NIDM Experiment Export directly from Tasks

[kastman]

In the same way that NI-DM Workflow provenance is being exported directly from FSL, SPM, etc., I'm thinking a little bit about exporting NI-DM Experiment tasks, contrasts, etc. from task software. Is anyone working on adding a little more resolution to the task / model part of the draft (2.1.1.[5-9])? I'm thinking specifically in concert w/ other ontologies like CogPo (@tsalo) / CogAtlas.

Also, is the gitter.io/incf-nidash room a good spot? It seems mostly abandoned but it may be a better spot for this kind of conversation than gh issues.

[jbpoline]

sounds a great idea - not sure where is the best discussion place, I suppose gitter should be fine - or the nidm email list ? cheers JB

On 7 June 2017 at 19:24, Erik Kastman notifications@github.com wrote:

In the same way that NI-DM Workflow provenance is being exported directly from FSL, SPM, etc., I'm thinking a little bit about exporting NI-DM Experiment tasks, contrasts, etc. from task software. Is anyone working on adding a little more resolution to the task / model part of the draft (2.1.1.[5-9])? I'm thinking specifically in concert w/ other ontologies like CogPo (@tsalo https://github.com/tsalo) / CogAtlas.

Also, is the gitter.io/incf-nidash room a good spot? It seems mostly abandoned but it may be a better spot for this kind of conversation than gh issues.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/incf-nidash/nidm/issues/411, or mute the thread https://github.com/notifications/unsubscribe-auth/AAQyaPCZl-JdfY1oV3_af8lD6tAmP3-Eks5sBtzQgaJpZM4NzB-T .

[kastman]

Also, which email list? I found this: https://groups.google.com/forum/#!forum/nidm-users listed in the spec itself, but it also seems abandoned.

[cmaumet]

@kastman: That's a great idea! Did you have specific task software in mind?

At some point, we discussed adding CogAtlas terms in NIDM-experiment but I don't think we've included them so far. @dbkeator and @khelm are currently focusing on NIDM-experiment and might have more updated info?

Regarding the mailing lists/gitter, I think @jbpoline was referring to the incf-nidash-nidm google group of which @dbkeator is the owner but emailing nidm-users is also a possibility (same people on both lists).

[kastman]

I've just finished a task battery for the HCP-Lifespan project, and would love to include structured, self-documenting descriptions of the tasks when we publicly release them. I've been a core maintainer of Psychopy since 2012, and used it for the HCP-L tasks, so that's what I'd use for implementation first.

The real motivation is to describe the tasks (basically a spec to directly link the abstract research questions with concrete regressor onset names and contrasts, for people who weren't involved with the task creation but may still want to use them). Exporting NIDM provenance from the task software itself is sort of secondary interest to developing the spec, but I figure that if I'm going to go to the trouble of writing something for myself I may as well make it convenient for others.

I've already had preliminary contact with both the CogPo and CogAtlas maintainers, and will be sending around some examples of the new HCP task spec proposals once I've finished them (possibly as a behavioral extension to BIDS, instead?), but I'm first trying to get a feel for anything that may already have been developed in order to avoid duplicating effort. I'd love to hear any updates on the NIDM Experiment side of things and would be happy to hop on the call if you send me info.

[jbpoline]

Hi Erik

You should now be on the google group - sounds like a useful project - Just to let you know, a small number (~10) of nidm + bids people will be at the OHBM conference and also 2 days after that - if you are free on Friday June 30 / Saturday July 1st let me know, we will be in one of the hotel conference room

cheers JB

On 8 June 2017 at 13:35, Erik Kastman notifications@github.com wrote:

I've just finished a task battery for the HCP-Lifespan project, and would love to include structured, self-documenting descriptions of the tasks when we publicly release them. I've been a core maintainer of Psychopy http://www.psychopy.org since 2012, and used it for the HCP-L tasks, so that's what I'd use for implementation first.

The real motivation is to describe the tasks (basically a spec to directly link the abstract research questions with concrete regressor onset names and contrasts, for people who weren't involved with the task creation but may still want to use them). Exporting NIDM provenance from the task software itself is sort of secondary interest to developing the spec, but I figure that if I'm going to go to the trouble of writing something for myself I may as well make it convenient for others.

I've already had preliminary contact with both the CogPo and CogAtlas maintainers, and will be sending around some examples of the new HCP task spec proposals once I've finished them (possibly as a behavioral extension to BIDS, instead?), but I'm first trying to get a feel for anything that may already have been developed in order to avoid duplicating effort. I'd love to hear any updates on the NIDM Experiment side of things and would be happy to hop on the call if you send me info.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/incf-nidash/nidm/issues/411#issuecomment-307219820, or mute the thread https://github.com/notifications/unsubscribe-auth/AAQyaCrVAPbH2RM3XsmgEhwXjww-TcWoks5sCFsNgaJpZM4NzB-T .

[kastman]

Thanks for adding me - looking forward to continuing the conversation. I'm sad to miss the hackathon - unfortunately I'm missing OHBM this year, but I'll keep my eye out for future events.