[SCHEMA] Change modalities to instruments

[tsalo]

Closes #593 and possibly closes #1141. The problem that this PR attempts to address is that the definition of "modality" in Common Principles conflicts with its use in the schema.

Related proposal from Robert Oostenveld: https://docs.google.com/document/d/1cY-iZL8rtgDJ1-YVs1f2h5ugOiuK144zu4ju9UXZEGY/edit

Changes proposed:

• Rename schema/objects/modalities.yaml to schema/objects/instruments.yaml.
• Rename schema/rules/modalities.yaml to schema/rules/instruments.yaml.
• Add a new "Instrument" element to the Common Principles.
• Reference instrument in rules, rather than modality.
• Change any cases where modality is checked against EEG or iEEG to "bioamp"

[codecov]

Codecov Report

Merging #1237 (19762bc) into master (69e1bde) will not change coverage.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master    #1237   +/-   ##
=======================================
  Coverage   88.09%   88.09%           
=======================================
  Files           6        6           
  Lines         974      974           
=======================================
  Hits          858      858           
  Misses        116      116           

Impacted Files	Coverage Δ
tools/schemacode/bidsschematools/validator.py	94.54% <100.00%> (ø)

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[tsalo]

This describes the new "instrument" common principle.

[tsalo]

This is a core change- now EEG and iEEG are both considered "biopotential amplification". At some point it would be nice to reorganize the specification to reflect this, but I don't think it's necessary.

[tsalo]

@nellh noted that OpenNeuro uses the term "Modality", and keeps EEG and iEEG separate. If we change it to "Instrument" and combine EEG and iEEG under "Biopotential Amplification", OpenNeuro would likely be impacted, and users probably want to search for either EEG or iEEG, not both.

[CPernet]

1. I think add 'modality' to glossary #1141 makes the point that modality is not well defined in BIDS, and as such is a separate issue of how OpenNeuro uses it.
2. I agree that people want to search iEEG, MEG, EEG on their own, and @robertoostenveld said so too. @nellh could rely on the datatype rather than instrument (previous modality) and everything should keep running as before
3. I'd have approved those changes

[effigies]

2. I agree that people want to search iEEG, MEG, EEG on their own, and @robertoostenveld said so too. @nellh could rely on the datatype rather than instrument (previous modality) and everything should keep running as before

Yes, OpenNeuro can maintain its own mapping. It has been nice to rely on BIDS validation outputs directly, as it reduces maintenance tasks when updating software for new BEPs and avoids getting out of sync.

I wonder if we could keep modality as an unexposed component of the validator that OpenNeuro (and other tools that want a more intuitive, albeit less logically consistent, grouping) could rely on, so that this information does stay single sourced. So it would get emitted by the validator JSON but not be exposed in UI or used in schema rules.