Providing a life-saving treatment safely and systematically.
A structured course on electroconvulsive therapy for psychiatry trainees, nurses, medical students and GPs. Ten modules, worked cases, dosing calculators and self-assessment, written for Australian practice and referenced to the RANZCP guideline.
Modules
How to use this course
Work through the modules in order, then test yourself on the cases. Use the Reading level switch to hide or show the advanced material. Each module ends with a short quiz that gives instant feedback, followed by a 10-question assessment whose answers are on a separate page. Your quiz scores are kept in this browser so you can see what you have finished.
Boxes marked From the author's research point to studies by the author that bear directly on the topic in front of you. The full list is on the Author's research page.
This site is for education only. It does not replace your hospital's ECT policy, device manufacturer instructions, anaesthetic advice or the law in your state or territory. Stimulus parameters differ between devices; check your machine's manual before applying any number on this site.
Indications and work-up
Who benefits from ECT, how to prepare them safely, and what consent and the law require.
Lisanby SH. Electroconvulsive therapy for depression. A concise clinical overview built around a case: indications, evidence, technique, adverse effects and relapse prevention. Read it before starting this module.
N Engl J Med 2007;357:1939–45. doi:10.1056/NEJMct075234 · PMID 17989386
What is ECT? Core
Definition
Electroconvulsive therapy is the delivery of a train of brief electrical current pulses through electrodes on the scalp to induce a therapeutic seizure, given under general anaesthesia and muscle relaxation. In the consensus nomenclature for transcranial electrical stimulation, modern ECT is classed as a form of transcranial pulsed current stimulation, set apart from techniques such as tDCS and tACS by its purpose: it is the one form of transcranial stimulation intended to produce a seizure (Bikson et al., 2019).
| Feature | Modern ECT |
|---|---|
| Waveform | Constant-current, rectangular brief pulses, alternating in polarity (bidirectional pulse pairs) |
| Current | 800–900 mA (lower amplitudes are experimental) |
| Pulse width | 0.25–1 ms (ultrabrief < 0.5 ms) |
| Frequency and train | 10–120 Hz pulse pairs, train ≤ 8 s |
| Electrodes | 5 cm stainless-steel discs with gel, or adhesive pads |
| Dosing | Individualised relative to the patient's seizure threshold |
Bikson M, Esmaeilpour Z, Adair D, et al. Transcranial electrical stimulation nomenclature. Brain Stimul 2019;12:1349–66. doi:10.1016/j.brs.2019.07.010 · PMID 31358456
Another way to see it: non-focal deep brain stimulation
Classical ECT can also be understood as a form of brain stimulation in which a small amount of electricity, typically tens to a few hundred millicoulombs over a few seconds, is used only to trigger a seizure. The current itself acts mostly at the cortex near the electrodes. The therapeutic work is done by the seizure, which generalises and engages deep diencephalic and centrencephalic structures such as the thalamus and hypothalamus. Seen this way, ECT is a non-focal deep brain stimulation: it reaches the same deep circuits that implanted electrodes target, but through a generalised seizure rather than a wire.
This idea has a long history. Bolwig's review of how ECT works sets out three theories: that a generalised seizure involving deep centrencephalic structures is essential; that ECT normalises the neuroendocrine (hypothalamic–pituitary) dysfunction of melancholic depression; and that it promotes hippocampal neurogenesis and synaptogenesis. Bolwig concluded that the neuroendocrine, diencephalic theory has the strongest foundation (Bolwig, 2011).
Bolwig TG. How does electroconvulsive therapy work? Theories on its mechanism. Can J Psychiatry 2011;56:13–8. doi:10.1177/070674371105600104 · PMID 21324238
| Deep brain stimulation | Classical ECT | |
|---|---|---|
| Access | Surgically implanted electrodes | Scalp electrodes, non-invasive |
| Reach | Focal: one chosen nucleus or tract | Non-focal: the seizure spreads to diencephalic and centrencephalic structures |
| Role of the current | Directly modulates the target | Initiates the seizure; the seizure is the treatment |
| Schedule | Continuous or near-continuous | Intermittent: a few seconds, two or three times a week |
This view explains several practical points in the later modules: why a seizure must occur for ECT to work, why seizure generalisation and quality matter more than duration (Module 5), and why electrode placement can be changed to spare memory regions without losing efficacy, since the deep structures are reached through the seizure rather than by the current (Module 2).
When to consider ECT Core
ECT is the most effective acute treatment for severe depression. It is used first-line when speed matters or other treatments carry more risk, and later when medication has failed.
| Setting | Typical indications |
|---|---|
| First-line | Depression with psychotic features or melancholia; high suicide risk; refusal of food or fluids; catatonia; need for a rapid response; previous good response to ECT; patient preference |
| After other treatments | Treatment-resistant unipolar or bipolar depression; severe or refractory mania; schizophrenia, especially clozapine-resistant illness |
| Other | Neuroleptic malignant syndrome; some cases of Parkinson's disease with severe "off" states or psychiatric complications |
Who responds best
A meta-analysis of clinical predictors found that psychotic features and older age were associated with better response (Haq et al., 2015). Response is lower when ECT follows failed medication trials than when it is used earlier, although many medication non-responders still respond (Heijnen et al., 2010).
In a randomised trial of drug-naive patients with major depression, bilateral ECT and imipramine gave comparable improvement over four weeks, and ECT caused fewer subjective side effects.
Selvan CP, Mayur PM, Gangadhar BN, et al. Indian J Psychiatry 1999;41:228–35. PMID 21455395
ECT has a role beyond depression. Case series from the author's unit describe rapid improvement of severe mania with right unilateral ultrabrief ECT, and of clozapine-resistant psychotic symptoms with high-dose ultrabrief ECT.
Sidorov A, Mayur P. Australas Psychiatry 2017;25:10–12. doi:10.1177/1039856216663736 · Mayur P, Chakos A. J ECT 2019;35:e16–17. doi:10.1097/YCT.0000000000000561
Special indications: catatonia Advanced
Karl S, Sartorius A, Aksay SS. Catatonia and ECT across the lifespan. A review of ECT for catatonia in children, adolescents and older adults, with practical advice for each age group.
Schizophr Res 2024;263:246–51. doi:10.1016/j.schres.2023.04.004 · PMID 37087393
Catatonia is a psychomotor syndrome (stupor, mutism, posturing, negativism, rigidity, echophenomena, refusal to eat or drink) that occurs across psychiatric and medical illnesses, not only in schizophrenia. ECT is a safe and highly effective treatment for it, with high response rates (Karl et al., 2024).
Why it is missed and under-treated
- Underdiagnosis. Long classified as a subtype of schizophrenia, catatonia is easily missed elsewhere. In the young it overlaps with autism and developmental conditions; in older people with delirium, dementia and medical illness.
- Underuse of ECT. Clinicians cite uncertainty about mechanism, the lack of controlled trials, and general reluctance to use ECT in children and adolescents.
- The evidence. Studies in children, adolescents and older adults consistently show ECT for catatonia is safe and effective, with tolerability and efficacy at least as good as in adults. The authors conclude the evidence does not justify restricting ECT by age.
Practice points
| Situation | Approach |
|---|---|
| Recognition | Screen with a structured scale such as the Bush–Francis Catatonia Rating Scale; look for an underlying medical or psychiatric cause |
| First line | A lorazepam challenge and trial. A clear response supports the diagnosis. |
| When to use ECT | No or partial response to benzodiazepines; and urgently in malignant catatonia (fever, autonomic instability, rigidity), where delay can be fatal. Neuroleptic malignant syndrome is closely related and also responds to ECT. |
| Benzodiazepines during ECT | Often continued for catatonia, unlike in depression. Expect a higher seizure threshold; withholding the morning dose or giving flumazenil at induction can help. |
| Children and adolescents | Seizure thresholds are usually lower, so start titration low and watch for prolonged seizures. Check the extra legal and consent requirements for minors in your state or territory (Module 8). |
| Older adults | Higher thresholds, more medical comorbidity and a greater risk of post-ictal delirium. Plan anaesthesia carefully and monitor cognition (Module 8). |
The first three bullet points condense Karl et al. (2024). The practice points draw on standard clinical guidance; check them against your local protocol.
An early report described threshold-stimulus ECT in an adolescent.
