{
  "id": "transcranial-direct-current-stimulation",
  "title": "Transcranial Direct-Current Stimulation",
  "lang": "en",
  "source_lang": "en",
  "summary": "Transcranial direct-current stimulation applies weak electrical currents through scalp electrodes to modulate brain activity, with applications in neuroscience and continuing clinical research.",
  "category": "health",
  "subcategory": "medicine",
  "type": "concept",
  "aliases": [
    "tDCS",
    "transcranial DC stimulation"
  ],
  "generated_at": "2026-11-06",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/transcranial-direct-current-stimulation/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/transcranial-direct-current-stimulation/",
      "json": "https://aiwiki.page/zh/transcranial-direct-current-stimulation.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Transcranial direct-current stimulation (tDCS) is a form of noninvasive brain stimulation in which a weak, sustained electrical current is delivered through electrodes placed on the scalp. The current produces an electric field within the brain, altering the conditions under which neurons and neural networks become active. Unlike stimulation methods designed to trigger neuronal firing directly, conventional tDCS primarily modifies ongoing activity and the response to other inputs.\n\nThe technique is important both as an experimental tool for studying human brain function and as a potential treatment for selected neurological and psychiatric conditions. Its relatively simple equipment, portability, and compatibility with behavioral training have supported extensive research. Nevertheless, ease of delivering a current does not imply that its biological effects are simple, reliably targeted, or clinically effective. Results depend on electrode arrangement, individual anatomy, stimulation parameters, concurrent activity, and the condition being studied.\n\nResearch has addressed depression, pain, stroke rehabilitation, cognition, and appetite-related behavior, among other applications. The strength of evidence differs substantially between these fields. In particular, claims that a fixed daily session reliably stabilizes blood glucose or produces a specified percentage of weight loss are not established general properties of tDCS."
    },
    {
      "heading": "Historical development and place among stimulation methods",
      "text": "Medical experimentation with electricity long predates modern neuroscience. After the development of practical batteries around 1800, investigators applied continuous electrical currents to the body and head for a wide range of proposed therapeutic purposes. These early practices lacked standardized dosing, reliable comparison groups, and the physiological measurements needed to distinguish genuine effects from spontaneous variation or expectation. Consequently, historical use of electricity is not evidence that present-day tDCS treats the conditions claimed for those earlier procedures.\n\nAn important experimental precursor appeared in animal research during the 1960s. Studies by D. J. Albert and by Olof Creutzfeldt and colleagues showed that weak polarizing currents could alter cortical neuronal activity. Their findings helped establish that electrical fields need not provoke a conventional action potential to influence nervous tissue. Depending on field direction and neuronal orientation, polarization could increase or decrease the likelihood of firing and could sometimes leave effects after current delivery ended.\n\nModern human tDCS developed around experiments that combined scalp stimulation with measurements of motor-system excitability. In 1998, Alberto Priori and colleagues reported changes in human motor-cortex excitability following weak transcranial direct-current stimulation. In 2000, Michael Nitsche and Walter Paulus demonstrated polarity-dependent changes using transcranial magnetic stimulation to measure motor-evoked responses. These studies provided an experimentally accessible way to investigate stimulation intensity, duration, and after-effects, helping launch the technique’s wider use.\n\ntDCS belongs to the broader family of transcranial electrical stimulation methods. Transcranial alternating-current stimulation uses oscillating currents, while transcranial random-noise stimulation applies fluctuating waveforms. Their mechanisms and parameter choices overlap with, but are not identical to, those of direct-current stimulation. Findings from one waveform cannot automatically be transferred to another.\n\nOther brain-stimulation procedures are more distinct. Transcranial magnetic stimulation uses rapidly changing magnetic fields to induce electrical currents in tissue. Electroconvulsive therapy deliberately induces a controlled seizure under anesthesia, whereas conventional tDCS does not aim to produce seizures. Deep brain stimulation involves surgically implanted electrodes and can reach targets inaccessible to direct focal stimulation from scalp electrodes. These differences affect indications, physiological effects, equipment requirements, and risks. Grouping all of these methods under “electrical brain treatment” obscures clinically important distinctions."
