ecogtools

functions for ecog data analysis

See example notebooks in pearsonlab/ecog_data_analysis_notebooks for best examples of analysis workflow.

See comments in ecogtools.py file for specifics about different functions/class.

Use

Class Data has the majority of the functions for analysis, while subclasses (e.g. ToM_Localizer) have task-specific parameters.

1. Calling e.g. ToM_Localizer class
   • accepts "patient_num" : an integer value assigned to each patient (usually in #2000s)
   • searches for analysis.json files with parameters for:
     • Patient
     • Task
   • Parameter files contain directory information for patients, event information and locations of relevant files.
2. Does following automatically:
   • Loads physiology data (ecog- fif/edf file)
   • Loads behavioral data (task - json file)
   • Loads trigger data (task/ecog - csv file)
   • Combines behavioral data and triggers
   • Defines events for MNE
   • Checks directories (for saving image files)
   • Sets triggers (task/class-specific) to differentiate conditions
3. Additional functions within Data Class:
   • Remove_irrelevant_channels – helpful for looping through all channels and not having Event, EKG or Stim channels in averages.
     • accepts list of channels to remove or removes default irrelevant channels
   • Initialize_epochs_object – used for creating epochs objects (loads data automatically)
     • accepts channels of interest and any kwargs used by MNE Epochs
     • saves epochs object as data.epochs
   • create_evoked - used for creating evoked objects
     • accepts "condition" from event_id list (as string)
     • returns evoked object and saves object to data.evoked_list
   • compute_power - used to calculate power with tfr_morlet
     • accepts "condition" and all kwargs for tfr_morlet
     • out-dated : better to use tfr_multitaper
   • compute_diff_power - used to calculate ratio of two TF objects
     • accepts two TF objects
     • returns combined (one/two) TF object