Transcranial Magnetic Stimulation: How TMS Tests Brain Function
Author: Ukrainian Psychological Hub · Published: October 12, 2026 · Editorial Policy
Transcranial magnetic stimulation (TMS) is a technique that uses a rapidly changing magnetic field outside the head to induce an electric field in brain tissue. Neuroscientists use this controlled intervention to investigate cortical excitability, the timing of information processing, interactions between brain areas, and the contribution of neural circuits to behavior. Unlike a recording that observes signals while a person performs a task, TMS can change an ongoing neural process and test what happens next. The strength of the resulting scientific claim depends on experimental design, the distribution of stimulation, appropriate controls, and the measured outcome.
The central research question is often precise: If a particular cortical system is perturbed at a particular moment, does perception, movement, memory performance, or another behavioral measure change? A carefully controlled result supplies evidence that the stimulation-sensitive circuit participates in the task under those conditions. It does not establish that a single anatomical region exclusively produces the psychological capacity. Brain function depends on interactions between multiple regions and on the participant's current physiological and cognitive state, as analyzed by Hobot and colleagues.
This article explains how TMS works as an experimental method, what scientists actually measure, which causal inferences are justified, why findings can vary across studies, and how laboratory TMS differs from prescribed repetitive TMS treatment. The broader context is provided by our guide to brain imaging methods, which explains how observational measurements complement interventions.
What is transcranial magnetic stimulation?
TMS is a form of noninvasive brain stimulation in which an electromagnetic coil positioned near the scalp delivers one or more magnetic pulses. Faraday's principle of electromagnetic induction explains the central physics: a time-varying magnetic field generates an electric field in conducting tissue. Under suitable conditions, that induced electric field changes the membrane potentials of excitable neural elements and can initiate or influence neural activity. The coil does not inject electrical current through scalp electrodes, and the magnetic pulse does not itself record a brain signal.
TMS is both a perturbation method and, when paired with an appropriate measurement, a neurophysiological probe. Over the motor cortex it can elicit a muscle response that researchers record using electromyography. Over another cortical target it may influence reaction time, perception, choices, or performance on a cognitive task. Combined with EEG or fMRI, TMS can investigate responses throughout connected systems. The physical stimulus is experimentally controlled; the biological effects depend on local anatomy, electric field direction and strength, circuitry, and brain state. An international review of transcranial stimulation dose emphasizes the need to report the coil and pulse parameters that shape the induced field.
The word transcranial describes delivery across the intact skull, magnetic describes induction through a changing magnetic field, and stimulation describes the ability to influence electrically excitable neural tissue. Experimental TMS encompasses several distinct protocols. Single-pulse TMS, paired-pulse TMS, repetitive TMS (rTMS), and patterned stimulation answer different questions and should be interpreted separately.
How TMS developed into a neuroscience research method
Modern human TMS emerged from work by Anthony Barker, Reza Jalinous, and Ian Freeston, whose 1985 report in The Lancet demonstrated noninvasive magnetic stimulation of human motor cortex. Early research used observable muscle responses to assess motor pathways and to identify relationships among stimulus intensity, coil location, and physiological response. Subsequent coil designs and stimulation protocols expanded its use in neurological physiology and cognitive neuroscience.
By the late 1990s, investigators were using TMS to create short disruptions or changes in task-related processing and to ask when particular cortical sites contribute to perception and cognition. Pascual-Leone, Walsh, and Rothwell's review framed three influential research uses: virtual lesions, temporal chronometry, and functional connectivity. The term virtual lesion became popular for a reversible interference experiment, although modern interpretation recognizes that TMS can facilitate, suppress, synchronize, or otherwise reorganize activity depending on context.
Today, structural MRI-based neuronavigation, electric-field simulations, TMS-EEG, and concurrent TMS-fMRI extend this original experimental approach. These additions improve targeting or reveal responses at connected sites, while introducing technical and inferential challenges that require their own controls. International TMS-fMRI consensus guidance published in 2025 illustrates how the field has moved toward explicit network-level experiments.
How does a magnetic pulse influence neural tissue?
