Transcranial Magnetic Stimulation (TMS): rTMS, iTBS, and Deep TMS — Protocols, Mechanisms, Evidence, and Clinical Outcomes

Transcranial magnetic stimulation (TMS) represents one of the most significant advances in non-invasive brain stimulation to emerge over the past three decades. First introduced by Anthony Barker and colleagues in 1985 at the University of Sheffield, TMS uses rapidly changing magnetic fields — generated by a coil placed against the scalp — to induce electric currents in underlying cortical tissue. These induced currents are sufficient to depolarize neurons, and when applied in repeated pulses (repetitive TMS, or rTMS), they can produce lasting changes in cortical excitability that outlast the stimulation period itself.

The clinical significance of TMS has grown substantially since the U.S. FDA cleared the first rTMS device (NeuroStar) for treatment-resistant depression (TRD) in 2008. Subsequent clearances have expanded approved indications to include obsessive-compulsive disorder (OCD, 2018), smoking cessation (2020), and anxious depression (2021). As of 2024, major professional organizations including the American Psychiatric Association (APA), the Canadian Network for Mood and Anxiety Treatments (CANMAT), and the National Institute for Health and Care Excellence (NICE) recognize TMS as an evidence-based treatment for major depressive disorder (MDD), particularly after failure of at least one adequate antidepressant trial.

This article provides an in-depth clinical review of TMS protocols — including conventional repetitive TMS (rTMS), intermittent theta-burst stimulation (iTBS), and deep TMS (dTMS) — covering their neurobiological mechanisms, comparative effectiveness, prognostic predictors, and evolving research frontiers. The scope and specificity are intended for clinicians, trainees, and informed patients seeking a rigorous understanding of the current evidence base.

TMS operates on the principle of electromagnetic induction described by Faraday's law. A brief, high-amplitude current pulse (typically 3,000–5,000 amperes lasting ~200 microseconds) passes through a coil, generating a rapidly time-varying magnetic field of approximately 1.5–2.0 Tesla at the coil surface. This magnetic field passes through the skull without attenuation and induces an electric field in the underlying cortex, primarily affecting neurons oriented parallel to the cortical surface — principally interneurons and the horizontally running axonal fibers in cortical layers II and III.

Synaptic Plasticity: LTP and LTD Analogues

The lasting clinical effects of rTMS are believed to arise through mechanisms analogous to long-term potentiation (LTP) and long-term depression (LTD) — the canonical forms of synaptic plasticity. High-frequency stimulation (≥5 Hz) tends to increase cortical excitability at the stimulation site, an effect resembling LTP, while low-frequency stimulation (≤1 Hz) reduces excitability, resembling LTD. These effects are NMDA receptor-dependent, as demonstrated by studies showing that the NMDA antagonist memantine blocks the aftereffects of theta-burst stimulation (Huang et al., 2007). This NMDA dependence directly parallels the molecular mechanisms of hippocampal LTP.

Neurotransmitter Systems

TMS exerts measurable effects across multiple neurotransmitter systems:

• Dopamine: PET studies have demonstrated that rTMS applied to the left dorsolateral prefrontal cortex (DLPFC) increases dopamine release in the ipsilateral caudate nucleus and the anterior cingulate cortex (ACC). Strafella et al. (2001) showed approximately 7–10% reductions in [¹¹C]raclopride binding in the caudate following left prefrontal rTMS, indicating increased endogenous dopamine release.
• Serotonin: Animal studies show that repeated rTMS sessions increase 5-HT_1A receptor sensitivity in the dorsal raphe nucleus and elevate extracellular serotonin in the hippocampus and frontal cortex. These effects parallel those seen with chronic antidepressant administration.
• GABA and Glutamate: Magnetic resonance spectroscopy (MRS) studies reveal that rTMS alters the balance of cortical GABA and glutamate concentrations at the stimulation site. High-frequency rTMS to the DLPFC has been associated with increased glutamate/glutamine ratios, while low-frequency stimulation may increase GABAergic tone.
• Brain-Derived Neurotrophic Factor (BDNF): Serum BDNF levels increase following a course of rTMS in depressed patients who respond to treatment. The BDNF Val66Met polymorphism (rs6265) has been studied as a potential predictor of TMS response, with Met carriers showing reduced LTP-like plasticity in motor cortex studies, though clinical prediction data remain inconsistent.

