the Declaration of Helsinki. The subject was seated on a chair in a relaxed position with the right leg on the foot rest with the ankle bent in a natural position. The left leg remained on the floor. Electroencephalogic signals were detected by the Active Two system with eight electrodes (F3, F4, C3, Cz, C4, P3, Pz, and P4) based on the conventional 10–20 method. Only alpha (8–12 Hz) and beta bandwidths (13–30 Hz) were used for FTM detection.
4.3 Experimental protocol
Using the system explained above, we first verified that the signal of motor imagery of ankle uplift movement was indeed detected, which in turn operated the machinery. The experiment was conducted in three stages:
(1) development and practice of motor imagery for the participant, (2) system tune-ups to determine parameters of motor imagery EEG, and (3) testing for motor imagery EEG to use the device for activating machinery.
For motor imagery development, voluntary movement of the right ankle uplift was used first (voluntary task: 30 trials). (I) Motor imagery of the voluntary movement immediately followed (imagery task: 30 trials). Furthermore, the machine produced autonomous ankle movement simulating human rehabilitation behavior (autonomous movement task: 30 trials). Each trial was initiated by a trigger LED, consisted of 4 s periods. EEG was measured during this session. (II) The second process of system tune-ups consisted of two sub-steps in which (a) EEG signals of the motor imagery were detected (2 min). For comparison, a baseline (resting) EEG signal was measured during the resting period (2 min). (b) The Fuzzy Template Matching Algorithm was applied to ascertain parameters to discriminate the two states of motor imagery EEG and the resting EEG. The FTM calculation took 60 s. (III) Finally, as the third step, the FTM logic was tested by detecting a real motor imagery EEG. A trigger signal was generated as the output FTM signal. Therefore, as soon as the system detected the motor imagery EEG, the triggering signal was presumed to be generated, which was to be transferred to the robot system. To verify this point, two tasks were used: (1) participants were asked to employ motor imagery when the LED light was illuminated; (2) participants were asked to rest (rest task) to test the resting EEG signal.
To analyze the tendency of brain activity, independent component analysis was employed using EEGLAB [24]. The EEG signal was first notch-filtered (60+/-0.5 Hz and 120+/-0.5 Hz bandwidth) to exclude artifacts from electrical current, followed by band-pass filtering with frequencies of 0.04–200 Hz (FIR filtering). Special care was taken to minimize the notch filter artifacts in determining an optimal bandwidth.
4.4 Results
Figure 9 depicts the FTM outputs and the system thresholds, triggering position, and rotation angle.