Evaluating the Feasibility of Visual Imagery for an EEG-Based Brain-Computer Interface

Area of Science:

• Neuroscience
• Human-Computer Interaction
• Cognitive Science

Background:

• Visual imagery, the mental simulation of visual information, is a potential control strategy for brain-computer interfaces (BCIs).
• Previous BCI studies using visual imagery often employed paradigms that mimic visual working memory rather than spontaneous mental imagery.
• Evaluating true spontaneous visual mental imagery is crucial for developing effective BCI control strategies.

Purpose of the Study:

• To compare the neural signatures of short-term visual imagery versus spontaneous long-term visual imagery for BCI control.
• To investigate the differences and commonalities between visual imagery and visual perception in terms of neural correlates.
• To assess the feasibility of using visual imagery as a BCI control strategy.

Main Methods:

• Electroencephalography (EEG) was used to record brain activity.
• Participants performed short-term visual imagery tasks (following a target image) and spontaneous long-term visual imagery tasks (cued by auditory stimuli).
• EEG data were analyzed to identify distinct neural signatures and predictive features for each imagery type and compared with visual perception data.

Main Results:

• Short-term visual imagery produced a stronger and more classifiable neural signature in EEG compared to spontaneous long-term visual imagery.
• Both short-term visual imagery and visual perception shared common predictive electrodes and spectral features.
• Visual imagery showed greater influence from frontal electrodes, suggesting contributions from memory and attention, while visual perception was primarily associated with occipital electrodes.

Conclusions:

• Short-term visual imagery is more suitable for BCI control than spontaneous long-term visual imagery due to its clearer neural signals.
• Distinct neural patterns for visual imagery (frontal influence) and perception (occipital influence) highlight differences in cognitive processing.
• This research provides valuable insights into the potential and challenges of employing visual imagery in brain-computer interface applications.