Improving the efficacy of ERP-based BCIs using different modalities of covert visuospatial attention and a genetic...

Area of Science:

• Neuroscience
• Human-Computer Interaction
• Biomedical Engineering

Background:

• Brain-computer interfaces (BCIs) offer alternative control methods for individuals with motor impairments.
• Event-related potentials (ERPs) are crucial neural signals for BCI operation.
• Visuospatial attention, both voluntary and automatic, influences ERPs and can be leveraged for BCI control.

Purpose of the Study:

• To investigate the efficacy of covert orienting of visuospatial attention in ERP-guided BCIs.
• To compare the performance of voluntary versus automatic attention-based visual interfaces.
• To evaluate different epoch classification procedures for optimizing BCI performance.

Main Methods:

• Three visual interfaces were developed: one utilizing voluntary attention and two using automatic attention.
• Two epoch classification methods were employed: online Independent Component Analysis (ICA) and offline genetic algorithm (GA) based feature extraction.
• Support Vector Machines (SVM) and logistic classifiers were used for signal categorization.

Main Results:

• Offline classification successfully differentiated between voluntary and automatic attention interfaces, unlike online classification.
• Participants demonstrated superior performance with the voluntary attention interface, supported by neurophysiological data.
• The GA classifier outperformed the ICA classifier in epoch analysis.

Conclusions:

• Voluntary orienting of visuospatial attention, combined with personalized feature extraction via GA classifiers, can significantly enhance the control efficiency of visual BCIs.
• Offline analysis and GA-based classification offer advantages for improving BCI performance compared to online methods.
• This research provides neurophysiological evidence for the benefits of voluntary attention in BCI applications.