The coil, magnetic field, and induced electric field
A stimulator rapidly changes current through a coil held on or close to the head. That changing current generates a brief magnetic field that passes through the scalp and skull and induces an electric field in brain tissue. The induced electric field, especially its direction and variation along neuronal processes, influences membrane polarization in axons and other excitable elements. The biological response can arise from direct recruitment of neural elements followed by synaptic transmission through local and distant circuits.
A figure-of-eight coil is often used when investigators want a relatively focused cortical intervention. Circular coils and other specialized designs produce different field distributions. The stimulated volume is neither a geometric point nor a perfect cylinder beneath the coil. It depends on coil geometry, orientation, distance from cortex, intensity, pulse waveform, skull-to-cortex spacing, head conductivity, and cortical folding. A nominal coordinate on the scalp is therefore an approximation to a physical field distribution in a three-dimensional brain.
This matters whenever investigators say that TMS stimulated a named brain area. More precise wording is that a coil was positioned to target an anatomical region and produced an electric field predicted to engage some tissue there. Field modeling, individual structural scans, recorded coil position, and physiological validation strengthen this statement. Research on coil orientation and neuronavigation accuracy shows why placement and direction must be documented.
From induced field to a brain-network response
The immediate electrical perturbation and the subsequent neural response are distinct stages. Activation in a local population can propagate along anatomical pathways, alter the activity of other cortical or subcortical populations, and interact with neural signals already supporting a task. Even when a behavioral result is reproducible, the direct site of initiation and the wider circuit explaining the result can differ.
These interactions make TMS valuable for investigating network participation. They also prohibit the simple inference that the brain area nearest the coil contains the entire mental function affected by the experiment. A change in performance can reflect disruption of information exchange, altered timing, recruitment of compensatory regions, or effects on other systems that receive input from the stimulated region. This network interpretation is central to the causal-inference analysis by Hobot and colleagues.
Brain state and stimulus history
The same nominal pulse can produce different responses depending on whether someone is resting, preparing a movement, attending to a stimulus, learning a task, or recovering from recent stimulation. Neural populations differ in their current excitability, ongoing oscillatory activity, and recent activity history. Medications, fatigue, sleep, time of day, and individual anatomy may also influence some measures. A 2024 review of target engagement and brain-state dependence describes the importance of these variables.
As a result, TMS effects should be understood as interactions between an intervention and an evolving biological system. A protocol's average effect in one state may not generalize to another task, participant, or target. For causal research this is a design consideration: timing, task instructions, and baseline state belong in the experimental model.
What kinds of TMS do researchers use?
Single-pulse TMS
Single-pulse TMS delivers an isolated magnetic pulse and measures an immediate physiological or behavioral consequence. Stimulating motor cortex may produce a motor-evoked potential in a peripheral muscle. In a cognitive experiment, an isolated pulse can be delivered at one of several moments after a visual or auditory event to examine when a targeted network influences task performance. Effects at a specific time point support a temporally constrained causal hypothesis, provided the control conditions address the pulse's sensory effects.
The pulse is not a direct reading of firing rates, thought content, or neurotransmitter release. The observable response belongs to the output system being sampled, and its interpretation depends on how that system is related to the stimulation site.
Paired-pulse TMS
Paired-pulse experiments use two pulses separated by a controlled interval. In a common motor-cortex design, a conditioning pulse changes the size of the response to a later test pulse. Different intensities and intervals yield measures often called short-interval intracortical inhibition, intracortical facilitation, and long-interval intracortical inhibition. These measures probe aspects of corticospinal and intracortical physiology, as reviewed by Reis and colleagues.
Pharmacological studies suggest partial relationships between particular paired-pulse phenomena and the activity of GABAergic or glutamatergic systems. Nevertheless, no single paired-pulse index measures an isolated neurotransmitter concentration or supplies a direct assay of one receptor population in an individual brain. Multiple neural pathways, spinal excitability, stimulation parameters, and methodological choices contribute to recorded outcomes. For this reason, results are often described as physiological indices associated with inhibitory or facilitatory processes rather than direct chemical measurements.
Repetitive TMS and patterned stimulation
Repetitive TMS delivers trains of pulses. Frequency, intensity, pulse number, train duration, spacing, and target can alter physiological responses during stimulation and sometimes afterward. Theta-burst stimulation groups pulses into bursts delivered according to a patterned schedule; commonly studied intermittent and continuous theta-burst protocols are distinct interventions. Researchers may use rTMS to test how a transiently altered cortical system changes subsequent behavior or to study plasticity-like effects across minutes or longer intervals.