Mayur PM, Kumar TC, Seshadri SP, Gangadhar BN. J Am Acad Child Adolesc Psychiatry 1997;36:1488. doi:10.1016/S0890-8567(09)66550-3
ECT in schizophrenia and mania Advanced
In clozapine-resistant schizophrenia, adding ECT to clozapine produced response in about half of patients, compared with none who continued clozapine alone (Petrides et al., 2015). ECT is also an established treatment for catatonia and for severe or refractory mania.
A historical commentary considers ultrabrief ECT in schizophrenia, and a case report describes rapid resolution of clozapine-resistant thought broadcast with high-dose RUL ultrabrief ECT.
Mayur P. J ECT 2020;36:e8–9. doi:10.1097/YCT.0000000000000647 · Mayur P, Chakos A. J ECT 2019;35:e16–17. doi:10.1097/YCT.0000000000000561
A report on threshold-stimulus ECT in an adolescent with depression.
Mayur PM, Kumar TC, Seshadri SP, Gangadhar BN. J Am Acad Child Adolesc Psychiatry 1997;36:1488. doi:10.1016/S0890-8567(09)66550-3
Contraindications and risk Core
There are no absolute contraindications. Some conditions raise the risk and need anaesthetic and medical input before treatment:
- Recent myocardial infarction, unstable angina, decompensated heart failure or severe valvular disease
- Raised intracranial pressure or a space-occupying lesion; recent stroke; known cerebral aneurysm or vascular malformation
- Phaeochromocytoma; high anaesthetic risk (ASA 4–5); severe pulmonary disease
- Retinal detachment, loose teeth or other problems made worse by a sudden rise in pressure
ECT is very safe. A systematic review estimated a mortality of about 2.1 deaths per 100,000 treatments (Tørring et al., 2017), lower than for many common procedures under general anaesthesia.
Pre-ECT work-up Core
| Domain | What to do |
|---|---|
| History | Psychiatric diagnosis and severity; prior ECT, including placement, dose and response; anaesthetic history and family history of anaesthetic problems; dental status |
| Examination | Full physical examination, cardiovascular and respiratory in particular; airway and dentition |
| Investigations | Full blood count, electrolytes and renal function; ECG for older patients or anyone with cardiac history; further tests guided by findings (e.g. chest X-ray, brain imaging if a neurological cause is suspected) |
| Baseline ratings | Depression rating (e.g. MADRS); cognitive screen (e.g. MoCA); record these so change can be measured |
| Medication review | Identify drugs that raise seizure threshold or interact with anaesthesia (see table below) |
| Anaesthetic review | Pre-anaesthetic assessment, fasting instructions, plan for regular medicines on treatment mornings |
Consent and the law Core
Informed consent covers the reason for ECT, the expected benefit, the number and frequency of treatments, electrode placement, common and serious side effects including memory effects, the alternatives, and the right to withdraw consent at any time. Give written information and record the discussion.
Each state and territory sets its own rules. In New South Wales, ECT for an involuntary patient, or for a voluntary patient who cannot give informed consent, requires a determination by the Mental Health Review Tribunal under the Mental Health Act 2007. Check the current legislation and your local policy before every course.
Medications during ECT Advanced
| Drug | Concern | Usual approach |
|---|---|---|
| Benzodiazepines | Raise seizure threshold; shorter, weaker seizures | Reduce or stop if possible; withhold the evening and morning doses before treatment; flumazenil at induction is an option with anaesthetic agreement |
| Anticonvulsants (valproate, lamotrigine, carbamazepine) | Raise seizure threshold | When used as mood stabilisers, consider reducing or withholding the dose before treatment; continue when prescribed for epilepsy |
| Lithium | Reports of post-ictal delirium and prolonged seizures | Keep levels in the low therapeutic range; many units withhold the dose the night before treatment |
| Antidepressants | Generally safe to continue | In OPT-ECT, nortriptyline improved ECT efficacy and reduced cognitive effects, while venlafaxine gave a smaller benefit and tended to worsen cognition (Sackeim et al., 2009). Continuing an antidepressant also eases the step to relapse prevention |
| Theophylline | Prolonged seizures and status epilepticus | Avoid or minimise |
| Cholinesterase inhibitors | Possible bradycardia and interaction with suxamethonium | Discuss with the anaesthetist |
In a randomised comparison, continuing antidepressants during a course of unilateral ECT gave no therapeutic advantage over a drug placebo, and side effects were similar apart from anticholinergic effects of tricyclics. OPT-ECT, a larger later trial in a medication-resistant sample, found the reverse for nortriptyline. Weigh both when deciding what to continue.
Mayur PM, Gangadhar BN, Subbakrishna DK, Janakiramaiah N. J Affect Disord 2000;58:37–41. doi:10.1016/s0165-0327(99)00096-8
A single case report describes an interaction between ECT and donepezil in an older patient.
Bhat RS, Mayur P, Chakrabarti I. Int J Geriatr Psychiatry 2004;19:594–5. doi:10.1002/gps.1085
Typical cardiovascular response Advanced
The stimulus causes an immediate vagal (parasympathetic) response, with bradycardia and sometimes brief asystole, most marked when no seizure follows, as during titration. The seizure then produces a sympathetic surge with tachycardia and hypertension that settles as the seizure ends. The rate–pressure product (heart rate × systolic pressure) is a simple index of myocardial oxygen demand.
Anticholinergic premedication
Practice differs between countries. In Australia, atropine or glycopyrrolate is not routinely used and is usually reserved for selected patients. In the United States an anticholinergic, usually glycopyrrolate, has long been given routinely; early US studies compared atropine with glycopyrrolate in routine use and found glycopyrrolate caused less supraventricular tachycardia (Swartz and Saheba, 1989), and that glycopyrrolate shortened post-stimulus pauses in cardiac rhythm (Shettar et al., 1989).
The case for selective use: atropine raises the post-stimulus rate–pressure product, increases tachyarrhythmias, and in a randomised crossover trial worsened post-ECT cardiac autonomic dysfunction. The authors recommended restricting it to patients susceptible to bradyarrhythmia (Jadhav et al., 2017). Consider an anticholinergic for:
- Titration sessions, or any stimulus likely to be subconvulsive
- Right unilateral placement, where stimulus-related asystole is more common than with bifrontal placement (Nagler, 2010)
- Patients on beta-blockers, or with a history of bradycardia, heart block or asystole at a previous treatment
- Excess secretions
Without atropine premedication, the post-stimulus rate–pressure product was significantly lower and there was no clinically significant bradyarrhythmia. When threshold titration is not needed, leaving out atropine can limit the cardiovascular surge.
Mayur PM, Shree RS, Gangadhar BN, et al. Br J Anaesth 1998;81:466–7. doi:10.1093/bja/81.3.466
Bilateral ECT raised the post-ictal rate–pressure product more than right unilateral ECT (31% vs 20%), and suprathreshold stimuli raised it more than threshold stimuli, suggesting the cardiovascular response tracks the therapeutic potency of the treatment. In the same NIMHANS group, a higher rate–pressure product without atropine predicted early antidepressant response, consistent with the response reflecting diencephalic seizure generalisation (Saravanan et al., 2002).
Mayur PM, et al. J ECT 1998;14:94–8. PMID 9641805 · Gangadhar BN, Mayur PM, et al. J ECT 2000;16:177–82. doi:10.1097/00124509-200006000-00009
Technique and dosing
Electrode placement, the stimulus, seizure threshold and how to choose a dose.
Sackeim HA, Prudic J, Devanand DP, et al. A prospective, randomized, double-blind comparison of bilateral and right unilateral electroconvulsive therapy at different stimulus intensities. The landmark trial showing that right unilateral ECT works only at high multiples of the seizure threshold, and that at six times threshold it matches bilateral ECT with fewer cognitive effects. It is the basis for threshold titration and dosing as taught in this module.
Arch Gen Psychiatry 2000;57:425–34. doi:10.1001/archpsyc.57.5.425 · PMID 10807482
The procedure in outline Core
Modern ECT uses a small dose of electricity, less than 1 ampere in amplitude, delivered to the brain through electrodes placed on the scalp to produce a therapeutic seizure. The patient is unconscious throughout, having received an ultrashort-acting anaesthetic and a muscle relaxant. The procedure is painless.
This module follows the treatment in the order it is planned: where to place the electrodes, how to set the electrical stimulus, how to find the seizure threshold and choose a dose, and how many treatments to give and how often. Anaesthesia is covered in Module 3.