Circuit-Level Effects

The therapeutic mechanism of TMS for depression is understood primarily through a circuit-level framework. The left DLPFC — the standard target — is hypoactive in MDD, while the subgenual anterior cingulate cortex (sgACC, Brodmann area 25) is hyperactive. The DLPFC and sgACC are anticorrelated within the default mode network (DMN) and frontoparietal control network. High-frequency stimulation of the left DLPFC is hypothesized to increase prefrontal activity and, through functional connectivity, suppress the pathologically hyperactive sgACC. This model was substantively validated by a landmark 2021 study from the Stanford group (Cole et al., published in American Journal of Psychiatry), which demonstrated that clinical response to TMS correlated with the degree of functional connectivity between the individual's DLPFC stimulation site and the sgACC.

Additional circuit effects include modulation of the cortico-striato-thalamo-cortical (CSTC) loop — particularly relevant for OCD applications where the medial prefrontal cortex and orbitofrontal cortex are targeted — and normalization of amygdala hyperreactivity in anxious depression.

Conventional High-Frequency rTMS (HF-rTMS)

The most widely studied TMS protocol for depression applies high-frequency (10 Hz) repetitive pulses to the left DLPFC at 120% of the resting motor threshold (rMT). A standard FDA-cleared protocol delivers 3,000 pulses per session (75 trains of 40 pulses, with 26-second inter-train intervals), over approximately 37.5 minutes. The standard acute course consists of 5 sessions per week for 4–6 weeks (20–30 total sessions), often followed by a taper period of 1–2 sessions per week for several additional weeks.

Low-frequency (1 Hz) rTMS to the right DLPFC is an alternative protocol based on the interhemispheric imbalance model — where right prefrontal hyperactivity is suppressed. Meta-analytic data suggest modestly lower effect sizes for right-sided 1 Hz protocols compared to left-sided 10 Hz protocols, though some patients who fail one laterality may respond to the other.

Intermittent Theta-Burst Stimulation (iTBS)

Theta-burst stimulation (TBS) mimics endogenous hippocampal theta rhythms. In iTBS, bursts of 3 pulses at 50 Hz are repeated at 5 Hz (theta frequency). The "intermittent" pattern delivers 2-second trains with 8-second inter-train intervals, totaling 600 pulses over approximately 3 minutes and 9 seconds. This dramatically shorter session time was validated as noninferior to conventional 10 Hz rTMS in the landmark THREE-D trial (Blumberger et al., 2018), a multisite, randomized, noninferiority study of 414 patients. Response rates were 49% for iTBS versus 47% for 10 Hz rTMS, and remission rates were 32% for both groups — meeting the prespecified noninferiority margin. iTBS received FDA clearance in 2018 and has substantially improved clinical throughput and patient access.

Stanford Neuromodulation Therapy (SNT) / Stanford Accelerated Intelligent Neuromodulation Therapy (SAINT)

An intensive accelerated iTBS protocol developed at Stanford delivers 10 sessions per day (1,800 pulses per session, 18,000 pulses/day) over 5 consecutive days, totaling 90,000 pulses across the full course. Target selection is individualized using functional connectivity MRI (fcMRI) to identify the DLPFC coordinate most anticorrelated with the patient's sgACC. In the initial open-label SAINT study (Cole et al., 2020), remission rates were 90.5% (19/21 participants). A subsequent double-blind, sham-controlled RCT (Cole et al., 2022, published in American Journal of Psychiatry) demonstrated a 78.6% remission rate for active treatment versus 15.4% for sham (p < 0.001). The FDA cleared this protocol (marketed as the Magnus system by Magnus Medical) in 2022. While these results are striking, they require replication at independent sites with larger, more diverse samples. The intensive nature of the protocol and the requirement for individualized fcMRI targeting limit widespread implementation at present.

