NeuralEnsemble/cobrawap

data_entry step fails with most IDIBAPS files

Closed this issue · 2 comments

apdavison commented

I've been trying to run the pipeline with the files from https://kg.ebrains.eu/search/instances/Dataset/2ead029b-bba5-4611-b957-bb6feb631396

For most of the files, it fails with the error:
AssertionError: Signal channel in groups do not have same size, use try_signal_grouping=False

Such files have channels that differ slightly in length, e.g.

ipdb> sig_sizes
array([638735, 638735, 638735, 638735, 638735, 638734, 638734, 638734,
       638734, 638734, 638734, 638734, 638734, 638735, 638735, 638735])

I think it would be fairly easy to adapt the load_neo() function so that it loads the data using try_signal_grouping=False, trims the signals to a common length, and then 'manually' groups them.

rgutzen commented

You are right, the script stage01_data_entry/scripts/curate_IDIBAPS_data.py tries to do exactly that. There was, however, still a small bug which should now be resolved.
Tested with WT subjects 32 channel recordings.

rgutzen commented

double checked with:

  • Sanchez-Vives, M. (2020) "Propagation modes of slow waves in mouse cortex" EBRAINS
  • Sanchez-Vives, M. (2019) "Cortical activity features in transgenic mouse models of cognitive deficits (Fragile X Syndrome)" EBRAINS
  • Sanchez-Vives, M. (2019) "Cortical activity features in transgenic mouse models of cognitive deficits (Williams Beuren Syndrome" EBRAINS