^RNeuMark: A Riemannian EEG Analysis Framework for Neuromarketing

Area of Science:

• Neuroscience
• Marketing Science
• Machine Learning

Background:

• Neuromarketing uses neuroimaging to improve traditional marketing tools like surveys.
• Electroencephalography (EEG) is a popular, non-invasive, low-cost neuroimaging technique increasingly used in wearable devices.
• Current methods for translating brainwave patterns into consumer attitudes have limitations, driving research for novel descriptors.

Purpose of the Study:

• To propose sample covariance matrices (SCMs) and Riemannian geometry as novel descriptors for EEG data in neuromarketing.
• To develop and evaluate a decoding scheme for predicting consumer choices using these new methods.
• To demonstrate the superiority of the proposed approach over existing techniques.

Main Methods:

• Utilized sample covariance matrices (SCMs) to capture coordinated neural activity across brain areas.
• Applied Riemannian geometry for processing and analyzing SCMs.
• Developed an ensemble classifier with a multi-view perspective, dedicating each view to a specific frequency band for decoding consumer choices.
• Trained classifiers using standard machine learning procedures on neuromarketing datasets.

Main Results:

• Established the suitability of the Riemannian approach for neuromarketing applications.
• Demonstrated the potential of SCMs as effective descriptors for EEG-based neuromarketing.
• Showcased the proposed decoder's superior performance compared to popular alternatives on two distinct datasets.

Conclusions:

• Sample covariance matrices and Riemannian geometry offer a promising new framework for neuromarketing analysis.
• The proposed ensemble decoder effectively predicts consumer choices by integrating multi-band EEG information.
• This research advances the field of neuromarketing by providing more accurate and robust methods for understanding consumer behavior.