Electrode placement Core
Standard electrodes are 5 cm (2-inch) discs. Mark the landmarks before the patient is anaesthetised: the outer (lateral) canthus of each eye, the tragus of each ear, the nasion, and the vertex, which is where the nasion–inion line crosses the line joining the two tragi.
| Placement | Electrode 1 | Electrode 2 |
|---|---|---|
| Bitemporal (BT) | One on each side. The lower edge of the electrode sits at the midpoint of the line from the tragus to the outer canthus, so its centre is about 2.5–3 cm above that point. | |
| Right unilateral, d'Elia (temporoparietal) | Right temporal, in the bitemporal position | Centred just to the right of the vertex (about 2.5 cm right of the midline), on the line joining the tragi |
| Bifrontal (BF) | Two concave electrodes, each centred 5 cm above the outer canthus, on a vertical line through the canthus | |
Placement trades speed and robustness of response against cognitive effects. Right unilateral ECT spares the dominant hemisphere's verbal memory systems, but works well only at high multiples of the seizure threshold. Bitemporal ECT works at lower multiples and often acts faster, but carries more cognitive burden. Bifrontal ECT sits between them. Because of that cognitive burden, this course treats bilateral placements as a last resort, after right unilateral ECT has been optimised (appendix). Keep at least a few centimetres of scalp between the electrode edges; electrodes placed too close let current shunt across the scalp instead of passing through the brain.
The landmark comparison showed that low-dose RUL ECT was markedly less effective, while RUL at six times threshold matched bilateral ECT in efficacy with fewer cognitive effects (Sackeim et al., 2000).
Urgent situations: the question of speed
The main argument for bitemporal ECT is speed. In a randomised trial of 230 patients, right unilateral ECT at six times threshold, bifrontal ECT and bitemporal ECT (both at 1.5 times threshold) gave similar remission rates: 55%, 61% and 64%. Bitemporal ECT produced a faster fall in symptoms early in the course, and the authors recommended it as the preferred placement for urgent situations (Kellner et al., 2010).
Set against this:
- End-of-course efficacy was the same, so the advantage is one of speed only.
- The trial gave right unilateral ECT at six times threshold. Response to right unilateral ECT keeps improving with dose through at least 12 times threshold (McCall et al., 2000), so a higher dose may narrow the gap. This has not been tested directly.
- In older adults, ultrabrief right unilateral ECT produced remission after a mean of 7.3 treatments (Kellner et al., 2016a).
This course's approach is to start with right unilateral ECT even in urgent cases, using brief pulses when speed matters most, to review response early, and to shift gears promptly (appendix). Know the Kellner trial, because many guidelines and units follow it and start with bitemporal ECT when life is at risk.
Newer placements Advanced
Two placements aim to keep the electric field away from the hippocampi, which computational modelling links to memory effects, while still stimulating frontal and midline regions linked to efficacy. Both are still being studied and are not yet standard practice.
| Placement | Anterior electrode | Posterior electrode |
|---|---|---|
| Fronto-medial (FM) | On the midline of the forehead, centred about 2.5 cm above the nasion (the "MFV2" position; a higher variant sits 5 cm above) | At the vertex, on the midline |
| Fronto-parietal (FP), right unilateral | Lower edge just above the midpoint of the right eyebrow | Just to the right of the vertex, as for d'Elia RUL |
Fronto-medial. Rotating the electrode axis into the sagittal plane puts the anterior cingulate between the electrodes and keeps both hippocampi far from them. In modelling, this placement cut hippocampal field strength two- to three-fold compared with bitemporal ECT and raised it in the dorsal anterior cingulate (Steele et al., 2022).
Fronto-parietal. First described by d'Elia, this moves the anterior RUL electrode from the temple to the brow. In a small comparison, reorientation was faster than with temporoparietal RUL, with similar seizures (Loo et al., 2015). A pilot randomised crossover study found less visual retrograde memory impairment, especially with ultrabrief pulses (Martin et al., 2021), and modelling shows less hippocampal stimulation than standard RUL (Martin et al., 2026).
A prospective pilot case series examined fronto-medial ultrabrief ECT in major depression.
Mayur P, Harris A, Serrao S, Fernandes J. Brain Stimul 2026;19:103237. doi:10.1016/j.brs.2026.103237
Electrical parameters Core
Modern devices deliver a constant-current train of rectangular pulses that alternate in direction. Four settings define the stimulus: current (usually 800–900 mA), pulse width, frequency and train duration. Together they set the total charge, in millicoulombs (mC), which is how dose is recorded and titrated.
Of the four, pulse width has the biggest effect on how efficient the stimulus is and on cognitive side effects. Four widths are in use today:
| Pulse width | Category | Where it fits |
|---|---|---|
| 0.25–0.3 ms | Ultrabrief | Cognition-sparing. Use with right unilateral placement at about six times threshold, not with bilateral placement. |
| 0.5 ms | Brief (the "orphan") | A common device default, but never tested in the trials that established brief-pulse or ultrabrief ECT. |
| 1.0 ms | Brief (standard) | The reference standard for efficacy, for any placement. |
| 1.5 ms | Brief (long) | Now an outlier. Used in some older trials and centres, but less efficient than 1.0 ms with no evidence that it works better. |
Sine-wave stimulation is obsolete. Use the charge calculator to see how the four settings combine.
Pulse width in depth Advanced
Why shorter pulses are more efficient
Only a small part of each pulse is needed to make a neuron fire. The parameter that matters is chronaxie, the pulse duration at which stimulation is most efficient. Working in slices of cortical grey matter, Nowak and Bullier showed that extracellular electrical stimulation activates axons, not cell bodies: the chronaxie for axonal activation was about 40 times shorter than the chronaxie for direct activation of a cell body (Nowak and Bullier, 1998). Axonal chronaxie is therefore a fraction of a millisecond. A pulse much longer than that delivers charge which adds little to neuronal firing but still spreads through the brain, where it is thought to add to cognitive effects.
Nowak LG, Bullier J. Axons, but not cell bodies, are activated by electrical stimulation in cortical gray matter. I. Evidence from chronaxie measurements. Exp Brain Res 1998;118:477–88. doi:10.1007/s002210050304 · PMID 9504843
The ECT stimulus has moved steadily toward that value: from sine wave (about 8 ms per half-cycle), to 1.0–1.5 ms brief pulses, to 0.25–0.3 ms ultrabrief pulses.
Because shorter pulses are more efficient, the seizure threshold in millicoulombs is usually lower at shorter widths. Re-titrating is the ideal whenever you change pulse width, because a multiple of threshold at one width does not carry over to another. It matters most when shortening the pulse. When widening it, for example from 0.3 ms to 0.5 or 1.0 ms with the other settings unchanged, the charge rises in proportion, so in practice re-titration may be unnecessary. On a Thymatron, where the dose is set as a percentage, the charge can stay the same, so re-titrate (appendix).
Each width in practice
- Ultrabrief (0.25–0.3 ms). Effective with high-dose right unilateral placement, with the least cognitive impairment. Ultrabrief bilateral ECT was clearly less effective in the key trial (35% remission) and should not be used routinely (Sackeim et al., 2008). It works best under set conditions: right unilateral placement, six times threshold, three times a week (Module 4).
- 0.5 ms. Technically brief pulse, but outside the range of the defining trials (1.0–1.5 ms) and too long to be ultrabrief. Observational studies disagree on whether it matches 1.0 ms; registry data suggest a lower response rate (Braithwaite and Jelovac, 2026). Choose it deliberately, not just because it is the default.
- 1.0 ms. The benchmark: bitemporal, bifrontal and right unilateral ECT were all established at 1.0–1.5 ms. Use it when efficacy matters most, or after a slow response to ultrabrief or 0.5 ms treatment.
- 1.5 ms. Used for the brief-pulse arms of the key pulse-width trial (Sackeim et al., 2008) and in some trials in schizophrenia (Ithal et al., 2019). It is now an outlier. It delivers more charge beyond the chronaxie than 1.0 ms, and there is no good evidence that it is more effective.
Module 4 sets out the trial evidence for each pulse width and the experimental work on lowering current amplitude.
Seizure threshold and dosing Core
The seizure threshold is the smallest charge that produces an adequate seizure. It varies up to forty-fold between people. It is higher in older people and men, with bilateral placement, and with threshold-raising drugs, and it usually rises during a course.