    },
    {
      "heading": "Equipment, electrode arrangements, and electrical dose",
      "text": "A typical research tDCS system contains a regulated constant-current source, two or more electrodes, conductive interfaces, and circuitry that monitors the connection between device and participant. Conventional studies commonly use currents of approximately 1–2 milliamperes for sessions lasting about 10–30 minutes, although protocols outside these ranges also exist. These figures describe frequent research practice, not a universal therapeutic dose or a boundary within which every possible application is established as safe.\n\nIn a simple two-electrode arrangement, one electrode is designated the anode and the other the cathode. Both participate in the electrical circuit and influence current distribution; the second electrode is not biologically irrelevant merely because it is called a “reference” or “return.” Electrodes may be positioned over different scalp regions, or one may be placed outside the head. Moving either electrode changes the current pathway and can alter effects across multiple brain regions.\n\nTraditional arrangements often use saline-soaked sponge electrodes with contact areas of approximately 25–35 square centimetres. Smaller gel-based electrodes are used in high-definition arrangements, including configurations with one central electrode and several surrounding returns. Such arrays can make the modeled field more spatially concentrated, but “high definition” does not mean that current reaches only a single brain structure or neuronal population. Electrode contact quality and the detailed arrangement remain consequential.\n\nSeveral physical quantities describe stimulation. If current is \\(I\\), electrode contact area is \\(A\\), and duration is \\(t\\), nominal electrode current density and delivered charge are\n\n\\[J_{\\mathrm{electrode}}=\\frac{I}{A}, \\qquad Q=It.\\]\n\nFor example, 2 milliamperes delivered through a 25-square-centimetre electrode corresponds to a nominal current density of \\(0.08\\) milliamperes per square centimetre. Over 20 minutes, that current delivers 2.4 coulombs. These calculations characterize the device–electrode exposure; they do not specify the electric field experienced by a particular cortical neuron.\n\nA complete protocol therefore includes electrode material, size, location, conductive medium, current intensity, duration, ramping, and session schedule. The activity performed during stimulation is also important. Gradually increasing and decreasing the current can reduce abrupt sensations, while contact monitoring helps identify an inadequate electrical connection. A report that gives only “20 minutes of tDCS” omits much of the information needed to reproduce the intervention or interpret its effects."
    },
    {
      "heading": "How weak currents influence neurons and networks",
      "text": "Most current applied at the scalp does not travel directly through the intended cortical target. It spreads through the scalp and other tissues, with its intracranial distribution shaped by skull resistance, the conductivity of cerebrospinal fluid, and the geometry of brain tissue. At commonly studied intensities, intracranial fields are generally weak and spatially distributed. Their magnitude varies by location and individual anatomy; a single universal conversion from milliamperes at the device to field strength in the cortex is not available.\n\nThe immediate neuronal effect involves polarization of the cell membrane. An external field can slightly change the voltage difference across different portions of a neuron, bringing some compartments closer to firing threshold and others farther from it. Dendrites, cell bodies, and axons need not polarize in the same direction. Consequently, the orientation of a neuron relative to the field is as important as the label attached to the scalp electrode.\n\nEarly motor-cortex experiments encouraged the shorthand that anodal stimulation is excitatory and cathodal stimulation inhibitory. That description is useful only within restricted experimental contexts. The folded geometry of the cerebral cortex means that adjacent cortical surfaces may have different orientations to the same field. Effects also depend on cell type, intensity, duration, and ongoing activity. Increasing current does not necessarily produce a proportionally larger effect, and polarity does not provide a simple on–off switch for a mental function.\n\nEffects persisting after stimulation are associated with changes in synaptic effectiveness and neuroplasticity. Pharmacological experiments implicate N-methyl-D-aspartate receptors in some tDCS after-effects, while studies of inhibitory neurotransmission and other signaling systems indicate additional contributions. Processes resembling long-term potentiation or depression are relevant models, but a whole-brain behavioral response cannot be reduced to a single synaptic mechanism.\n\nNetwork state is another major determinant. Stimulation delivered during motor practice, language therapy, rest, or a demanding cognitive task acts on different patterns of activity. It may influence how active connections are modified rather than simply raising or lowering the activity of an isolated region. This state dependence helps explain why apparently similar protocols can yield different outcomes across laboratories or populations.\n\nPeripheral effects also complicate interpretation. Scalp nerves and sensory receptors are exposed to current, and stimulation sensations can alter attention or expectations. Changes in behavior or brain measurements therefore require experimental controls before they can be attributed specifically to direct polarization of a proposed cortical target."
