tfrStats.mvtfr_classifier module

tfrStats.mvtfr_classifier.mvtfr_classifier(jobs, cond, fband, split)[source]

Linear Discmininant Analysis classification

Function to decode stimulus conditions in electrophysiological data using a Linear Discriminant Analysis classifier from sklearn. At the moment it relies on the python package ACME, an efficient SLURM manager to use in high performance clusters (HPC), to speed up computations. Given the inputs, selects the data: conditions x repetitions x channels x frequency x time points and computes, for each frequency and channel, the classification accuracy matrix: conditions x conditions x time points. Because SLURM and ACME only accepts integers as inputs, the data selection occurs inside the function. Paths, conditions and frequency bands.

Todo

  • Add the ACME free version (using batch SLURM job arrays)

Parameters:

  • jobs (int) – a 1d numpy array specifying the number of jobs

  • cond (int) – condition index(gratings, natural images)

  • fband (int) – frequency band index (low, high and higher frequencies)

:param int split : index to split (even, odd or whole)

Returns:

classification accuracy.

Return type:

numpy array

@author: Nicolas Gravel, 19.09.2023