Abstract:Mental stress and associated heart disorders are some of the considerable causes of death in India and globally, as reported by the World Health Organization (WHO). The long-term presence of mental stress and hypertension in lifestyle can lead to significant disorders, including cardiac arrhythmias. Some researchers in the past have proposed methods to assess mental health by considering various physiological signals and artificial intelligence methods. Most studies have used complex algorithms and a combination of features extracted from bio-signals such as electroencephalogram (EEG), electrocardiogram (ECG), etc. This work aims to develop a simple and efficient method to classify mental stress from Heart Rate Variability (HRV) features of ECG signal. The proposed method for stress assessment is implemented using 1-min segments of real-time ECG signals from the Stress Recognition in Automobile Drivers (SRAD) database. We derived the HRV signal from the de-noised ECG signal and extracted time-domain statistical features. Correlation analysis is applied to statistical HRV features to establish the relationship among different metrics of HRV. The main contribution is to develop a Multiple Linear Regression (MLR) model to predict Heart Rate (HR) using statistically correlated HRV features. Another contribution is the design of a multinomial logistic classifier to classify the mental stress level into three classes: low, medium, and high stress. Simulation results demonstrate that time-domain HRV features are statistically correlated, and with a suitable choice of parameters, the proposed method can achieve a classification accuracy of 90.32%. The findings suggest that using fewer relevant features of ECG, the proposed method provides improved classification performance comparable to the state-of-the-art methods. Because of the non-invasive and simple approach, the proposed technique is suitable for accurately assessing mental stress in real-time environments. The presented approach strongly promotes the development of remote and automatic methods of mental stress assessment for real-world applications using wearable sensors and the Internet of Things.

Proposing a convolutional neural network for stress assessment by means of derived heart rate from functional near infrared spectroscopy

Accurate Stress Assessment based on functional Near Infrared Spectroscopy using Deep Learning Approach

Stress Assessment with Convolutional Neural Network Using PPG Signals

Early Detection of Mental Stress Using Advanced Neuroimaging and Artificial Intelligence

Symmetric Convolutional and Adversarial Neural Network Enables Improved Mental Stress Classification From EEG

Real-Time Psychological Stress Detection According to ECG Using Deep Learning

Multimodal Hierarchical CNN Feature Fusion for Stress Detection

Multimodal Affective State Assessment Using fNIRS + EEG and Spontaneous Facial Expression

Multimodal Stress Detection Using Facial Landmarks and Biometric Signals

A Novel Method for Mental Stress Assessment Based on Heart Rate Variability Analysis of Electrocardiogram Signals

Stress Detection Using PPG Signal and Combined Deep CNN-MLP Network

Enhancing Cognitive Workload Classification Using Integrated LSTM Layers and CNNs for fNIRS Data Analysis

Fabrication of a portable device for stress monitoring using wearable sensors and soft computing algorithms

StressNet: Detecting Stress in Thermal Videos

Driving Stress Detection Using Multimodal Convolutional Neural Networks with Nonlinear Representation of Short-Term Physiological Signals

Stress detection with encoding physiological signals and convolutional neural network

Shuffled ECA-Net for stress detection from multimodal wearable sensor data

Multi-level Stress Assessment Using Multi-domain Fusion of ECG Signal

A Fusion Shallow and Deep Features Network for Facial StO2 Stress Classification

An Automatic Stress Analysis and Emotion Detection Framework for Humans Based on Several Brain Signals

A deep convolutional neural network for estimating hemodynamic response function with reduction of motion artifacts in fNIRS