FBIRN:FBIRN Ontology plans Oct 2005

Summary of the FBIRN Ontology discussions, Oct 17-19, 2005

1) Our aim for next October is: Be able to query the FBIRN HIDs and BrainMap to “Find the healthy control datasets doing fMRI and a working memory task.” That is a boring question (but there aren’t any non-healthy control data in the FBIRN HIDs, actually); but it gives us a concrete achievable, and the capability to answer a toy question should bring with it the ability to answer other, much better queries. That requires:

      i.      BrainMap being wrapped/registered as noted below

      ii.      Marking up both the HID and BrainMap database(s) as noted below

      iii.      An interface that can access all the data sources; map the query to concept IDs; and answer the query at least by finding an “exact match” in the concept IDs (just to get us started)

      iv.      Maybe, if we’re lucky, we can have the ability to generalize beyond an “exact match” with a particular concept ID. (We have to get there eventually! but possibly not by next October)

      v.      Identify by that point what the issues are regarding better/improved/non-toy queries, more interesting search algorithms, etc.

2) Angela to standardize the BrainMap database schema and begin discussion with Amarnath/Jeff regarding their structure/interoperability

3) Angie, Jess, and Chris to review the “cognitive primitives” list: Compare it to APA thesaurus, bounce it off the FBIRN Cog group, discuss with Cog Phenomics people at UCLA if possible

4) By March: Angie, Jess, and Chris—to agree on the final (for now) cognitive primitives list; start mapping them to UMLS and seeing what works, and what relationships get lost or added that we don’t want. The interaction with Chris is needed to include the aspects of the clinical issues that MBIRN is working on. Report on that progress at FBIRN and MBIRN March AHMs.

5) Jess: Sit down with David, Dingying and Jeff asap and get the scope of what “marking up the database” entails for the FBIRN HIDs.—maybe over drinks at SFN? This discussion has to be held substantially prior to the Markup Boot Camp. I am NOT sure how much of the HID needs to be marked up in order to achieve #1.

6) Markup Boot Camp in December—8 hours of working through the HID and/or BrainMap to find appropriate concepts or identify needed concepts. From FBIRN that will probably be Jess and Dingying. Angie and Angela offered to host all of us at San Antonio if we want to get out of town.

7) And of course, FBIRN starts advertising for a Real Ontologist™. Jess will be in touch with everyone regarding what should go in the job ad and where we should recruit from, in order to get a person who can be truly helpful. Hopefully that can go out in early December.

---