A Spiking Neural Network Approach to Electroencephalography based Consumer Preference Modeling

Neuromarketing is an emerging interdisciplinary field that applies neuropsychology in marketing to study consumer sensory-motor actions such as cognitive and affective responses to marketing stimuli through Brain Computer Interface (BCI) technology. While marketers spend over 750 billion dollars annually on traditional marketing procedures such as surveys, interviews, and consumer’s feedback, these methods are often criticized for their inability to capture genuine consumer preferences. Neuromarketing promises to overcome such issues by analyzing neural responses directly. This paper presents a novel framework for predicting consumer preferences by analyzing Electroencephalography (EEG) signals. EEG signals are acquired from 25 volunteers while administering 14 products with three different variations. The EEG signals are preprocessed using Modified Wavelet Thresholding (MWT) to remove noise while preserving neural activity patterns. A third-generation network, Spiking Neural Network (SNN) is designed to recognize consumer preferences based on EEG frequency bands. Unlike conventional models, SNN captures temporal dynamics through spike timing, which is crucial for EEG signals. The efficacy of the model is tested across individual EEG bands to identify the most influential frequency band in decision-making. Simulation outcomes demonstrate that the proposed model can effectively predict consumer preferences. The model achieved an accuracy of 90.91%, recall of 90.7%, a precision of 91.14%, a specificity of 91.12%, and an F1-score of 90.92%. The outcomes highlight the potential of EEG based neuromarketing systems to decode subconscious consumer responses, enabling brands and businesses to design more targeted marketing strategies based on objective neural data.