Our study introduces a novel two-stage AI framework that detects and localises pain using Electrodermal Activity (EDA) signals. We developed Multi-Domain Binary Patterns (MDBP) to extract hidden patterns from EDA signals, achieving 77.9% accuracy in pain detection and 69.67% in pain localisation. This work could pave the way for non-verbal pain assessment.
What other biosignals do you think could enhance pain recognition? Additionally, are there other publicly available physiological signal datasets for pain research that could help validate AI-based pain assessment models?
Read more about our work here: A Two-Stage Architecture for Identifying and Locating the Source of Pain Using Novel Multi-Domain Binary Patterns of EDA. (DOI:10.1016/j.bspc.2024.107454)
For pain research, various indicators and datasets are available to support different research approaches. These include:
### **Indicators:**
1. **Types of Pain:**
- Acute Pain
- Chronic Pain
- Neuropathic Pain
- Nociceptive Pain
2. **Pain Intensity:**
- Visual Analogue Scale (VAS)
- Numeric Rating Scale (NRS)
- Wong-Baker Faces Pain Scale
3. **Etiological Factors:**
- Surgery, injury, or infection
- Cancer and other chronic conditions
- Autoimmune disorders
4. **Impact on Quality of Life:**
- Sleep disturbances
- Functional limitations
- Emotional and psychological effects
5. **Biomarkers:**
- Inflammatory markers in blood (e.g., Cytokines: IL-6, TNF-alpha)
- Neurotransmitter levels (e.g., Serotonin, Dopamine)
- Hormonal changes (e.g., Cortisol levels)
6. **Cognitive and Emotional Factors:**
- Depression and anxiety scales
- Stress and pain perception
---
### **Datasets:**
1. **Open Pain Datasets:**
- **Open Pain Project:** Data on pain intensity and neuropathic indicators.
- **Brain Imaging Data Structure (BIDS):** Data on pain and brain functional connectivity.
2. **Clinical Trial Datasets:**
- **ClinicalTrials.gov:** Research data and outcomes on pain management.
- **ICPSR Database:** Health-related data, including pain management.
3. **Other Datasets:**
- **UK Biobank:** Data on pain and associated disorders.
- **National Health and Nutrition Examination Survey (NHANES):** Data on health and pain-related factors.
- **MIMIC-III Database:** Data from ICU patients.
4. **Specialized Datasets:**
- **The Pain Phenotype Repository:** Data on different types of pain.
- **SENSELab:** Neurobiological aspects of pain.
---
### **Applications in Research:**
- Development of new drugs or therapies for pain management.
- Assessment of the effects of yoga, meditation, and Ayurvedic treatments.
- Studying the neurological and psychological aspects of pain.
If you need detailed information or analysis on any specific indicator or dataset, feel free to ask!
Pain research involves a wide range of signals and datasets derived from both biological and computational sources to analyze the mechanisms, perception, and management of pain. Here are some key categories of signals and datasets used in pain research:
---
**1. Physiological Signals:**
- **Electroencephalogram (EEG):**
Captures brainwave activity to study pain perception and modulation. Often used to investigate pain-related neural oscillations and connectivity patterns.
- **Electromyography (EMG):**
Measures muscle activity and can assess pain-induced muscle tension or spasms.
- **Heart Rate Variability (HRV):**
Reflects autonomic nervous system responses to pain, often used to study acute and chronic pain effects.
- **Skin Conductance (SC):**
Measures autonomic arousal, which can indicate the intensity of pain.
- **Functional Magnetic Resonance Imaging (fMRI):**
Maps brain regions activated by pain stimuli, providing insights into pain processing pathways.
- **Near-Infrared Spectroscopy (NIRS):**
Measures oxygenation changes in the brain, often related to pain and its cognitive modulation.
---
### **2. Behavioral and Self-Reported Data:**
- **Pain Intensity Scales (e.g., Visual Analog Scale - VAS):**
Quantitative self-report measures of pain severity.
- **Behavioral Observation:**
Video and movement analysis datasets that record grimacing, body posture, or other pain-related behaviors.
- **Questionnaires and Surveys:**
Data on quality of life, functional impairment, and psychological factors related to chronic pain (e.g., McGill Pain Questionnaire, SF-36).
---
### **3. Clinical and Experimental Pain Datasets:**
- **Thermal, Mechanical, and Electrical Pain Threshold Data:**
Measurements of responses to calibrated stimuli used in experimental pain models.
- **Pharmacological Response Datasets:**
Data on the efficacy of pain-relieving medications in clinical trials.
- **Chronic Pain Patient Registries:**
Longitudinal datasets that track the progression and management of conditions such as neuropathic pain or fibromyalgia.
---
### **4. Genetic and Molecular Datasets:**
- **Genomics Data:**
Includes datasets from studies investigating genetic predispositions to pain (e.g., single nucleotide polymorphisms (SNPs) related to pain sensitivity).
- **Proteomics and Transcriptomics Data:**
Datasets analyzing biomarkers and signaling pathways involved in pain mechanisms.
- **Inflammatory Biomarkers:**
Cytokines, neuropeptides, and hormones measured in blood, cerebrospinal fluid, or tissue samples.
---
### **5. Digital Health and Wearables Data:**
- **Wearable Sensors:**
Data on activity levels, sleep patterns, and physiological responses from devices like smartwatches or dedicated health trackers.
