[ ] Trial-based-Segmentation should be renamed ophys, since it is a ProcessingModule to store optophysiology data.
[ ] [dFoF_commitIncorrAl, dFoF_firstSideTryAl, etc…] should be named [DfOverF_0, DfOverF_1]
[ ] Time dimension should go first.
[ ] Time series should be stored as one continuous series, with time gaps (if present) properly represented on timestamps. Trials should be defined within trials field, with add_trial() and add_trial_column() methods.
[ ] timestamps seems to be in milliseconds, they should be in seconds.
[ ] timestamps start with negative values, I believe a good practice would be to use only positive values.
[ ] The sequence of DfOverF containers are different experimental epochs? If so, that can be explicitly represented with the add_epoch() method.
processing/Image-Segmentation:
[ ] the content of this ProcessingModule is the same category as the above, ophys. Its content (ImageSegmentation) should be stored at ophys together with other optophysiology data.
[ ] img_seg should be named ImageSegmentation
[ ] pln_seg should be named PlaneSegmentation
[ ] the full [0, 1] matrices stored as image_mask could be stored more efficiently as pixel_mask.
[ ] The number of ROIs within PlaneSegmentation (653) do not match the number of ROIs at RoiResponseSeries objects (576). Also, the PlaneSegmentation ROIs are not properly referenced at the RoiResponseSeries from the DfOverF containers. To properly associate each ROI defined in a PlaneSegmentation with the data within RoiResponseSeries, use the method create_roi_response_series() from DfOverF containers.
intervals/trials:
[ ] trials should be sequential in time, belonging to the same timeline, e.g.: trial1:{start_time=10, stop_time=15} and trial2:{start_time=20, stop_time=25}.
processing/Behavior:
[ ] Behavior should be renamed behavior
[ ] Epoched_behavioral_series should be renamed BehavioralTimeSeries and it should contain continuous recordings (e.g. wheel speed/position).
[ ] Irregular, binary events (e.g. licking events) should be stored within BehavioralEvents instead of BehavioralTimeSeries.
[ ] [firstSideTryAl_lick, firstSideTryAl_wheelRev] should be renamed [TimeSeries_0 , TimeSeries_1]
[ ] Both TimeSeries should be one dimensional arrays. The corresponding trial times should be stored in trials.
Here's a list of suggested improvements:
Inspecting: mouse2_fni17_151005_001-002_ch2-PnevPanResults-170719-024122.nwb
processing/Trial-based-Segmentation:
add_trial()andadd_trial_column()methods.add_epoch()method.processing/Image-Segmentation:
create_roi_response_series()from DfOverF containers.intervals/trials:
processing/Behavior: