Demo data

Sample EEG dataset

A sample EEG dataset has been compiled to test all the MVPAlab main functionalities. Three different EEG data files have been selected from the original work.

For each participant, two different main conditions (condition_a vs. condition_b) have been selected for decoding analyses. Additionally, four subconditions (condition_1, condition_2, vs. condition_3 and condition_4) have been selected for multivariate cross-classification analyses.

During the original study, high-density EEG was recorded from 65 electrodes. The TP9 and TP10 electrodes were used to record the electrooculogram (EOG) and were removed from the dataset after the preprocessing stage. Impedances were kept below 5kΩ and EEG recordings were average referenced, downsampled to 256 Hz, and digitally filtered using a low-pass FIR filter with a cutoff frequency of 120 Hz, preserving phase information. No channel was interpolated for any participant. Continuous data were epoched [−1000, 2000ms centered at onset of the stimulus] and baseline corrected [−200, 0ms]. Independent Component Analysis (ICA) was computed to remove eye blinks from the signal, and the artifactual components were rejected by visual inspection of raw activity of each component, scalp maps and power spectrum. Finally, an automatic trial rejection process was performed, pruning the data from non-stereotypical artifacts.

The final compiled dataset consists of an EEGlab data structure per subject and condition with [63 x 768 x ntrials] EEG data matrices and it is freely available for download in the following repository:

Download demo dataset

Readers interested on the experimental details of these data should refer to the original publication.