Single-trial detection of EEG error-related potentials in serial visual presentation paradigm

Area of Science:

• Neuroscience
• Cognitive Science
• Human-Computer Interaction

Background:

• Error-Related Potentials (ErrPs) are brain responses to unexpected outcomes, crucial for performance monitoring.
• Existing methods struggle to record ErrPs when there's a mismatch between an object and its description.
• A novel experimental paradigm is needed to capture ErrPs in object-label dissociation scenarios.

Purpose of the Study:

• To propose a Serial Visual Presentation (SVP) paradigm for recording ErrPs during image-label dissociation.
• To introduce a novel single-trial method for detecting ErrPs using electrode-averaged features.
• To evaluate the efficacy of the proposed method in distinguishing correct and incorrect image labels.

Main Methods:

• Designed an SVP paradigm presenting labeled images (bike, car, flower, fruit, cat, dog) serially.
• Utilized text or audio clips as labels for the presented images.
• Employed novel electrode-averaged features for single-trial ErrP detection.

Main Results:

• ErrP data from 11 subjects showed characteristics consistent with existing literature.
• The novel feature extraction method achieved classification accuracies of 69.09±4.70% for audio labels and 63.33±4.56% for visual labels.
• The proposed method outperformed two existing feature extraction techniques in classification accuracy.

Conclusions:

• The proposed SVP paradigm and single-trial ErrP detection method are effective for identifying image-label mismatches.
• This research paves the way for Brain-Computer Interface (BCI) applications in cognitive assessment and image annotation.
• The findings hold potential for quantitative evaluation and treatment of mild cognitive impairment.