EEG Dataset for Visual Imagery

1.Research contextThis dataset contains high-resolution EEG recordings collected from 22 healthy adult participants performing visual imagery (VI) tasks.Participants imagined ten commonly encountered images across three semantic categories: geometric figures (circle, square, pentagram), animals (dog, fish, bird), and objects (cup, chair, watch, scissors). Each participant completed up to two sessions on separate days (≥48 hours apart). However, sub-09 and sub-10 completed only the first session (ses-01) due to personal reasons, and thus have no available data for ses-02. Additionally, the animal imagery data of sub-08 in ses-02 was excluded from analysis because of poor signal quality, which introduced substantial noise contamination and led to unreliable classification results.EEG was recorded with a 34-channel cap (32 active EEG channels + AFz ground + CPz reference) at 1000 Hz.2.Data overviewSubjects: 22 healthy adults(5 females, age 20-23)Task:AVI(animal visual imagery): Imagine animal images(dog, fish, bird; 120 trials/session)FVI(animal visual imagery): Imagine figure images(circle, square, pentagram; 120 trials/session)OVI(animal visual imagery): Imagine object images(cup, chair, watch, scissors; 160 trials/session)3. Experimental DesignThe paradigm was implemented using Psychtoolbox-3 and presented on a screen with a resolution of 1920×1080 pixels. Each recording session lasted between 37 and 56 minutes and comprised four VI blocks for each task category. Flexible rest periods of 2 to 4 minutes were provided between blocks to mitigate the effects of fatigue. In the figure and animal categories, each image was presented 40 times, yielding a total of 120 trials per category.Each trial lasted 17 seconds and followed a structured sequence：Fixation Cross​: 3 s. A central fixation cross ("+") is displayed on the screen. To cue participants to remain alert and relaxed, preparing them for the trial.Visual perception: 4 s. Participants are instructed to observe and memorize the image.Mask: 2 s. A visual mask is briefly displayed.To eliminate any residual visual aftereffects from the presented image.Visual imagery: 4 s.The screen turns black. Participants perform mental imagery of the previously shown image.Rest: 4 s. The word "Rest" is displayed on the screen.Participants are instructed to stop any mental imagery and rest.4.File structureThe data are organized according to the EEG-BIDS convention to maximize interoperability. Raw EEG (<code>*.bdf</code>) and BIDS sidecars: <code>*_channels.tsv</code>, <code>*_events.tsv</code>, <code>*_eeg.json</code>, etc.<br>