Two approaches set the first dose:
- Empirical titration at the first session: start low and restimulate at increasing doses until a seizure occurs. The treatment dose is then a multiple of that threshold. This is the approach recommended in Australian practice.
- Age-based or fixed dosing: the dose is estimated from age. Simpler, but less precise, and patients often receive more charge than they need.
| Placement and pulse | Usual dose relative to threshold |
|---|---|
| RUL, ultrabrief (0.3 ms) | 6 × threshold |
| RUL, brief pulse | 5–6 × threshold |
| Bitemporal, brief pulse | 1.5–2.5 × threshold |
| Bifrontal, brief pulse | 1.5–2 × threshold |
In 80 patients receiving unilateral ECT, age was the only significant predictor of threshold and explained just 15% of its variance. Applied prospectively, the age formula would have worked for 73% of patients, but most would have received more than the recommended dose. The authors concluded that titration is preferred.
Girish K, Mayur PM, Saravanan ES, et al. J ECT 2000;16:258–62. doi:10.1097/00124509-200009000-00006
Length of a course Core
Treatments are usually given two or three times a week. A typical acute course for depression is 6 to 12 treatments. Continue until remission, or until improvement has plateaued over two or three treatments, and reassess with a rating scale at least weekly.
Course length also depends on the diagnosis:
- Mania usually needs a shorter course. A 50-year review found that manic patients do not need high-frequency or prolonged treatment to respond, and that their seizure threshold tends to be lower than in depression (Mukherjee et al., 1994).
- Schizophrenia usually needs a longer course, sometimes 20 to 24 treatments. In a large prospective series of treatment-resistant schizophrenia treated with ECT and flupenthixol, index courses ran to 20 treatments, with seizure threshold re-measured at the 20th treatment (Chanpattana et al., 2000). In a clinical series, courses for schizophrenia averaged 13.3 treatments against 9.4 for mania, with a range up to 24 (Ittasakul et al., 2020).
Frequency per week Advanced
The best weekly frequency depends on the type of ECT being given.
Bitemporal and brief-pulse ECT: twice weekly is enough
For brief-pulse ECT, twice weekly is as effective as three times weekly and causes fewer cognitive effects.
- In a double-blind trial of bilateral brief-pulse ECT, twice and three times weekly gave the same final outcome. Three times weekly worked faster but caused more cognitive impairment (Lerer et al., 1995). A follow-up trial from the same group confirmed this and concluded that twice weekly is the optimum schedule for bilateral ECT unless a rapid response is needed (Shapira et al., 1998).
- In melancholia, twice and three times weekly ECT gave the same outcome over four weeks and at six months, even though the thrice-weekly group received more treatments (Gangadhar et al., 1993).
- Once weekly is too infrequent. It was less effective than three times weekly, and a higher stimulus dose did not make up for it (Janakiramaiah et al., 1998).
- A meta-analysis confirmed similar efficacy for twice and three times weekly. Twice weekly needed slightly fewer treatments but took about five to six days longer to complete (Charlson et al., 2012).
In practice, use three times weekly when speed matters most: high suicide risk, refusal of food or fluids, catatonia, or severe physical deterioration. Otherwise twice weekly is the usual choice.
Ultrabrief right unilateral ECT: the question is open
This evidence may not apply to ultrabrief RUL ECT. Ultrabrief treatment already spares cognition and works more slowly, so reducing the frequency may cost efficacy without much cognitive gain. The optimum frequency for ultrabrief ECT has not yet been worked out systematically.
- In a double-blind trial that gave RUL ECT twice weekly at eight times threshold, ultrabrief pulses gave a lower remission rate than brief pulses (49% vs 68%) and needed more treatments, with no cognitive advantage (Spaans et al., 2013). This suggests twice weekly may be too infrequent for ultrabrief RUL.
- Accelerated schedules. Ultrabrief RUL ECT has been given five times a week ("daily") for up to two weeks. In an open pilot study, response and remission rates matched those of standard schedules, with no clinically significant effect on retrograde memory (Rasmussen et al., 2016). In a retrospective comparison, daily ultrabrief RUL ECT finished the course 6.5 days sooner than bitemporal ECT, with similar efficacy and a similar number of treatments (Shafi et al., 2018). Randomised trials are still needed.
Three times weekly remains the usual starting schedule for ultrabrief RUL ECT. Move to twice weekly as the patient improves, or if cognitive effects build up.
Anaesthesia
Why ECT needs a brief general anaesthetic and muscle relaxation, how oxygenation is maintained, the agents and doses used, and anaesthetic techniques that improve the seizure.
Anaesthesia and oxygenation Core
Treatment is given under brief general anaesthesia with full muscle relaxation.
- Why muscle relaxation. Without it, the seizure produces an intense tonic contraction of the whole body followed by clonic jerking, which can fracture vertebrae and long bones, dislocate joints and injure the mouth. The relaxant softens the seizure to a mild, often barely visible movement, which is why a cuffed limb is used to watch it.
- Why general anaesthesia. The relaxant also paralyses the breathing muscles. The anaesthetic makes sure the patient is never aware of being paralysed or of being unable to breathe, and is not awake for the stimulus. It only needs to last a few minutes, until the relaxant wears off.
Common induction agents are propofol, thiopentone, etomidate and ketamine (and methohexitone in the United States). Suxamethonium is the usual relaxant.
Pre-oxygenation and ventilation
Once the relaxant takes effect, the patient cannot breathe until it wears off, usually for several minutes. The seizure itself sharply increases the brain's oxygen consumption and blood flow (Gaines and Rees, 1992). Oxygen therefore has to be built up beforehand and maintained throughout:
| Stage | What happens |
|---|---|
| Monitoring | Pulse oximetry, ECG and blood pressure from the start; EEG and the cuffed limb for the seizure. Capnography is increasingly used. |
| Pre-oxygenation | 100% oxygen through a close-fitting face mask for about three minutes while the patient breathes normally, before the anaesthetic is given. This fills the lungs with an oxygen reserve. |
| Induction and relaxation | Anaesthetic, then suxamethonium. The anaesthetist ventilates by bag and mask until the relaxant has taken full effect. |
| Stimulus | Ventilation pauses, a bite block is placed to protect the teeth and tongue, and the stimulus is given. |
| Seizure and recovery | Bag-and-mask ventilation resumes as soon as the stimulus ends and continues until spontaneous breathing returns. The patient recovers in the recovery area with oxygen and monitoring until awake and oriented. |
During the COVID-19 pandemic, anaesthetic induction for ECT was treated as aerosol-generating. The author's unit described changes to PPE and anaesthetic technique to reduce transmission risk.
Ramakrishnan VS, Kim YK, Yung W, Mayur P. Australas Psychiatry 2020;28:527–9. doi:10.1177/1039856220953705
Anaesthetic agents and doses Advanced
Choice and dose are decided by the anaesthetist, but the psychiatrist needs to know how each agent affects the seizure. Use the lowest dose that gives reliable unconsciousness: most induction agents are anticonvulsant, and a deeper anaesthetic means a shorter, weaker seizure.
The doses and ratings below are those given in the APA Task Force report on the practice of ECT, 3rd edition (2024). The rating scores each agent's anticonvulsant effect: the higher the number the more anticonvulsant, 0 is neutral, and a negative number means the agent is proconvulsant. Agents are listed from most to least anticonvulsant.
| Agent | IV dose | Anticonvulsant rating | Notes |
|---|---|---|---|
| Propofol | 1.0–2.0 mg/kg | +3 | The most anticonvulsant of the induction agents, and the most common in Australia and the UK. Shortens seizures in a dose-dependent way; blunts the hypertensive surge. |
| Thiopentone | 2–3 mg/kg | +2 | More bradycardia and ectopic beats than methohexitone (Mokriski et al., 1992). |
| Methohexitone (methohexital) | 0.5–1.2 mg/kg | +1 | The traditional agent of choice in the United States (Gaines and Rees, 1992). Least anticonvulsant of the barbiturates, with fewer arrhythmias than thiopentone. Not readily available in Australia. |
| Etomidate | 0.15–0.4 mg/kg | 0 | Neutral, and so gives longer seizures. Useful when seizures are short or the threshold is high. Myoclonus and adrenal suppression. |
| Ketamine | 1.0–2.0 mg/kg | −1 | Proconvulsant. More hypertension and tachycardia. A randomised trial found no faster antidepressant response than methohexitone (Rasmussen et al., 2014). |
| Remifentanil (opioid adjunct) | 2.5–3.5 µg/kg | −1 | Added to allow a lower dose of the induction agent, and so a longer seizure (Nishikawa et al., 2011). |
| Alfentanil (opioid adjunct) | 10–20 µg/kg | −1 | Used in the same way as remifentanil. |
The two opioids are adjuncts, not agents to be used alone. Reading down the table gives a practical ladder: when seizures are short or the threshold is rising, move from propofol toward a less anticonvulsant agent, or add an opioid so that less induction agent is needed.
