Related Experiment Video
Updated: May 15, 2025
08:45
Mapping Cortical Dynamics Using Simultaneous MEG/EEG and Anatomically-constrained Minimum-norm Estimates: an Auditory Attention Example
Published on: October 24, 2012
14.6K
Comparing MEG and EEG measurement set-ups for a brain-computer interface based on selective auditory attention.
Dovilė Kurmanavičiūtė1, Hanna Kataja1, Lauri Parkkonen1,2
1Department of Neuroscience and Biomedical Engineering, Aalto University, Finland.
Plos One
|April 10, 2025
Summary
Tracking auditory attention using electroencephalography (EEG) is feasible with fewer channels. Accuracy decreases with fewer EEG channels and when training data is limited to the recording
Area of Science:
- Neuroscience
- Biomedical Engineering
- Signal Processing
Background:
- Auditory attention significantly impacts brain responses to sounds.
- Whole-scalp magnetoencephalography (MEG) with classification algorithms accurately tracks auditory attention.
- The impact of reduced channel count and limited training data on EEG-based attention tracking is not fully understood.
Purpose of the Study:
- To investigate the decline in classification accuracy when transitioning from whole-scalp MEG to lower-channel EEG for auditory attention tracking.
- To assess the effect of training data selection (entire recording vs. initial part) on classifier performance.
- To determine the feasibility of implementing auditory-attention-based brain-computer interfaces using limited EEG channels.
Main Methods:
- Simultaneous MEG (306 channels) and EEG (64 channels) data were recorded from 18 healthy volunteers.
- Participants listened to concurrent spoken "Yes"/"No" words and were instructed to attend to one stream.
- Support vector machine classifiers were trained on MEG and varying numbers of EEG channels (64, 30, 9, 3) using different trial extraction strategies.
Main Results:
- MEG achieved the highest accuracy (73.2%) when training trials were randomly extracted throughout the recording.
- EEG classification accuracy decreased with fewer channels: 69% (64), 69% (30), 66% (9), and 61% (3).
- Training classifiers solely on initial recording data reduced accuracy by an average of 11%-units, with 3-channel EEG falling below chance level.
Conclusions:
- While whole-scalp MEG offers higher accuracy, EEG-based auditory attention tracking is viable with a reduced channel count.
- Optimally placed EEG channels can support the development of practical auditory-attention-based brain-computer interfaces.
- The temporal selection of training data significantly impacts classifier performance, highlighting the importance of diverse trial selection.
