Constructing an Emotion Estimation Model Based on EEG/HRV Indexes Using Feature Extraction and Feature Selection Algorithms

Abstract

In human emotion estimation using an electroencephalogram (EEG) and heart rate variability (HRV), there are two main issues as far as we know. The first is that measurement devices for physiological signals are expensive and not easy to wear. The second is that unnecessary physiological indexes have not been removed, which is likely to decrease the accuracy of machine learning models. In this study, we used single-channel EEG sensor and photoplethysmography (PPG) sensor, which are inexpensive and easy to wear. We collected data from 25 participants (18 males and 7 females) and used a deep learning algorithm to construct an emotion classification model based on Arousal–Valence space using several feature combinations obtained from physiological indexes selected based on our criteria including our proposed feature selection methods. We then performed accuracy verification, applying a stratified 10-fold cross-validation method to the constructed models. The results showed that model accuracies are as high as 90% to 99% by applying the features selection methods we proposed, which suggests that a small number of physiological indexes, even from inexpensive sensors, can be used to construct an accurate emotion classification model if an appropriate feature selection method is applied. Our research results contribute to the improvement of an emotion classification model with a higher accuracy, less cost, and that is less time consuming, which has the potential to be further applied to various areas of applications.

EEG Emotion Classification Network Based on Attention Fusion of Multi-Channel Band Features

“…Anderson et al proposed that facial muscle movements can represent emotional states, in which the support vector machine (SVM) was used to identify six basic emotions commonly associated with facial expressions [ 5 ]. The second type of research is based on the neurophysiological state, that is, the acquisition of various physiological signals [ 6 , 7 , 8 , 9 , 10 ], such as electrocardiogram (ECG), photoplethysmography (PPG), and electroencephalogram (EEG), among many others. Although this type of research requires subjects to wear certain appropriate physiological signal acquisition equipment, compared with the former external behavioral research, focusing on neurophysiological states is a more objective method of representing emotions.…”

Neuroorganoleptics: Organoleptic Testing Based on Psychophysiological Sensing

“…After the selection and processing of the ECG samples, feature extraction was performed for relevant sections of the tests (mainly the sections where subjects tasted the food and answered the questions) to check the relation between the results of the self-report questionnaire (more specifically, the classification obtained in the global acceptance question) and the extracted features. The focus of this study was on the frequency features HF (High Frequency power) and LF (Low Frequency power), since they have been shown in previous studies to be valid in short-term recordings [ 34 ], and because of their relevance in emotional studies [ 39 ].…”

The Implementation and Evaluation of Individual Preference in Robot Facial Expression Based on Emotion Estimation Using Biological Signals

“…The emotion estimation methods in [ 10 , 18 , 19 ] correlate the values obtained by two physiological sensors, i.e., the brain waves (EEG) and heart rate variability (HRV), to the Arousal and Valence axes, respectively, in Russell’s circumplex model of affect [ 23 ]. The EEG can be used to measure the state of concentration and is reported to be negatively correlated with the subjectively evaluated level of arousal [ 24 , 25 ]. The HRV is a relatively reliable index for detecting stress and negative emotions [ 26 ].…”