The older heuristic that low-frequency stimulation invariably inhibits the brain while high-frequency stimulation invariably excites it is scientifically unreliable. Those labels arose largely from particular motor-evoked potential findings under particular experimental conditions. A 2026 systematic review and comparative meta-analysis of 202 studies found weaker and more variable effects than common simplified descriptions imply: changes to motor-evoked potentials were modest, suppression by traditionally inhibitory protocols did not survive sham normalization, and paired-pulse or TMS-EEG outcomes showed no consistently replicated pattern across protocols. This evidence strongly argues against exporting a binary excitatory/inhibitory label to every cortical region or psychological function.
An individual's response to repetitive stimulation is not established merely by the published name of the protocol. Direct physiological or behavioral verification remains important, especially outside motor cortex.
Paired associative and network stimulation
Some studies combine peripheral nerve stimulation or sensory events with TMS at carefully selected delays to investigate timing-dependent changes in neural responsiveness. Other studies stimulate two cortical sites in sequence to examine whether conditioning one region changes a response evoked from another. These protocols offer more specific hypotheses about interregional interactions and plasticity-related processes, but physiological interpretations remain model-dependent and susceptible to variable recruitment of the larger circuit.
How does TMS test what a brain region contributes to behavior?
Perturbation and causal inference
Observation-based neuroscience can show that an area is active when a participant performs a task. Perturbation-based neuroscience asks whether deliberately changing activity related to that area modifies a measurable outcome. A randomized experiment comparing active TMS with suitably controlled conditions can support a causal claim about the effect of that intervention on behavior.
Suppose researchers hypothesize that a cortical site contributes to identifying a visual object at a particular stage of processing. They could deliver pulses at several defined latencies after the object appears, compare performance with a control site or sham condition, and test whether accuracy or response time changes selectively. A temporally specific effect can help distinguish alternative information-processing models. The conclusion is that this intervention affected task performance at that time, consistent with participation of the stimulated circuit.
A rigorous causal claim must remain at the level warranted by the manipulation. The intervention has a distributed electrical field and can affect a network, as well as peripheral sensory pathways. Changing behavior after a pulse does not logically establish a one-to-one correspondence between the stimulated site and the entire mental ability. Hobot and colleagues examine this important distinction.
The virtual-lesion idea and its limits
A virtual-lesion experiment aims to interfere reversibly with normal processing at a targeted moment or site. Its value comes from experimentally manipulating neural dynamics without producing an actual anatomical injury. The expression virtual lesion is metaphorical: the induced effect need not be silencing, and it may change timing, signal-to-noise relations, competition between representations, or communication through a circuit.
An observed performance impairment may support the hypothesis that the perturbed system normally makes a contribution under the tested conditions. A null result does not prove that the area is irrelevant. The stimulation may miss the effective neural elements, arrive outside the critical time window, have insufficient strength, or be counteracted by network compensation. Conversely, performance improvement after stimulation does not show that a region was previously excessive or dysfunctional.
Chronometry: when a brain process matters
In online TMS, pulses are applied while the person completes a task. Systematically varying stimulation timing relative to a sensory event, movement cue, or decision allows researchers to test time-resolved hypotheses about cortical participation. This approach is especially informative when paired with behavioral experiments that already isolate stages of processing.
Chronometric conclusions require care. A time window during which TMS changes performance indicates a window of intervention sensitivity; it does not identify the exact moment a representation first emerged or disclose the detailed internal code of the circuit. The induced physiological effects and the task may extend beyond the nominal pulse instant.
Dissociations, interactions, and replication
Stronger functional claims arise when a study evaluates multiple tasks, targets, and stimulation times. A selective interaction between task, site, and timing can help distinguish explanations based on general distraction, discomfort, or performance slowing. Independent replication, preregistered outcomes where appropriate, adequate sample size, reporting of uncertainty, and transparent analysis strengthen the evidence.
An anatomical dissociation should be interpreted alongside task difficulty and network organization. Two tasks can depend on overlapping circuits at different levels, and a stimulated network can reorganize its activity when challenged. The experimental question is therefore about the contribution of a circuit under specified conditions, rather than possession of a psychological faculty by an isolated region.
