Transcranial magnetic stimulation (TMS)

Transcranial magnetic stimulation (TMS) is a neurotechnological practice developed by Anthony Barker and colleagues in Sheffield in 1985. It is among the most extensively validated non-invasive brain stimulation methods, with FDA clearance for treatment-resistant depression, obsessive-compulsive disorder, and smoking cessation. It works primarily through focal electromagnetic induction — brief magnetic pulses that depolarize neurons in targeted cortical regions.

The patient sits in a treatment chair while a clinician positions an electromagnetic coil against the scalp over a predetermined cortical target, most commonly the left dorsolateral prefrontal cortex. The device delivers patterned magnetic pulses calibrated to a fraction of the individual's motor threshold. A standard session runs 19 to 37 minutes; theta burst protocols compress to 3 minutes.

Sessions are administered five days per week for four to six weeks. During stimulation, patients feel a tapping sensation at the coil site and hear a loud clicking, mitigated by earplugs. Common acute experiences include scalp discomfort, mild headache, and transient facial muscle twitching. The procedure produces no anesthesia, no sedation, and no immediate alteration of consciousness — therapeutic effects emerge cumulatively over the treatment course.

The principal serious risk is seizure, with an estimated incidence near 1 in 30,000 sessions in screened populations.^[1] Personal or family history of epilepsy, intracranial metallic implants, and ferromagnetic devices in the head are absolute or relative contraindications. TMS is the most extensively studied non-pharmacological intervention for treatment-resistant depression — meta-analyses of dozens of sham-controlled trials establish efficacy,^[2] and FDA-cleared protocols are now reimbursed by major US insurers.

Origins

TMS rests on electromagnetic induction, characterized by Michael Faraday in 1831: a time-varying magnetic field induces an electric current in a nearby conductor. The principle sat largely unused for neurological application for over a century. Earlier non-invasive brain stimulation methods relied on scalp electrodes (transcranial direct and alternating current), which were limited by skin impedance and discomfort, and could not achieve focal cortical activation.

Anthony T. Barker, Reza Jalinous, and Ian L. Freeston at the Royal Hallamshire Hospital in Sheffield, United Kingdom, published the first demonstration of TMS in The Lancet in 1985.^[3] A purpose-built coil delivered a brief magnetic pulse over the motor cortex and produced contralateral hand muscle contractions, with a peak magnetic field of approximately 2 Tesla sustained for roughly 100 microseconds. Through the early 1990s, TMS served primarily as a research tool — mapping motor cortex representation, measuring corticospinal conduction, and probing cortical excitability in healthy subjects and patients.

Transition into psychiatry

The pivot from neurophysiology to psychiatric therapy came from Mark George, Eric Wassermann, and colleagues at the National Institutes of Health and the Medical University of South Carolina. George et al. published a small open-label pilot in 1995 reporting that daily left prefrontal rTMS produced mood elevation in patients with depression.^[4] The hypothesis was that left dorsolateral prefrontal cortex (DLPFC) was hypoactive in depression and that excitatory high-frequency stimulation could reverse this.

In 2007, O'Reardon and colleagues published a multicenter, randomized, sham-controlled trial of the NeuroStar device in 301 medication-free patients with major depressive disorder who had failed at least one antidepressant trial.^[5] Response rates were 24.4% (active) versus 15.2% (sham); remission 14.2% versus 5.5%. An NIH-funded independent replication followed in 2010.^[6] The FDA cleared the NeuroStar system under 510(k) in October 2008 for adults with depression who had failed one adequate antidepressant trial in the current episode.^[7]

Contemporary context

The acute treatment protocol diversified through the 2010s and 2020s. Theta burst stimulation (TBS), introduced by Huang et al. in 2005, compressed the active stimulation window from 37 to 3 minutes per session;^[8] the THREE-D trial in 2018 established non-inferiority of intermittent TBS to standard 10 Hz rTMS.^[9] BrainsWay's H-coil deep TMS received FDA clearance for OCD in 2018^[10] and for smoking cessation in 2020. In 2022, the FDA cleared the Stanford accelerated theta burst (SAINT/SNT) protocol developed at Stanford,^[11] compressing a full course into five days. NeuroStar received clearance for adolescents ages 15-21 in 2023.

