Website of Philip Gouverneur

Selected Publications

27 September 2024

The PainMonit Database: An Experimental and Clinical Physiological Signal Dataset for Automated Pain Recognition

Scientific Data

Access to large amounts of data is essential for successful machine learning research. However, there is insufficient data for many applications, as data collection is often challenging and time-consuming. The same applies to automated pain recognition, where algorithms aim to learn associations between a level of pain and behavioural or physiological responses. Although machine learning models have shown promise in improving the current gold standard of pain monitoring (self-reports) only a handful of datasets are freely accessible to researchers. This paper presents the PainMonit Dataset for automated pain detection using physiological data. The dataset consists of two parts, as pain can be perceived differently depending on its underlying cause. (1) Pain was triggered by heat stimuli in an experimental study during which nine physiological sensor modalities (BVP, 2×EDA, skin temperature, ECG, EMG, IBI, HR, respiration) were recorded from 55 healthy subjects. (2) Eight modalities (2×BVP, 2×EDA, EMG, skin temperature, respiration, grip) were recorded from 49 participants to assess their pain during a physiotherapy session.

Download page

9 February 2023

Explainable Artificial Intelligence (XAI) in Pain Research: Understanding the Role of Electrodermal Activity for Automated Pain Recognition

MDPI Sensors

Philip Gouverneur Sensors Publication 2023

Artificial intelligence and especially deep learning methods have achieved outstanding results for various applications in the past few years. Pain recognition is one of them, as various models have been proposed to replace the previous gold standard with an automated and objective assessment. While the accuracy of such models could be increased incrementally, the understandability and transparency of these systems have not been the main focus of the research community thus far. Thus, in this work, several outcomes and insights of explainable artificial intelligence applied to the electrodermal activity sensor data of the PainMonit and BioVid Heat Pain Database are presented. For this purpose, the importance of hand-crafted features is evaluated using recursive feature elimination based on impurity scores in Random Forest (RF) models. Additionally, Gradient-weighted class activation mapping is applied to highlight the most impactful features learned by deep learning models. Our studies highlight the following insights: (1) Very simple hand-crafted features can yield comparative performances to deep learning models for pain recognition, especially when properly selected with recursive feature elimination. Thus, the use of complex neural networks should be questioned in pain recognition, especially considering their computational costs; and (2) both traditional feature engineering and deep feature learning approaches rely on simple characteristics of the input time-series data to make their decision in the context of automated pain recognition.

Link

15 July 2021

Comparison of Feature Extraction Methods for Physiological Signals for Heat-Based Pain Recognition

MDPI Sensors

Philip Gouverneur Sensors Publication 2021

While even the most common definition of pain is under debate, pain assessment has remained the same for decades. But the paramount importance of precise pain management for successful healthcare has encouraged initiatives to improve the way pain is assessed. Recent approaches have proposed automatic pain evaluation systems using machine learning models trained with data coming from behavioural or physiological sensors. Although yielding promising results, machine learning studies for sensor-based pain recognition remain scattered and not necessarily easy to compare to each other. In particular, the important process of extracting features is usually optimised towards specific datasets. We thus introduce a comparison of feature extraction methods for pain recognition based on physiological sensors in this paper. In addition, the PainMonit Database (PMDB), a new dataset including both objective and subjective annotations for heat-induced pain in 52 subjects, is introduced. In total, five different approaches, including techniques based on feature engineering and feature learning with deep learning, are evaluated on the BioVid and PMDB datasets. Our studies highlight the following insights: (1) Simple feature engineering approaches can still compete with deep learning approaches in terms of performance. (2) More complex deep learning architectures do not yield better performance compared to simpler ones. (3) Subjective self-reports by subjects can be used instead of objective temperature-based annotations to build a robust pain recognition system.

Link

All Publications

2025

eVAS: A user-friendly electronic Visual Analogue Scale

Journal of Open Source Software, 10.107, 6876.

Link

A comprehensive survey and comparative analysis of time series data augmentation in medical wearable computing

PloS one, 3, 20.

Link

2024

An Experimental and Clinical Physiological Signal Dataset for Automated Pain Recognition

Scientific Data, 11, 1051.

Link

2023

Explainable Artificial Intelligence (XAI) in Pain Research: Understanding the Role of Electrodermal Activity for Automated Pain Recognition

Sensors, 23(4), 1959.

Link

Objective Measurement of Subjective Pain Perception with Autonomic Body Reactions in Healthy Subjects and Chronic Back Pain Patients: An Experimental Heat Pain Study

Sensors, 23(19), 8231.

Link

Psychological Mechanisms of Offset Analgesia: The Effect of Expectancy Manipulation

PLOS ONE, 18(1), 1-16.

Link

Differential Effects of Thermal Stimuli in Eliciting Temporal Contrast Enhancement: A Psychophysical Study

The Journal of Pain, Elsevier.

Link

Exploring the Benefits of Time Series Data Augmentation for Wearable Human Activity Recognition

iWOAR '23, 20, 1-7.

