Transcranial Magnetic Stimulation for Developmental Neuropsychiatric Disorders with Inflammation

Transcranial magnetic stimulation (TMS) is a form of noninvasive brain stimulation that is FDA-cleared for treatment-resistant major depression, migraines, obsessive-compulsive disorder (OCD), and smoking cessation in the USA [1]. TMS relies on the principle of electromagnetic induction, with a time-varying magnetic field leading to electrical current changes in the brain and subsequent neuronal firing or inhibition [2].

In this paper, we will discuss the mechanisms of TMS in further detail and focus on the synaptic plasticity changes induced by TMS. As there is evidence that TMS modulates synaptic plasticity in the brain, and plasticity can be modulated by inflammation, there are theoretical reasons to believe that TMS may be useful in inflammatory brain conditions, including in the developing brain [3]. In this review, we will focus on evidence for TMS effectiveness in humans with conditions that are thought to have an inflammatory component, particularly neurodevelopmental disorders such as autism, Tourette syndrome (TS), and OCD.

TMS is a means of inducing electrical activity in the brain by using magnetic pulses on the scalp. The advantage of using this method instead of using electricity directly on the scalp to induce electrical activity in the brain is that the magnetic field induced by TMS passes unimpeded through the scalp and skull to the brain (unlike electricity) [4]. The fact that the magnetic field can pass unimpeded to the brain has the advantage that TMS is relatively painless, though there can still be scalp discomfort [4]. Using electrical current at the scalp directly has to overcome the resistance of the scalp and skull, leading to increased pain.

The concepts behind TMS are based on the fundamental physics of charge and magnetism. Charges are a fundamental characteristic of nature. Charges move due to a difference in electric potential (also known as voltage). Moving charges induce a current; current is defined as the amount of charge that passes through a given location per unit time. These moving charges induce changes in the surrounding electric field that impact other charges. In the context of TMS, the changes in the electric field induced by the current in the stimulating coil induce a time-varying magnetic field.

As Michael Faraday discovered in the 19th century [5], a time-varying, changing magnetic field causes electromagnetic induction to occur, which leads to the induction of an electrical current [5]. In the context of TMS for brain stimulation, the induction of electrical current occurs in the cortex of the brain.

A TMS device thus needs a charging circuit and capacitor, as well as a stimulating coil [2]. A time-varying current in the stimulating coil leads to a time-varying magnetic field, which, via electromagnetic induction, induces current in the conductor (in this case, cortical tissue). Current in the brain tissue leads to action potentials in the cortex and subsequent changes in neuronal firing.

There are a large number of coils available for research, but there are two basic coil types available clinically, figure-of-8 coils and H-coils. A figure-of-8 coil is composed of two circular coils placed side by side, where the two coils are touching. The location where the two coils meet is where the magnetic field energy is the greatest and where brain stimulation is the greatest. H-coils are also available, with complicated windings that lead to deeper and, in particular, more diffuse effects on brain tissue.

TMS was designed originally to stimulate peripheral nerves [6], but Barker in 1985 was the first to use the device on the brain (motor cortex). When TMS was first used, it was with single pulses over the motor cortex, leading to movement.

However, over time, TMS was utilized with repeated pulses (a train of pulses with a break in between called the intertrain interval). The benefit of repetitive TMS is that it can cause long-term changes in synaptic and network plasticity, leading to therapeutic benefit. When we use the term TMS now in the context of clinical conditions, it is to be understood that we are referring to repetitive TMS.

Plasticity can be categorized as developmental, adaptive (experience-dependent), and reactive [7]. TMS can be thought of as affecting adaptive plasticity by modulating synapses at a dendritic level, with effects on dendritic sprouting or pruning. At a molecular level, this plasticity may be mediated by changes in calcium levels, AMPA receptor concentrations, GABA, glutamate, and n-acetylaspartate [8, 9].

The current thinking is that high-frequency TMS (greater than or equal to 5 Hz) causes AMPA receptor changes and consequent ejection of the magnesium plug from the AMPA receptor, leading to opening of calcium channels and phosphorylation and delivery of more AMPA receptors to the synapse. This, in turn, leads to bigger dendritic spines and increased synaptic strength and can be considered part of long-term potentiation or LTP [9]. Low-frequency TMS (less than 5 Hz) is thought to lead to decreased calcium, removal of AMPA receptors from the synapse, a smaller dendritic spine, and consequently decreased synaptic strength (long-term depression or LTD) [9].