What does TMS actually measure?
TMS itself delivers an intervention. What a study measures comes from the accompanying recording or observation. In motor physiology, the outcome can be a muscle electromyography signal, a motor threshold, or a period of suppressed voluntary activity. In cognition it can be response accuracy, reaction time, confidence, perceptual discrimination, or another operationally defined task measure. In TMS-EEG studies it can be scalp voltage changes following stimulation; in TMS-fMRI studies it can be changes in the blood-oxygen-level-dependent signal.
A measurement is evidence for its own defined variable first. Motor-evoked potential amplitude is an output of a multistage corticospinal and neuromuscular pathway. A behavioral score reflects the person's task performance. EEG provides electrical signals influenced by neural and non-neural sources. fMRI measures a hemodynamic correlate of aspects of neural activity. None of these outputs directly reveals a full mental state, a particular neurotransmitter concentration, or a unique location where a complex behavior is stored.
Motor thresholds and motor-evoked potentials
When a suitable motor-cortex site is stimulated, descending activity through motor pathways can produce electrical activity in a muscle, typically recorded with surface electromyography. The motor-evoked potential, or MEP, is quantified by its amplitude, timing, probability, or related parameters. Researchers can estimate a resting or active motor threshold as an operational stimulus level that elicits a defined response under standardized conditions.
MEPs are particularly useful because they give a measurable output linked to stimulation. Yet their amplitude varies from trial to trial and reflects spinal and peripheral contributors as well as cerebral ones. A change in MEP amplitude is not synonymous with a change in the number of active neurons or a universal increase in overall brain activity. Motor physiology reviews and an international methodological checklist discuss these sources of variability.
What about nonmotor regions?
Outside motor cortex, researchers generally lack a direct and simple muscle-output calibration for the targeted psychological operation. They must validate the stimulation site and hypothesized neural response using anatomical targeting, simulations, converging physiology, or carefully controlled task effects. Motor threshold may be used as a reference for stimulator intensity, but the same percentage of motor threshold does not guarantee the same electric field or biological impact in a different cortical target.
This difference matters when comparing studies of memory, language, decision-making, and attention. The apparent simplicity of an intensity setting conceals substantial variation in anatomy and target engagement.
TMS and EEG: observing an evoked electrical response
TMS-EEG combines an intervention with electrodes that record voltage differences at the scalp. Investigators may analyze evoked potentials, oscillatory changes, or responses outside the targeted location. This allows the timing and distribution of a measured response to be examined on a millisecond scale, making it useful for hypotheses about cortical reactivity and network dynamics.
The TMS pulse introduces substantial artifacts into the recording system. Coil discharge, electrode movement, facial and scalp muscle activation, and recovery of amplifiers can affect the recorded trace. The coil also produces an audible click and somatosensory sensations, which elicit their own brain responses. A TMS-evoked potential is therefore not automatically a clean recording of direct activity in the target cortex.
The comprehensive 2023 TMS-EEG recommendations emphasize equipment, preprocessing, appropriate control conditions, and careful interpretation. A 2023 experimental study disentangling TMS and sensory responses illustrates why masking noise and optimized sham procedures are scientifically significant. Researchers need to show which portions of a measured response plausibly arise from direct cortical perturbation, which arise from sensory consequences, and where uncertainty remains.
TMS and fMRI: testing distributed network responses
Concurrent TMS-fMRI pairs magnetic stimulation with MRI acquisition so that investigators can examine how stimulation affects hemodynamic signals at cortical targets and connected regions. Other designs acquire fMRI before or after TMS. These strategies can link perturbation parameters with distributed signal changes and help test hypotheses about functional connections or task-dependent modulation.
The fMRI readout remains indirect. Blood-oxygen-level-dependent signals reflect physiological changes in local blood oxygenation and related vascular processes; they are influenced by neural activity but are not direct measures of spikes, excitation, inhibition, or mental content. A remote fMRI change after stimulating one region can support a network-effect interpretation, but cannot alone prove a direct monosynaptic anatomical connection.
Combining stimulation with MRI requires equipment compatibility, timing controls, artifact management, participant safety checks, and careful modeling. The international 2025 consensus on concurrent TMS-fMRI sets out detailed use cases and procedural considerations. Its value is especially clear when the experiment distinguishes local target engagement from downstream effects on a larger circuit.
