A hybrid neuromarketing approach exploiting EEG graph signal processing and gaze dynamic patterning

Area of Science:

• Neuroscience
• Cognitive Science
• Data Science

Background:

• Consumer intent classification is crucial for marketing.
• Existing methods often struggle with imbalanced datasets and individual variability.
• Integrating neurophysiological and eye-tracking data offers a richer understanding of decision-making.

Purpose of the Study:

• To develop and evaluate a hybrid decoding scheme for binary consumer intent classification (Buy vs. NoBuy).
• To leverage simultaneous electroencephalography (EEG) and eye-tracking data for enhanced classification accuracy.
• To assess the scheme's performance at the individual subject level using appropriate metrics for imbalanced data.

Main Methods:

• A hybrid decoding framework combining graph signal processing features from EEG functional connectivity and eye movement pattern statistics.
• Utilizing Cohen's kappa and F1-score metrics for performance evaluation on imbalanced datasets.
• Subject-level analysis to account for individual differences in neural and behavioral patterns.

Main Results:

• The proposed hybrid scheme demonstrated statistically significant superiority over competing approaches.
• Averaged Cohen's kappa and F1-scores exceeded benchmarks by 0.08-0.30 and 0.06-0.23, respectively.
• Connectivity analysis revealed consistent couplings between distinct brain regions and significant individual variability in functional connections.

Conclusions:

• Hybrid decoding approaches integrating multimodal data show significant potential for advancing consumer decision behavior classification.
• Subject-specific connectivity patterns are key to understanding and predicting individual consumer intent.
• The developed framework offers a robust method for analyzing complex consumer choices.