graph will assume it can apply a paired test if two groups have equal size. This can be wrong, and suprises the users. Its default behavior should change to not do paired testing unless requested by an option. Perhaps extra_option 'paired', which could be false (default) or true.

move all experiment specific things from XXprocessparams and tpstacktypes,tpstacklabels to experiment/dec database

when params.get_intensities is false, still intensities are given but this are not updated and should be set to NAN by analyse_tprecord

*Faster switch zstack without reloading preview (fewer preview reloads in general)

hard returns in comment fields give problems for parsing in other functions. they should be directly removed when the comment field is edited or saved

for analyse_ps or results_ectestrecord write a fit for orientation tuning (using a von_mises or something else)

in tptestdb measures should faithfully reflect ROIs.

When data has moved to Experiment folders, clean up tp-, oi- and ecprocessparams.m from dec-protocol specific settings and put these in m-files in the Experiment folders.

all intensity stuff in record.ROIs should go into measures in tptestdb