Why targeting and electric-field modeling matter
Scalp coordinates, MRI guidance, and neuronavigation
Traditional targeting can use anatomical landmarks, standardized scalp coordinates, or distances from a motor hotspot. Such approaches are practical but can differ in accuracy because head anatomy and cortical folding vary across individuals. MRI-guided neuronavigation connects the participant's head position with a structural scan and displays the estimated coil location and orientation relative to anatomy.
A controlled comparison found substantially better coil-placement accuracy with neuronavigation than with cap-based methods, including less variation in modeled prefrontal electric fields. This finding concerns precision of delivery, not proof that neuronavigation necessarily improves every behavioral outcome or clinical endpoint. See the 2022 positioning comparison.
Modeling physical dose within an individual brain
Computational modeling estimates the electric field created inside a head model by a specified coil and pulse. This can improve interpretation of spatial spread and orientation. Electric-field simulations are useful physical models with assumptions about anatomy, tissue conductivities, segmentation, and numerical methods. They are not direct recordings of every individual neuron's membrane response.
Newer experiments illustrate both potential and limits. A 2025 repeat-scan study found that single-run fMRI-based target localization could vary substantially across sessions, while using multiple runs improved stability. A 2026 demonstration of real-time electric-field navigation showed technical progress in field estimation and target visualization. These studies support better measurement and delivery methods; they do not establish that a particular individualized targeting method will reliably improve cognition or clinical outcomes in all populations.
How to design a trustworthy TMS experiment
A compelling TMS experiment links a narrowly defined hypothesis to a target, an intervention schedule, a measurement, and one or more control conditions that distinguish the hypothesized mechanism from competing explanations. The target region should be motivated by converging evidence or an explicit theoretical model, the stimulation parameters should be reportable and reproducible, and the primary outcomes should be connected to the proposed psychological process.
Sensory confounding deserves special emphasis. Active TMS generates sound, scalp sensations, and sometimes muscle twitches. A participant may infer which condition they received, change attention in response to the click, or alter behavior because of discomfort. Sham coils, angled-coil conditions, stimulation of a control site, sensory masking, or factorial combinations can help but are not interchangeable and rarely eliminate every alternative explanation. The correct control follows from the specific causal claim.
Study quality also depends on pulse intensity calibration, coil position and orientation, number of trials, intertrial intervals, task timing, baseline state, outcome selection, statistical analysis, sample size, and handling of missing or excluded trials. An international motor-TMS methodological consensus and the TMS-EEG reporting recommendations provide concrete examples of information that should be reported.
Researchers should analyze within-person variability and between-person differences where the design supports them. A statistically significant mean effect does not imply a uniform effect in every participant. A null result is informative when experimental sensitivity and uncertainty are adequately characterized. Research conclusions improve when multiple outcome measures converge, replications use comparable protocols, and competing sensory or network explanations are directly tested.
What does current evidence establish?
The physical ability of a rapidly changing magnetic field to induce electrical currents in conducting tissue is established. TMS reliably elicits measurable outputs from appropriate motor pathways under suitable conditions, and controlled studies demonstrate that interventions can alter task-related physiology and behavior. The method has decades of research use, supported by established motor physiology, experimental brain-behavior studies, and expert methodological and safety guidance.
The exact direction, magnitude, spatial distribution, and persistence of physiological changes from a given repetitive protocol are more variable. The 2026 meta-analysis by Yassi and colleagues pooled 202 studies and 8,079 healthy participants across commonly used protocols, cortical targets, and measures. It challenges routine assumptions that a named protocol has a stable excitatory or inhibitory effect across contexts. Its findings do not mean that TMS cannot alter brain function. They narrow which effects can be presumed without direct verification.
Evidence about cognitive effects is also task- and design-dependent. Some well-controlled experiments support specific intervention-sensitive time windows and network contributions. Other results are difficult to reproduce, reflect mixed mechanisms, or depend on participant state and target localization. Reviews of causal logic and brain-state effects explain why this variability must be incorporated into interpretation.
Claims about TMS as a treatment belong to a separate clinical evidence base. A protocol that changes a laboratory motor outcome or task score is not automatically an effective medical intervention. Conversely, evidence that a specified treatment protocol benefits a clinical population does not establish a detailed causal theory about every aspect of the disorder.