Deep TMS (dTMS)

Deep TMS uses an H-coil (developed by Brainsway) designed to stimulate deeper and broader cortical and subcortical structures compared to the standard figure-8 coil. The H1 coil stimulates the left DLPFC at greater depth (~3–4 cm from the scalp surface, compared to ~1.5–2 cm for figure-8 coils). The standard dTMS depression protocol uses 18 Hz stimulation, delivering 1,980 pulses per session over approximately 20 minutes. The pivotal trial for dTMS in MDD (Levkovitz et al., 2015) was a multisite RCT of 212 patients demonstrating response rates of 38.4% for active versus 21.4% for sham (p = 0.013), and remission rates of 32.6% versus 14.6% (p = 0.008).

For OCD, the FDA-cleared dTMS protocol uses an H7 coil targeting the medial prefrontal cortex and anterior cingulate cortex at 20 Hz, preceded by a personalized symptom provocation procedure. The pivotal multicenter trial (Carmi et al., 2019) showed a response rate (≥30% reduction in Y-BOCS) of 38.1% for active versus 11.1% for sham (p = 0.003).

Bilateral and Sequential Protocols

Some clinicians use bilateral sequential protocols — typically high-frequency left DLPFC stimulation followed by low-frequency right DLPFC stimulation in the same session. A meta-analysis by Berlim et al. (2013) suggested that bilateral protocols may have slightly higher response rates than unilateral protocols, though head-to-head RCT data are limited.

Major Depressive Disorder — The Primary Indication

MDD affects approximately 8.3% of U.S. adults annually (~21 million people; NIMH, 2021 data) and has a lifetime prevalence of approximately 17–20%. Of these, roughly 30% meet criteria for treatment-resistant depression, typically defined as failure to respond adequately to at least two antidepressant trials of adequate dose and duration. This translates to approximately 6–7 million Americans with TRD — a population for which TMS has the strongest evidence base.

In the landmark STAR*D study (Sequenced Treatment Alternatives to Relieve Depression; Rush et al., 2006), cumulative remission rates fell sharply with each successive treatment step: ~37% after Step 1 (citalopram), ~31% after Step 2, ~14% after Step 3, and ~13% after Step 4. These diminishing returns underscore the clinical need for non-pharmacological approaches like TMS in treatment-resistant populations.

Obsessive-Compulsive Disorder

OCD has a lifetime prevalence of approximately 2–3% (DSM-5-TR). TMS for OCD targets the medial prefrontal cortex (mPFC) or the supplementary motor area (SMA) rather than the DLPFC. The evidence base is smaller than for depression, but dTMS with the H7 coil targeting mPFC/ACC is FDA-cleared. A 2019 Cochrane-style meta-analysis found a modest effect size (Hedges' g ≈ 0.5) for low-frequency rTMS over the SMA in OCD.

Other Emerging and Off-Label Indications

• Post-Traumatic Stress Disorder (PTSD): Several RCTs support right DLPFC low-frequency or bilateral rTMS for PTSD, with response rates in the 40–55% range. A meta-analysis by Kan et al. (2020) found a significant overall effect (g = 0.56).
• Generalized Anxiety Disorder: Limited but growing evidence, primarily when comorbid with MDD.
• Substance Use Disorders: The FDA cleared dTMS for smoking cessation (2020) following a multicenter RCT (Zangen et al., 2021) targeting the bilateral lateral prefrontal cortex and insula with the H4 coil. Emerging research targets alcohol, cocaine, and methamphetamine use disorders.
• Chronic Pain and Fibromyalgia: High-frequency rTMS to the primary motor cortex (M1) has Level A evidence (definite efficacy) in the European guidelines for neuropathic pain.
• Auditory Hallucinations in Schizophrenia: Low-frequency rTMS to the left temporoparietal junction shows modest efficacy with effect sizes around d = 0.4–0.5 in meta-analyses, though results are inconsistent.

