Predicting cortical response during transcranial magnetic stimulation in humans.

摘要

Transcranial magnetic stimulation (TMS) is capable of noninvasively activating neurons in the brain. TMS can induce persistent effects and is being increasingly used in both clinical and research applications. Despite this growing interest, the relationship between TMS-generated electric fields (E-fields) and specific cortical electrophysiological responses is not well understood. Most analytical approaches focus on applied magnetic field strength in the target region as the primary factor, placing activation on the gyral crowns. However, imaging studies show cortical targets are typically located in the sulcal banks. To study this causal relationship, we combined subject-specific detailed finite element modeling (FEM) to calculate induced E-field information and employed concurrent TMS-PET data as a measure of cortical response. The research presented in this work is divided into three main parts, each one building on the results of the previous: (1) We determined that neocortical surface orientation was a critical determinant of regional activation by studying the locations of activation during TMS on the cortical surface. Results indicated that brain activations were primarily sulcal for both the TMS and task activations. This study provided further evidence that a major factor in cortical activation during TMS is the orientation of the cortical surface with respect to the induced E-fields. This was demonstrated by the fact that the sulcal bank of the primary motor cortex had larger cerebral blood flow (CBF) responses during TMS despite the gyral crown of the cortex being subjected to a larger magnetic field magnitude. (2) We sought to determine the E-field characteristics that lead to cortical activation. We found that decomposing the E-field into orthogonal vector components based on the cortical surface geometry (and hence, cortical neuron directions) led to significant differences between the regions of cortex that were active and non-active. Specifically, active regions had significantly higher E-field components in the normal orthodromic direction (i.e., parallel to pyramidal neurons in the dendrite-to-axon orientation) and in the tangential direction (i.e., parallel to interneurons) at high spatial gradient. This provides important new understanding of the factors by which TMS induces cortical activation necessary for predictive and repeatable use of this noninvasive stimulation modality (3) Finally, two different but related algorithms were formulated using different optimization approaches that provide a means for predicting topographical maps of cortical activation in humans. This is the first study to produce an algorithm for predicting the electrophysiological responses of neurons in the cortex based on both gross and microscopic brain anatomy correlated to relevant experimental recordings. This new innovation could provide an invaluable tool for predicting regions of cortical activation that may permit, among other benefits, improved prescriptive TMS protocols to optimize therapeutic response to TMS treatment.