- **Mobile App Data:**
Includes patient-reported outcomes, daily pain diaries, and medication adherence data.
---
### **6. Computational and Simulation Datasets:**
- **Neuromodulation Models:**
Computational simulations of spinal cord stimulation, transcranial magnetic stimulation (TMS), or other interventions.
- **Machine Learning Datasets:**
Labeled datasets used for training AI models to predict pain intensity or classify chronic pain conditions.
- **Pain Expression Databases:**
Image and video datasets used to analyze facial expressions or vocalizations linked to pain (e.g., CASME II for micro-expressions).
---
### **7. Publicly Available Pain Research Datasets:**
- **Open Pain Datasets:**
- **SPM Pain Dataset:** Brain imaging data related to pain stimuli.
- **Neuropain Database:** Information on gene targets involved in neuropathic pain.
- **Opioid Use and Outcomes Data:** Tracks opioid prescriptions and outcomes in chronic pain management.
- **Human Connectome Project (HCP):**
Data on brain connectivity, useful for understanding pain networks.
- **UK Biobank:**
Includes genetic, imaging, and phenotypic data, which can be used for pain-related research.
---
### **8. Animal Model Data:**
- **Behavioral Pain Models:**
Data on responses to stimuli in rodent models of neuropathic, inflammatory, or cancer pain.
- **Electrophysiology and Imaging:**
Neural recordings and brain activity maps from animal studies.
Pain research is a complex field, and researchers utilize a wide range of signals and datasets to understand and quantify pain. Here's an overview of some key areas:
1. Physiological Signals:
- Electroencephalography (EEG): Measures brain activity, providing insights into how the brain processes pain signals. Can reveal patterns associated with different types and intensities of pain.
- Electromyography (EMG):Measures muscle activity, which can indicate pain-related muscle tension or reflexes. Useful for studying musculoskeletal pain.
- Electrocardiography (ECG):Measures heart activity, which can be affected by pain-induced stress responses. Changes in heart rate variability can indicate pain.
- Skin Conductance (EDA/GSR): Measures changes in skin conductivity, reflecting sympathetic nervous system activity.Increased skin conductance is often associated with pain-related stress.
- Facial Expressions: Facial action coding system (FACS) is used to quantify facial expressions related to pain. Computer vision techniques can automate the analysis of facial pain expressions.
- Thermal Imaging: Measures skin temperature, which can change in response to pain. Useful for detecting localized pain, such as in inflammatory conditions.
2. Behavioral Data:
- Self-Report Questionnaires: Pain scales (e.g., visual analog scale, numerical rating scale) are used to quantify subjective pain intensity. Questionnaires assess pain-related disability, emotional distress, and quality of life.
- Observational Data: Researchers observe and record pain behaviors, such as guarding, limping, or facial grimacing.This can be particularly useful for studying pain in infants or non-verbal individuals.
- Movement Analysis: Motion capture systems and wearable sensors can track movement patterns, revealing pain-related changes in gait or posture.
3. Neuroimaging Data:
- Functional Magnetic Resonance Imaging (fMRI):Measures brain activity by detecting changes in blood flow. Reveals brain regions involved in pain processing.
- Positron Emission Tomography (PET):Uses radioactive tracers to measure brain activity and neurotransmitter function. Provides insights into the neurochemical basis of pain.
4. Clinical and Demographic Data:
- Medical Records:Provide information on diagnoses, treatments, and medical history.
- Demographic Data: Age, sex, and other demographic factors can influence pain perception and experience.
Key Dataset Resources:
- The PainMonit Database:Focuses on physiological signals for automated pain recognition.
- Datasets related to facial expression analysis: Many datasets exist containing video recordings of facial expressions during induced pain.
- NIH Minimal Dataset for Chronic Low Back Pain:Contains a wide range of self-reported data relating to chronic lower back pain.
It is important to note that ethical considerations are paramount in pain research, particularly when involving human subjects.
Sources and related content
The Radical Development of an Entirely New Painkiller
The opioid crisis has made it even more urgent to come up with novel approaches to treating suffering. Finally there’s something effective...
"A young street performer in Pakistan who became famous for not feeling pain. A group of people in Alabama who felt excruciating burning in response to the slightest heat. Both had mutations in the SCN9A gene, which acts on the function of the tiny sodium channel called NaV1.7 that helps to conduct electrical signals in pain-sensing nerves. “We all went crazy, because people without NaV1.7 were pain-free but otherwise normal,” says molecular neurobiologist John Wood. “It was unbelievably exciting.” Eventually, this discovery helped to kick off the development of a promising new painkiller: suzetrigine. In January, the drug (marketed as Journavx), became the first new non-opioid painkiller in more than twenty years to receive approval for acute-pain treatment from the US Food and Drug Administration..."
https://www.newyorker.com/magazine/2025/06/02/the-radical-development-of-an-entirely-new-painkiller
Why Do Headaches Feel So Different From Other Kinds of Pain?
Why does a severe headache — from the grind of a migraine to the searing cluster attack that throws a person into fits of rocking and pacing — feel so different from any other kind of physical anguish? Essentially, excruciating head pain seems to collapse the separation between body and self...
https://undark.org/2025/07/18/book-excerpt-headache/
- Badges
- Science topic
- Similar topics
- Agriculture
- Horticulture
- Seeds