Patient Selection and Contraindications

Absolute contraindications include ferromagnetic material in or near the head (e.g., cochlear implants, metallic hardware, shrapnel) due to the risk of movement or heating. Seizure risk is the most significant safety concern, occurring in approximately 0.01–0.1% of sessions with standard protocols when safety guidelines (Rossi et al., 2009, updated 2021) are followed. A personal or family history of epilepsy is a relative contraindication requiring careful risk-benefit analysis. TMS is generally safe in pregnancy and is being actively studied in perinatal depression, as it avoids systemic pharmacological exposure.

Conventional rTMS for MDD

A comprehensive 2017 network meta-analysis by Brunoni et al. (published in JAMA Psychiatry) pooled data from 81 RCTs (n = 4,233) and found that high-frequency left DLPFC rTMS was significantly superior to sham, with an odds ratio (OR) of approximately 3.3 for response compared to sham. The number needed to treat (NNT) for response with conventional left-sided 10 Hz rTMS in treatment-resistant populations typically ranges from 4 to 8, depending on the population and outcome definition.

Real-world effectiveness data from large observational registries provide important context. The NeuroStar Outcomes Registry, encompassing over 5,000 patients treated in routine clinical practice, reported that approximately 58% of patients met response criteria (≥50% reduction in PHQ-9) and 37% achieved remission (PHQ-9 ≤ 4) at the end of the acute treatment course. These real-world numbers tend to be modestly higher than pivotal RCT figures, likely because clinical practice allows protocol flexibility (dose optimization, extended courses) not available in fixed-protocol trials.

iTBS for MDD

The THREE-D trial established noninferiority of iTBS to 10 Hz rTMS, with nearly identical response (~49%) and remission (~32%) rates. This was confirmed by a subsequent individual patient data meta-analysis. The clinical significance of the 3-minute iTBS session (versus the 37-minute conventional session) is enormous for patient compliance and clinic throughput, effectively allowing a clinic to treat 3–4 times as many patients per treatment chair per day.

Deep TMS for MDD

The pivotal dTMS trial (Levkovitz et al., 2015) reported a response rate of 38.4% versus 21.4% for sham, yielding an NNT of approximately 5.9 for response and 5.6 for remission. Effect sizes in meta-analyses comparing dTMS to sham for depression are in the moderate range (Cohen's d ≈ 0.5–0.7). There is no definitive evidence from head-to-head trials that dTMS is superior to standard figure-8 coil rTMS for depression, though the deeper stimulation field is theoretically advantageous for certain targets and conditions.

Stanford Accelerated Protocol (SAINT/SNT)

The double-blind RCT results — 78.6% remission for active versus 15.4% for sham — represent by far the most impressive acute outcomes in the TMS literature. However, several caveats apply: the sample size was small (n = 29), participants had moderate (not extreme) treatment resistance, and the protocol required individualized fcMRI targeting. Durability data showed that approximately 60% of initial remitters maintained remission at 4 weeks post-treatment without maintenance sessions. Larger, multisite replication trials are ongoing.

Durability and Relapse

Naturalistic follow-up studies suggest that the durability of TMS response is comparable to that of antidepressant pharmacotherapy. Approximately 50–70% of acute responders maintain their response at 6 months, and 40–50% at 12 months. The Dunner et al. (2014) 12-month follow-up study reported that 67.7% of acute responders maintained their response at 12 months, with many receiving intermittent reintroduction ("booster") sessions as needed. Predictors of relapse include greater number of prior failed antidepressant trials, comorbid anxiety, and younger age at onset.

