[WIP] Eyetrack-EEG dataset

[Remi-Gau]

for BEP20

from @arnodelorme suggestion

see here https://github.com/bids-standard/bids-specification/pull/1128#issuecomment-1621201447

TODO

• [ ] merge events files from eyetrack and eeg
• [ ] add missing required metadata
• [ ] update eyetrack.json to document what other columns are
• [ ] update dataset listing

[arnodelorme]

Thanks @Remi-Gau I assume all the checks need to pass (is that running the BIDS validator?).

[Remi-Gau]

Thanks @Remi-Gau I assume all the checks need to pass (is that running the BIDS validator?).

Don't worry about the checks for now because they will only ever pass once the validator has been updated to validate eyetracking data: so failure is expected for now.

[arnodelorme]

Do you have an example we could follow? I have looked at all of the ones listed on https://github.com/mszinte/bids-specification/issues/3

However, none contains EEG and eye-tracking data (or even eye-tracking as a secondary measure). I am happy to move files around, but I am also unclear on what you mean by "add missing required metadata" and "update dataset listing"

If there is no example and no existing documentation, I will do what I can and you can provide feedback. Thanks

[Remi-Gau]

Do you have an example we could follow? I have looked at all of the ones listed on mszinte/bids-specification#3

No example or precedent I am afraid. We are making history: the thrill of working on BEP... Sigh...

However, none contains EEG and eye-tracking data (or even eye-tracking as a secondary measure). I am happy to move files around

but I am also unclear on what you mean by "add missing required metadata"

Will flag the TODOs in comments in the PR

and "update dataset listing"

We need to update this file (not urgent for now)

https://github.com/bids-standard/bids-examples/blob/master/dataset_listing.tsv

If there is no example and no existing documentation, I will do what I can and you can provide feedback. Thanks

[arnodelorme]

Thanks @Remi-Gau @deepagupta please use the files in https://drive.google.com/drive/folders/1kzI5Lwg__BvDKGup37QO2lLLzBTlHe0L?usp=drive_link And modify them according to Remi's advises on this tread. We can check them together before showing them to Remi

[arnodelorme]

@Remi-Gau , one more question. The eye-tracking and EEG are at different sampling rates, yet they will share the same event file. I assume this means that the eye-tracking data need to be upsampled x10. This is far from ideal (and a lot of people will complain about that). Any other suggestions?

[Remi-Gau]

@Remi-Gau , one more question. The eye-tracking and EEG are at different sampling rates, yet they will share the same event file. I assume this means that the eye-tracking data need to be upsampled x10. This is far from ideal (and a lot of people will complain about that). Any other suggestions?

Silly suggestion: I think that then the value and sample columns would then have to be dropped: those columns are be moved from optional to arbitrary. See: https://github.com/bids-standard/bids-specification/pull/1457

I am sure that there are potential issues I am missing with this approach so let me know what you think.

[arnodelorme]

Following your advice @Remi-Gau, I have finally worked out the alignment of EEG and eye-tracking using both the SMI and the Gazepoint eyetrackers. The BIDS repo passes the validator except for the eye-tracking files.

The export (including data) is available at https://drive.google.com/drive/folders/1jEKySJdLc2cdO12y_MENxUHC95cjVIq9?usp=sharing