As dendritic spines increase with high-frequency TMS, the dendrites become thicker and have more branches (a process called arborization). This in turn leads to plasticity at a network level that endures over time [9].

TMS has been shown to change dendritic spine density and morphometry [10]. Changes in these dendritic factors can lead to, at a larger scale, changes in connectivity between networks in the brain. These concepts have been worked out to a significant degree for neuropsychiatric disorders such as major depression (for which TMS is FDA-cleared). TMS is thought to improve clinical outcomes by modulating brain networks that have become dysregulated in the disease.

In treatment-resistant major depressive disorder (MDD), functional neuroimaging studies have shown that there is hyperconnectivity within an intrinsic connectivity network called the default mode network (DMN) [11]. The DMN is active during periods of introspection or self-reflection [12]. Treatment of MDD with TMS leads to decreased connectivity within the DMN [13]. It also leads to increased connectivity between the central executive network and the DMN [13]. These changes in brain network connectivity are correlated with clinical improvement of MDD [13].

TMS may be a potential treatment in brain disorders with an inflammatory component via its effects on synaptic plasticity. Inflammation can cause synaptic plasticity dysregulation, perhaps via effects of microglia on regulation of synaptic plasticity [14].

Microglia may affect synapse formation via direct contact with dendrites [15]. Microglia may also affect synapse formation with release of cytokines, particularly IL-10 [16]. Microglia may also cause synapse elimination by direct dendritic phagocytosis [17].

Neurodevelopmental disorders may involve microglial dysfunction [18, 19], with decreased synaptic elimination by microglia. In Fragile X syndrome, where patients can have autistic symptoms such as decreased sociality and repetitive behaviors, lack of proper phagocytosis of synapses by microglia may be a factor [20]. Something similar may be found in Rett syndrome [21]. There could also be excessive synaptic pruning by microglia in autism [18]. Overall, there is evidence that microglia are important in proper dendritic pruning and that this can go awry in neurodevelopmental disorders.

As microglia respond to inflammatory signals, inflammation at critical periods of brain development can lead to life-long consequences [18]. Prenatal infections seem to increase risk of autism [18]. Indeed, there is evidence of excessive inflammation in mothers who have children with autism [22]. There is also evidence of increased inflammatory cytokines such as TNF-alpha and IL-6 in the cerebrospinal fluid of ASD patients [23].

TMS can affect inflammation directly [3], perhaps with reduction of TNF-alpha [24]. High concentrations of TNF-alpha (as in local or systemic inflammation) may impair synaptic plasticity via decrease of long-term potentiation [24]. TMS may also decrease expression of genes in astrocytes related to inflammation and calcium signaling [25]. Finally, TMS may act directly on microglia (which are a source of TNF-alpha) by modulating microglial cytokine expression [24].

We will now discuss the evidence for the use of TMS in neurodevelopmental disorders that may have an inflammatory component. In this review, we will focus on autism spectrum disorder, TS, and OCD. It is important to note that these disorders are complex disorders, with likely multiple etiologies, and inflammation is likely one of many factors that lead to disease manifestation.

ASD in patients can have a number of symptoms, but the core symptoms involve impairments in social communication and restricted and repetitive interests and behaviors. ASD can also involve abnormalities in sensory reactivity and cognition, particularly executive function [26]. Given the diffuse range of symptoms and the likelihood of multiple inflammatory aspects to the development of ASD [18‒20], dysfunction in multiple brain networks is likely [26].

Current thinking regarding ASD suggests that there may be aberrant synaptic plasticity and an abnormal excitation/inhibition ratio in the brain. Most TMS studies in autism have used low-frequency (1 Hz) stimulation (which is thought to be inhibitory) to the dorsolateral prefrontal cortex (DLPFC) due to the notion that an imbalance of excitatory to inhibitory activity is important in the causation of the disorder [26, 27]. GABAergic dysfunction, leading to decreased inhibition and excess frontal activation and abnormal connectivity to other regions of the brain, may be occurring in ASD [26]. Further, the subsequent excess glutaminergic activity can lead to glutaminergic cortical excitotoxicity [28].