TMS versus Pharmacotherapy

Direct head-to-head comparisons between TMS and antidepressant medications are surprisingly limited. The most notable is a study by Schutter (2009) whose meta-analysis estimated that the effect size of left-sided HF-rTMS for depression (d ≈ 0.55) is roughly comparable to that of antidepressant medications over placebo in meta-analyses (d ≈ 0.3–0.5 in intent-to-treat analyses). However, it is critical to note that TMS studies typically enroll patients who have already failed antidepressants, making the populations and comparisons nonequivalent. A more recent network meta-analysis by Mutz et al. (2019, published in The Lancet Psychiatry) included 113 trials and found that bilateral rTMS and HF left rTMS were among the most effective non-invasive brain stimulation modalities, with effect sizes comparable to pharmacological augmentation strategies.

TMS versus Electroconvulsive Therapy (ECT)

ECT remains the most effective acute treatment for severe, treatment-resistant depression, with response rates of 60–80% and remission rates of 50–60%. TMS is generally considered less effective than ECT for the most severely ill and treatment-resistant patients, but it has significant advantages in tolerability: no need for general anesthesia, no cognitive side effects (particularly no retrograde amnesia), and suitability for outpatient administration. A meta-analysis by Ren et al. (2014) confirmed ECT's superiority over rTMS in direct comparisons, but effect size differences were moderate (d ≈ 0.3–0.5). The Stanford accelerated protocol may narrow this gap, but direct head-to-head comparisons with ECT have not yet been completed.

TMS versus Esketamine (Spravato)

No published head-to-head trials compare TMS to intranasal esketamine. Both are positioned for treatment-resistant depression after pharmacotherapy failure. Esketamine's pivotal trials showed response rates of approximately 65–70% and remission rates of 36–38% in acute-phase studies, but these were conducted as adjuncts to new oral antidepressants, complicating comparison. The cost profiles differ substantially: a full course of TMS ($6,000–$15,000 in the U.S.) is typically a one-time investment, while esketamine requires ongoing twice-weekly to monthly maintenance sessions. Insurance coverage for both remains variable.

TMS versus Psychotherapy

No large RCTs have directly compared TMS to evidence-based psychotherapies (e.g., CBT) for treatment-resistant depression. Some evidence supports combination approaches — TMS paired with concurrent CBT or behavioral activation — based on the theoretical premise that TMS-induced neuroplasticity may enhance the brain's capacity to benefit from psychotherapeutic learning. Several pilot studies suggest additive effects, but this remains an active area of investigation.

Identifying reliable predictors of TMS response is a major clinical priority, given that approximately 40–50% of patients in standard protocols do not achieve meaningful improvement. Research has identified several demographic, clinical, neurobiological, and genetic factors that influence outcomes.

Clinical Predictors

• Degree of treatment resistance: This is the single most robust predictor. Patients who have failed fewer prior antidepressant trials respond at significantly higher rates. In the NeuroStar pivotal trial, patients who had failed only one adequate antidepressant had response rates approximately double those who had failed four or more.
• Episode duration: Shorter duration of the current depressive episode predicts better outcomes. Chronic episodes (>2 years) are associated with reduced response rates.
• Age: Some analyses suggest younger patients may respond somewhat better, though findings are inconsistent. Older adults (>65) may have reduced cortical excitability, but TMS remains effective in geriatric populations with appropriate dosing.
• Comorbid anxiety: Anxious depression (MDD with prominent anxiety features) was specifically included in FDA labeling for the Brainsway dTMS device. Some evidence suggests that comorbid anxiety does not reduce — and may slightly enhance — TMS response rates, particularly with bilateral or deep protocols.
• Comorbid personality disorders: Presence of borderline or other cluster B personality disorders has been associated with reduced response rates in some but not all studies.

