Source code for allensdk.brain_observatory.behavior.session_metrics

import numpy as np
from allensdk.brain_observatory.behavior import trial_masks as masks

def response_bias(trials, detect_col, trial_types=("go", "catch")):
    """
    Calculate the response bias for a subset of trial types from a behavioral
    training dataframe.
    Args:
        trials (pandas.DataFrame): Dataframe containing trial-level information
            from a behavioral training session. Required columns:
            "trial_type", `detect_col`.
        detect_col (str): Name of column containing boolean 
            or numeric codings (0/1) for whether or not the mouse had a 
            response.
        trial_types (iterable<str>): Iterable containing string trial types
            to check for the response bias. Trials of types not included in this
            iterable will be ignored. Default=("go", "catch")
    Return:
        The response bias (or average value of the `detect_col`)
        for trials in `trial_types`.
    """
    mask = masks.trial_types(trials, trial_types)
    return trials[mask][detect_col].mean()


def num_contingent_trials(session_trials):
    """
    Returns the number of "go" and "catch" trials in a training session
    dataframe.
    Args:
        session_trials (pandas.DataFrame): a pandas.DataFrame describing 
        behavior training trials, with the string column "trial_type"
        describing the type of trial.
    Returns (int): Number of "go" and "catch" trials
    """
    return session_trials["trial_type"].isin(["go", "catch"]).sum()