Development of stress biomarkers through physiological analysis using MATLAB software

Abdul Razak, Ahmad Fakhrullah (2025) Development of stress biomarkers through physiological analysis using MATLAB software. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Text (Full Text) Development of stress biomarkers through physiological analysis using MATLAB software.pdf - Submitted Version Download (2MB)

Abstract

Stress is one of the most common contributors to mental illnesses, affecting people worldwide. Recent studies have highlighted the potential of electrocardiogram (ECG) signals as biomarkers for stress detection, offering a non-invasive approach to identifying and managing these conditions. While ECG signaling characteristics have demonstrated high accuracy in identifying biomarkers, existing methods for biomarker identification remain limited. This project aims to study ECG signal biomarkers most likely to be influenced by stress, develop stress detection through ECG biomarkers using baseline methods and Bazett’s formula in MATLAB, and analyze stress episodes in QT intervals of ECG signals using statistical approaches. The project employs the KL75001 ECG module and K&H software for ECG data acquisition. Mental arithmetic and the Stroop Color Word Test (SCWT) were used as stimuli to induce changes in ECG waveforms under stress conditions. QT intervals were calculated using Bazett’s formula and baseline methods, while statistical analysis involving mean, variance, and standard deviation was conducted to quantify signal characteristics during stress episodes. The findings show that a QRT peak detection method was successfully implemented, enabling stress detection through ECG signals. Using Bazett’s formula, it was determined that QTc intervals exceeding the baseline threshold of 0.45 indicate stress with accuracy of 70%. Positive skewness in QT intervals suggests a longer tail on the right side of the distribution, indicating more variability and a tendency toward longer QT intervals which means the data are under stress conditions. Additionally, high kurtosis reflects a distribution with heavy tails and a sharp peak, implying that extreme QT interval values, both long and short, occur more frequently under stress condition. The integration of controlled stimuli, baseline thresholding methods, and computational tools in MATLAB demonstrates the potential of this approach for applications in health monitoring, stress management, and personalized interventions aimed at mitigating stress-related health risks.

Actions (login required)

View Item