New denoising modules and miscellaneous fixes from WASHU - #274

This is a re-RP of #262 with conflict resolution, as seen below:

This PR primarily contains new code for epi denoising. This includes aamod_scrub_epis to generate frame censoring (scrubbing) regressors, and a collection of related functions like aamod_DVARS to calculate DVARS and make it available as a stream, aamod_GLOBALMEAN for global mean, aamod_greyplot to generate a greyplot (ala Power), and aamod_lossplot to generate data loss summaries resulting from various thresholding.

The PR also contains a new module aamod_AROMA_denoise that makes FSL's untrained ICA available for denoising. Also, Jorn Diedrichsen's robust weighted least squares has been added as an option to aamod_firstlevel_model (toolbox class name: wls). Also, aamod_waveletdespike has been modified to work with the new toolbox class interface (class name wds).

There's one example script aa_example_denoising.m/xml to demonstrate how to use the new functionality. Testing has been broken into separate scripts, aatest_ds000114_scrubbing, aatest_ds000114_AROMA, and aatest_ds000114_wavelet.

Some of the new code requires a third-party installs: AROMA requires FSL/AROMA, wavelet despiking requires BrainWavelet toolbox, and rWLS requires the rWLS toolbox.

A few miscellaneous fixes also included:

• Some of the test scripts got broken in the last go-round when subject and session parameters being passed to aas_processBIDS were replaced with aap.acq_details.selected_subjects and aap.acq_details.selected_visits. aas_processBIDS ignores these settings, so I change the code back to the way I originally wrote it.

• I modified aa_test so you can pass in a parameter file. I find that useful, but it's totally optional.

• I edited the comments in aatest_ds000114_TEMPLATE.m to (hopefully) be clearer.

• I fixed the aamod_biascorrect_meanepi typo mentioned in discussion Experiences with Getting Started #254

• I added a bit of QA to aamod_firstlevel_model: It now saves the design matrix (previously this required you to run reporting) and it also now saves a plaintext file of regressors and event onsets.

• added aamod_rescale_epi to rescale an epi

• modified aamod_realign to calculate and output FD.

• added another pause(1) to aas_getfiles_bystream.m trying to fix random failures under OS-X PCT that felt like a race condition issue

• a few other minor things that I'm probably forgetting