[WIP] Eyetrack-EEG dataset #389
Conversation
|
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. |
|
Do you have an example we could follow? I have looked at all of the ones listed on mszinte/bids-specification#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 |
No example or precedent I am afraid. We are making history: the thrill of working on BEP... Sigh...
Will flag the TODOs in comments in the PR
We need to update this file (not urgent for now) https://github.com/bids-standard/bids-examples/blob/master/dataset_listing.tsv
|
| @@ -77,3 +77,4 @@ ieeg_visual_multimodal [@irisgroen](https://github.com/irisgroen) anat, fmap, | |||
| genetics_ukbb multiple tasks, T1w, DTI, BOLD, genetic info [@cpernet](https://github.com/cpernet) anat, dwi, fmap, func FLAIR, T1w, bold, dwi, events, info, magnitude1, phasediff | |||
| motion_dualtask older and younger participants walking while performing discrimination task [@sjeung](https://github.com/sjeung) eeg, motion channels, eeg, events, motion, scans | |||
| motion_spotrotation participants rotated heading using full-body motion or joystick [link](https://openneuro.org/datasets/ds004460) [@sjeung](https://github.com/sjeung) eeg, motion channels, coordsystem, eeg, electrodes, events, motion, scans | |||
| eyetrack_eeg eye-tracking (plus simultaneous EEG) dataset https://drive.google.com/drive/folders/1kzI5Lwg__BvDKGup37QO2lLLzBTlHe0L?usp=sharing [@arnodelorme](https://github.com/arnodelorme) eeg channels, coordsystem, eeg, electrodes, events, eyetrack | |||
There was a problem hiding this comment.
@arnodelorme not sure if the dataset will be on google drive forever
if the actual data is stored somewhere in the end, the link in the tsv will have to be updated
| "ScreenDistance": "REQUIRED", | ||
| "ScreenResolution": "REQUIRED", | ||
| "ScreenSize": "REQUIRED" |
There was a problem hiding this comment.
REQUIRED metadata missing
| "SampleCoordinateUnit": "REQUIRED", | ||
| "SampleCoordinateSystem": "REQUIRED", |
There was a problem hiding this comment.
REQUIRED metadata missing
There was a problem hiding this comment.
not checked this file but it may be good to truncate it at least partially to avoid getting bloating the repo
also this tsv has a lot of extra columns: this is fine but it would be a better examples if their content was described in its side car json
|
Thanks @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 I am sure that there are potential issues I am missing with this approach so let me know what you think. |
| 0.074626866 226 10 93 0 | ||
| 1.76119403 46 236 4 1 | ||
| 2.104477612 348 282 8 1 | ||
| 4.701492537 76 630 4 1 | ||
| 5.268656716 433 706 8 1 |
There was a problem hiding this comment.
@arnodelorme
are those durations in seconds?
There was a problem hiding this comment.
yes, first column is in second. The second column is in samples for the EEG (EEG files) and samples for eye-tracking (eye-tracking file) - but at different sampling rates.
There was a problem hiding this comment.
duration column MUST be in seconds
https://bids-specification.readthedocs.io/en/latest/glossary.html#objects.columns.duration
samples will have to be dropped from the events.tsv if we are merging data with different sampling scheme: this was the old recommendation in the spec before the samples column got dropped altogether from the spec
There was a problem hiding this comment.
OK, I understand about dropping the "sample" column.
|
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 |
for BEP20
from @arnodelorme suggestion
see here
bids-standard/bids-specification#1128 (comment)
TODO