Typical M/EEG experiments involve presentation of stimuli and responses based on some form of task demands. The occurrence of each stimulus or response can can be used to define an epoch which captures the brain signals preceding the stimulus or response as well as the response following them. Depending on the experimental paradigm and the analysis employed, an epoch is typically 500 ms to 2 s long. Epochs of different experimental conditions obtained from one subject are stored in MNE-Python in an instance of the Epochs class. An Epochs instance is created by specifying one or more instances of Raw to operate on, the event/stimulus type(s) of interest, and the time window to include. The Epochs object has various parameters for preprocessing single trial data, such as baseline correction, signal detrending, and temporal decimation. Epochs can be averaged to form evoked data containing the MEG and EEG signals known respectively as event related fields (ERFs) and event related potentials (ERPs). The averaged data are stored in instances of the Evoked class, and can be created simply by calling the average