TMS is delivered exclusively in clinical settings under physician supervision. Principal device manufacturers include Neuronetics (NeuroStar), BrainsWay, MagVenture, Magstim, and Magnus Medical. The procedure is reimbursed by most major US insurers for depression after documented antidepressant failure; access remains concentrated in urban specialty centers, and a typical US course costs $6,000 to $15,000.

Technique

Setting & Preparation

TMS is delivered exclusively in clinical settings — outpatient psychiatric clinics, hospital-based neuromodulation centers, or specialized research facilities. The treatment room contains the TMS console, the stimulating coil on an articulated arm, a treatment chair, and resuscitation equipment with personnel trained in basic life support for the rare possibility of a seizure.^[1]

Required equipment includes the TMS device (e.g., NeuroStar, BrainsWay, MagVenture, Magstim, Magnus), a stimulation coil matched to the target (figure-of-eight for cortical targets, H-coil for deep targets), surface EMG electrodes for motor threshold determination at initiation, and disposable foam earplugs for hearing protection. For individualized targeting (SAINT/SNT), pre-treatment structural and resting-state fMRI scans are required.

The session is administered by a trained operator with supervised TMS experience and current basic life support certification; a prescribing physician oversees indications, contraindication review, and motor threshold determination at initiation.^[1] The patient does not need to prepare in any specific way and can return to normal activities immediately after the session.

TMS is delivered only in clinical settings and only by trained operators under physician oversight, so harm reduction at the patient level focuses on pre-treatment screening rather than session-level environment control. The patient's principal contribution is honest disclosure during the pre-treatment evaluation: personal and family history of seizures, intracranial implants, current medications (especially seizure-threshold-lowering agents), alcohol and other substance use, sleep patterns in the days preceding sessions, and pregnancy status all bear directly on safety planning.^[1]

Pre-session, patients are advised to maintain regular sleep, avoid alcohol for at least 24 hours, and take prescribed medications as scheduled unless the prescribing physician has specified a hold. Earplugs should be confirmed in place before each session; the audible click can cause hearing threshold shifts with chronic unprotected exposure. The patient should report any new neurological symptoms — recent head injury, new headache pattern, unusual sensory experiences — to the treating clinician before each session, not only at the initial intake.

Coil-to-cortex induction

A TMS pulse generates a rapidly changing magnetic field at the coil, which induces an electric field in nearby tissue oriented perpendicular to the magnetic field. In cortical gray matter, this electric field depolarizes axons aligned parallel to the induced current, producing action potentials in superficial neurons within a focal volume of roughly 2-3 cm³ for figure-of-eight coils. The primary motor cortex (M1) is the standard target for dose calibration: a suprathreshold pulse over M1 elicits a contralateral motor evoked potential (MEP) measurable by EMG, and the motor threshold (MT) — the minimum intensity producing an MEP in 50% of trials — sets the per-patient dose for therapeutic protocols. The H-coil geometry developed by Zangen extends the effective depth of stimulation to 3-4 cm, accessing medial prefrontal and anterior cingulate targets.^[10]

LTP/LTD-like plasticity

The enduring effects of repetitive TMS outlast the stimulation period, implying synaptic plasticity rather than acute neuromodulation alone. High-frequency rTMS (≥5 Hz) produces facilitatory effects on cortical excitability consistent with long-term potentiation (LTP); low-frequency rTMS (≤1 Hz) produces inhibitory effects consistent with long-term depression (LTD). Theta burst stimulation, designed by Huang et al. to mimic the theta-gamma coupling patterns associated with LTP induction in hippocampal slice preparations, achieves comparable plasticity induction in roughly one-tenth the pulse count of standard 10 Hz protocols.^[8]