Most TMS studies in ASD have been open label, and most of the ones that were controlled used a wait-list control rather than a sham-controlled arm [28]. The studies have shown improvement in different aspects of autism, depending on the area of the brain targeted. For example, with DLPFC targeting, executive function, irritability, and repetitive behaviors have been shown to improve, while motor behavior has been shown to improve with supplementary motor area (SMA) targeting [26]. In a recent meta-analysis of the extant literature, with 12 clinical trials incorporated, TMS showed mild to moderate behavioral improvements and large effects on cognition [28]. There are neurophysiological changes that may occur with TMS that correspond to clinical improvement, particularly increased peak alpha frequency and increased alpha coherence on EEG [27]. This may be particularly relevant as EEG studies have shown that children with ASD may have decreased long-range connectivity in the alpha band between the frontal lobes and other brain regions [29].

While promising, there are a number of questions yet to be answered before TMS may be used clinically for ASD. There are limited randomized sham-controlled trials (only 1 as of 2022) [28]. It is also not yet clear what the optimal target, frequency of stimulation, number of sessions, number of pulses per session, and durability of effects are; there is significant heterogeneity in functioning and intellectual disability in autism, and it is not clear who the best candidates would be for TMS treatment [30]. Seizures are a theoretical risk with TMS, especially with high-frequency stimulation. Given that patients with ASD have higher seizure rates, this is also a factor to be considered, though preliminary data are reassuring [28]. Finally, the optimal age of intervention with TMS has not yet been established, but there may be greater effects of TMS earlier in life, when the brain is more plastic [31].

TS may also have inflammatory aspects and may be improved with TMS. TS is a disorder where patients have motor and vocal tics. There often is an urge for the patient to act on a tic, with subsequent time-bound relief. Tics can involve sudden large movements but can also be more subtle or primarily vocal [32].

There are a number of lines of evidence that suggest that there is an inflammatory component to TS. Increased antiphospholipid and antineural antibodies directed against the basal ganglia have been reported in TS patients [33]. Increased IgE levels have also been reported [33]. Further, there may be increased levels of IL-12 and TNF-alpha in TS [34].

There is also some evidence that TS can occur after infection, though there is some controversy regarding this. PANDAS (Pediatric Autoimmune Neuropsychiatric Disorders Associated with Streptococcal Infections) was conceptualized as an autoimmune disorder with cross-reactivity of antibodies to streptococcus A epitopes with basal ganglia antigens [35]. In particular, antibodies against N-acetyl-beta-d-glucosamine (GlcNAc), the dominant epitope of the Streptococcus A carbohydrate, may cross-react to neuronal antigens like lysoganglisode-GMI and beta-tubulin [36].

PANDAS may involve tics or OCD with a temporal correlation with recent streptococcus A infection. However, it is not clear whether this is a truly separate clinical entity. There is some evidence, for example, that non-streptococcus infections may also be correlated with TS [37].

Infection in general may lead to increased microglial activation in the striatum (caudate, putamen) of the basal ganglia and increased release of pro-inflammatory cytokines like TNF-alpha [35, 36]. Inflammatory cytokines may affect dopamine levels in the striatum and affect the cortico-striato-thalamic-cortical circuit, leading to motor and neuropsychiatric symptoms of TS [35, 36]. TS may thus involve a heightened immune response and a hyperdopaminergic state in the CSTS [35, 36].

TMS for TS has often targeted the SMA, as the SMA connects to all 4 networks thought to play significant roles in TS [38]. The SMA may have excess activity in TS. In an open-label trial, 1 Hz stimulation of the left and right SMA for 15 daily sessions led to decreased tic severity [39]. Another open-label study targeting the SMA showed significant improvements in TS (as well as OCD symptoms) [40]. A randomized sham-controlled trial targeting the parietal lobe bilaterally showed significant improvement in terms of tics [41]. While TMS for TS has shown some encouraging results, there are limited data, and the optimal target and stimulation parameters are not yet clear.

OCD is a chronic disorder characterized by obsessions (unwanted, intrusive, and distressing thoughts) and compulsions (repetitive acts that the patient feels compelled to do) [42]. OCD can often be comorbid with TS and may share similarities in terms of etiology [32]. Similar to TS, with OCD, there may be a hyperactive cortico-striato-thalamic-cortical circuit. There can be an inflammatory component to OCD. Indeed, OCD has been shown to occur after von Economo’s encephalitis [43], with insults to the basal ganglia (particularly the globus pallidus and caudate) leading to OCD symptoms [43].