Safety, participant screening, and research ethics
TMS is generally well tolerated when delivered under established research and clinical safety protocols by trained personnel. Common experiences include tapping or discomfort at the scalp, transient headache, facial muscle twitching, and the loud sound produced by coil discharge. Hearing protection is important. Serious adverse events are uncommon under appropriately screened and monitored conditions, but seizures are a recognized risk and should be addressed explicitly in protocol-specific risk assessment.
Risk assessment includes the participant's medical and neurological history, medicines that may affect seizure threshold, prior seizures, pregnancy-related considerations where relevant, and the presence and location of implants, conductive or ferromagnetic material, or implanted electronic devices. Safety decisions depend on device type, stimulation parameters, implant characteristics, and individual circumstances. Researchers should not infer general compatibility from a simple label such as metal implant or from the fact that TMS is noninvasive.
The international IFCN expert safety guidelines discuss stimulation parameters, screening, seizure response procedures, novel devices, training, ethical obligations, and clinical versus experimental settings. Institutional review and informed consent are essential elements of research involving human participants. A research protocol requires its own risk assessment and should never be reconstructed or self-administered from a general-audience description.
Particular caution applies to prolonged or patterned protocols, combinations with other stimulation or imaging modalities, special populations, and experiments in which aftereffects may persist. Investigators should use current device-specific instructions and institutional procedures rather than relying only on general descriptions of TMS safety.
Research TMS versus therapeutic TMS
Experimental TMS typically aims to test a mechanistic hypothesis: whether changing stimulation at a defined cortical location and time alters a neurophysiological signal or a specific behavior. The central outcomes are designed to test brain-function models. Such studies may enroll healthy volunteers or clinical participants and commonly use single pulses, paired pulses, brief trains, or controlled stimulation sessions.
Therapeutic repetitive TMS is a medical intervention delivered according to a clinical protocol with a defined treatment target, patient selection process, dosing schedule, monitoring, and symptom outcomes. In the United States, FDA-cleared TMS devices have defined indications and labeling. The National Institute of Mental Health's brain-stimulation guidance describes clinical uses, including certain forms of depression and obsessive-compulsive disorder, and explains how rTMS differs from electroconvulsive therapy.
Clinical consensus recommendations updated in 2025 support the effectiveness of appropriate rTMS protocols for depressive disorders and discuss emerging methods. That clinical evidence does not permit unlimited generalization across stimulation targets, diagnoses, ages, schedules, or consumer devices. Specific treatment questions are addressed separately in our TMS for depression guide and TMS for OCD guide.
The practical dividing line is the purpose and evidence of the protocol. A laboratory intervention may reveal a network's causal involvement without showing therapeutic benefit. An evidence-based clinical treatment may reduce symptoms while important details of its underlying biological mechanism remain under investigation. Decisions about individual treatment eligibility, risks, or interactions belong with qualified clinicians familiar with the patient's health and the exact device protocol.
TMS compared with MRI, EEG, tDCS, and brain-computer interfaces
Structural MRI primarily characterizes anatomy and tissue contrast. Functional MRI detects hemodynamic signals associated with aspects of neural activity. EEG records voltage differences at the scalp. These are recording or imaging methods; by themselves they do not constitute a direct experimental perturbation of the brain. TMS supplies an intervention, and combined methods can examine its consequences. For a wider comparison, read Brain Imaging Methods: What MRI, fMRI, PET, and EEG Measure.
Transcranial electrical stimulation, such as transcranial direct current stimulation (tDCS), applies electrical currents through scalp electrodes and operates through a different physical delivery method and dose profile. It should not be treated as interchangeable with TMS. Deep brain stimulation uses implanted hardware to influence neural circuits and has its own clinical indications and risks. Electroconvulsive therapy intentionally induces a generalized seizure under anesthesia as a medical treatment; conventional TMS protocols do not aim to produce such a seizure.
Brain-computer interfaces acquire neural or related physiological signals to decode or control an external system. That signal-acquisition and output purpose differs from stimulating a network to test its causal contribution to cognition. Some research systems combine sensing with stimulation, but the scientific status of each component must be evaluated independently.