    },
    {
      "heading": "Individual anatomy, targeting, and physiological measurement",
      "text": "Electrode position is often described using landmarks from electroencephalography, such as the international 10–20 system. These coordinates make placement more reproducible, but they are scalp locations rather than exact maps of the underlying cortex. Differences in head shape, cortical folding, and the position of functional networks mean that the same coordinates do not produce identical stimulation in every participant.\n\nComputational modeling provides a more detailed account of current flow. Structural magnetic resonance imaging can be used to distinguish scalp, skull, cerebrospinal fluid, gray matter, and white matter. Models then estimate electric fields using assumed tissue conductivities and electrode properties. Finite-element calculations can show, for example, that an electrode placed above one region also exposes distant tissue or produces stronger fields near boundaries between tissues.\n\nThese estimates have limitations. Tissue conductivities are imperfectly known, segmentation can be inaccurate, and small differences in electrode placement or contact can matter. A model predicts physical exposure, not necessarily therapeutic response. Even if two people receive similar estimated fields, differences in disease, neural circuitry, medication, or task engagement may produce different effects.\n\nStructural changes caused by stroke, surgery, or cerebral atrophy are especially relevant. A lesion containing fluid can redirect current, while increased cerebrospinal-fluid spaces can change its distribution. This is one reason that protocols developed in healthy young adults cannot simply be assumed to behave identically in older adults or people with neurological disease. Individualized modeling is an active research approach, but improved field targeting still requires confirmation through clinical outcomes.\n\nPhysiological measurements assess different aspects of the response. Motor-evoked potentials provide an indirect measure of corticospinal excitability. Functional magnetic resonance imaging can examine changes in activity and connectivity, while EEG and magnetoencephalography can characterize aspects of neural timing and oscillations. Measurements made during electrical stimulation face particular technical difficulties because the stimulation artifact may be much larger than the biological signal.\n\nNo single physiological biomarker has become a universal indicator of whether tDCS has worked. A change in motor excitability does not establish improved movement, and altered connectivity does not by itself demonstrate relief of depression. Mechanistic studies and clinical studies answer related but different questions: the former examine how exposure affects a system, while the latter test whether those effects produce meaningful benefits for a defined population."
