Objective Neural Biomarkers for Pain, using Non-Invasive EEG
Where Pain Meets Proof
SynapSense is a research-driven neurotechnology initiative developing EEG-based biomarkers that capture pain-related cortical activity and translate it into quantitative metrics. We focus on primary somatosensory cortex centered signals and validate candidate features against controlled experimental pain tasks and standardized self-report measures.
SynapSense is research-stage and investigational. It is not a diagnostic or therapeutic medical device.
What is SynapSense Technologies?
SynapSense studies whether non-invasive EEG can reliably capture brain activity associated with the experience of pain, and whether those signals can be processed into an objective pain-related index that tracks pain over time.
Near-Term Goal
Scientific validation under controlled conditions, establishing reproducible neural features that correlate with pain states.
Long-Term Vision
Translation into clinically meaningful decision support tools that complement self-report and improve continuity of pain measurement.
What "Objective" Means
Measurable neural features that change systematically with pain states, evaluated for correlation with repeated self-report labels and generalization.
Scientific Definition
SynapSense develops and validates EEG-based biomarkers that capture pain-related brain activity in a continuous and clinically meaningful way. The approach combines high-density EEG recordings during experimentally controlled pain states with signal processing pipelines designed to isolate reproducible pain-related components while reducing confounds such as motion and artifact contamination.
Why this matters
Pain is a pervasive neurological and physiological condition that disrupts daily life, impairs functioning, and contributes to significant social and economic burden.
In the United States, more than 50 million adults live with chronic pain, and acute or recurrent pain affects tens of millions more.
When pain is misassessed, patients can experience under-treatment, inequitable care, and avoidable escalation of interventions. At the system level, poor measurement contributes to inappropriate opioid exposure and impairs clinical trials.
Current Measurement Limitations
Most clinical pain metrics rely on subjective, intermittent self-reports such as VAS or numeric rating scales. These are valuable, but they can fail when:
Non-verbal patients
Sedated, cognitively impaired, or non-verbal patients cannot self-report pain accurately.
Rapid fluctuations
Intermittent scores miss dynamics across minutes and hours as pain changes rapidly.
Provider variation
Interpretation varies and documentation becomes inconsistent across settings.
Recall bias
Patient recall biases distort reported averages rather than capturing true time courses.
The Unmet Need
There is no validated continuous, objective biomarker of pain in routine use. SynapSense exists because this gap blocks both scientific progress and clinical translation.
Our Research Thesis
Pain intensity and pain quality are encoded in cortical activity. If this encoding produces reproducible patterns in EEG, then we can extract features that track pain states and build a quantitative pain-related index that complements self-report and improves continuity of measurement.
Why Cortical Signals?
Neuroimaging and electrophysiology support that pain signals are represented in cortical networks, particularly the primary somatosensory cortex (S1) and motor-adjacent regions. This motivates focusing on sensorimotor electrode sites (C3, Cz, C4) aligned with S1 in the 10–20 system as practical access points for pain relevant dynamics.
Why S1 is a Priority Target
- S1 encodes spatial precision and intensity features closely linked to physical pain experience
- S1 shows rapid temporal responses at pain onset, enabling continuous measurement
Why EEG, and Why Now?
EEG captures millisecond-level dynamics and can be portable and scalable relative to MRI-based approaches. Pain is dynamic, and objective measurement needs high temporal resolution to detect onset, adaptation, and recovery. fMRI, while spatially precise, is unsuitable for continuous monitoring.
Targeted EEG Features
Alpha Band Suppression
Alpha suppression over sensorimotor cortex is commonly reported in pain paradigms and can correlate with pain features.
Beta Dynamics
Beta shifts at pain onset and during resolution suggest potential markers of transition states.
Connectivity Changes
Increased local coherence with reduced long-range connectivity can reflect network reorganization during pain.
Primary Research Hypothesis
SynapSense hypothesizes that oscillatory activity recorded from primary somatosensory cortex using a 32-channel wet-electrode EEG system can serve as a reliable, objective neural biomarker of pain intensity.
What we are doing right now
This is a mixed-method, cross-sectional investigation with repeated measures. Each participant completes a single laboratory session including baseline EEG, controlled pain tasks, recovery, and debrief.
Cold Pressor Test
120 seconds max
Blood Pressure Cuff Occlusion
5 minutes max
Transcutaneous Electrical Stimulation
10 minutes max
Each task is time-limited with strict safety caps. Participants can stop at any time.
Session Length
90–120 minutes
EEG System
32-channel wet electrode
Pain Labeling
VAS every 20 seconds
Questionnaires
McGill Pain, Pain Catastrophizing Scale
Recovery
Paced breathing + EEG
Debrief
Semi-structured usability interview
Real-Time Labeling
Participants provide VAS ratings every 20 seconds during tasks to produce a time series label stream aligned to EEG features.
Multidimensional Characterization
McGill Pain Questionnaire captures qualitative and multidimensional descriptors beyond simple intensity ratings.
Usability Feedback
Semi-structured interviews on comfort, tolerability, and usability support human factors learning for future iteration.
Data & Analysis Approach
Our pipeline combines rigorous EEG preprocessing with machine learning approaches that allow for comparison between interpretability and performance.
EEG Feature Extraction
- Spectral power (theta, alpha, beta, gamma)
- Entropy measures
- Coherence analysis
- Artifact rejection preprocessing
Model Families
- Support Vector Machines
- Random Forests
- Convolutional Neural Networks
Ground Truth
Self-reported VAS time series are used as ground truth labels for supervised learning. We explicitly train and evaluate association between neural features and reported pain intensity, rather than claiming pain can be inferred without labels.
Evaluation Philosophy
Performance is measured with metrics like balanced accuracy and AUC, complemented by correlation analyses (e.g., Spearman's rho) that test monotonic relationships between features and pain intensity ratings. Features should distinguish baseline, pain, and recovery phases while correlating with repeated VAS labels.
Safety, Ethics & Data Protection
Participant Safety
- All tasks are time-limited and supervised
- Participants can stop at any time
- Strict maximum exposure caps for all procedures
- Trained research staff present throughout
Privacy & Data Security
- Unique numerical subject IDs assigned
- Identifiers separated from research data
- Encrypted UT-approved storage
- Access limited to authorized personnel
Research Legitimacy
- Conducted in UT Austin research facilities
- IRB oversight and trained staff
- Active pilot study in progress
Recognition & IP
- Dell Medical School "Hacking Pain" finalist
- Nucleate Activator, Spring 2026
- Kendra Scott WELI Founder Program, Spring 2026
- Patent pending: Non-Invasive Neural Biomarker Quantification of Pain
Interested in participating?
We are actively recruiting participants for our pilot study. Your participation helps advance the science of objective pain measurement and could contribute to better pain care in the future.
What to Expect
Initial Screening
Brief eligibility check and scheduling
Lab Session
90–120 minute session with EEG recording during controlled pain tasks
Recovery & Feedback
Paced breathing recovery period followed by brief usability interview