MOBI_QC is a python quality control pipeline designed for efficient, automated assessment of multimodal datasets (in xdf format), including EEG, Eye-tracking, Physiological signals, Audio and Video data. The pipeline performs preprocessing, computes a comprehensive quality control metrics for each modality and generates a PDF report summarizing the results for easy review and documentation.
- Multimodal Compatibility: Supports EEG, Eye-Tracking, ECG, EDA, Respiration, Audio, Video and Behavior data.
- Automated Preprocessing: Standardized cleaning and preprocessing for each modality.
- Quality Control Metrics Computation:
- EEG: sampling rate, mean peak-to-peak amplitude, standard deviation of amplitude, absolute voltage range, high and low frequency noise, flat channels, line noise ratio, dropout percentage, signal-to-noise ratio, alpha power, spectral entropy, correlations between channels, number of missing samples.
- Eye-tracking: validity of gaze point, gaze origin and pupil diameter, mean difference in percent valid data between right and left eyes.
- ECG: average heart rate, signal-to-noise ratio, signal quality indices - kurtosis, power spectrum distribution, relative power in baseline.
- EDA: signal integrity check, average skin conductance level (SCL), SCL standard deviation, SCL coefficient of variation, average amplitude of skin conductance response (SCR), SCR validity, signal-to-noise ratio.
- Respiration: signal-to-noise-ratio, breath amplitude - mean, standard deviation and range, respiration rate - mean standard deviation and range, peak-to-peak interval - mean, standard deviation and range, baseline drift, autocorrelation at typical breath cycle.
- Audio: percent of missing data, distribution of microphone samples - first quartile, third quartile, mean, standard deviation, minimum and maximum.
- Video: sampling rate, percentage of frames with faces detected.
- Generation of PDF Report: A PDF report is generated for each subject data summarizing quality control metric and visualizations for each modality for review and documentation.
- Export to csv: Results from entire dataset are summarized and exported in a csv file to facilitate feature identification and descriptive statistics for the dataset.
- pyxdf: a python library for importing Extensible Data Format(XDF)
- MNE-Python: a python pacakge used for preproecessing of EEG
- Neurokit2: a python toolbox used for processing of physiological signals
- librosa: a python package for music and audio analysis
- opencv-python: a open-source python library for computer vision and machine learning
Install this package via :
pip install MOBI_QCOr get the newest development version via:
pip install git+https://github.com/childmindresearch/MOBI_QCShort tutorial, maybe with a
import MOBI_QC
MOBI_QC.short_example()