    },
    {
      "heading": "Depression and other psychiatric applications",
      "text": "Depression is among the most extensively investigated clinical applications of tDCS. A common approach places an anode over the left dorsolateral prefrontal cortex, with the return electrode over another frontal or supraorbital region. The rationale concerns prefrontal networks involved in cognitive control, emotional processing, and mood regulation. It does not imply that depression is caused by a single inactive spot or that applying current restores a universally normal activity pattern.\n\nStudies usually deliver repeated sessions over several weeks and assess symptoms with standardized depression-rating scales. Some also measure response, remission, functioning, and durability after the treatment course. These outcomes are not interchangeable: a reduction in average symptom score does not establish that most participants reached remission, and an immediate post-treatment improvement does not establish a lasting effect.\n\nThe 2013 SELECT-TDCS trial enrolled 120 participants and used a factorial design to investigate active or sham stimulation alongside sertraline or placebo medication. It provided evidence that tDCS could have antidepressant effects under the tested conditions and examined its combination with medication. The 2017 ELECT-TDCS trial enrolled 245 participants and compared tDCS, escitalopram, and placebo conditions. Although active tDCS outperformed placebo on the trial’s principal depression measure, it did not meet the prespecified criterion for noninferiority to escitalopram.\n\nAcross trials, results vary with participant selection, treatment resistance, stimulation schedule, medication use, and study design. Evidence of average benefit in a pooled analysis does not establish that every device or electrode arrangement is effective. Comparisons between studies are also complicated when treatment courses differ in session number or when stimulation is combined with psychotherapy or other interventions.\n\nOther psychiatric applications remain heterogeneous. In schizophrenia, some protocols have targeted networks associated with auditory verbal hallucinations using frontotemporal electrode arrangements. Research has also examined symptoms related to substance use, obsessive–compulsive disorder, and anxiety. The evidence and proposed mechanisms differ across these conditions; a positive study for one symptom cannot establish efficacy for an entire diagnostic category.\n\nMood changes require particular attention in psychiatric studies. Hypomania or mania has been reported during depression-treatment research, including studies involving antidepressant medication. Such events affect both safety assessment and interpretation of symptom improvement. Clinical status also depends on jurisdiction, device, and indication: regulatory authorization, professional guidance, and experimental efficacy are separate matters, and recognition of one application does not validate unrelated uses."
    },
    {
      "heading": "Neurological rehabilitation, pain, and cognitive performance",
      "text": "In rehabilitation research, tDCS is usually intended to modify the effects of training rather than replace it. After stroke, investigators have paired stimulation with movement practice, physical therapy, or language treatment. The central hypothesis is that changing the excitability or plasticity of a relevant network may help practice produce more useful learning.\n\nMotor-rehabilitation protocols have frequently targeted the motor cortex of the damaged hemisphere, the opposite hemisphere, or both. Some were motivated by an interhemispheric-competition model, in which activity in the unaffected hemisphere was thought to suppress recovery. Subsequent work has shown that the role of the opposite hemisphere is more complex: it may be maladaptive in some circumstances but compensatory in others. Lesion size, residual pathways, impairment severity, and time since stroke can therefore influence the relevance of a particular arrangement.\n\nLanguage studies have paired stimulation with speech-language therapy for aphasia. Some report improvements in practiced naming tasks, but transfer to untrained words, everyday communication, and durable functional recovery is a separate question. Better performance on a laboratory task is not equivalent to reduced disability. Rehabilitation trials consequently need outcomes that reflect activities and participation as well as impairment.\n\nResearch on chronic pain has often targeted motor or prefrontal cortex. Proposed effects involve distributed pain-processing and modulatory networks rather than direct treatment of the injured tissue or underlying disease. Trials have examined conditions such as neuropathic pain and fibromyalgia, but their results cannot be treated as a single estimate for all pain disorders. Pain intensity, physical function, sleep, medication use, and health-related quality of life may change differently.\n\nIn healthy volunteers, studies have investigated working memory, attention, language learning, decision-making, and motor learning. Findings are mixed, and positive effects are often specific to a task, timing, or participant subgroup. Improvement in one task component may coexist with no change elsewhere. A stimulation arrangement that benefits performance under one set of conditions can be ineffective under another because the relevant neural activity and baseline performance differ.\n\nThese distinctions limit claims of general cognitive enhancement. Faster responses on a short experimental task do not demonstrate increased intelligence, improved educational attainment, or protection against dementia. Studies of healthy performance and studies of neurological rehabilitation also involve different goals and populations. Their findings require separate evaluation rather than being combined into a broad claim that stimulation makes the brain function better."
