mne.preprocessing.eyetracking.find_blinks

mne.preprocessing.eyetracking.find_blinks(inst, *, chs_src=None, method='dropout', dropout_value=None, description='BAD_blink', chs_dest=None, verbose=None)

Find blinks in eye-tracking data and create corresponding annotations.

Parameters:

instinstance of Raw

The data instance to use for finding blinks. Must contain pupil channels.

chs_srclist | None

A list of channel names that will be used to find blinks. None (default) will result in selecting all channels of type pupil. See Notes for more information.

methodstr

Which method to use to find blinks in inst. Currently the only supported method is 'dropout' (default), which uses the value specified via dropout_value to identify blinks.

dropout_valuefloat | None

Which value in the data denotes a dropout. In Eyelink data, this is 0, whereas for other eye-tracking data this may be np.nan, or something else. Defaults to None, which sets the dropout_value to np.nan.

descriptionstr

Which description to use for the blink annotations. Defaults to 'BAD_blink'.

chs_destlist | str | None

Which channels to associate the blink annotations with. None (default) associates each blink annotation with the source channels (chs_src) it was detected from. Passing 'all' associates the annotations with all channels in inst. Alternatively, a list of channel names can be passed (e.g., EEG channels, for subsequent ocular artifact cleaning).

verbosebool | str | int | None

Control verbosity of the logging output. If None, use the default verbosity level. See the logging documentation and mne.verbose() for details. Should only be passed as a keyword argument.

Returns:

annotinstance of Annotations

The annotations object containing blink annotations. It is not attached to inst; use raw.set_annotations to do so.

Notes

If multiple channels are used to find blinks in inst, resulting overlapping blink annotations are always merged over all channels. That is, if a left and a right pupil channel would be used for blink detection, and each on their own would produce overlapping blink annotations with onsets [1.5, 1.6] and durations [0.2, 0.3], respectively, then passing both channels into this function will result in a single blink annotation with onset 1.5 and duration 0.4. Note that this corresponds to the minimum onset and the maximum offset between the two annotations.

Examples using mne.preprocessing.eyetracking.find_blinks

Working with eye tracker data in MNE-Python

Working with eye tracker data in MNE-Python