The dataset is freely available on http://bci.med.tsinghua.edu.cn/download.html. The dataset was the raw continuous data without any processing. It contains 128 MATLAB MAT files corresponding to data from all 64 subjects (approximately 15 GB in total). Data were stored as double-precision floating-point values in MATLAB. Each MAT file covers a group of EEG data. There are two sets of EEG data (group A and B) for subjects from 1 to 64. The files were named as subject and group indices (i.e., sub1A.mat, sub1B.mat, . . . , sub64A.mat, sub64B.mat). For each file, the data loaded in MATLAB generate two 2-D matrices named ¡°EEGdata1¡± (block1) and ¡°EEGdata2¡± (block2) with dimensions of [64, L] (the two dimensions indicate ¡°Electrode index¡±, ¡°Time points¡±, respectively) and two 2-D matrices named ¡°class_labels¡± and ¡°trigger_positions¡± with dimensions of [2, 4000]. The parameter of L (the length of time points) might be different for different blocks. The two dimensions indicate ¡°class labels¡±, in which ¡°2¡± and ¡°1¡± indicate ¡°non-target images¡± and ¡°target images¡±, respectively. Each circle corresponds to the EEG data of a visual stimulus image. For each group, the data matrix consists of 8000 circles (100 circles ¡Á 40 trials ¡Á 2 blocks) and each circle consists of 64 channels of EEG data.