    },
    {
      "heading": "Appetite, body weight, and glucose regulation",
      "text": "Interest in metabolic applications arises partly from the brain’s role in eating behavior and energy regulation. The hypothalamus integrates signals related to energy availability, while prefrontal and reward networks contribute to food selection, inhibition, and responses to food cues. Brain activity can also influence autonomic and hormonal processes relevant to metabolism. These relationships provide biological reasons to investigate neuromodulation, but they do not establish that stimulating the scalp produces a predictable metabolic benefit.\n\nMany tDCS studies related to eating have targeted the dorsolateral prefrontal cortex rather than directly stimulating hypothalamic nuclei. Outcomes have included subjective food craving, attention to food images, performance on inhibitory-control tasks, and food consumed during a test meal. These measures address different questions. Reduced craving on a rating scale does not necessarily lead to a sustained reduction in energy intake, and lower intake at one meal may be offset later.\n\nBody-weight change requires longer observation. Trials must distinguish transient changes in fluid balance from sustained changes in body tissue and account for concurrent dietary interventions, physical activity, medication, and attrition. Results in healthy volunteers exposed to food cues are not equivalent to treatment evidence in people with obesity. Similarly, an observed average weight change in a small, selected sample does not establish a reproducible response for the wider population.\n\nGlucose-related claims require further distinctions. Fasting glucose, post-meal glucose, insulin resistance, and long-term glycemic control are different outcomes. A short-term physiological change does not establish improved control of type 2 diabetes. Glycated hemoglobin reflects glycemic exposure over approximately two to three months, while continuous glucose monitoring can characterize daily patterns, variability, and time spent within defined glucose ranges.\n\nThere is also an anatomical limitation to simplistic explanations. Conventional scalp tDCS produces distributed fields and does not selectively activate the deep hypothalamus in the way an implanted electrode might target a specific structure. Any effects on metabolic regulation may involve indirect network pathways, behavioral changes, peripheral stimulation, or combinations of these factors. Demonstrating one of these mechanisms requires more than observing that glucose changed after a session.\n\nA claim such as “one hour of electrical stimulation each day stabilizes blood sugar and reduces weight by 10%” cannot be treated as an established description of tDCS. A percentage weight change needs a defined starting population, comparison group, duration, and method of analysis; “stable blood sugar” needs a specified clinical measure. Findings from other electrical-stimulation methods, implanted devices, or animal experiments are not interchangeable with evidence for scalp-applied tDCS in humans."
    },
    {
      "heading": "Adverse effects, safety assessment, and home delivery",
      "text": "The most commonly reported adverse effects in conventional tDCS research are local or transient: tingling, itching, warmth, redness beneath electrodes, and discomfort. Headache and fatigue are also reported. Some symptoms occur in sham groups as well, making systematic collection and comparison important. The presence of a sensation does not prove effective brain stimulation, and the absence of sensation does not prove that no current is being delivered.\n\nSkin injury is an important preventable technical hazard. Poor electrode contact, uneven current distribution, drying conductive material, damaged skin, or unsuitable equipment can concentrate current at the skin interface. Electrode material and electrochemical reactions also matter. Consequently, findings from monitored studies using specified equipment cannot be generalized to arbitrary electrodes, improvised devices, or substantially different exposure schedules.\n\nSafety depends on more than current intensity. Duration, electrode area, placement, repeated exposure, and participant characteristics all affect the interpretation of risk. Extending a session from a studied duration to an hour changes total charge and may change biological effects in ways that cannot be predicted by simple linear scaling. Evidence about short treatment courses also does not resolve all questions about frequent use over months or years.\n\nStudies commonly screen for factors such as implanted electronic devices, relevant cranial surgery or hardware, skin disease at electrode sites, seizure history, and medications that may affect cortical excitability. These factors do not all represent identical or universal contraindications; rather, they influence eligibility and device-specific risk assessment. Research involving children, pregnancy, or particular neurological conditions requires population-specific evidence rather than extrapolation from healthy adults.\n\nSerious neurological events have not been common in conventional supervised research, but this observation does not establish that every protocol is harmless. Trials may exclude vulnerable populations and may be too small to quantify very rare events. Safety monitoring therefore includes neurological and psychiatric symptoms, skin inspection, adherence, and documentation of all adverse events, not merely those investigators initially consider device-related.\n\nHome delivery ranges from remotely supervised clinical research to unsupervised consumer use. Structured programs may incorporate device lockouts, programmed dosing, placement checks, contact monitoring, and electronic session records. These features help preserve protocol fidelity and enable oversight. They do not establish efficacy by themselves. A portable device can make repeated treatment more accessible, but clinical benefit, long-term safety, and the quality of supervision remain separate questions."