Common misconceptions about TMS
Myth: TMS reveals exactly where a thought is stored
A stimulation-induced behavioral change provides evidence about a circuit's contribution under defined conditions. Thoughts and complex psychological operations involve interacting processes, and the electric field affects neural elements that may communicate widely. TMS is a tool for testing causal hypotheses, not a device that reads or pinpoints an entire thought.
Myth: TMS switches brain regions off
The effects of a pulse depend on stimulated neural elements, intensity, orientation, timing, ongoing activity, and the broader network. The virtual-lesion label describes a class of interference experiments rather than a literal shutdown. Even when task performance declines, physiological activity may be reorganized rather than simply absent.
Myth: High-frequency TMS always excites, and low-frequency TMS always inhibits
These labels summarize historical averages in particular motor-cortex measures. A broad 2026 evidence synthesis found much greater variability and weaker general effects than simple slogans suggest. Direct testing of the actual target and outcome is needed.
Myth: An MEP or TMS-EEG trace measures one neurotransmitter
Physiological recordings reflect the activity of interacting neural and peripheral systems. Pharmacological sensitivity can help constrain hypotheses about receptors or inhibition, but no routine TMS measure uniquely identifies a person's GABA, glutamate, serotonin, or dopamine concentration. Clinical or cognitive conclusions require additional independent evidence.
Myth: A TMS result automatically proves causation by the targeted brain area
A randomized stimulation experiment can show the causal effect of its intervention. Assigning that effect uniquely to one anatomical area requires evidence about the induced field, sensory confounds, remote network responses, and the outcome measure. Methodological analyses emphasize that these are distinct levels of inference.
Myth: Noninvasive means risk-free
TMS avoids implanted electrodes in routine protocols, but magnetic stimulation can cause discomfort, trigger peripheral effects, interact with some implants, and in rare circumstances be associated with seizures. Research ethics and current safety guidance remain necessary.
Practical implications for reading neuroscience research
When a news article claims that scientists used TMS to prove the function of a brain region, identify what was actually manipulated, the size and location of the likely induced field, the control condition, the task measure, and the specific statistical comparison. Determine whether the study used an acute single-pulse experiment, paired-pulse physiology, patterned stimulation, or a clinical treatment course. These paradigms answer different questions.
Next, ask whether the outcome was a motor signal, an EEG response, an fMRI measure, or behavior. Examine how the authors addressed auditory and tactile stimulation, whether their claim concerns a local effect or a connected network, and whether independent evidence converges with the interpretation. For treatment claims, check whether the population, clinical endpoint, and device protocol match those in rigorous trials and professional guidelines.
The most scientifically informative TMS evidence links a clearly specified intervention with a replicable consequence and a carefully delimited interpretation. Used this way, TMS helps connect neural dynamics with behavior while preserving the complexity of the brain systems that make cognition possible.
Frequently asked questions
How does transcranial magnetic stimulation work in simple terms?
A coil near the scalp produces a rapidly changing magnetic field, which induces an electric field in brain tissue. That field can alter electrical activity in responsive neural elements and their connected circuits. Researchers then measure a muscle response, brain signal, or behavioral change to study what the intervention affected.
Does TMS directly measure brain activity?
TMS delivers a stimulus. The response is measured with another tool such as electromyography, EEG, fMRI, or behavioral testing. Without a separately defined measurement, a pulse does not provide a recording of brain activity.
Can TMS demonstrate that a brain region causes a behavior?
Controlled TMS experiments can support causal claims about how a specific intervention affects task performance. Strong attribution to a precise neural substrate additionally requires accurate targeting, suitable sham or control conditions, network-aware interpretation, and replication.
What is the difference between TMS and rTMS?
TMS refers to the overall technique. Single-pulse and paired-pulse methods probe responses to isolated or closely timed pulses. Repetitive TMS delivers trains of pulses and may produce effects that outlast individual pulses; therapeutic rTMS is a distinct use involving clinically specified protocols.
Is TMS the same as electroconvulsive therapy?
No. Conventional TMS induces a localized electric field without the intention of producing a seizure and usually does not require anesthesia. Electroconvulsive therapy is an anesthetized clinical procedure designed to induce a controlled seizure. Their indications, procedures, and mechanisms differ.
