Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Multi-task Neural Networks for Pain Intensity Estimation using Electrocardiogram and Demographic Factors

Stefanos Gkikas, Chariklia Chatzaki, Manolis Tsiknakis

TL;DR

This work tackles objective pain intensity estimation by leveraging ECG signals and demographic factors (age and gender). It introduces a multi-task neural network that jointly learns pain level along with age and/or gender, either directly or via feature augmentation with demographic data, to capture demographic-related variations in pain expression. Using the BioVid Heat Pain Database and HRV-derived ECG features, the study demonstrates that demographics substantially influence pain manifestation and that a demographic-aware MT-NN (especially with age+gender) yields superior performance compared to single-task models and prior methods. The findings support incorporating patient demographics into clinical pain assessment tools and motivate further integration of additional biosignals for robust, real-time pain estimation in healthcare settings.

Abstract

Pain is a complex phenomenon which is manifested and expressed by patients in various forms. The immediate and objective recognition of it is a great of importance in order to attain a reliable and unbiased healthcare system. In this work, we elaborate electrocardiography signals revealing the existence of variations in pain perception among different demographic groups. We exploit this insight by introducing a novel multi-task neural network for automatic pain estimation utilizing the age and the gender information of each individual, and show its advantages compared to other approaches.

Multi-task Neural Networks for Pain Intensity Estimation using Electrocardiogram and Demographic Factors

TL;DR

This work tackles objective pain intensity estimation by leveraging ECG signals and demographic factors (age and gender). It introduces a multi-task neural network that jointly learns pain level along with age and/or gender, either directly or via feature augmentation with demographic data, to capture demographic-related variations in pain expression. Using the BioVid Heat Pain Database and HRV-derived ECG features, the study demonstrates that demographics substantially influence pain manifestation and that a demographic-aware MT-NN (especially with age+gender) yields superior performance compared to single-task models and prior methods. The findings support incorporating patient demographics into clinical pain assessment tools and motivate further integration of additional biosignals for robust, real-time pain estimation in healthcare settings.

Abstract

Pain is a complex phenomenon which is manifested and expressed by patients in various forms. The immediate and objective recognition of it is a great of importance in order to attain a reliable and unbiased healthcare system. In this work, we elaborate electrocardiography signals revealing the existence of variations in pain perception among different demographic groups. We exploit this insight by introducing a novel multi-task neural network for automatic pain estimation utilizing the age and the gender information of each individual, and show its advantages compared to other approaches.
Paper Structure (14 sections, 4 equations, 6 figures, 8 tables)

This paper contains 14 sections, 4 equations, 6 figures, 8 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methods
  4. Dataset details
  5. ECG processing and feature extraction
  6. Neural network
  7. Single-task neural network
  8. Multi-task neural network
  9. Experiments and results
  10. Demographic groups
  11. Augmentation of feature vectors
  12. Multi-Task Neural Network
  13. Comparison with existing methods
  14. Conclusion

Figures (6)

  • Figure 1: The PQRST complex
  • Figure 2: The flow diagram of the pre-processing procedure of the Pan-Tompkins algorithm gkikas_chatzaki_2022
  • Figure 3: ECG pre-processing with Pan-Tompkins algorithm gkikas_chatzaki_2022
  • Figure 4: Illustration of the MTL network. The output vectors' size of the network is: for Pain classifier $nx1$ where n the number of pain estimation tasks (i.e. 2 for binary classification, 5 for multi-class classification), for Age classifier $36x1$ where 36 is equal to the number of possible values of subjects' age, for Gender classifier $2x1$ where 2 the number of possible values (i.e males, females)
  • Figure 5: Classification results on different Schemes
  • ...and 1 more figures