    },
    {
      "heading": "Clinical-trial design and unresolved evidence",
      "text": "A central challenge in tDCS research is constructing a convincing comparison condition. Sham stimulation often briefly ramps current up and then down to imitate initial sensations without delivering the full active exposure. Participants may nevertheless distinguish conditions, especially after repeated sessions or when active stimulation causes sustained sensations. Conversely, a sham arrangement that delivers more prolonged current may not be physiologically inert. Successful blinding must therefore be assessed rather than assumed.\n\nA well-designed randomized controlled trial separates stimulation effects from expectations, natural symptom fluctuation, and co-interventions. Allocation concealment prevents knowledge of the upcoming assignment from influencing enrollment. Blinded outcome assessment and standardized contact reduce additional biases. These details are especially important for subjective outcomes such as mood, craving, and pain, where expectations and interactions with research staff can influence reports.\n\nParameter flexibility creates another difficulty. Investigators can vary polarity, intensity, electrode arrangement, session timing, treatment length, and behavioral task. They can also select among many outcomes and follow-up points. Without prespecified hypotheses and analyses, this flexibility increases the chance that a positive result reflects selective reporting or chance. Small exploratory studies can identify possibilities, but they generally provide imprecise estimates of effect size.\n\nTrials must distinguish statistical significance from clinical importance. A small average difference on a symptom scale may be statistically detectable yet have limited implications for daily functioning. Confidence intervals describe uncertainty around estimates, while responder proportions, remission, participation, and sustained outcomes provide complementary information. Physiological measures can support a mechanistic interpretation, but a surrogate endpoint does not automatically establish patient benefit.\n\nRepeated-session studies also face adherence and attrition problems. Participants who complete a demanding course may differ from those who withdraw. Intention-to-treat analysis helps preserve the comparison created by randomization, although missing-data assumptions still matter. In crossover trials, after-effects may persist into later periods, making carryover and washout central design concerns.\n\nSeveral unresolved questions concern personalization. Investigators are examining whether anatomy-based field estimates, baseline network activity, symptoms, or treatment history can identify likely responders. Such prediction requires validation in participants outside the dataset used to develop the model. A reliable predictive marker would need to improve decisions beyond simpler clinical information, not merely correlate retrospectively with improvement.\n\nLong-term treatment questions remain distinct from short-term efficacy. Maintenance schedules, durability after stopping stimulation, interactions with medication and training, and the consequences of repeated home use require extended observation. Multicenter replication and standardized reporting of equipment, electrode arrangements, delivered exposure, adverse effects, and clinically meaningful outcomes are necessary to determine which findings generalize beyond a particular laboratory or treatment protocol."
    }
  ],
  "links": [
    "electric-field",
    "brain",
    "anesthesia",
    "cerebrospinal-fluid",
    "cell-membrane",
    "cerebral-cortex",
    "neuroplasticity",
    "electroencephalography",
    "magnetic-resonance-imaging",
    "gray-matter",
    "white-matter",
    "stroke",
    "functional-magnetic-resonance-imaging",
    "magnetoencephalography",
    "biomarker",
    "depression",
    "psychotherapy",
    "schizophrenia",
    "rehabilitation",
    "physical-therapy",
    "speech-language-therapy",
    "aphasia",
    "chronic-pain",
    "health-related-quality-of-life",
    "dementia",
    "hypothalamus",
    "metabolism",
    "obesity",
    "insulin-resistance",
    "type-two-diabetes",
    "glycated-hemoglobin",
    "continuous-glucose-monitoring",
    "headache",
    "fatigue",
    "pregnancy",
    "blinding-clinical-trials",
    "randomized-controlled-trial",
    "allocation-concealment",
    "confidence-interval",
    "surrogate-endpoint",
    "intention-to-treat-analysis",
    "crossover-trial"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}