Suxamethonium dosing
The dose is a balance. Too little leaves a strong convulsion; too much prolongs apnoea without benefit. The guidelines frame it differently:
| Source | Suxamethonium | Emphasis |
|---|---|---|
| APA Task Force report (US, 3rd ed, 2024) | 0.5–1.0 mg/kg | A fixed weight-based range, unchanged from the 2nd edition and consistent with US anaesthetic practice (Gaines and Rees, 1992). Aim for a substantially but not fully modified motor seizure, visible in a cuffed limb. |
| RCPsych ECT Handbook (UK) | 0.5–1.0 mg/kg | Similar range; individualise and adjust from one treatment to the next according to how well the convulsion was modified. |
| RANZCP (Weiss et al., 2019) | Leaves the agent and dose to the anaesthetist | Joint psychiatrist–anaesthetist responsibility; adequate modification and monitoring rather than a prescribed dose. |
The research supports the upper end of the range and wide individual variation:
- In a randomised crossover trial, the minimum effective dose of suxamethonium was 0.85 mg/kg for 50% of patients and 1.06 mg/kg for 90%. Rocuronium, reversed with neostigmine or sugammadex, was an alternative at 0.41–0.57 mg/kg (Mirzakhani et al., 2016).
- In 500 patients, the mean optimal dose was 0.9 mg/kg, but over a third needed adjustment after the first treatment, and outliers needed anywhere from 0.29 to 2.1 mg/kg (Bryson et al., 2018).
- In a randomised crossover comparison, 1.0 mg/kg modified the convulsion acceptably in 85% of treatments against 59% with 0.5 mg/kg, with no difference in side effects (Aaron et al., 2020).
Use more (around 1 mg/kg or above) when there is osteoporosis, recent fracture, or other risk from a strong convulsion. Avoid suxamethonium, and use rocuronium instead, in hyperkalaemia, recent burns or crush injury, prolonged immobility, neuromuscular disease, malignant hyperthermia susceptibility, or known pseudocholinesterase deficiency. Guideline wording changes between editions, so check the current versions.
Oxygenation and advanced anaesthetic techniques Advanced
Oxygenation and airway
- How much pre-oxygenation. In a small pilot study using the oxygen reserve index, three minutes of pre-oxygenation built a larger oxygen reserve than one minute or none. Non-obese patients did not desaturate even without it (Kadoi et al., 2023). Obese patients, older patients and those with lung disease desaturate faster, so standard pre-oxygenation is the safe default for everyone.
- High-flow nasal oxygen (THRIVE). In a randomised trial of 70 patients, high-flow humidified nasal oxygen kept the lowest oxygen saturation higher than face-mask ventilation, with no difference in seizure duration, stimulus dose or haemodynamics (Deng et al., 2026). It has also prevented desaturation in a patient who had previously dropped to 50% saturation during ECT (Toyota et al., 2024). It is an option for patients who desaturate despite good mask technique.
- Machine-assisted ventilation. In a randomised trial of 110 patients, bilevel positive airway pressure (BPAP) instead of hand-held mask ventilation shortened turnover time and reduced strain on the anaesthetist, with no increase in hypoxaemia (Li et al., 2026).
- Hyperventilation. Not needed routinely: in a randomised trial, the ventilation rate during induction did not affect seizure quality (Taylor et al., 2020). It can still be trialled as a troubleshooting step when seizures are missed or short (Module 5).
Timing: the anaesthetic–ECT interval
Most induction agents are anticonvulsant, and their effect is greatest when their brain concentration peaks shortly after injection. Giving the stimulus later produces a better seizure. In 771 sessions with propofol, longer intervals between the anaesthetic and the stimulus gave significantly better seizure amplitude, regularity and post-ictal suppression (Gálvez et al., 2016). The same was found with thiopentone (Taylor et al., 2019), and a randomised trial confirmed that a longer interval improved seizure quality (Taylor et al., 2020). Agree a target interval with the anaesthetist and record it, especially when interpreting a poor EEG before raising the dose.
Depth of anaesthesia
- BIS monitoring. In a meta-analysis of seven studies (1,283 sessions), a higher, lighter pre-stimulus bispectral index was associated with longer EEG and motor seizures (Guerrier and Gianni, 2019). Using the lightest anaesthesia that ensures unconsciousness helps the seizure.
- Lighter anaesthesia in practice. When anaesthetic doses were reduced during the COVID-19 pandemic (methohexitone 0.82 vs 0.87 mg/kg; suxamethonium 0.33 vs 0.53 mg/kg), response and cognitive outcomes were unchanged. However, insufficient initial sedation and post-ictal agitation were about twice as common (Espinola et al., 2025). There is a lower limit.
Anaesthetic combinations
- Ketofol (ketamine with propofol). In 1,077 sessions, ketofol and etomidate were associated with a smaller fall in seizure duration and a smaller rise in stimulus charge over a course than propofol or thiopentone. Ketofol also needed fewer restimulations (Gurel et al., 2022). A combination of S-ketamine with propofol has been used successfully across more than 900 sessions (Sartorius et al., 2021). Consider it for patients whose seizures are short or whose threshold is high.
- Remifentanil allows a lower propofol dose (see table above).
Controlling the cardiovascular surge
A short-acting beta-blocker given before induction blunts the post-ictal rise in heart rate and blood pressure. In a crossover trial, landiolol 0.1 mg/kg did this without shortening the seizure, but 0.2 mg/kg shortened it (Nomoto et al., 2006). Use the lowest effective dose. Esmolol and labetalol are commonly used.
Relaxant alternatives and recovery
- Rocuronium with sugammadex. When suxamethonium is contraindicated, rocuronium can be reversed rapidly with sugammadex, giving relaxation comparable to suxamethonium in duration (Arora et al., 2025).
- Muscle pain. Suxamethonium fasciculations cause myalgia after treatment. In a randomised trial, low-dose dexmedetomidine (0.25 µg/kg) reduced fasciculations and myalgia without haemodynamic change (Sriramka et al., 2024).
- Post-ictal agitation. Agitation on waking is more likely with too little anaesthetic (Espinola et al., 2025). Treat with a small dose of propofol or a benzodiazepine, and give it prophylactically at later treatments if it recurs.
Pulse width and stimulus refinements
The trial evidence behind the pulse widths introduced in Module 2: from 1 ms brief pulses to 0.3 ms ultrabrief pulses, the uncertain place of 0.5 ms, and experimental work on lowering the current itself.
Sackeim HA, Prudic J, Nobler MS, et al. Effects of pulse width and electrode placement on the efficacy and cognitive effects of electroconvulsive therapy. The four-arm trial that defined ultrabrief right unilateral ECT and showed that ultrabrief bilateral ECT does not work well.
Brain Stimul 2008;1:71–83. doi:10.1016/j.brs.2008.03.001
From 1 ms to 0.3 ms: the evidence Core
In the key trial, patients were randomised to right unilateral ECT at six times threshold or bilateral ECT at 2.5 times threshold, each with a 1.5 ms or 0.3 ms pulse. Remission was 73% with ultrabrief RUL, 65% with brief-pulse bilateral, 59% with brief-pulse RUL, and only 35% with ultrabrief bilateral ECT. Ultrabrief RUL had the fewest cognitive effects on almost every measure (Sackeim et al., 2008).
Two lessons follow. Ultrabrief pulses can be cognition-sparing and effective when paired with high-dose right unilateral placement. Ultrabrief bilateral ECT should not be used routinely.
