Development of Unified Neuro-Affective Classification Tool (UNACT)

Abstract

Brain signals have been analysed to understand the affective state of different cognitive and mental conditions. For example, through the analysis, we can visualize the changes of emotion while driving, identify an autistic kid, understand the conditions that stimulate attention while studying, and many more, because emotion has a strong impact on cognitive processes in humans’ activities. This can be done through a machine learning technique, which includes data acquisition, pre-processing, feature extraction, and training. However, no existing tool integrates all supervised machine learning processes for affective state classification, which makes the process tedious and time-consuming for an analyst by doing programming. Therefore, this project aims to develop a brain analysis tool, namely Unified Neuro-Affective Classification Tool (UNACT). It consists of 3 main functions including training, classifying, and analysis. In the study of affective state electroencephalogram (EEG) signals have used, which measures brain signals. UNACT uses the Butterworth Bandpass filter for EEG signal filtering, the Power Spectral Density method for feature extraction, and the Multi-layer perceptron (MLP) for emotion classification. This tool can be used by a non-technical person to perform affective-emotional state analysis without having programming knowledge.

Estimating Systemic Cognitive States from a Mixture of Physiological and Brain Signals

“…The PSD of EEG is one of the most widely used features of EEG signals (Qin, Zheng, & Chen, 2019; Hossain, Yaacob, & Nordin, 2021; Hamzah, Norhazman, Zaini, & Sani, 2016; Al‐Nafjan, Hosny, Al‐Wabil, & Al‐Ohali, 2017). Specifically, PSD measures the power distribution of a given signal for each frequency in a given time‐frequency transform (Stoica et al., 2005).…”