ENH: Enable models for sparsely sampled fMRI series

[effigies]

Initial commit is all but the Convolve portions of #376 squashed onto master @ 4315865375a6eb811ba9f7eadbc1d621ecb2de96. There may be some parts that are no longer appropriate.

Supersedes and closes #376 (leaving open for now for reference). Closes #252.

[codecov[bot]]

Codecov Report

Merging #414 into master will decrease coverage by 0.29%. The diff coverage is 44.23%.

@@            Coverage Diff            @@
##           master     #414     +/-   ##
=========================================
- Coverage   62.37%   62.08%   -0.3%     
=========================================
  Files          27       27             
  Lines        4564     4602     +38     
  Branches     1174     1185     +11     
=========================================
+ Hits         2847     2857     +10     
- Misses       1433     1455     +22     
- Partials      284      290      +6

Flag	Coverage Δ
#unittests	62.08% <44.23%> (-0.3%)	:arrow_down:

Impacted Files	Coverage Δ
bids/variables/kollekshuns.py	83.57% <100%> (ø)	:arrow_up:
bids/variables/entities.py	87.77% <100%> (+0.13%)	:arrow_up:
bids/variables/variables.py	83.54% <36.36%> (-4.85%)	:arrow_down:
bids/variables/io.py	72.24% <37.5%> (-3.01%)	:arrow_down:
bids/analysis/analysis.py	86.91% <50%> (-1.87%)	:arrow_down:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 4315865...0729028. Read the comment docs.

[effigies]

Ah, saw you said we should set up a call. Looking at this again, I agree. Hacking sparse sampling into or next to the current approach is probably not worth it, and we should work out a strategy for resampling more broadly, with discontinuous integration windows in mind. Might also be worth thinking through the clustered acquisition approach at the same time, to avoid settling on a solution that will need to be refactored again when we get around to that.

@satra Do you think you or anyone in your group should be on the call?

[effigies]

@satra Bump.

[tyarkoni]

Any status update on this, post resampling changes? (Fine if not, this is a low priority from my perspective.)

[satra]

@tyarkoni - i haven't had a look at this in a while. can you point me to where the design matrix is being built and i can backtrack from there?