Neurobiological and Neuroimaging Predictors

• Baseline DLPFC-sgACC functional connectivity: Stronger negative (anticorrelated) connectivity between the stimulation site in the DLPFC and the sgACC at baseline predicts better clinical response. This finding, replicated across multiple studies, supports the circuit-based rationale for TMS and the potential for connectivity-guided targeting.
• Cortical thickness: Greater prefrontal cortical thickness at baseline has been associated with improved outcomes in structural MRI studies, suggesting that more preserved cortical architecture facilitates the TMS response.
• EEG biomarkers: Higher baseline frontal theta power (6–8 Hz) on quantitative EEG has been identified as a positive predictor of TMS response in several studies. This biomarker may reflect preserved anterior cingulate function.

Genetic Factors

Pharmacogenomic predictors of TMS response are not yet clinically actionable, but several candidate gene findings have emerged. The BDNF Val66Met polymorphism has been associated with variable TMS-induced plasticity, with Val/Val homozygotes showing more robust motor cortex LTP-like effects. The serotonin transporter gene-linked polymorphic region (5-HTTLPR) has been studied with mixed results. The COMT Val158Met polymorphism, which influences prefrontal dopamine metabolism, has shown preliminary associations with differential TMS outcomes. However, none of these genetic markers have sufficient sensitivity or specificity for routine clinical use at present.

What Predicts Poor Outcomes

Factors consistently associated with reduced TMS efficacy include: high degree of treatment resistance (≥4 failed adequate trials), very long episode duration (>3 years), active substance use disorder, severe insomnia that remains untreated, and benzodiazepine use (which may attenuate LTP-like plasticity through GABAergic enhancement). Anticonvulsant medications, similarly, may theoretically reduce TMS efficacy, though clinical evidence is limited.

In clinical practice, patients referred for TMS rarely present with uncomplicated, single-diagnosis MDD. Understanding how common comorbidities affect TMS outcomes is essential for informed treatment planning.

Anxiety Disorders

Comorbid anxiety disorders occur in approximately 50–60% of MDD patients. Generalized anxiety disorder is the most common co-occurring condition, followed by social anxiety disorder and panic disorder. Several studies suggest that TMS for depression concurrently improves anxiety symptoms, even when anxiety is not the primary target. In the THREE-D trial, both iTBS and 10 Hz rTMS produced significant reductions in GAD-7 scores. Specific anxiety reduction may be partly mediated by amygdala deactivation downstream from DLPFC stimulation.

Post-Traumatic Stress Disorder

PTSD comorbidity is present in approximately 20–35% of TRD patients, depending on the population. Emerging evidence supports TMS targeting for PTSD-specific symptoms, including right DLPFC stimulation to reduce hyperarousal and re-experiencing symptoms. The presence of PTSD comorbidity does not appear to substantially reduce TMS efficacy for the depressive component.

Substance Use Disorders

Comorbid substance use disorders are present in approximately 20–30% of patients with MDD and pose challenges for TMS treatment. Active alcohol or benzodiazepine dependence increases seizure risk and may require stabilization before initiating TMS. However, TMS applied to the DLPFC has shown preliminary efficacy in reducing craving for alcohol, nicotine, cocaine, and methamphetamine — suggesting potential dual-benefit applications. The FDA clearance for smoking cessation using dTMS with the H4 coil (targeting bilateral prefrontal cortex and insula) supports this emerging indication.

Bipolar Depression

The use of TMS in bipolar depression remains an area of evolving evidence. DSM-5-TR distinguishes bipolar I and II disorder from MDD, and most pivotal TMS trials excluded bipolar patients. However, several open-label studies and small RCTs suggest comparable efficacy in bipolar II depression, with low rates of treatment-emergent mania or hypomania (<1–2%). Given these data, some clinicians use TMS off-label for bipolar II depression, typically with mood stabilizer coverage. TMS for bipolar I depression requires greater caution due to higher theoretical switch risk, though empirical evidence of TMS-induced mania is sparse.