Link

2022

SenseHunger: Machine Learning Approach to Hunger Detection Using Wearable Sensors

Sensors, 22(20), 7711.

Link

Temporal Properties of Pain Contrast Enhancement using Repetitive Stimulation

European Journal of Pain, Wiley Online Library.

Link

To Calibrate or not to Calibrate? A Methodological Dilemma in Experimental Pain Research

The Journal of Pain, Elsevier.

Link

2021

Comparison of Feature Extraction Methods for Physiological Signals for Heat-Based Pain Recognition

Sensors, 21(14), 4838.

Link

Classification of Heat-Induced Pain Using Physiological Signals

Information Technology in Biomedicine, 239-251.

Link

2020

AI Approaches Towards Prechtl's Assessment of General Movements: A Systematic Literature Review

Sensors, 20(18), 5321.

Link

My-AHA: Software Platform to Promote Active and Healthy Ageing

Information, 11(9), 438.

Link

2019

Adaptiv-lernende, technische Alltagsbegleiter im Alter: Abschlussbericht Cognitive Village: Berichtszeitraum: 01.09. 2015-30.11. 2018

Project report.

Link

2018

The My Active and Healthy Aging (My-AHA) ICT Platform to Detect and Prevent Frailty in Older Adults: Randomized Control Trial Design and Protocol

Alzheimer's & Dementia: Translational Research & Clinical Interventions, 4(1), 252--262.

Link

2017

Classification of Physiological Data for Emotion Recognition

In Artificial Intelligence and Soft Computing: 16th International Conference, ICAISC 2017, Zakopane, Poland, June 11-15, 2017, Proceedings, Part I 16 (pp. 619-627). Springer International Publishing.

Link

Publication list

Citations

Projects

Multimodale Plattform zum Schmerzmonitoring in der Physiotherapie (PainMonit)

BMBF project

Accurate pain monitoring is of paramount importance in many medical applications. As pain can be both a symptom and a disease and is also one of the most common reasons why people seek medical help, the accurate detection of pain plays a crucial role in the day-to-day history taking in hospitals, clinics and doctors' offices. The aim of PainMonit is to develop a multimodal sensor-based pain monitoring tool for the holistic monitoring of pain in a physiotherapeutic treatment context. The aim is to make the otherwise subjective perception of pain objectively quantifiable and reliable by combining simultaneously measured physiological data from different sensor-based systems. Different machine learning approaches were used, comparing traditional learning algorithms with artificial neural networks. Explainable AI methods helped to better understand the physiological responses to pain.

Tasks: Writing of project proposal; Lead implementation WP1

View Project

My Active and Healthy Aging (myAHA)

Horizon 2020 project

Interrelated frailties often manifest in older people, encompassing cognitive decline, physical weakness, depression, anxiety, and poor sleep quality. These conditions contribute to social isolation and impose a social and economic burden on both ageing adults and healthcare systems. Detecting these frailties early is crucial for promoting active and healthy ageing (AHA) and mitigating further decline. The EU-funded my-AHA project aims to employ advanced analytical concepts to facilitate early and precise health monitoring and disease prevention through personalised detection, recommendations, feedback and support. The project seeks to reduce the risk of frailty by enhancing overall wellbeing in ageing individuals, encompassing aspects such as physical activity, cognitive function, psychological well-being, nutrition and sleep.

Tasks: Lead of WP6 "Data Fusion, Analytics and other Services": organisation and implementation of a Decision Support System (DSS)

View Project

Sensor platform for automatic detection of fidgety movements (ScreenFM)

BMBF project

The ScreenFM study was conducted at the Department of Paediatrics and Adolescent Medicine at the University Medical Centre Schleswig-Holstein, Campus Lübeck. The researchers involved are developing and evaluating an app to analyse spontaneous movement patterns, known as fidgety movements (FM) in babies aged three to five months. The app is intended to improve screening for developmental abnormalities and enabling paediatricians and paediatricians across the country to do so. Learning-based AI algorithms have been developed to process the multimodal sensor signals and automatically assess the babies' fidgety movements to provide an early assessment.

Tasks: Writing of project proposal and report

View Project

eVAS: A user-friendly electronic Visual Analogue Scale

Open source software

Ratings from subjects are often obtained using simple Visual Analogue Scales (VAS) where subjects are asked to rate a particular outcome on a straight line or by a number in a given range with specific anchors. While these self-reports are easy to do with pen and paper, open-source software that can easily digitise and quantify the results of self-reports on an ongoing basis is rare. This is particularly true for researchers who are unfamiliar with writing code and software themselves. Thus, we are introducing the electronic Visual Analogue Scale (eVAS), an easy-to-use electronic visual analogue scale that continuously records subjects' feedback. Initially introduced to record pain levels for automated pain detection, but not limited to this use case, eVAS aims to provide a ready-to-use tool that is highly configurable, even for non-computer scientists, that can be used for various applications.

Tasks: lead implementation

Download page GitHub page