Fig. 1. Scheme for calculating event-related heart rate changes (data from subject S2, VE condition). (A) Original ECG signal with marked QRS complexes. (B) Interbeat interval between the QRS complexes and calculation of instantaneous heart rate (IHR). (C) IHR time course in beats-per-minute (bpm) of one run, with indicated cue-stimuli (start of the motor imagery). (D) Zoomed version of (C), additionally the type of cue, either right hand imagination R or feet imagination F is exemplarily indicated. (E) Average of event-related heart rate changes with 3 s before and 5 s after the cue-stimulus presentation. Mean HR is plotted in solid and standard deviation (SD) in dotted. (F) Relative event-related HR change in percent compared to the mean HR in a 1-s reference interval between second 1 and 2 (normalized HR change). (G) Timing of the one trial. At second 3 the cue-stimulus appears and between second 4 and 8 feedback is given to the subject. After each trial a pause of random duration between 0.5 and 3 seconds occurs.

Fig. 2. Results of one control session composed of 4 runs (subject S1) are displayed in the upper panel (A). Time courses of the classification error in percentage are plotted in the top row. The vertical lines indicate the time point of the cue presentation. The bottom row displays the HR time courses in beats per minute (bpm). The thick line represents the mean over the trials and the two thin lines represent the corresponding inter-trial SD. The last column on the right side (framed) is the outcome of one session, which equals the averaged result over the 4 runs and 160 trials, respectively. The results of one VE experiment of the same subject (S1) are shown in the lower panel (B).

Fig. 3. Normalized HR time courses (mean ± SE) for VE (plotted in black) and control (PC, plotted in orange) condition. For calculation 160 trials are used in the control condition and 280/240 (S1 and S2/S3) trials (exclusive of the first run) in the VE condition. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4. Normalized HR time courses (mean ± SE) calculated for correct classified foot MI trials (126 in S1 and 129 in S2, plotted in black) and false classified hand MI trials (35 in S1 and 28 in S2, plotted in orange). Correct classification resulted in forward moving and false classification in backward moving. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)