Where ultrabrief ECT fits
Ultrabrief right unilateral ECT is a first-line option in depression when treatment has to balance protecting cognition against producing a clinical response. It should be trialled under the conditions in which it is known to work. The patient conditions below reflect well-established clinical opinion:
| Condition | Detail |
|---|---|
| Psychotic depression | The good response of psychotic depression to ECT (Haq et al., 2015) holds with ultrabrief pulses. |
| Younger patients | Ultrabrief pulse is an imperative first line, to preserve cognition. |
| Older patients | In the PRIDE study, 240 older adults received ultrabrief right unilateral ECT at six times threshold, three times a week, with venlafaxine. Remission was 62%, reached after a mean of 7.3 treatments, with good safety and tolerability (Kellner et al., 2016a). |
| Placement | Right unilateral |
| Dose | 6 × seizure threshold |
| Frequency | Three times a week |
If these conditions are met and the patient has still not responded, defined as less than 25% change in the depression score after about six to eight treatments, optimise the delivery of ECT. The first step is to shift to brief-pulse treatment: 0.5 ms first, then 1.0 ms. Bilateral placements are a last resort. The appendix sets out the full sequence for shifting gears.
The author's double-blind randomised trials compared 0.3 ms ultrabrief with 1.0 ms brief-pulse RUL ECT, both at six times threshold, thrice weekly. Depression improved equally in both arms, with similar median time to remission (26 vs 28 days). On autobiographical memory, semantic memory improved with ultrabrief treatment but worsened with brief-pulse treatment after eight sessions, and a small advantage for early-adult memory persisted at three months.
Mayur P, Harris A. J ECT 2011;27:e59–60. doi:10.1097/YCT.0b013e318218945e · Mayur P, Byth K, Harris A. J Affect Disord 2013;149:426–9. doi:10.1016/j.jad.2012.12.005 · Mayur P, Byth K, Harris A. J ECT 2013;29:277–82. doi:10.1097/YCT.0b013e3182941baf
The trade-off: efficacy against cognition Advanced
Later trials and meta-analyses have questioned how well ultrabrief ECT performs:
- A meta-analysis of six trials (689 patients) found brief-pulse RUL slightly more effective than ultrabrief RUL, with a higher remission rate and about one fewer treatment (8.7 vs 9.6), but with more cognitive effects in every domain examined. The authors advised weighing efficacy against cognitive sparing for each patient (Tor et al., 2015).
- A double-blind trial giving RUL ECT twice weekly at eight times threshold found lower remission with ultrabrief pulses (49% vs 68%), more treatments, and no measurable cognitive advantage (Spaans et al., 2013). The twice-weekly schedule differs from the conditions set out above.
- A 2025 meta-analysis of 30 studies found a pooled remission rate of only 32% with high-dose ultrabrief RUL ECT. More than a quarter of patients (28%) were switched to brief-pulse ECT, and 44% of remitters relapsed or withdrew within six months. The authors concluded that ultrabrief ECT cannot be regarded as "state of the art" (Jelovac and McLoughlin, 2025).
The author's reading of this evidence
It is premature to discard ultrabrief ECT as a first-line option on the strength of one meta-analysis that has a number of critical limitations. For example, it pooled remission rates from randomised and observational studies together, rather than comparing treatments head to head. Trials that gave right unilateral ECT at six times threshold, including the author's own thrice-weekly trials, found an antidepressant response equal to brief-pulse ECT with a cognitive advantage (Sackeim et al., 2008; Mayur et al., 2013).
The practical answer is to start ultrabrief ECT under the right conditions, measure response, and shift gears promptly when there is non-response (appendix).
The orphan category: 0.5 ms Advanced
A pulse width of 0.5 ms falls between the two categories that have been tested. It is technically "brief pulse" (0.5 ms or more), but it was never evaluated in the trials that established brief-pulse efficacy, which used 1.0–1.5 ms, and it is too long to be ultrabrief. It is nevertheless the default setting on a commonly used ECT device, and in Sweden 72% of patients treated with unilateral ECT received 0.5 ms (Ernstsson et al., 2024).
The evidence is mixed and mostly observational:
- No difference. In a matched comparison of bilateral ECT for depression, 0.5 ms and 1 ms gave similar improvement, response (96% vs 98%), remission (87% vs 81%) and number of treatments (de Arriba-Arnau et al., 2019). Swedish register data showed no difference in remission or subjective memory at six months between <0.5, 0.5 and >0.5 ms (Tornhamre et al., 2020).
- Lower response. In UK and Irish registry data from 1,956 patients, mostly on bilateral ECT, both 0.5 ms and 0.25–0.3 ms were associated with lower odds of response than 1.0 ms. Predicted response was 73% with 1.0 ms, 64% with 0.5 ms and 61% with ultrabrief. The authors recommended 1.0 ms as standard unless cognitive sparing is paramount (Braithwaite and Jelovac, 2026).
- Threshold. With RUL titration, seizure thresholds at 0.5 ms and 0.3 ms overlapped substantially (Rosa et al., 2013).
- Schizophrenia. A randomised trial has compared 0.5 ms with 1.5 ms bilateral ECT in schizophrenia (Ithal et al., 2019), with comparable cardiovascular responses (Ithal et al., 2020).
Without randomised trials in depression, 0.5 ms cannot be assumed to be as effective as 1.0 ms. If 0.5 ms is the device default, choose the pulse width deliberately rather than by default.
Experimental: low-amplitude stimulation Advanced
Current amplitude, not total charge, sets the strength of the electric field in the brain. Clinical ECT is fixed at 800–900 mA, a value chosen for historical and engineering reasons rather than evidence. Lowering the amplitude produces a weaker, more focal field, which may reduce cognitive effects.
- In non-human primates, seizures could be induced at 112–174 mA across bilateral, bifrontal, right unilateral and fronto-medial placements, far below clinical amplitudes (Peterchev et al., 2015). At these titrated amplitudes the stimulated brain volume fell from 94–99% to 23–63% (Lee et al., 2016).
- In an open-label proof-of-concept trial in 22 patients, low-amplitude seizure therapy at 500–600 mA reliably induced seizures, with a median reorientation time of 4.5 minutes and improvement in both depressive and psychotic symptoms (Youssef and Sidhom, 2017).
A proof-of-concept report described successful ECT at 500 mA, below the usual 800–900 mA.
Mayur P, Harris A, Gangadhar B. J ECT 2015;31:e23–6. doi:10.1097/YCT.0000000000000214
Experimental: amplitude-based seizure threshold Advanced
Conventional titration varies the charge at a fixed current. An alternative is to fix the train and increase the current amplitude until a seizure occurs, giving a threshold in milliamperes that reflects how strongly each patient's brain is stimulated.
- In non-human primates, a single-pulse motor threshold measured through the ECT electrodes explained 63% of the variance in amplitude-titrated seizure threshold, more than any other known predictor. Motor threshold could become a faster, non-convulsive way to individualise dosing (Peterchev et al., 2015).
- In 29 older adults with depression, amplitude-determined seizure threshold was measured at the first treatment, with standard 800 mA RUL ECT thereafter. Threshold related to the electric field strength from MRI-based modelling, and to antidepressant outcome at mid-course and the need to switch to bitemporal placement. The authors propose individualising amplitude using threshold or field modelling (Abbott et al., 2024).
None of these approaches is yet ready for routine use. Standard devices do not allow amplitude titration outside research settings.
Experimental: combining stimulation methods Advanced
tDCS with ECT. In a sham-controlled trial, adding prefrontal tDCS to ultrabrief RUL ECT did not speed response or change cognition at two weeks, and it increased the need for restimulation at titration.
Mayur P, Howari R, Byth K, Vannitamby R. J ECT 2018;34:291–5. doi:10.1097/YCT.0000000000000479
Seizure monitoring and EEG
Confirming the seizure, judging its quality, and handling missed, short and prolonged seizures.
Mayur P. Ictal electroencephalographic characteristics during electroconvulsive therapy: a review of determination and clinical relevance. A practical review of what the ictal EEG can tell you beyond seizure length, from electrode placement and dose to predicting response.
J ECT 2006;22:213–7. doi:10.1097/01.yct.0000235922.14623.39 · PMID 16957539
Why monitor Core
A generalised seizure is necessary for ECT to work. The stimulus alone, without a seizure, has little therapeutic effect. Each treatment is monitored with a two-channel frontal EEG and by watching the motor seizure in a limb isolated from the muscle relaxant by a blood-pressure cuff (the "cuff method").
