Definition of "expected Number of xxx": reference to Random Field Theory (pending final vetting)

[cmaumet]

This issue is open to keep track of a discussion started at #309.

The question is whether we should add a mention of "Random Field Thewory" in the following definitions (given that those terms will also be useful in the non-parametric setting, see https://github.com/incf-nidash/nidm/issues/309#issuecomment-93897763):

• nidm:'expected Number Of Clusters': Expected number of clusters in an excursion set under the null hypothesis.
• nidm:'expected Number Of Voxels Per Cluster': Expected number of voxels in a cluster under the null hypothesis.
• nidm:'expected Number Of Vertices Per Cluster': Expected number of vertices in a cluster under the null hypothesis.

[tiborauer]

How about cluster mass thresholding?

[nicholst]

Right! We need that too! It's in FSL's randomise (and, down the line, SnPM).

[khelm]

So: nidm:'expected Number Of Clusters': Expected number of clusters in an excursion set under the null hypothesis using Random Field Theory? Seems kind of extraneous to me, but I'll defer to others if they have a compelling reason for inclusion.