Decoding covert shifts of attention induced by ambiguous visuospatial cues

Area of Science:

• Neuroscience
• Cognitive Science
• Human-Computer Interaction

Background:

• Covert shifts of visual attention enhance stimulus processing at attended locations.
• Alpha-band oscillations over parietal and occipital regions are modulated by attention shifts.
• Decoding these modulations from electroencephalography (EEG) signals on a single trial basis is possible.

Purpose of the Study:

• To investigate whether covert attention shifts occur in complex environments with ambiguous cues.
• To explore the feasibility of decoding covert attention shifts using EEG in such conditions.
• To assess the robustness of decoding methods across different levels of cue predictability.

Main Methods:

• Utilized random-dot stimuli to cue target locations, contrasting predictive and ambiguous conditions.
• Employed the Common Spatial Patterns (CSPs) algorithm to extract alpha-band EEG power features.
• Applied cross-validation for decoding accuracy assessment on single trials.

Main Results:

• Behavioral data confirmed attention shifts in anticipation of targets under both predictive and ambiguous conditions.
• Significant decoding accuracy of attended locations was achieved using CSPs in the predictive condition (7/10 subjects).
• Remarkably, similar decoding accuracy was obtained when CSPs trained on predictive data were applied to ambiguous data (5/10 subjects).

Conclusions:

• Covert attention shifts occur and can be decoded even with ambiguous spatial information in complex visual environments.
• Alpha-power features derived from EEG show potential for brain-computer interfaces (BCIs) to interpret attention in real-world scenarios.
• The findings suggest that attention decoding strategies may generalize across varying levels of environmental predictability.