SMI and gazepoint dataset

[Remi-Gau]

Below is an EEG and Eye-tracking dataset

https://drive.google.com/drive/folders/1jEKySJdLc2cdO12y_MENxUHC95cjVIq9?usp=drive_link

The EEG and eye-tracking are perfectly synchronized and are at the same sampling frequency. There is only one event file for both.

Originally posted by @arnodelorme in #3 (comment)

[Remi-Gau]

@arnodelorme

Am moving this to a separate issue to discuss.

[Remi-Gau]

Things to change on the dataset according to the state of this BEP

https://bids-specification--1128.org.readthedocs.build/en/1128/modality-specific-files/physiological-recordings.html#eye-tracking (as of 12th april 2024)

Only talking about sub-01 here, you should get the logic to extend to subject 2 even if the eyetracker is a different one.

DATA

• split data by eye: one file for left eye, the other for the right eye
• rename file from sub-001_task-HBNdata_eyetrack.tsv to:
  • sub-001_task-HBNdata_recording-eye1_physio.tsv.gz
  • sub-001_task-HBNdata_recording-eye2_physio.tsv.gz
  • given the name of the columns it seems that some of them refer to only the left eye or the right eye so they should be put in the correct file (for example LEYEX | LEYEY | LEYEZ | LPUPILD | LPUPILV | REYEX | REYEY | REYEZ | RPUPILD | RPUPILV)
• because these are now physio files, the files must be tsv.gz and have no headers
• eye_timestamp and eye1_y_coordinate columns seems to be duplicate of each other

METADATA

• the name of the columns should be in sub-001_task-HBNdata_recording-eye1_physio.json and sub-001_task-HBNdata_recording-eye2_physio.json
• eye_timestamp becomes timestamp

For example

  "Columns": ["timestamp", "x_coordinate", "y_coordinate", "pupil_size"],

• make sure you got a description of each column to help data users know what each file is

    "column_name": {
      "Description": "foo bar",
      "Units": "proper unit in SI if possible"
    },

• The following metadata should go in an sub-001_task-HBNdata_events.json not in the eyetracking metadata "ScreenSize", "ScreenResolution", "ScreenDistance". The rationale is that you could still report those metadata even if you had no eyetracking: but we have made them required when working with eyetrakcing.
• Missing metadata: StartTime, PhysioType should be "eyetrack", RecordedEye must be "left" or "right".

[arnodelorme]

Thank you @Remi-Gau for taking the time. I will make the necessary changes.