Neurocognitive Disorders

Emerging research explores TMS for cognitive symptoms in MDD, mild cognitive impairment (MCI), and early Alzheimer's disease. High-frequency DLPFC rTMS has shown modest improvements in working memory and executive function in depressed patients, and several trials are investigating multi-site TMS protocols for Alzheimer's disease using the nBrainX or other navigated systems.

Common Side Effects

TMS is well-tolerated overall, with a highly favorable side-effect profile compared to antidepressant medication or ECT. The most common adverse effects include:

• Scalp discomfort or pain at the stimulation site: Reported by 30–50% of patients, typically mild, and tending to diminish over the first week of treatment. This is caused by stimulation of the superficial scalp muscles and nerves.
• Headache: Mild, tension-type headache occurs in approximately 20–30% of patients and is usually responsive to over-the-counter analgesics.
• Transient lightheadedness: Uncommon, occurring in <5% of patients.

Serious Adverse Events

The most serious risk is seizure, which occurs at a rate of approximately 0.01–0.1% per course of treatment when established safety parameters (pulse frequency, intensity, train duration, inter-train interval) are followed according to the Rossi et al. (2009, updated 2021) international safety guidelines. Seizure risk is modestly elevated by concurrent use of medications that lower seizure threshold (e.g., bupropion, clozapine, theophylline), sleep deprivation, and alcohol withdrawal. No fatalities attributable to TMS have been reported.

Cognitive Safety

Unlike ECT, TMS does not impair memory or cognitive function. In fact, meta-analyses of neuropsychological outcomes show either no change or modest improvement in executive function and processing speed following a course of rTMS for depression — likely reflecting improvement in depression-related cognitive impairment rather than a direct cognitive-enhancing effect of stimulation.

Practical Administration

Motor threshold determination is performed at the beginning of treatment (and periodically rechecked) by stimulating the primary motor cortex to identify the minimum intensity required to produce a visible thumb twitch in the contralateral abductor pollicis brevis in 5 out of 10 trials. Treatment intensity is then set as a percentage of this threshold (typically 120%). DLPFC targeting is performed using either the "5 cm rule" (measuring 5 cm anterior to the motor cortex hotspot along the scalp — an older, less precise method), the Beam F3 scalp-based localization system, or neuronavigation using structural MRI. Neuronavigation and fcMRI-based targeting represent the most anatomically precise approaches but are not universally available.

Accelerated Protocols

The Stanford SAINT protocol has generated enormous interest in accelerated delivery schedules. Multiple groups worldwide are now testing variations: some using iTBS without individualized fcMRI targeting, some using 5 sessions per day rather than 10, and some testing whether similar outcomes can be achieved with simplified targeting methods. If the core findings replicate — that intensive, brief courses can achieve rapid remission — this could fundamentally transform TMS from a 6-week treatment to a 5-day intervention.

Biomarker-Guided Targeting and Personalized Protocols

The most impactful research frontier is the development of biomarker-guided treatment personalization. This includes: (1) using resting-state fcMRI to identify individualized optimal stimulation targets based on each patient's unique connectivity map; (2) using EEG-based measures (e.g., frontal theta cordance, alpha asymmetry) to predict response before treatment begins; and (3) developing closed-loop TMS systems that adjust stimulation parameters in real time based on concurrent EEG or fMRI feedback.

Novel Targets and Indications

Research is expanding beyond the DLPFC to include: the dorsomedial prefrontal cortex (dmPFC) for depression with rumination; the right orbitofrontal cortex for depression with anhedonia; the posterior parietal cortex for attention and working memory; and the cerebellum for affective dysregulation. Dual-site and multi-site stimulation protocols are also under investigation.

Combination Therapies

The combination of TMS with concurrent psychotherapy (particularly CBT), pharmacotherapy augmentation, or even concurrent psychedelic-assisted therapy represents a convergent research area based on the premise that TMS-induced neuroplasticity creates a window of enhanced learning and synaptic reorganization.

