A Predictive Model for Emotion Recognition Based on Individual Characteristics and Autonomic Changes.

Abstract

Introduction: Studies have repeatedly stated the importance of individual differences in the problem of emotion recognition. The primary focus of this study is to predict Heart Rate Variability (HRV) changes due to affective stimuli from the individual characteristics. These features include age (A), gender (G), linguality (L), and sleep (S). In addition, the best combination of individual variables was explored to estimate emotional HRV.

Methods: To this end, HRV indices of 47 college students exposed to images with four emotional categories of happiness, sadness, fear, and relaxation were analyzed. Then, a novel predictive model was introduced based on the regression equation.

Results: The results show that different emotional situations provoke the importance of different individual variable combinations. The best variables arrangements to predict HRV changes due to emotional provocations are LS, GL, GA, ALS, and GALS. However, these combinations were changed according to each subject separately.

Conclusion: The suggested simple model effectively offers new insight into emotion studies regarding subject characteristics and autonomic parameters.

Highlights: HRV affective states was predicted using the individual characteristics.A novel predictive model was proposed utilizing the regression.Distinctive emotional situations provoke the importance of different individual variable combinations.The close association exists between gender and physiological changes in emotional states.

Plain language summary: In everyday life, emotions play a critical role in health, social relationships, and daily functions. Among physiologicalmeasures, the ANS activity, especially Heart Rate Variability (HRV), plays an important role in many recent theories of emotion. Many studies have analyzed HRV differences in the physiological mechanism of emotional reactions as a function of individual variables such as age, gender, and linguality, as well as other factors like sleep duration. It is the first study that explored the importance of individual characteristic's involvements and combinations was explored in the problem of emotion prediction based on an HRV parameter. To this effect, an emotion predictive model was proposed based on the linear combinations of individual differences with acceptable performance.