Network connectivity and the DLPFC-sgACC axis

The left DLPFC — the principal antidepressant target — does not directly regulate emotion. Its therapeutic relevance is understood through network connectivity. Fox et al. demonstrated in a meta-analysis of TMS-neuroimaging studies that the optimal DLPFC stimulation site is anticorrelated at rest with the subgenual anterior cingulate cortex (sgACC), a region consistently hyperactive in depression.^[12] Stimulating the DLPFC is hypothesized to propagate excitation through the network and secondarily suppress sgACC activity, normalizing the hyperactive limbic tone. Individualized fMRI-guided targeting in the SAINT/SNT protocol selects the patient-specific DLPFC voxel most anticorrelated with sgACC, with response rates substantially exceeding non-individualized protocols.^[11]

TMS-EEG and biomarker development

Concurrent TMS-EEG records cortical oscillatory responses to single pulses — TMS-evoked potentials (TEPs) — with components (N15, P30, N45, P60, N100, P180) reflecting distinct inhibitory and excitatory cortical circuits. Cortical silent period (CSP) and short-interval cortical inhibition (SICI) derived from TMS-EMG assess GABAergic inhibitory function.^[13] Multi-modal concurrent TMS-EEG-fMRI protocols have demonstrated that the brain's oscillatory state at the moment of pulse delivery — for example, alpha-band phase over the target — modulates the amplitude of the evoked cortical response, opening the door to closed-loop, state-dependent TMS paradigms.^[14] No TEP-based biomarker has yet been validated for clinical response prediction.

TMS does not produce acute alteration of consciousness during a session. The patient remains fully alert and oriented throughout administration; no sedation, dissociation, or perceptual change is intended or observed. Acute experience is dominated by physical sensation at the coil site — rhythmic tapping synchronized to the pulse train, audible clicking attenuated by earplugs, and incidental twitching of facial or scalp muscles when the induced field crosses peripheral nerves. Mild headache, scalp discomfort, and fatigue are common and typically resolve within hours of a session.^[1]

Therapeutic effects accumulate over a treatment course rather than emerging acutely. In depression protocols, mood improvement typically begins between sessions 10 and 20 of a 20-30 session course, with full response or remission generally assessed at course completion. In a meta-analysis of 29 sham-controlled trials of high-frequency rTMS for depression, pooled response was 29.3% versus 10.4% sham and remission 18.6% versus 5.0%, with a number needed to treat of 6 for response.^[2] The Stanford accelerated SAINT/SNT protocol compressed this timeline to one week with substantially higher remission in a small severe-treatment-resistant sample.^[11]

Cognitive effects during and after a treatment course are minimal. Unlike electroconvulsive therapy, rTMS has not been associated with significant anterograde or retrograde amnesia in controlled trials; some neuropsychological batteries detect transient attentional changes during a course without lasting deficits.^[1]

Adverse Effects

Other mild effects include neck pain from head positioning, post-session fatigue, transient lightheadedness, and rare vasovagal syncope. The audible click of the coil discharge approaches 120 dB at peak and necessitates earplug use throughout the session; documented hearing threshold shifts have occurred in patients and operators who did not use adequate ear protection.^[1]

Contraindications

Substance Interactions

The primary interaction concern with TMS is pharmacologically mediated reduction of seizure threshold. Bupropion lowers seizure threshold in a dose-dependent manner and is the most frequently cited interaction in TMS safety guidelines;^[1] it is not an absolute contraindication but requires documented risk assessment and, in higher-dose regimens, parameter adjustment. Tricyclic antidepressants show a similar dose-dependent reduction in seizure threshold. Clozapine carries a well-characterized seizure-threshold reduction and requires careful protocol planning. Stimulants (amphetamines, methylphenidate) elevate seizure risk and are assessed case by case.