In 232 patients monitored with both methods at the first session, the motor seizure was inadequate in 7% despite an adequate EEG seizure, and the motor seizure failed to show a prolonged EEG seizure in 39% of cases where one occurred. Motor monitoring alone is unreliable; EEG monitoring should be routine.
Mayur PM, Gangadhar BN, Janakiramaiah N, Subbakrishna DK. Br J Psychiatry 1999;174:270–2. doi:10.1192/bjp.174.3.270 · Mayur PM, et al. Can J Psychiatry 1999;44:191. PMID 10097847
Anatomy of the ictal EEG Core
Use the simulator to compare a well-formed seizure with a poor one. The trace is schematic, but the phases and features are the ones to look for on a real recording.
| Phase | What you see |
|---|---|
| Stimulus artefact | Large artefact for the duration of the pulse train |
| Recruitment | Low-voltage fast activity building in amplitude |
| Polyspike | High-frequency, high-amplitude spikes |
| Slow-wave (spike-and-wave) | Regular, high-amplitude waves around 3 Hz, slowing toward the end |
| Termination | Ideally abrupt |
| Post-ictal suppression | A flat or near-flat line straight after the seizure |
Judging adequacy Core
Duration alone is a weak guide. Traditional cut-offs are an EEG seizure of at least 25 seconds and a motor seizure of at least 20 seconds, but quality matters more. Features linked to good outcome are:
- High mid-ictal amplitude and a well-formed slow-wave phase
- Symmetry between the two channels (bilateral generalisation)
- Abrupt termination followed by marked post-ictal suppression
- A clear heart-rate rise during the seizure
Seizure duration tends to shorten across a course as the threshold rises. A short seizure is not a reason to increase the dose by itself; look at the whole picture, including clinical progress.
When things go wrong Core
| Problem | Action |
|---|---|
| Missed seizure (no seizure after stimulus) | Check electrode contact and impedance; keep the patient oxygenated; restimulate at a higher dose, usually after about 20–30 seconds and within the same anaesthetic. Consider hyperventilation and threshold-raising drugs. |
| Short or abortive seizure | Review anaesthetic agent and dose, benzodiazepines and anticonvulsants, hyperventilation, and the dose relative to threshold. Restimulation is a judgement call. |
| Prolonged seizure (EEG over 120 s) | Terminate with an intravenous benzodiazepine (e.g. midazolam) or an anaesthetic agent (e.g. propofol); maintain oxygenation; review lithium, theophylline and other contributors; lower the next dose. |
| Tardive (late) seizure | Monitor in recovery; treat as a prolonged seizure |
Quantitative ictal EEG Advanced
Beyond duration, the ictal EEG can help distinguish electrode placements and doses, predict response, track threshold changes across a course and guide dosing. Post-ictal suppression is the most studied predictor. Automated indices built into devices (such as suppression and coherence indices) put some of this at the bedside, though their clinical use is still debated.
A review of ictal EEG during ECT set out four practical uses: discriminating electrode positions and doses, predicting response (with post-ictal suppression the most studied measure), indicating threshold changes across a course, and informing clinical algorithms.
Mayur P. J ECT 2006;22:213–7. doi:10.1097/01.yct.0000235922.14623.39
Using multitaper spectral analysis, ultrabrief-pulse ECT produced lower mid-ictal spectral power than brief-pulse ECT, but regularity and post-ictal suppression were comparable. A lower-amplitude trace with ultrabrief treatment does not by itself mean an inadequate seizure.
Mayur P, Harris A, Rennie C, Byth K. J ECT 2012;28:229–33. doi:10.1097/YCT.0b013e318255a3b6
Adverse effects
Recognising and limiting the adverse effects of ECT, with a deeper look at memory: what is known, what is contested, and how to measure it.
Semkovska M, McLoughlin DM. Objective cognitive performance associated with electroconvulsive therapy for depression: a systematic review and meta-analysis. The key synthesis on the time course of cognitive effects, and the evidence to draw on when counselling patients and families about memory.
Biol Psychiatry 2010;68:568–77. doi:10.1016/j.biopsych.2010.06.009 · PMID 20673880
Common adverse effects Core
| Effect | Notes |
|---|---|
| Headache | Common; simple analgesia before or after treatment |
| Muscle aches | Often from suxamethonium fasciculations; usually settle after the first treatments |
| Nausea | Antiemetic if needed |
| Post-ictal confusion | Usually brief; post-ictal agitation or delirium may need management in recovery |
| Cardiovascular | Transient arrhythmia and hypertension; see Modules 1 and 3 |
| Dental and oral injury | Prevented by a well-placed bite block |
Cognitive effects Core
Patients and families worry most about memory. Expect three kinds of effect:
- Acute disorientation after each treatment. Time to reorientation is a useful bedside marker.
- Anterograde amnesia: difficulty learning new information during the course, which recovers within weeks after it ends.
- Retrograde amnesia: gaps for events before and during the course, mostly autobiographical. This recovers more slowly and can persist for some events.
A meta-analysis found that objective cognitive deficits mostly resolved within about two weeks of the last treatment, and that some measures then improved beyond baseline (Semkovska and McLoughlin, 2010). In community settings, bilateral placement and sine-wave stimulation were linked to more persistent retrograde amnesia (Sackeim et al., 2007).
Two further findings are worth knowing at this level. Repeated courses do not appear to accumulate: in 498 assessments of 199 patients who had received up to 186 treatments, the number of previous treatments had no effect on cognitive performance. What did affect it was age, the severity of depression at the time of testing, and the number of days since the last treatment (Kirov et al., 2016). And some of what patients report is the depression, not the treatment: before ECT, depressed patients rate their memory worse than controls do, and ratings improve as mood improves (Coleman et al., 1996).
Reducing cognitive impact
- Prefer right unilateral, ultrabrief-pulse ECT where clinically appropriate
- Dose relative to the measured threshold, not more
- Space treatments if impairment builds up; with bitemporal or brief-pulse ECT, twice weekly is as effective as three times weekly (Module 2)
- Review lithium, benzodiazepines and anticholinergic drugs
- Monitor reorientation time, a cognitive screen, and subjective memory through the course
- Identify the vulnerable before starting: poorer baseline cognition and prolonged post-ictal disorientation both predict retrograde amnesia (Sobin et al., 1995)
Why the cognitive evidence is weak Advanced
The efficacy of ECT rests on dozens of randomised trials with a common outcome measure. The cognitive literature has no such foundation, for six reasons:
- No standard battery. Studies use different instruments across different domains, so results cannot be pooled with confidence.
- Depression itself impairs memory. Depressed patients perform worse at baseline and recall fewer specific personal memories, and that difficulty can persist into euthymia (Semkovska and McLoughlin, 2013).
- Autobiographical memory is measured as consistency over time, yet normal recall becomes less consistent with the passage of time, and time affects different kinds of personal memory unequally. Few studies separate the effect of ECT from normal forgetting and from mood-related forgetting (Semkovska and McLoughlin, 2013).
- Subjective and objective measures disagree, and what patients report depends on how they are asked (Berman et al., 2008).
- Timing confounds everything. Performance depends on how many days have passed since the last treatment and on the mood state at the moment of testing (Kirov et al., 2016).
- Most trials are powered for efficacy, with cognition as a secondary outcome and follow-up rarely beyond six months.
The practical consequence is that a clinician cannot quote a single number for the risk of persistent memory loss. What can be done is to measure each patient systematically, which is what the framework below sets out.
Autobiographical and personal memory Advanced
Retrograde amnesia for personal events is the effect patients care about most, and the one that has driven the refinements in technique taught in Modules 2 and 4.
Mathiassen AB, Semkovska M, Lundsgaard CC, Gbyl K, Videbech P. Autobiographical memory after electroconvulsive therapy: systematic review and meta-analysis. The current synthesis: nine studies, 432 patients and 173 controls.
Br J Psychiatry 2025;228:263–73. doi:10.1192/bjp.2025.2 · PMID 40357797
The classic studies
- Personal memory is spared relative to impersonal memory. In 55 patients randomised to right unilateral or bilateral ECT at low or high dose, the amnestic effect was greatest for knowledge about the world and least for knowledge about the self. Bilateral ECT caused more marked amnesia than right unilateral ECT, especially for impersonal memories, independently of electrical dose and of clinical outcome. At two months the retrograde amnesia had receded, but deficits for recent impersonal events persisted (Lisanby et al., 2000).
