Neuromarketing and decision-making: Classification of consumer preferences based on changes analysis in the EEG signal of brain regions

Abstract

Neuromarketing involves the study of brain responses that focuses on understanding how consumers' brains respond to products and services, and how these responses influence their choice. Evidence suggests that electroencephalography (EEG) can provide valuable insight into consumer preferences and decision-making processes. This study aims is to assess the relative importance of right/left brain regions (including hemispheres, frontal, temporal, parietal, and occipital lobes) in the consumer choice towards E-commerce products. Also, this study aims to distinguishes the EEG characteristics of consumers’ preference using a classification system.Using a publicly available EEG neuromarketing dataset, the change in EEG signals has been evaluated by a mixed model for repeated measures for all brain regions. Four classification algorithms (k-NearestNeighbor, Random Forest, Neural Network, and Gradient Boosting) were used to distinguish like and dislike preferences.Greater EEG activity in the right hemisphere, right parietal, right occipital, and left occipital was related to like responses. Except for both sides of the temporal lobe, all the subdivisions of the brain considered showed a significant decrease of activity at 4000 ms for like-related responses. However, no significant change in the activity was related to the dislike response. The highest AUC of the four classifiers used was as follows: 76.61% for the right parietal lobe with Neural Network, 75.33% for the left parietal lobe with Gradient Boosting, 73.55% for the right frontal lobe with k-Nearest Neighbor and 72.62% for the right frontal lobe with Random Forest. Considering the significant difference between like and dislike responses at 4000 ms, Neural Network showed the best performance followed by Gradient Boosting.Our framework suggests that the formation of preferences (like and dislike) requires different patterns of brain activity and that Neural Network and Gradient Boosting are valuable tools for distinguishing consumer preference.