Resting-state networks link invasive and noninvasive brain stimulation across diverse psychiatric and neurological diseases

Abstract

Brain stimulation, a therapy increasingly used for neurological and psychiatric disease, traditionally is divided into invasive approaches, such as deep brain stimulation (DBS), and noninvasive approaches, such as transcranial magnetic stimulation. The relationship between these approaches is unknown, therapeutic mechanisms remain unclear, and the ideal stimulation site for a given technique is often ambiguous, limiting optimization of the stimulation and its application in further disorders. In this article, we identify diseases treated with both types of stimulation, list the stimulation sites thought to be most effective in each disease, and test the hypothesis that these sites are different nodes within the same brain network as defined by resting-state functional-connectivity MRI. Sites where DBS was effective were functionally connected to sites where noninvasive brain stimulation was effective across diseases including depression, Parkinson's disease, obsessive-compulsive disorder, essential tremor, addiction, pain, minimally conscious states, and Alzheimer's disease. A lack of functional connectivity identified sites where stimulation was ineffective, and the sign of the correlation related to whether excitatory or inhibitory noninvasive stimulation was found clinically effective. These results suggest that resting-state functional connectivity may be useful for translating therapy between stimulation modalities, optimizing treatment, and identifying new stimulation targets. More broadly, this work supports a network perspective toward understanding and treating neuropsychiatric disease, highlighting the therapeutic potential of targeted brain network modulation.

Keywords: TMS; clinical application; human connectome project; neurosurgery; tDCS.

Fig. 1.

Methodological approach for linking sites for invasive and noninvasive brain stimulation. (A) An ROI is created at a DBS site with reported efficacy for a given disease, in this case the subgenual cingulate for depression. (B) For each of 1,000 normal subjects, spontaneous modulations in the fMRI signal are extracted from this DBS ROI. (C) This time course is correlated with all other brain voxels and then averaged across subjects to create a DBS correlation map. (D) An ROI is created at the site where noninvasive stimulation is reported effective in the given disease, in this case the left DLPFC. (E) The site of noninvasive brain stimulation is illustrated on the DBS correlation map using a circle centered over the site.

Fig. 2.

Sites for invasive and noninvasive brain stimulation with the best evidence of therapeutic efficacy in each disease are functionally connected. For each disease, the site at which DBS is most effective is shown in red. Resting-state functional connectivity with this site is shown along with the correspondence to the site at which noninvasive stimulation is most effective in each disease (circles). Black circles indicate sites at which noninvasive excitatory stimulation (>5 Hz TMS or anodal tDCS) has been reported to be efficacious. White circles indicate sites where inhibitory stimulation (<1 Hz TMS or cathodal tDCS) has been reported to be efficacious.

Fig. 3.

Resting-state functional connectivity between sites of invasive and noninvasive brain stimulation is significantly higher than expected by chance. (A) For each disease, functional connectivity between the sites at which invasive and noninvasive stimulation are most effective is shown minus the connectivity between the same DBS site and random noninvasive sites (black bars). This analysis was repeated including all stimulation sites with evidence of efficacy rather than just the best site in each disease (gray bars). (B) Across diseases, resting-state functional connectivity between the site where DBS is most effective and the site where noninvasive stimulation is most effective (black bar) or between all sites where stimulation was effective (gray bar) was significantly greater than DBS connectivity with random sites. *P < 0.01, **P < 0.005.

Fig. 4.

Resting-state functional connectivity differentiates sites where brain stimulation is effective from sites where it is ineffective. (A–D) Diseases in which a specific site of noninvasive brain stimulation has been reported to be ineffective. For each disease, there is a lack of resting-state functional connectivity between the best DBS site for that disease and the site where noninvasive brain stimulation is ineffective (circle). In all cases, connectivity with the ineffective site was at or below chance levels and was significantly less than connectivity with the site at which noninvasive brain stimulation was effective for that disease (graphs). (E) One disease, Parkinson's disease, in which a specific DBS site (the VIM) has been found to be ineffective for most symptoms. Resting-state functional connectivity between the site where DBS is ineffective and the site where noninvasive brain stimulation is most effective (M1) was below chance and was significantly less than the connectivity between M1 and the sites where DBS was effective (graph). Black circles denote sites of excitatory noninvasive stimulation. White circles indicate sites of inhibitory noninvasive stimulation, as in Fig. 2. **P < 0.0001.

Fig. 5.

Positive versus negative resting-state functional connectivity with DBS sites relates to whether excitatory (>5 Hz TMS or anodal tDCS) or inhibitory (<1 Hz TMS or cathodal tDCS) noninvasive brain stimulation has been found to be more beneficial. (A) Resting-state functional connectivity with the STN shows negative correlation with M1 (black circle) and positive correlation with the SMA (white circle), consistent with the double dissociation in clinical benefit seen with excitatory versus inhibitory noninvasive brain stimulation at these sites (table). (B and C) Across diseases, sites at which inhibitory noninvasive brain stimulation was reported to be beneficial were more likely to be positively correlated with the DBS site; sites at which excitatory stimulation was reported to be beneficial were more likely to be negatively correlated. Results are shown for the best stimulation sites in each disease (B) and for all stimulation sites (C). **P < 0.005.