BUG: Somato BIDS dataset derivatives need to be updated

[larsoner]

Turns out there was a bug with the conversion in #6414 conversion. The subject used to be somato but now it's 01 and the -src.fif was never updated. We thus end up with:

>>> fwd['src'][0]['subject_his_id']
'somato'

which is wrong (and is now arguably correctly treated as an error downstream later by MNE). I have a quick fix for this but it would be good @jasmainak @sappelhoff if someone could re-create the src, fwd, and (if applicable) inv for this subject because they will all have the old, incorrect subject_his_id value.

[jasmainak]

@larsoner how urgent is this? can this wait till the sprint? @sappelhoff any chance you can fix it in the coming few days?

[sappelhoff]

Unfortunately I am completely swamped and won't be able to submit a fix for this :(

[larsoner]

It can wait until 0.20, I pushed a workaround in #6771

[wmvanvliet]

Also, dataset_description.json needs to be updated with a new link to relevant docs:

http://martinos.org/mne/stable/manual/datasets_index.html?#somatosensory ->
https://mne.tools/stable/overview/datasets_index.html#somatosensory

[larsoner]

@sappelhoff @jasmainak any chance to get to this in the next week or two to make it into 0.20?

[sappelhoff]

This is the conversion script.

If I understand you correctly, the error is with these lines:

https://github.com/mne-tools/mne-scripts/blob/9e4c77b724bb04a6d20e93720b5dd6cf110ff8eb/bids_conversion/convert_somato_data.py#L227-L231

where we simply rename the existing fwd file, instead of recomputing it. As far as I see it, there are no src or inv files.

Could a fix be as simple as loading the fwd, changing subject_his_id, and saving under a new name?

[agramfort]

yes should be as simple as this and then reupload to OSF.

…

[larsoner]

Could a fix be as simple as loading the fwd, changing subject_his_id, and saving under a new name?

I would just overwrite the old one with the correct one

[sappelhoff]

@larsoner to create a new (correct) one, I assume that I would use mne.make_forward_solution

What should I supply for the src and bem parameters?

These are all the files I have (printed up to depth 3):

|MNE-somato-data/
|--- subjects/
|------ somato/
|--------- scripts/
|--------- stats/
|--------- label/
|--------- touch/
|--------- surf/
|--------- mri/
|--------- trash/
|--------- tmp/
|--------- src/
|--------- bem/
|--- MEG/
|------ somato/
|--------- sef_raw_sss-trans.fif
|--------- somato-meg-oct-6-fwd.fif
|--------- sef_raw_sss.fif

[larsoner]

No I would not create a new one, just use read_forward_solution, modify s['subject_his_id'] for each s in fwd['src'], then write_forward_solution. It ensures that everything stays the same except the subject_his_id