Graph theory-based machine learning for automated fibromyalgia diagnosis using fMRI during an emotional task

Background Fibromyalgia (FM) is a multifaceted chronic pain disorder presenting with pain throughout the musculoskeletal system, alongside chronic fatigue,sleep disturbances, and cognitive impairments. Current diagnostic criteria primarily rely on symptom questionnaires, with no specific laboratory markers or definitive tests, making diagnosis subjective and often delayed. Moreover, the overlap with other central pain syndromes further complicates accurate diagnosis. Neuroimaging studies, particularly functional…

By the Truvace deskOctober 15, 20262 min read1 sourcePeer-reviewedTRV-2026-1463· cite & verify

In brief

Researchers analysed functional MRI scans collected during an emotional task from 32 people with fibromyalgia and 30 healthy controls. They computed graph-theory metrics for 13 regions of the emotion regulation network and used a genetic algorithm to pick 84 features to feed several machine learning classifiers.

Fibromyalgia is currently diagnosed mainly through symptom questionnaires, so an objective, imaging-based aid could reduce subjectivity and delay. The work is a model-development study on a small sample, and the source does not report whether it works in routine clinical settings or on other patient groups.

Main points

Fibromyalgia diagnosis relies mainly on symptom questionnaires, with no specific laboratory markers or definitive tests.
The data came from 32 patients and 30 healthy controls, using fMRI recorded during an emotional task.
Graph theory metrics were taken from 13 regions of the emotion regulation network.
Polynomial SVM had the highest overall accuracy, while the mixture-kernel SVM showed superior sensitivity and AUC.

The gain

A mixture-kernel support vector machine trained on graph-theory features from emotion-regulation brain regions best distinguished 32 fibromyalgia patients from 30 controls on sensitivity and AUC, suggesting a route to less subjective diagnosis.

The rundown

The authors note that fibromyalgia overlaps with other central pain syndromes, which complicates diagnosis. Although fMRI has shown promise in revealing brain abnormalities linked to the disorder, turning those findings into reliable diagnostic tools has been difficult, especially with small samples.

To handle the limited data, the team used a genetic algorithm for feature selection and trained SVMs with polynomial and RBF kernels, Random Forest, Gradient Boosting Machine, AdaBoost and a mixture-kernel SVM within a cross-validation framework to support robustness.

They chose the final model mainly on sensitivity and AUC because of the clinical value of reducing false negatives. The authors conclude the approach shows promising potential for improving diagnostic accuracy and supporting psychological and psychiatric interventions.