Limitations of Current Evidence

Several important limitations merit emphasis:

• Sham condition challenges: Active TMS produces scalp sensation and clicking sounds that are difficult to perfectly replicate in sham conditions. Although modern sham coils include electrical scalp stimulation to mimic the sensation, blinding integrity remains imperfect and may inflate effect sizes in RCTs.
• Publication bias: Meta-analyses have detected evidence of publication bias favoring positive TMS studies, particularly in smaller trials.
• Heterogeneity of protocols: The proliferation of different coils, targets, frequencies, and dosing schedules makes cross-study comparison challenging and complicates the development of definitive treatment guidelines.
• Limited diversity in study populations: Most pivotal TMS trials were conducted in predominantly White, English-speaking populations. Data on efficacy across racial, ethnic, and socioeconomic groups remain insufficient.
• Cost and access: Despite growing insurance coverage, TMS remains inaccessible to many patients due to cost ($6,000–$15,000 per course in the U.S.), geographic availability, and the time commitment of daily sessions over several weeks.

Based on current evidence and guideline recommendations, TMS occupies a well-defined position in the depression treatment algorithm:

• First-line treatment: Pharmacotherapy and/or evidence-based psychotherapy (CBT, IPT, behavioral activation) remain first-line for MDD per APA, CANMAT, and NICE guidelines.
• Second-line or augmentation: After failure of 1–2 adequate antidepressant trials, TMS is a guideline-endorsed option alongside pharmacological augmentation strategies (e.g., lithium, atypical antipsychotics, thyroid augmentation). CANMAT guidelines (2016, updated 2023) rate rTMS as a Level 1 (highest evidence) recommendation for MDD after pharmacotherapy failure.
• Before ECT: For patients with moderate (not immediately life-threatening) treatment-resistant depression, TMS is typically tried before ECT due to its superior tolerability and absence of cognitive side effects. ECT remains preferred for severe, imminently dangerous depression with acute suicidality, psychotic features, or catatonia.
• Concurrent with pharmacotherapy: Most patients continue their antidepressant medication during TMS treatment. There is no strong evidence that any specific antidepressant class enhances or diminishes TMS efficacy, though clinical caution is warranted with medications that lower seizure threshold.

For OCD, TMS (specifically dTMS with the H7 coil) is positioned after failure of adequate trials of SSRIs and exposure-response prevention (ERP) therapy. For smoking cessation, TMS is an option when first-line treatments (counseling, nicotine replacement, varenicline, bupropion) have failed or are contraindicated.

The emergence of accelerated protocols like SAINT/SNT may shift TMS earlier in the treatment algorithm if their remarkable acute outcomes are confirmed in larger, independent samples. A 5-day intensive protocol with 80%+ remission rates, if validated, would represent a paradigm shift in depression treatment.

Transcranial magnetic stimulation has matured from an experimental neuroscience technique to an established, FDA-cleared treatment modality with a robust evidence base, particularly for treatment-resistant depression. The core therapeutic mechanism — modulation of prefrontal-limbic circuits through electromagnetic induction of cortical neuroplasticity — is supported by convergent evidence from neuroimaging, neurochemical, and electrophysiological studies.

Clinicians now have several validated TMS protocol options: conventional 10 Hz rTMS (the most extensively studied), iTBS (equivalent efficacy in ~3 minutes versus ~37 minutes), deep TMS (broader stimulation field for deeper targets), and emerging accelerated protocols that compress treatment into days rather than weeks. Response rates across these modalities range from approximately 40–60% in treatment-resistant populations, with NNTs of 4–8 — figures that compare favorably with pharmacological alternatives in comparable populations.

The field is advancing rapidly toward personalized, biomarker-guided treatment. Functional connectivity-based targeting, EEG-predicted response phenotypes, and accelerated intensive protocols represent the most promising near-term developments. As access improves, evidence accumulates, and targeting precision increases, TMS is likely to assume an increasingly central role in psychiatric treatment algorithms — not only for depression, but potentially for a broad range of circuit-based psychiatric and neurological conditions.