Does TMS increase BOLD activity at the site of stimulation?

TL;DR

This review examines whether TMS increases local BOLD activity at stimulation sites, finding effects in motor and visual cortices likely due to downstream effects, but no consistent increase elsewhere at rest.

Contribution

It synthesizes existing TMS-fMRI studies to clarify the neurophysiological effects of TMS on local BOLD activity, highlighting the role of downstream effects and neuronal firing patterns.

Findings

TMS increases BOLD in motor and visual cortices due to downstream effects.

No consistent BOLD increase at other sites during rest.

Neuronal firing changes may cancel out, resulting in no net BOLD change.

Abstract

Transcranial magnetic stimulation (TMS) is widely used for understanding brain function in neurologically intact subjects and for the treatment of various disorders. However, the precise neurophysiological effects of TMS at the site of stimulation remain poorly understood. The local effects of TMS can be studied using concurrent TMS-fMRI, a technique where TMS is delivered during fMRI scanning. However, although concurrent TMS-fMRI was developed over 20 years ago and dozens of studies have used this technique, there is still no consensus on whether TMS increases blood-oxygen-level-dependent (BOLD) activity at the site of stimulation. To address this question, here we review all previous concurrent TMS-fMRI studies that reported analyses of BOLD activity at the target location. We find evidence that TMS increases local BOLD activity when stimulating the primary motor and visual cortices but that these effects are likely driven by the downstream consequences of TMS (finger twitches and phosphenes). However, TMS does not appear to increase BOLD activity at the site of stimulation for areas outside of the primary motor and visual cortices when conducted at rest. We examine the possible reasons for such lack of BOLD signal increase based on recent work in non-human animals. We argue that the current evidence points to TMS inducing periods of increased and decreased neuronal firing that mostly cancel each other out and therefore lead to no change in the overall BOLD signal.