Model for RDF graph with participant node

[smpadhy]

Currently, a plan, say Plan_1, has several attributes <ID,NumberOfSession,Name,Description, SessionCollections>. A Session Collection has members <Session1, Session2, ...>. Each Session has attributes<name, Instrument collections >. Instrument Collection has members <Instrument1, instrument2, ...>. An instrument has attributes <task name, description, assignee, time estimate, instrument name>.

While adding a participants node, it should be added to the plan, Plan_1, as another attribute with 'wasAssociatedWith' predicate. Since there can be several participants for an experiment plan, corresponding to each participant a node will be added to the RDF Plan graph.

Another alternative could be to have a participants Collection associated to the plan and each participant could be member of that collection.

Another question - should participant be part of plan or part of data collection activity? As part of plan, there might be prescreen session to recruit participants.

[smpadhy]

@satra @dbkeator @khelm Any thoughts?

[khelm]

Hi @smpadhy We've decided to use Subject rather than Participant in NIDM-Experiment see here.

We should also talk about the number of Collections to be created - on an earlier discussion on a NIDM call the feeling was to minimize the number of collections. (I do include several in the ontology though).

I would keep prov:Plan as not directly involving Subjects. There is the term "Group" in the ontology that is used for collections of prov:Person's who have prov:role nidm:Subject and are associated with nidm:DataAcquisition.

The prescreening of subjects would be a separate activity, with two Agents: the person who administered the prescreening and the Subject. It would look the same as the InformedConsent activity in the latest version of the lucidchart in the NIDASH folder.

[satra]

since collections can be built from the underlying graph, i would rather avoid collections when possible.

[smpadhy]

@khelm @satra Thank you for the clarifications. @khelm In the NIDASH/Specifications folder, which is the latest lucidchart version? I am unable to open NIDM-Experiment-Hierarchy.

[khelm]

I agree that Collections should be minimized, but I am not totally against them. For example, one could envision a collection of "all dicom tags for this AcquisitionObject" or an "AuxiliaryFile" collection that would include the b-val and b-vec files. Yes, you could build them each time from the underlying graph each time, but there is something attractive about it conceptually.

The lucid chart you want is NIDM-Experiment_Draft_8a in the NIDASH folder