Identification of reproducible individualized targets for treatment of depression with TMS based on intrinsic connectivity

Abstract

Transcranial magnetic stimulation (TMS) to the left dorsolateral prefrontal cortex (DLPFC) is used clinically for the treatment of depression however outcomes vary greatly between patients. We have shown that average clinical efficacy of different left DLPFC TMS sites is related to intrinsic functional connectivity with remote regions including the subgenual cingulate and suggested that functional connectivity with these remote regions might be used to identify optimized left DLPFC targets for TMS. However it remains unclear if and how this connectivity-based targeting approach should be applied at the single-subject level to potentially individualize therapy to specific patients. In this article we show that individual differences in DLPFC connectivity are large, reproducible across sessions, and can be used to generate individualized DLPFC TMS targets that may prove clinically superior to those selected on the basis of group-average connectivity. Factors likely to improve individualized targeting including the use of seed maps and the focality of stimulation are investigated and discussed. The techniques presented here may be applicable to individualized targeting of focal brain stimulation across a range of diseases and stimulation modalities and can be experimentally tested in clinical trials.

Keywords: Depression; Dorsolateral prefrontal cortex; Individual differences; Intrinsic connectivity; MRI; Resting state functional connectivity; Seed map; Subgenual; TMS; Transcranial magnetic stimulation; Variability.

Copyright © 2012 Elsevier Inc. All rights reserved.

Figures

Figure 1

Outline of datasets, analyses, and figures used in the current article. MRI data from dataset 1 (red), dataset 2 (blue), and dataset 3 (green) were run through different analyses (black boxes) and the results presented in several figures (right). The color of the box around each figure indicates the dataset used to generate that figure. SG = Subgenual

Figure 2

Identification of connectivity-based TMS targets in the left dorsal lateral prefrontal cortex (DLPFC) at the group and single subject level. Resting state functional connectivity maps are shown for the population (group) and two individual subjects (subject 1, 2) for a seed region in the subgenual cingulate (A) a seed map based on subgenual connectivity (B) and a seed map based on connectivity differences between effective and ineffective DLPFC TMS sites (C). Surface-based maps are masked to show only voxels in the left DLPFC. Black circles identify a potential stimulation site at the group level that is different between subjects 1 and 2.

Figure 3

Reproducibility of individual differences in dorsal lateral prefrontal cortex (DLPFC) connectivity across scanning sessions for two example subjects. Resting state functional connectivity maps are shown for two subjects (subject 1, 2) scanned on two separate days (day 1, 2) using a seed region in the subgenual cingulate (A) a subgenual-based seed map (B) and an efficacy-based seed map (C).

Figure 4

Similarity of dorsal-lateral prefrontal cortex (DLPFC) functional connectivity within individual subjects scanned on two separate days compared to the similarity between different subjects or between individual subjects and the group average. Bar graphs show the similarity (spatial correlation) between functional connectivity maps in the left DLPFC for individual subjects compared to themselves scanned on a different day (blue) other subjects (orange) or the group map (red) for the small subgenual seed region (A), the subgenual-based seed map (B), and the efficacy-based seed map (C). * P–4, **P<10−12.

Figure 5

Example of TMS targets identified based on group versus individualized functional connectivity for a single subject. Functional connectivity with the subgenual-based seed map was used to identify an optimal TMS target (i.e. strongest anticorrelation within a 6 mm sphere) in the left DLPFC based on the group map (red box) or individualized single-subject data from day 1 (blue box). The relative utility of these two targets was then tested using the independent single-subject fcMRI data from day 2 (black box). The average voxel value within the individualized target (blue bar) was compared with the average value within the group-based target (red bar). For this subject and sphere size, individualized targeting identifies a better anticorrelated node in the independent dataset than group-based targeting.

Figure 6

The difference between individualized and population-based targeting of TMS varies depending on the modeled size of the stimulated tissue. Graphs show the average voxel value from day 2 within presumptive TMS targets of various sizes (spheres of radii 1–30 mm and a 12 mm cone) identified on the basis of either individualized functional connectivity from day 1 (blue bars) or the group map (red bars). Results are shown for the small subgenual ROI (A), the subgenual-based seed map (B), and the efficacy based seed map (C). Significant differences between individualized and population-based targeting are identified. * P−5

Figure 7

Replication of primary findings in two patients with medication-refractory depression. DLPFC functional connectivity with the subgenual-based seed map is qualitatively more similar within patients scanned on two separate days (day 1, day 2) than across patients (A). Computation of the spatial correlation coefficient confirms this qualitative impression and also shows that patients are more similar to themselves scanned on a different day than they are to the group map (B). Similar to normal subjects, selecting a TMS target (12 mm cone) based on individualized functional connectivity identifies a more anticorrelated stimulation site in an independent dataset than selecting a target based on the group map (C).

Figure 8

Proposed strategy for individualized targeting of TMS to the left DLPFC based on resting state functional connectivity (fcMRI). *The ideal stimulation site is the volume most anticorrelated with the two seed maps. **The model of the volume of tissue activated by TMS is likely to vary depending on the particular TMS coil, stimulation intensity, coil orientation, underlying anatomy, and many other factors that could potentially be included to increase the accuracy of individualized targeting. DLPFC = dorsal lateral prefrontal cortex.