In a study of mothers of children with OCD and tics, mothers had a greater number of proinflammatory conditions than mothers of neurologically autoimmune and healthy controls [44]. Inflammatory cytokines such as interferon-alpha were elevated in this population. There are also data that suggest microglia and astrocytes produce high levels of pro-inflammatory cytokines in OCD [44], particularly IL-6, TNF-alpha, and IL-8 [45]. Further, there may be an increase in monocytes and microglial dysregulation [45], with microglial dysregulation leading to alterations of synaptic pruning and affecting neuronal plasticity.

TMS studies for OCD have targeted various areas implicated in OCD, including the medial prefrontal cortex and anterior cingulate, SMA, DLPFC, and orbitofrontal cortex [4]. 18 studies used a figure-of-8 coil with various targets such as the SMA and DLPFC, and per a metanalysis of these studies, there was significant improvement in OCD symptoms with TMS [46].

The H-coil, which is a different coil from the standard figure-of-eight TMS coil, has been used to target the anterior cingulate and medial prefrontal cortex in double-blind sham-controlled studies. In a study using 20 Hz stimulation with 5 days/week treatment for 5 weeks as well as exposure response prevention, active treatment showed significant improvement compared to sham [47]. Another double-blind sham-controlled trial targeting the same areas with the same protocol showed significant benefit from active versus sham TMS [48]. These studies led to FDA clearance for the H-coil for OCD. These studies were in adults only, and thus the FDA clearance is not for children.

TMS is a technology that may be well-suited for developmental neuropsychiatric disorders where inflammation may play a role. TMS may work by modulating synaptic plasticity and hence change network connectivity, potentially countering the deleterious effects of inflammation on plasticity and connectivity. This may be particularly important in the developing brain, where the trajectory of the illness may be able to be more modifiable, given that children and adolescents have greater inherent plasticity capacity than adults.

Clinical studies have shown some benefit on multiple sclerosis, OCD, TS, MDD, and autism. However, it is still early in the field, and optimal TMS treatment parameters are not yet clear. The target location, frequency, number of pulses per session, number of sessions, interval between trains of pulses, durability of treatment, and need for retreatment or maintenance are all factors that need to be further studied.

Major depression and OCD are FDA-cleared indications for TMS, with the greatest clarity so far in terms of target and treatment protocol. However, even in these conditions, it is not yet clear if children or adolescents may have a different outcome or need modifications of the protocol for optimal outcomes. Indeed, a large randomized controlled study of adolescent depression using parameters from adult trials did not separate from sham treatment [49].

Neuronavigation and, in particular, using functional connectivity MRI (fcMRI) to individualize target selection may be an important consideration in enhancing outcomes. fcMRI-guided TMS with theta burst was recently shown to achieve more than a 78% remission rate in adults with treatment-resistant major depression [50]. Similar future studies in the developing brain may show enhanced results as well.

The optimal frequency of stimulation is another area that is ripe for further discovery. Currently, for MDD and OCD, studies have typically used frequencies of 10–20 Hz. Theta burst, which is now FDA-cleared for MDD, utilizes 50 Hz triplet bursts, with the 3 bursts repeating every 200 ms [51]. The treatment length is only around 3 min per session (as opposed to 18–20 min with conventional 10 Hz stimulation). It is not clear yet if theta burst, based as it is on the firing pattern in the hippocampus, an area of great plasticity, will show increased efficacy for brain disorders. However, this is an area of significant research and optimism.

Overall, TMS is a therapeutic technology platform. If the network connectivity dysfunction is known in any particular brain condition, then TMS could at least in theory be used to modulate and improve the dysfunction. TMS may in the future play a role with neuroinflammatory disorders, as there is evidence that inflammation affects plasticity, and this can be improved with TMS. Further, TMS may have direct anti-inflammatory effects. TMS may in the future be a useful addition to our armamentarium for neuropsychiatric disorders that involve inflammation, likely as an adjunct to more traditional treatments.

The author would like to acknowledge Milly Skiles for helping with this manuscript.

The author has had research supported by the National Institutes of Health, the US Department of Defense, Otsuka, Akili Interactive, Clarigent, Janssen, Lumos, and Brainsway. The author has been on the Speaker’s Bureau of Brainsway and is on the Advisory Board for Otsuka.

No funding sources were utilized for this review.

The author takes full responsibility for the contents of this paper and was the only contributor.