GuttmannFlury2025 — Eye-BCI multimodal dataset: Motor Imagery (re-hosted from Synapse, EEG converted to BDF)

Re-hosted EEG multimodal data (Motor Imagery) from Guttmann-Flury, Sheng & Zhu (2025), originally deposited on Synapse (DOI: 10.7303/syn64005218) under a CC0 public domain dedication. EEG data converted from Neuroscan CSV to BDF format for ~3x space reduction with negligible loss (max 0.006 μV error). Paper E. Guttmann-Flury, X. Sheng, and X. Zhu, “Dataset combining EEG, eye-tracking, and high-speed video for ocular activity analysis across BCI paradigms,” Scientific Data, 12, 587, 2025.DOI: 10.1038/s41597-025-04861-9 Paradigm: Motor Imagery Classes2 classes: left hand, right hand (grasping imagery)Trials40 trials/session (20 left + 20 right)Timing2 s fixation + 4 s imagery + 1-1.5 s restDuration~6-7 min per recordingEventsTrig code 1 = left hand, 2 = right hand Participants 31 healthy subjects (20 male, 11 female) Age: 20–57 years (mean 28.3) 25 right-handed, 2 left-handed, 4 ambidextrous 1–3 sessions per subject (63 sessions total) Recording setup ParameterValueSystemNeuroscan Quik-Cap 65-ch + SynAmps2Channels62 EEG + 2 mastoid (M1/M2) + 1 EOG (HEO) + 1 STIM (Trig) = 66Sampling rate1000 HzReferenceRight mastoid (M1)Line frequency50 HzFile formatBDF (24-bit, converted from CSV) Multimodal data per recording FileContentMIXXX.bdfEEG (66 channels, BDF)MIXXX_sync.csvMetadata sidecar (timestamps, cues, video sync, blinks)MIXXX_annotations.jsonRich trial annotations with E-Prime timing metadataMIXXX_tobii.csvTobii eye-tracking gaze dataMIXXX_phantom.aviHigh-speed eye video (~500 fps)MIXXX_phantom.xmlEye video annotations (pupil tracking)MIXXX_eprime.txtE-Prime stimulus timing/cues File structure 31 per-subject ZIP files (S01.zip–S31.zip). Each contains 1–3 session folders: S01.zip Sess01/ MI011.bdf MI011_sync.csv MI011_tobii.csv MI011_phantom.avi MI011_phantom.xml MI011_eprime.txt ... S31.zip Conversion details EEG: Neuroscan CSV → BDF (24-bit), ~3.2x compression Round-trip error: max 0.006 μV (0.000003% of signal range) Non-EEG CSV columns (timestamps, cues, video sync, blinks) preserved in _sync.csv sidecar Eye-tracking, video, E-Prime files: copied byte-for-byte (no conversion) Reading the data import mne raw = mne.io.read_raw_bdf('MI011.bdf', preload=True) print(raw.info) # 66 channels, 1000 Hz License CC0 1.0 Universal (Public Domain Dedication), as stated in the original paper (Section “Data Records”).