Reliability of Subject-Specific BOLD Response to Different TMS Protocols in the DLPFC

Abstract

Introduction

Transcranial magnetic stimulation (TMS) is a non-invasive technique for modulating neural activity, used in combination with functional MRI (fMRI) to study the causal effects of brain stimulation and its effects on cognition. Previous studies have shown that within-session TMS produces consistent brain activation [1], but to our knowledge, no studies have systematically investigated the reliability of blood-oxygen-level-dependent (BOLD) responses to TMS across multiple sessions in the timescale of days and weeks. We focus on the dorsolateral prefrontal cortex (DLPFC), a region implicated in cognitive control and emotion regulation, and a frequent target in clinical TMS interventions for treatment-resistant Major Depressive Disorder (MDD). Understanding the consistency of DLPFC activation across time is crucial for both basic neuroscience and the development of reliable neurostimulation protocols. We hypothesize that intra-subject BOLD response patterns will be more consistent than inter-subject patterns, though some variability across sessions will arise due to fluctuations in individual brain states and excitability. 

Methodology

This study leverages interleaved TMS-fMRI, with stimulation synchronized to scanner triggers to avoid imaging artifacts and carefully randomized to avoid behavioral anticipation. Following phantom validation, we conduct in-vivo studies beginning with an intake session for anatomical imaging, resting-state connectivity mapping for individualized DLPFC targeting, and motor threshold determination. Stimulation intensities are personalized using electric-field simulations to account for anatomical differences between the motor cortex (used to determine the threshold of activation through a finger twitch) and the DLPFC, aiming for consistent stimulation intensity across regions [2]. During subsequent TMS-fMRI sessions, neuronavigation ensures the precision of coil placement relative to the stimulation target, using infrared tracking of the coil and a subject-mounted mouthpiece. Each session includes resting-state scans and the stimulation protocols, alternating between singlet and triplet stimulation patterns, to examine protocol-specific effects. Functional MRI data is analyzed using subject-specific general linear models to assess intra- and inter-subject reliability of TMS-evoked BOLD responses. 

Motivation

Our expected insights include informing future TMS-fMRI research by quantifying protocol-dependent variations and assessing the reproducibility of TMS-induced brain activation over time. A deeper understanding of the effects of stimulation on the DLPFC target sheds light on the consistency of effects, providing important insights for MDD treatment. Therefore, this project bridges neuroscience, clinical psychology, and cognitive science, testing theories of brain network stability and the functional architecture underlying cognition, while providing empirical data relevant to further establishing TMS-fMRI as a predictor for clinical TMS treatment response [3]. Moreover, the methodology enhances our ability to measure causal brain network dynamics and assess the reproducibility of non-invasive brain stimulation over time. While primarily methodological, this research informs cognitive neuroscience by contributing to more reliable tools for probing brain-behavior relationships.

References

[1] C. Hawco et al., “Within‐subject reliability of concurrent TMS-fMRI during a single session,” Psychophysiology, vol. 60, no. 7, 2023. doi: 10.1111/psyp.14252.

[2] O. Numssen et al., “Electric-field-based dosing for TMS,” Imaging Neuroscience, vol. 2, pp. 1–12, 2024. doi: 10.1162/imag_a_00106.

[3] R. Duprat et al., “Resting fMRI-guided TMS evokes subgenual anterior cingulate response in depression,” NeuroImage, vol. 305, p. 120963, 2025. doi: 10.1016/j.neuroimage.2024.120963.