EEG-ExPy: Democratizing the Cognitive Neuroscience Experiment
Poster No:
1658
Submission Type:
Abstract Submission
Authors:
John Griffiths1,2, Taha Morshedzadeh3, Sorenza Bastiaens3, Parsa Oveisi4, Ore Ogundipe5, Erik Bjäreholt6, Daniele Marinazzo7, Yannick Roy8, EEG-ExPy Team .9
Institutions:
1Departments of Psychiatry, Medical Sciences, & Biomedical Engineering, University of Toronto, Canada, 2Krembil Centre for Neuroinformatics, Centre for Addicton and Mental Health, Canada, 3Institute of Medical Sciences, University of Toronto, Canada, 4Institute of Biomedical Engineering, University of Toronto, Canada, 5Fusion Research Inc., Canada, 6Lund University, Sweden, 7University of Ghent, Belgium, 8University of Montreal, Canada, 9Github, The World
First Author:
John Griffiths, PhD
Departments of Psychiatry, Medical Sciences, & Biomedical Engineering, University of Toronto|Krembil Centre for Neuroinformatics, Centre for Addicton and Mental Health
Canada|Canada
Departments of Psychiatry, Medical Sciences, & Biomedical Engineering, University of Toronto|Krembil Centre for Neuroinformatics, Centre for Addicton and Mental Health
Canada|Canada
Co-Author(s):
Introduction:
Cognitive neuroscience experiments using EEG and other neuroimaging techniques have traditionally been restricted to lab settings, with dedicated (typically university or hospital) spaces, expensive hardware, manned by professional technical and academic staff. Fortunately however, recent years have seen alternatives to these conventions begin to emerge. It is now becoming possible to run a wide range of classic experimental paradigms in a highly affordable fashion with minimal specialist equipment and expertise. This shift represents to many a democratizaton of the cognitive neuroscience experiment.
It is in this spirit that we introduce here EEG-ExPy - A Python-based platform for cognitive neuroscience experimentation and education.
Formerly known as eeg-notebooks (and before that muse-lsl), the EEG-ExPy project (github.com/NeuroTechX/EEG-ExPy) is an international open science initiative, driven by members of the NeuroTechX community. EEG-ExPy allows sophisticated visual, auditory, and other experimental paradigms to be run using a standard personal computer and a minimal, affordable, consumer-grade mobile EEG device.It is also fully compatible and usable with research-grade EEG systems in a traditional lab setting.
EEG-ExPy's ease of use makes it attractive for a wide audience, including research scientists, clinicians, educators, and hobbyists. Use cases to date span a wide range of settings, from high school outreach programs, hackathons, hands-on university-level cognitive neuroscience teaching, brain stimulation clinical trials in psychiatry, and bedside recordings in neurology patients.
Here we describe the motivation, design, and usage of EEG-ExPy, demonstrating example data from several featured experiments.
It is in this spirit that we introduce here EEG-ExPy - A Python-based platform for cognitive neuroscience experimentation and education.
Formerly known as eeg-notebooks (and before that muse-lsl), the EEG-ExPy project (github.com/NeuroTechX/EEG-ExPy) is an international open science initiative, driven by members of the NeuroTechX community. EEG-ExPy allows sophisticated visual, auditory, and other experimental paradigms to be run using a standard personal computer and a minimal, affordable, consumer-grade mobile EEG device.It is also fully compatible and usable with research-grade EEG systems in a traditional lab setting.
EEG-ExPy's ease of use makes it attractive for a wide audience, including research scientists, clinicians, educators, and hobbyists. Use cases to date span a wide range of settings, from high school outreach programs, hackathons, hands-on university-level cognitive neuroscience teaching, brain stimulation clinical trials in psychiatry, and bedside recordings in neurology patients.
Here we describe the motivation, design, and usage of EEG-ExPy, demonstrating example data from several featured experiments.
Methods:
EEG-ExPy contains functionality for the three main pieces of an eeg experiment: data streaming, stimulus presentation, and analysis. It is an installable Python package that can be run from an OS command line, a Python command line, or via Python scripts and jupyter notebooks, on a Windows, Mac, or Linux laptop/desktop.
The primary use case is with wireless mobile EEG devices such as the InteraXon Muse, Neurosity Crown, OpenBCI Cyton, and G.tec Unicorn. These are handled by a Device Class, which initiates a bluetooth data stream using third-party libraries such as BrainFlow. If data streaming is to be handled by external software, triggers can be delivered via internal or external hardware ports.
After initiating a streaming connection with the EEG device, experiments are launched. Currently all featured experiments use PsychoPy for visual and auditory stimulus presentation, key press recognition, and instruction presentation; although this is not required. Featured experiments that have been tested extensively include N170 RSVP w/ faces, P300 RSVP w/ animals, auditory oddball, SSVEP, SSAEP, visual cueing, rest.
Analyses draw primarily on the MNE-Python library, focusing on computing and visualizing ERPs, frequency-domain responses, and ML-based trial classification. An automated analysis report tool generates HTML files for fast inspection of data.
The primary use case is with wireless mobile EEG devices such as the InteraXon Muse, Neurosity Crown, OpenBCI Cyton, and G.tec Unicorn. These are handled by a Device Class, which initiates a bluetooth data stream using third-party libraries such as BrainFlow. If data streaming is to be handled by external software, triggers can be delivered via internal or external hardware ports.
After initiating a streaming connection with the EEG device, experiments are launched. Currently all featured experiments use PsychoPy for visual and auditory stimulus presentation, key press recognition, and instruction presentation; although this is not required. Featured experiments that have been tested extensively include N170 RSVP w/ faces, P300 RSVP w/ animals, auditory oddball, SSVEP, SSAEP, visual cueing, rest.
Analyses draw primarily on the MNE-Python library, focusing on computing and visualizing ERPs, frequency-domain responses, and ML-based trial classification. An automated analysis report tool generates HTML files for fast inspection of data.
Results:
Usage examples and key software elements are shown in Figure 1. Figure 2 shows results from the faces vs. houses and SSVEP experiments, showing the face-selective N170 ERP component and flicker-frequency peaks in the power spectrum. These two examples demonstrate clearly the ability to reproduce canonical experimental findings from the EEG cognitive neuroscience literature with mobile EEG systems.
Conclusions:
In conclusion, EEG-ExPy is a powerful tool for the amateur and expert-trained cognitive neuroscientist alike, lowering the barrier-to-entry for conducting meaningful research with established paradigms, reducing the cost and time for setup, iteration, and implementation of new ideas and initiatives, from education through research to clinical monitoring.
Modeling and Analysis Methods:
EEG/MEG Modeling and Analysis 1
Neuroinformatics and Data Sharing:
Informatics Other
Novel Imaging Acquisition Methods:
EEG 2
Perception, Attention and Motor Behavior:
Perception: Visual
Keywords:
Data analysis
Design and Analysis
Electroencephaolography (EEG)
ELECTROPHYSIOLOGY
1|2Indicates the priority used for review
