Autism spectrum disorder (ASD) is the chronic ailment of the central nervous system that causes the degradation of emotional and cognitive abilities. Long-term continuous monitoring with neurofeedback of human emotions for patients with ASD is crucial in mitigating its harmful effect. ASD is highly intimidating chronic disorder because of its sharp increase in the number of patients due to late detection. Lack of emotions’ regularization worsens their daily life by certain negative outbursts of emotions. These negative emotion outbursts cause self-injuries and degrade learning abilities. The timely emotion’s prediction can be used to regulate the emotions by controlling these outbursts. Noninvasive electroencephalogram (EEG) can be used for the emotion’s prediction using the valence and arousal scales. This work proposes a real-time wearable on-chip processor for the early prediction of the emotions. The miniaturized low-power processor can be embedded in a headband (patch sensor) for timely prediction of negative emotion’s outbursts. A linear support vector machine classifier is used with power spectral density, logarithmic interhemispheric power spectral ratio and the interhemispheric power spectral difference of eight EEG channel locations suitable for a wearable noninvasive classification system. A lookup table–based logarithmic division unit (LDU) is proposed to represent the division features in machine learning (ML) applications. The proposed LDU minimizes the cost of integer division by 34% for ML applications. The proposed emotion’s classification processor achieved accuracies of 72.96% and 73.14%, respectively, for the valence and arousal classification on multiple publicly available databases.
Neural Engineering Techniques for Autism Spectrum Disorder Volume 1: Imaging and Signal Analysis 2021, Pages 287-313,