Tracking feature-based attention

Area of Science:

• Cognitive Neuroscience
• Human-Computer Interaction
• Virtual Reality

Background:

• Feature-based attention (FBA) globally enhances specific visual features, allowing attention tracking using peripheral stimuli.
• Steady-state visually evoked potentials (SSVEPs) are commonly generated using foveal flicker in brain-computer interface (BCI) research.
• Tracking attention in dynamic virtual environments presents unique challenges for current BCI methods.

Purpose of the Study:

• To explore the use of peripherally presented flicker to generate SSVEPs for tracking central visual attention.
• To evaluate the robustness of using electroencephalography (EEG) to track FBA within a dynamic 3D virtual environment using a head-mounted display (HMD).

Main Methods:

• Two experiments were conducted: first, a visual search task in a stationary virtual environment to confirm peripheral flicker-induced SSVEPs during HMD use.
• Second, observers navigated a dynamic virtual racetrack while peripheral flicker tracked attention to a target color in a visual search task.
• EEG data was analyzed to assess the correlation between peripheral flicker stimuli and attended features.

Main Results:

• Peripheral flicker successfully generated SSVEPs that were influenced by attention in both stationary and dynamic virtual environments.
• Offline analysis demonstrated the ability to track an observer's attended color, indicating the potential for real-time FBA tracking.
• Attention tracking was robust across different viewpoints and movement conditions within the virtual environment.

Conclusions:

• Peripherally presented flicker can effectively track attention in dynamic virtual environments, offering a more flexible alternative to foveal stimulation.
• This approach provides a viable method for real-time FBA tracking, advancing the development of practical BCIs.
• The findings suggest a promising direction for future BCI research leveraging naturalistic visual search tasks in immersive environments.