ssmvep accuracy test

Experiment DesignThis was an offline Brain-Computer Interface (BCI) experiment designed to evaluate the accuracy of the Task Discrimination Component Analysis (TDCA) algorithm using the Steady-State Motion Visual Evoked Potential (SSMVEP) paradigm. The experiment was divided into a training phase (20 trials) and a testing phase (40 trials).The visual stimulation consisted of two composite stimuli, one positioned on the left and one on the right of the screen. Each stimulus featured a character image overlaid on a corresponding circular background. Like, the left character performed a left leg-lifting action and corresponding circular backgrounds performed a zoom motion. The two composite stimuli moved at distinct frequencies: 7 Hz (left) and 8 Hz (right). The stimuli were presented in a pseudorandom order throughout the experiment.The procedure for each trial, which lasted a total of 5.5 seconds, was as follows:Cue Phase (2.0 s): A yellow inverted triangle appeared, visually cueing the target stimulus (left or right). Subjects were instructed to quickly shift their gaze to the target.Stimulation Phase (2.5 s): The corresponding character figure and background circle began to move concurrently 0.5 seconds after the cue onset.Training Stage: No feedback was provided.Testing Stage: A green inverted triangle was displayed below the character image to facilitate continuous visual fixation. <i>Crucially, the position of this green feedback was dependent on the real-time TDCA decoding results.</i>Rest Phase (1.0 s): Immediately following the stimulation offset, the screen remained static. A sound cue was given, indicating a brief rest before the next trial began.In this experiment, 20 subjects participants were involved in the test. Electroencephalography (EEG) data were acquired using a custom-made 8-channel device at a sampling rate of 250 Hz. Eight electrodes were placed over the parietal and occipital areas (PO3, PO4, PO7, PO8, POz, Oz, O1, and O2) to specifically record SSMVEPs. The Task Discrimination (TDCA) algorithm was used for data decoding.The raw data were stored in a CSV file (or corresponding label files) with 11 columns, where the first eight columns represent the raw EEG signals. The remaining columns contain metadata: corresponding to trial, phase, and position separately.ResultsThe average decoding accuracy achieved using the TDCA algorithm across all participants was 77.5%.<br>