Alcohol use around the treatment window elevates seizure risk through acute intoxication and withdrawal rebound, and through pharmacokinetic interactions with co-prescribed medications such as bupropion. Pre-treatment screening routinely includes alcohol consumption assessment; regular heavy use is a relative contraindication pending clinical risk review.^[1] Acute discontinuation of anticonvulsants or benzodiazepines substantially raises seizure risk and should be reviewed before each session if dosing has recently changed.

A naturalistic cohort study of 1,753 rTMS patients found that concomitant antidepressant use did not reduce clinical response and that no antidepressant class was associated with differential outcome.^[15] SSRIs, SNRIs, and MAOIs have not shown major adverse interactions with rTMS in published trials, though the pivotal O'Reardon 2007 trial was medication-free by design.^[5]

TMS itself is under active investigation as a treatment for substance use disorders, with small-to-moderate effect sizes for craving reduction reported in alcohol, nicotine, and cocaine indications.^[16][17] No FDA clearance has been issued for any substance use indication. For patients presenting for depression TMS with comorbid active substance use, the interaction concerns above apply with particular force.

Habituation

TMS carries no documented habituation, tolerance, or compulsive-use pattern. The procedure is administered exclusively in clinical settings under physician supervision; patients cannot self-administer it, and there is no recreational use pattern. No withdrawal syndrome has been characterized on discontinuation of a treatment course.

The relevant long-term clinical question is durability of response rather than dependence. Most randomized trials follow patients only to the end of the acute course or to roughly six months; the proportion of responders who relapse and require re-treatment over one to two years is not precisely characterized in controlled data.^[18] Continuation or maintenance TMS protocols — tapering from weekly to monthly sessions over three to six months — are offered by some centers, with evidence weaker than for the acute phase. Patients who relapse after an initial successful course often respond to a repeat acute course, though this pattern has not been studied in controlled trials with the rigor of first-course data.

If Distress Arises

TMS sessions can be paused or stopped at any time by the patient or the operator. The patient remains fully conscious and can communicate throughout. Mild scalp discomfort, tapping sensations, and facial twitching are expected and do not warrant stopping the session. Persistent or worsening headache, sudden severe pain at the coil site, lightheadedness, nausea, or new neurological symptoms (unilateral weakness, visual disturbance, slurred speech) are indications to stop and report.

In the very rare event of a seizure during a session, the operator stops stimulation immediately, positions the patient safely, and follows standard seizure-response protocol. All TMS-induced seizures reported in the published literature have been self-limited and have not produced permanent neurological sequelae.^[1] Resuscitation equipment and personnel trained in basic life support are required on site for this contingency.

If a session is missed, the standard practice is to resume the protocol without dose escalation; missed sessions do not require restarting the course. Patients who develop new manic symptoms — elevated mood, decreased sleep need, racing thoughts, impulsive behavior — should contact the prescribing physician before the next session.

TMS has no patient-side skill progression — the patient is passive during administration, and there is no published model of patient-acquired competence. Progression is defined by the therapeutic protocol structure and clinical response trajectory.

A standard acute course consists of 20 to 30 sessions over four to six weeks for depression.^[5] Pre-treatment includes clinical evaluation, contraindication screening, baseline symptom scoring, and motor threshold determination by EMG-guided titration. Symptom scales are re-administered every three to five sessions; non-responders at mid-course (session 15) are assessed for protocol adjustment. Response is operationally defined as ≥50% reduction in baseline depression score and remission as a score below an established threshold (e.g., PHQ-9 below 5, HAMD-17 below 7). Partial or full responders may be offered tapering "continuation TMS" from weekly to monthly sessions over three to six months, though evidence for maintenance is less robust than for the acute phase.^[18]

