Fig 1.
Stimulus presentation for the motor imagery paradigm (A) and the flickering MR/WG paradigm (B), along with (C) the multimodal system setup captured during a data collection session [45] and (D) a schematic illustrating EEG electrodes and fTCD transducers placement.
Fig 2.
Flowchart illustrating the temporal sequence of events in each data collection session.
Fig 3.
fTCD envelope calculation: (A) raw fTCD signal of a sample trial, (B) spectrogram of raw fTCD signal, and (C) the calculated envelope signal.
Fig 4.
A complete list of features provided by TSFEL.
Table 1.
Maximum accuracy achieved for each subject using weighted fusion and the corresponding accuracies obtained using EEG only and fTCD only for MI paradigm.
Table 2.
P-values showing accuracy significance of weighted probabilistic fusion compared to EEG only for the MI paradigm.
Fig 5.
Optimal alpha values: (A) the values of alpha yielding the highest accuracy for each subject in Baseline Vs Left, Baseline Vs Right, and Left Vs Right tasks, and (B) average values of alpha yielding the highest accuracy for each task (Baseline Vs Left, Baseline Vs Right, and Left Vs Right).
Table 3.
Maximum accuracy achieved for each subject using weighted fusion and the corresponding accuracies obtained using EEG only and fTCD only for flickering MR/WG paradigm.
Table 4.
P-values showing accuracy significance of weighted fusion compared to EEG only for flickering MR/WG paradigm.
Fig 6.
Optimal alpha values: (A) the values of alpha yielding the highest accuracy for each subject in Baseline Vs MR, Baseline Vs WG, and MR Vs WG tasks, and (B) average values of alpha yielding the highest accuracy for each task (Baseline Vs MR, Baseline Vs WG, and MR Vs WG).
Table 5.
Comparison between the EEG-fTCD BCI performance with proposed analysis approach and the approaches introduced previously.
Table 6.
Comparison between MI and flickering WG/MR EEG-fTCD multimodal BCIs with the proposed analysis approach and the state-of-the-art multimodal EEG-fNIRS in literature.