A probabilistic transcranial magnetic stimulation localization method

Transcranial magnetic stimulation (TMS) offers a safe and noninvasive way to activate brain tissue. However, the specific parameters of this activation have remained largely unclear. In this study, we introduce a new model for neuronal activation and develop a method to determine its parameters using motor evoked potential signals.

Approach:

We model the relationship between neuronal activation induced by an electric field and the measured motor threshold. Using Bayes' formula, we infer the posterior distribution of the model parameters from measurement data. The measurements include active motor thresholds obtained with various stimulating coil locations. The parameters of the model consist of the location, preferred direction of activation, and threshold electric field value at the activation site. We sample the posterior distribution using a Markov chain Monte Carlo method. The model's plausibility is quantified by calculating the marginal likelihood of the measured thresholds. We validate the method with synthetic data and apply it to motor threshold measurements from the first dorsal interosseus muscle in five healthy participants.

Main Results:

The method generates a probability distribution for the activation location, which allows us to identify a minimal volume where activation occurs with 95% probability. For eight or nine stimulating coil locations, this volume was approximately 100 mm³. This region intersected the pre-central gyrus and the anterior wall of the central sulcus, with the preferred activation direction perpendicular to the central sulcus—findings consistent with previous research. It was also not possible to definitively determine whether the activation occurred in white or grey matter. In one participant, we identified two distinct activation sites, while others exhibited a single site.

Significance:

Our method is both versatile and robust, providing a foundation for a comprehensive framework that accurately analyzes and characterizes TMS activation mechanisms.

J Neural Eng. 2021 Sep 3;18(4). doi: 10.1088/1741-2552/ac1f2b. PMID: 34475274.