The accelerated SAINT/SNT protocol compresses the standard timeline to five days — 10 iTBS sessions per day separated by roughly 50-minute intervals, with individualized fMRI-guided targeting. In a double-blind sham-controlled trial of 52 patients with severe treatment-resistant depression, remission was 78.6% (active) versus 13.6% (sham) at one week, with the substantial caveats of small sample size, short follow-up, and concentration of severe-TRD patients.^[11] Operator training follows IFCN guidelines: supervised administration of at least 10 procedures before independent operation, demonstrated motor-threshold inter-rater reliability, and current basic life support certification.^[1]

The evidence base for rTMS in major depressive disorder is among the most developed in psychiatric neuromodulation. The IFCN evidence-based guidelines assign Level A (definite or probable efficacy) to left high-frequency rTMS over the DLPFC for depression, high-frequency M1 rTMS for neuropathic pain, and left DLPFC rTMS for post-stroke motor rehabilitation.^[18]

Depression

A meta-analysis of 29 double-blind sham-controlled RCTs (n=1,371) of high-frequency rTMS for depression reported pooled response 29.3% (active) versus 10.4% (sham) and remission 18.6% versus 5.0%, with an odds ratio of 3.3 for both endpoints and a number needed to treat of 6 for response and 8 for remission.^[2] A network meta-analysis of 37 RCTs (n=2,120) comparing parameter sets for treatment-resistant depression found high-frequency left DLPFC rTMS and iTBS most efficacious and comparable to each other.^[19] A 2023 meta-analysis of 19 RCTs in patients with ≥2 prior antidepressant failures reported a risk ratio of 2.25 for response and 2.78 for remission versus sham.^[20]

Theta burst and accelerated protocols

The THREE-D trial (n=414) established non-inferiority of intermittent TBS to standard 10 Hz rTMS, with response rates of 49% (iTBS) versus 46% (10 Hz).^[9] The SAINT/SNT double-blind sham-controlled RCT (n=52) reported remission of 78.6% (active) versus 13.6% (sham) at one week with individualized fMRI-guided accelerated iTBS, in a severe-treatment-resistant sample with short follow-up.^[11]

OCD and other indications

The OCD pivotal trial (n=99) of deep TMS targeting the anterior cingulate and medial prefrontal cortex reported response (Y-BOCS reduction ≥30%) of 38% (active) versus 11% (sham), forming the basis of FDA clearance.^[10] For auditory verbal hallucinations in schizophrenia, low-frequency rTMS over the left temporoparietal cortex has shown superiority to sham in small trials but is not FDA-cleared.^[21] In youth depression, a systematic review supports the 2023 FDA clearance for ages 15-21 with effect sizes broadly comparable to adults, though individual trials are small.^[22]

Limitations

Effect sizes in rTMS trials are moderate, and the clinical significance of response rates near 29% versus 10% sham must be interpreted against the treatment-resistant nature of most enrolled samples.^[2] In community practice with less treatment-resistant patients, response rates are typically higher; in severe treatment-resistant cohorts, they are lower.

Sham condition validity is an unresolved methodological challenge across the field. Standard sham coils produce scalp sensation and acoustic click, but blinding is imperfect — patients can sometimes distinguish active from sham, and the placebo response is correspondingly inflated. Active-sham coils with surface electrical stimulation address this partially. Most early positive trials were industry-sponsored, with independent NIH-funded replication (OPT-TMS) and large independent trials (THREE-D) providing critical corroboration.^[6][9]

Durability of response beyond the acute course and the optimal maintenance strategy are incompletely characterized; randomized trials rarely follow patients beyond six months, and continuation protocols vary widely across clinical practice without controlled-evidence support.^[18] No validated pretreatment biomarker reliably predicts response, and individualized targeting (as in SAINT/SNT) remains a research practice rather than a routine clinical adjustment. For non-depression indications, evidence ranges from FDA-cleared (OCD, smoking cessation, adolescent depression) to early-validation (PTSD, substance use disorders, schizophrenia hallucinations), with the latter category supported by small trials and exploratory meta-analyses rather than large pivotal RCTs.