Health profiling using event-related potential (ERP) brain signals and spiking neural network (SNN)

• المؤلفون الرئيسيون: Saffiera, Cut Amalia, Hassan, Raini, Ismail, Amelia Ritahani
• التنسيق: مقال
• اللغة: English
• منشور في: 2020
• الموضوعات: QA75 Electronic computers. Computer science
• الوصول للمادة أونلاين: http://irep.iium.edu.my/86317/1/152-Article%20Text-992-1-10-20201214.pdf; http://irep.iium.edu.my/86317/; https://journals.iium.edu.my/kict/index.php/IJPCC; https://doi.org/10.31436/ijpcc.v6i2.152

Unhealthy lifestyles, especially on nutritional factors have become a major problem causing many diseases in Malaysians in recent years. Identification of lifestyle profiles such as preventive for individuals who adopt healthy and curative for individuals who do not maintain their lifestyle is needed to increase their awareness regarding their lifestyle. Because self-assessment is known to be vulnerable to produce response biases that lead to misclassification, identification of profiles based on brain responses needs to be done. An Event-related potential (ERP) is the main tools of cognitive neurologists and make ideal techniques for studying perception and attention. This research captured brain activity using electroencephalography (EEG) during receiving images of healthy and unhealthy foods that act as health-related stimuli. These EEG signals converted mathematically into the ERP signals and entered into the classification interface as input. In terms of classification, the methodology used is a dynamic developing Spiking Neural Network (deSSN) based on the Neucube architecture. ERP analysis results show the mean amplitude of the LPP component in the Parietal and Occipital lobes is higher for healthy food in the preventive group. Whereas in the curative group it has been shown to be higher for unhealthy foods. This result is thought to reflect their preference in choosing food in their daily lifestyle. However, the results of the classification have shown that unhealthy food stimulation in the LPP wave showed superior results compared to data analysis in other conditions. Classification with ERP data is believed to support the results of self-assessment and build methods of making profiles that are more accurate and reliable.