Shredding artifacts: extracting brain activity in EEG from extreme artifacts during skateboarding using ASR and ICA
Daniel E Callan1,2, Juan Jesus Torre-Tresols1,2, Jamie Laguerta1,3
1Brain Information Communication Research Laboratory, Advanced Telecommunications Research Institute International, Kyoto, Japan.
Frontiers in Neuroergonomics
|July 11, 2024
Summary
The ASRICA method effectively cleans electroencephalography (EEG) artifacts during skateboarding, enabling accurate brain activity analysis in real-world conditions. This technique allows for studying neural processes during complex physical activities.
Area of Science:
- Neuroscience
- Signal Processing
- Biomedical Engineering
Background:
- Understanding brain function in natural environments requires robust electroencephalography (EEG) data acquisition, even amidst significant noise and artifacts.
- Traditional EEG analysis faces challenges due to environmental and physiological artifacts, necessitating advanced signal processing techniques for artifact removal.
Purpose of the Study:
- To demonstrate the efficacy of Artifact Subspace Reconstruction (ASR) and Independent Component Analysis (ICA) in extracting brain activity from EEG data collected during skateboarding.
- To evaluate different artifact cleaning pipelines, including ASR and ICA combinations, for their ability to isolate neural signals during physically demanding tasks.
Main Methods:
- A dual-task paradigm was employed, presenting auditory stimuli to participants during both skateboarding and rest conditions.
- Five artifact cleaning pipelines were assessed: minimal cleaning, ASR only, ICA only, ICA followed by ASR (ICAASR), and ASR preceding ICA (ASRICA).
- Support Vector Machine (SVM) was used to classify the presence or absence of auditory stimuli in single-trial EEG data to evaluate pipeline effectiveness.
Main Results:
- The ASRICA pipeline significantly outperformed minimal cleaning and other ICA-containing pipelines in single-trial classification accuracy during skateboarding.
- ASRICA demonstrated superior performance in identifying brain components compared to ICA alone or ICAASR, suggesting improved artifact removal.
- While ASRICA showed slight improvements during rest, its primary advantage was evident in the high-artifact skateboarding condition.
Conclusions:
- The ASRICA pipeline is highly effective for cleaning EEG artifacts and extracting single-trial brain activity during physically demanding activities like skateboarding.
- This study validates the feasibility of conducting EEG research in real-world, high-movement scenarios, paving the way for studying neural correlates of complex motor behaviors.
