validate_flash_blank_durations failing on cluster D

[dougollerenshaw]

The blank durations are longer than expected on the rig. We expect them to be 500 ms, but the mean is 516.84 ms. I built a one frame tolerance into the validation function that checks that, so it's falling just outside of the tolerance.

Here's a histogram of blank durations:

How should we deal with this?

[dougollerenshaw]

See http://w7dtmj0580ds:1111/notebooks/Doug/2018.06.08_Troubleshoot_failing_validation_functions_M363140_180607.ipynb

[dougollerenshaw]

@derricw: your camstim update seems to have over-compensated for this problem. Now the mode of the distribution has shifted to one frame below the expected. This is true for both flashes (expected value of 0.25) and blanks (expected value of 0.5):

[dougollerenshaw]

The plots above are from mouse M373115 on 6/19/18.

[dougollerenshaw]

Also, mean blank duration = 0.485 seconds. Mean flash duration = 0.237 seconds.

[nickponvert]

@dougollerenshaw is this still and issue we need to work on?