Skip to content

MNE-preprocessing is a python repository to reduce artifacts based on basic and unanimous approaches step by step from electroencephalographic (EEG) raw data.

Notifications You must be signed in to change notification settings

AGhaderi/MNE-Preprocessing

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

21 Commits
 
 
 
 

Repository files navigation

MNE-Preprocessing

Scientists usually record EEG data as the continuous (Raw) signal which consists of irrelevant signal (brain signal) and irrelevant signal (artifact), therefore we need to attenuate the effects artifacts to make clear signals.

For preprocessing EEG data, the MNE-Preprocessing repocitory is implemented by the MNE-package in python.

The preprocessing is performed as follows:

256 points as the sampling rate, band-pass filter in the range 1 Hz - 30 Hz, re-referencing with an average of sensors, visual inspection to remove abnormal frequency, interpolation fully corrupted signal, discarding range of -100 mV to +100 mV amplitude, extracting Epochs from - 100 ms (baseline) of cue’s onset to 600 ms after stimulus appearance (or any period time you want), running Independent Component Analysis (ICA) to remove manually irrelevant-task (e.g. eye movement, head motion, and muscular activity), detecting automatically eye-blinking EOG component and heartbeat peak ECG component.

Citations

Ghaderi-Kangavari, A., Rad, J. A., Parand, K., & Nunez, M. D. (2022). Neuro-cognitive models of single-trial EEG measures describe latent effects of spatial attention during perceptual decision making, Journal of Mathematical Psychology, 111; doi: https://doi.org/10.1016/j.jmp.2022.102725

Prerequisites

About

MNE-preprocessing is a python repository to reduce artifacts based on basic and unanimous approaches step by step from electroencephalographic (EEG) raw data.

Topics

Resources

Stars

Watchers

Forks