Does TMS permanently rewire the brain?
Some protocols can produce measurable aftereffects, and neural plasticity is one possible contributor. The magnitude, persistence, and meaning of those aftereffects vary, and an experimental change in a motor or behavioral measure cannot by itself demonstrate permanent structural rewiring. Long-term claims require direct longitudinal evidence.
Can TMS improve memory, focus, or intelligence in healthy people?
Individual studies explore cognitive performance under specific stimulation conditions, but findings are variable and cannot establish a general, reliable cognitive-enhancement benefit for healthy people. Effects on a laboratory task are not equivalent to durable improvements in everyday cognition. Personal use outside supervised research or clinical practice raises safety and evidence concerns.
Can TMS diagnose a psychiatric disorder?
TMS-based physiological indices are investigated as research measures in clinical populations. They are not established as standalone tests for diagnosing depression, OCD, ADHD, or other psychiatric disorders in individuals. Diagnosis depends on clinically validated criteria and appropriate assessment.
Is TMS safe for people with metal implants or a history of seizures?
Eligibility depends on the exact implant, its location and characteristics, the stimulation device, medical history, medications, and protocol. A qualified team must assess these risks using current safety guidance. General descriptions cannot establish individual suitability.
Why do different TMS studies reach different conclusions?
Results can vary with target anatomy, coil orientation, electric field strength, stimulation waveform, participant state, experimental timing, task design, sensory controls, sample size, and statistical uncertainty. Some protocols also have weaker physiological effects than their traditional labels suggest. Variation is an empirical question to measure, not something to conceal by averaging incompatible study designs.
A complementary approach to causal inference examines the effects of permanent neurological injury through brain lesion studies.
Related Articles
References
Barker, A. T., Jalinous, R., & Freeston, I. L. (1985). Non-invasive magnetic stimulation of human motor cortex. The Lancet, 1(8437), 1106–1107. https://doi.org/10.1016/S0140-6736(85)92413-4
Hernandez-Pavon, J. C., et al. (2023). TMS combined with EEG: Recommendations and open issues for data collection and analysis. Brain Stimulation, 16(2), 567–593. https://doi.org/10.1016/j.brs.2023.02.009
Hobot, J., Klincewicz, M., Sandberg, K., & Wierzchoń, M. (2021). Causal inferences in repetitive transcranial magnetic stimulation research: Challenges and perspectives. Frontiers in Human Neuroscience, 14, 586448. https://doi.org/10.3389/fnhum.2020.586448
National Institute of Mental Health. (2026 access). Brain stimulation therapies. https://www.nimh.nih.gov/health/topics/brain-stimulation-therapies/brain-stimulation-therapies
Pascual-Leone, A., Walsh, V., & Rothwell, J. (2000). Transcranial magnetic stimulation in cognitive neuroscience—Virtual lesion, chronometry, and functional connectivity. Current Opinion in Neurobiology. https://doi.org/10.1016/S0959-4388(00)00081-7
Peterchev, A. V., et al. (2012). Fundamentals of transcranial electric and magnetic stimulation dose: Definition, selection, and reporting practices. Brain Stimulation. https://pubmed.ncbi.nlm.nih.gov/22305345/
Reis, J., et al. (2008). Contribution of transcranial magnetic stimulation to the understanding of cortical mechanisms involved in motor control. The Journal of Physiology, 586(2), 325–351. https://doi.org/10.1113/jphysiol.2007.144824
Rossi, S., et al. (2021). Safety and recommendations for TMS use in healthy subjects and patient populations, with updates on training, ethical and regulatory issues: Expert Guidelines. Clinical Neurophysiology, 132(1), 269–306. https://doi.org/10.1016/j.clinph.2020.10.003
Wang, H., et al. (2025). Variability and repeatability of personalized rTMS targeting methods combining repeated fMRI scans and electric field simulations. NeuroImage, 317, 121346. https://doi.org/10.1016/j.neuroimage.2025.121346
Yassi, I. E., et al. (2026). Reassessing the neurophysiological effects of repetitive transcranial magnetic stimulation: A systematic review and comparative meta-analysis across protocols, outcome measures and cortical sites. Neuroscience & Biobehavioral Reviews, 185, 106648. https://doi.org/10.1016/j.neubiorev.2026.106648