683: Wearable EEG estimation of Blood Flow Restriction Walking Pain Intensity: Impact of Pain Sensitivity

Topic: Experimental (Non-clinical) Human Studies
Subtopic: Pain Category (ICD-11), Acute musculoskeletal pain, Electrophysiology/Clinical neurophysiology (e.g. microneurography, EEG, MEG, OPMs, TMS), Artificial Intelligence/Machine Learning, Pain models (e.g. saline injection, capsaicin, UVB, HFS, incision, exercise, itch)

Background and Aims: Pain assessment still relies largely on self report, which limits precise evaluation in clinical care. Objective brain based measures have been explored, but fMRI and conventional electroencephalography are constrained to resting conditions and are difficult to deploy in everyday clinical settings. We previously validated a wearable electroencephalography (EEG) system developed at our institution that can record stable signals during movement[1]. Using this device, we tested whether pain intensity during walking can be classified from spectral EEG features and whether individual pain sensitivity alters model performance under lower limb blood flow restriction. We hypothesized higher sensitivity yields clearer neural signatures.

Methods: We enrolled 47 healthy participants. To elicit gradually increasing pain, participants walked while ischemia was induced in lower limbs using blood flow restriction. During walking, EEG was recorded with the wearable EEG at 0, 2, 4, 6, and 8 minutes from onset. Signals were transformed with fast Fourier transform to obtain band power features across standard frequency ranges. Machine learning models were trained to estimate pain intensity classes, and performance was quantified by area under the ROC curve. Participants were stratified by pain sensitivity threshold into low and high threshold groups, and all analyses were repeated within each subgroup.

Results: In two class estimation comparing 0 and 8 minutes, overall AUROC reached 0.87, whereas the low threshold group achieved 1.00. In three class estimation at 0, 4, and 8 minutes, overall AUROC was 0.69 and increased to 0.91 in the low threshold group. In five class estimation using all time points, overall AUROC was 0.65, compared with 0.86 in the low threshold group. Across tasks, performance was consistently higher among participants with lower pain thresholds, indicating clearer separability of EEG features with greater sensitivity. Accuracy declined as the number of classes increased, yet sensitivity related gains remained in subgroup analyses.

Conclusions: Wearable EEG combined with machine learning enabled accurate estimation of pain intensity during walking under lower-limb blood flow restriction, demonstrating feasibility of objective pain monitoring during movement. Performance was highest in participants with low pain thresholds, suggesting that individual pain sensitivity modulates how separable pain-related EEG features become during walking. This is consistent with prior work indicating that multivariate EEG patterns carry information about inter-individual pain sensitivity[2], supporting sensitivity-aware or personalized modeling strategies. Future work should validate generalization in patient populations, evaluate probability calibration and clinical utility, and refine artifact-robust pipelines for real-time deployment.

References: [1] Imai H, Kanie Y, Yoshimoto S, Yamamoto N, Furuya M, Fujimori T, Okada S. Classification accuracy of pain intensity induced by leg blood flow restriction during walking using machine learning based on electroencephalography. Sci Rep 2025;15:27955.
[2] Schulz E, Zherdin A, Tiemann L, Plant C, Ploner M. Decoding an individual’s sensitivity to pain from the multivariate analysis of EEG data. Cereb Cortex 2012;22:1118–1123.

Presenting Author: Yuya Kanie
Submitting Author: Yuya Kanie

Senior Author: Seiji Okada

IASP 2026 Global Year theme of Neuropathic Pain: No

• Yuya Kanie: MD - The University of Osaka
• Hirotatsu Imai: MD - The University of Osaka
• Natsuki Yamamoto: MD - The University of Osaka
• Seiji Okada: MD, PhD - The University of Osaka