Predicting TMS Success for Tinnitus with Brain Biomarkers

A two-week course of repetitive transcranial magnetic stimulation (rTMS) helps about half of people with subjective tinnitus. Researchers have now identified a specific brain feature that appears to predict who those responders will be.

Predicting Treatment Success with Machine Learning

Led by Zhongling Ding, Bo Peng, and Mengfang Gong, a team sought to find a reliable predictor for rTMS outcomes. They enrolled 64 patients with chronic subjective tinnitus and 18 healthy controls. All patients received a standard two-week rTMS treatment protocol. Based on their symptom improvement, 36 patients (56.25%) were classified as responders.

Before treatment began, each participant underwent a high-resolution structural MRI scan. The researchers extracted 242 distinct measurements of brain morphology from these scans, covering cortical thickness, surface area, and gray matter volume across the entire brain. They then used a machine learning approach to see if any of these pre-treatment features could distinguish future responders from non-responders.

Ten regional features, spanning prefrontal, limbic, sensorimotor, and parietal brain networks, showed significant differences. The team fed these into an ExtraTrees machine learning model, evaluating its performance with rigorous 5-fold cross-validation.

The Right Frontal Lobe Holds the Key

The predictive model performed well, achieving an area under the curve (AUC) of 0.85 and an accuracy of 77%. To understand which brain feature drove the model’s decisions, the researchers used SHapley Additive exPlanations (SHAP) analysis. This technique identified a single region as the top predictor: the gray matter volume of the right pars triangularis of the inferior frontal gyrus (IFGtriang-R). A larger volume in this area was strongly associated with a positive treatment outcome.

To interpret this finding, the researchers compared the brain scans of three groups: responders, non-responders, and healthy controls. They discovered that responders had a significantly larger IFGtriang-R volume (0.90 ± 0.08) than both non-responders (0.86 ± 0.07) and healthy individuals (0.86 ± 0.06). This pattern suggests the feature is not simply a marker of having tinnitus, but a specific signature linked to a brain’s capacity to benefit from rTMS. Interestingly, the volume of this region did not correlate with the degree of symptom improvement, indicating it acts more as a threshold or gateway for potential success. You can read more about the promise of predicting TMS success for tinnitus in our related coverage.

A Threshold for Neuroplastic Change

The pars triangularis is part of the brain’s broader frontal network, which is heavily involved in cognitive control, attention, and executive function. In the context of tinnitus, which is often maintained by maladaptive attention and memory networks, this region’s role becomes clearer. The researchers propose that a larger gray matter volume in the right IFGtriang may represent a greater “neuroplastic reserve.”

This structural reserve could make certain brains more capable of undergoing the functional and connectivity changes that rTMS aims to induce. For these individuals, the magnetic stimulation may successfully modulate hyperactive auditory and attention networks, leading to reduced tinnitus perception. For those without this specific structural characteristic, the brain may be less responsive to the treatment’s neuroplastic push. This concept of a pre-treatment biomarker determining outcomes mirrors findings in other neuromodulation fields, such as research showing how baseline depression predicts long-term results for cognitive behavioral therapy for insomnia (CBT-I).

Practical Implications for Precision Treatment

This study, published with the DOI 10.3389/fneur.2026.1808769, moves the field toward precision neuromodulation. The practical implication is straightforward: a routine structural MRI scan before starting rTMS could help clinicians and patients make more informed decisions.

For patients showing the favorable brain signature, clinicians could proceed with rTMS with greater confidence in its likely benefit. For those without it, clinicians might discuss the lower probability of success and consider prioritizing or combining other evidence-based interventions first, such as the manual therapy and jaw exercises shown to reduce tinnitus severity in cases linked to temporomandibular dysfunction. This stratification prevents unnecessary cost, time commitment, and disappointment for patients unlikely to respond.

The finding also sharpens the scientific focus on the right frontal lobe’s role in tinnitus perception and treatment. Future therapies, whether advanced rTMS protocols or behavioral strategies like cognitive reappraisal used in misophonia treatment, might be designed to specifically engage and strengthen this cognitive control region. By identifying who is most likely to benefit, this research helps make targeted, effective tinnitus management a more tangible reality.