A Sparse Representation Classification Scheme for the Recognition of Affective and Cognitive Brain Processes in...

Area of Science:

• Neuroscience
• Cognitive Science
• Marketing Science

Background:

• Neuromarketing research often relies on understanding brain responses to stimuli.
• Electroencephalography (EEG) is a key tool for measuring brain activity.
• Accurate classification of cognitive and affective states from EEG is crucial for neuromarketing insights.

Purpose of the Study:

• To propose a novel framework for recognizing cognitive and affective brain processes during neuromarketing stimuli using EEG signals.
• To develop and evaluate a new classification algorithm based on sparse representation for EEG data.
• To enhance the accuracy of brain state recognition in neuromarketing applications.

Main Methods:

• A novel classification algorithm based on sparse representation classification is proposed.
• The algorithm assumes EEG features lie on a linear subspace, representing test signals as linear combinations of training data.
• The Sparse Bayesian Framework with graph-based priors is employed for determining class membership, using residuals for the classification rule.

Main Results:

• Experiments were conducted on a publicly available neuromarketing EEG dataset.
• The proposed classification scheme achieved higher accuracy than baseline and state-of-the-art methods.
• An improvement of over 8% in classification accuracy was observed for both affective and cognitive state recognition tasks.

Conclusions:

• The proposed framework and sparse representation classification algorithm are effective for recognizing cognitive and affective states from EEG signals in neuromarketing.
• The method demonstrates significant improvements in classification accuracy, offering a valuable tool for neuromarketing research.
• This approach advances the analysis of brain responses to marketing stimuli, providing more reliable insights.