- The effect is graded by technique. An early systematic study of stimulus parameters established the pattern that later work confirmed (Weiner et al., 1986). A systematic review of 15 studies found autobiographical memory loss is reduced by brief pulse rather than sine wave, by unilateral rather than bilateral placement, and by titrating the dose to the patient's own threshold; it falls most heavily on events in the six months before treatment (Fraser et al., 2008).
- Objective and subjective time courses differ. The same review found objective loss was relatively short term, under six months after treatment, while subjective accounts described amnesia lasting beyond six months (Fraser et al., 2008).
What the current meta-analysis adds
Pooling nine studies, autobiographical memory loss was greater in ECT patients than controls (standardised mean difference 0.55). Placement mattered: the effect was small with right unilateral ECT (SMD 0.32) and large with bilateral ECT (SMD 0.82). Higher age was associated with a smaller effect, which the authors call a surprising finding. Autobiographical memory was stable at long-term follow-up, which they read as suggesting that lost memories are not regained. The evidence is limited by small, mainly observational studies (Mathiassen et al., 2025).
In a double-blind randomised trial using the Kopelman Autobiographical Memory Interview and the Squire Subjective Memory Questionnaire, semantic autobiographical memory improved with 0.3 ms ultrabrief right unilateral ECT but worsened with 1.0 ms brief-pulse ECT at the same dose. The advantage for early-adult semantic memory persisted at three months.
Mayur P, Byth K, Harris A. J ECT 2013;29:277–82. doi:10.1097/YCT.0b013e3182941baf
How it is measured
The standard instrument is the Autobiographical Memory Interview (Kopelman et al., 1989), which separates personal semantic memory (facts about one's own life) from autobiographical incident memory (specific remembered events). The short form used in most ECT research is the Columbia AMI-SF. Interpret a fall in consistency cautiously, for the reasons in the section above.
Subjective memory: how you ask decides what you hear Advanced
What patients say about their memory depends heavily on the question put to them, and it is the single biggest source of apparent contradiction in this literature.
| Study | Finding |
|---|---|
| Squire, Wetzel and Slater, 1979 | The Squire Subjective Memory Questionnaire was developed for this purpose. Complaints one week after bilateral ECT differed both in degree and in kind from complaints before it, and six months later still resembled the post-treatment pattern in diminishing form. |
| Squire and Chace, 1975 | Six to nine months after treatment, six objective tests of delayed retention and remote memory showed no persisting impairment, yet patients who had received bilateral ECT rated their own memory as impaired significantly more often than the comparison groups. The authors offered two readings: either the tests were not sensitive enough, or the early impairment had made these patients more alert to ordinary memory failures. |
| Coleman et al., 1996 | Before ECT, patients rated their memory worse than controls did. Ratings improved markedly after treatment, and at every time point the severity of depression was strongly associated with the complaint. Shortly after ECT there was no relation between self-ratings and objective tests; at two months, greater retrograde amnesia for autobiographical memories went with worse self-rating. |
| Berman et al., 2008 | The decisive methodological study. A questionnaire about specific cognitive failures showed marked improvement after ECT and tracked depression severity. A direct global question about the overall impact of ECT on memory showed a deleterious effect at both one and 24 weeks, and that global rating, unlike the specific one, was associated with treatment technique and with long-term retrograde amnesia. This was the first report of concordance between subjective and objective measures. |
| Rose et al., 2003 | A systematic review of patients' own accounts found that at least one third reported persistent memory loss, and concluded that the then-current patient statement from the Royal College of Psychiatrists was unfounded. |
| Brus et al., 2017 | In 1,212 patients in the Swedish national register, subjective memory worsening within a week of treatment was reported by 26%. It was more common in women than men (31% vs 18%) and in those aged 18–39 than those over 65 (32% vs 22%). Patients in remission were less likely to report it, and brief pulse carried a higher risk than ultrabrief pulse (odds ratio 1.61). |
In interviews with ten patients receiving long-term continuation or maintenance ECT, subjective cognitive effects ranged from minimal to pervasive. Patients described adapting to them and judged them an acceptable trade-off for staying well and out of hospital.
Mayur P, Du N, Harris A, et al. J ECT 2026. doi:10.1097/YCT.0000000000001310
What this means in clinic
- Ask a global question as well as a checklist: "overall, what has ECT done to your memory?" That question detects what the specific ones miss.
- Record depression severity at the same sitting, since mood drives the complaint.
- A patient who reports memory loss is not simply mistaken because the tests are normal. The discrepancy is a known property of the measures, not evidence that the patient is wrong.
- Young women are the group most likely to report subjective worsening, and ultrabrief pulse is the technique least likely to produce it.
A framework for cognitive testing Advanced
No battery is universally agreed. The framework below is a practical minimum, built from the evidence above. It is designed so that each measurement can be compared with the same patient's own baseline.
| When | What to measure | Instrument or method |
|---|---|---|
| Baseline, before the first treatment | Global cognition; subjective memory; autobiographical memory; depression severity | MoCA (or MMSE); Squire Subjective Memory Questionnaire and a global self-evaluation question; AMI-SF where available; MADRS or HAM-D |
| Every treatment | Time to reorientation: name, age, place, date of birth, day of week | Minutes from the end of the seizure to correct answers. Prolonged disorientation is a warning sign. |
| After about three treatments | Early screen for emerging deficits | A brief screening battery. Change on such a battery after three treatments predicted later cognitive change, where change on the MMSE did not (Martin et al., 2013). |
| Mid-course, about six treatments | Global cognition; anterograde memory; subjective memory | Repeat MoCA; a word-list learning task; repeat the subjective questionnaire |
| End of the course | Full repeat of the baseline measures | As at baseline |
| Follow-up, at about two weeks and again at two to six months | Recovery of anterograde function; retrograde amnesia; subjective memory | Repeat the battery, the AMI-SF and the global question. Most objective deficits resolve by about two weeks (Semkovska and McLoughlin, 2010). |
Identify the patients at risk before you start
Two variables available at the bedside predicted the magnitude of retrograde amnesia for autobiographical memories, both in the week after a course and at two months, and did so largely regardless of the technical parameters used: poorer global cognitive status before treatment, and prolonged disorientation in the acute post-ictal period (Sobin et al., 1995). Both are in the framework above for that reason. A patient with either should prompt a shift toward the most cognition-sparing technique available.
Five rules for interpreting the numbers
- Test at a constant interval from the last treatment. Days since the last session independently affect performance (Kirov et al., 2016).
- Always score mood at the same time. Depression severity affects both objective performance and subjective complaint.
- Compare the patient with themselves, not with a published norm.
- Do not reassure on the basis of a normal screen alone. A MoCA can be normal while autobiographical memory and the patient's own account are not.
- A rising number of treatments is not in itself a reason to stop. Cumulative deficits were not demonstrated across courses (Kirov et al., 2016); judge the patient in front of you.
This framework is a teaching synthesis, not a validated protocol or a guideline recommendation. Check it against your local policy and the instruments available in your service.
After the acute course
What to do when ECT has worked — preventing relapse with continuation and maintenance treatment, and running an outpatient service — and what to do when it has not.
Jelovac A, Kolshus E, McLoughlin DM. Relapse following successful electroconvulsive therapy for major depression: a meta-analysis. Thirty-two studies, up to two years of follow-up. The paper that quantifies the problem this module exists to solve.
Neuropsychopharmacology 2013;38:2467–74. doi:10.1038/npp.2013.149 · PMID 23774532
Preventing relapse and recurrence
An acute course of ECT treats the episode. It does nothing to prevent the next one. Planning what follows is part of prescribing ECT, not an afterthought.
Definitions Core
The terms are used loosely in conversation but have distinct meanings, and the distinction decides how long you treat and what you are trying to prevent.
| Term | Meaning | Period |
|---|---|---|
| Relapse | The return of the same episode, before it has fully resolved. The illness was suppressed, not finished. | Conventionally within about six months of remission |
| Recurrence | A new episode after genuine recovery from the last one. | Beyond about six months |
| Continuation ECT (C-ECT) | ECT given after the acute course to prevent relapse, alongside medication. | Up to about six months after remission |
| Maintenance ECT (M-ECT) | ECT continued to prevent recurrence in people with recurrent, severe illness. | Beyond six months, sometimes for years |
In practice the two phases run together as a single tapering schedule, and the literature often writes C/M-ECT for both. What matters clinically is that the first six months carry the highest risk and need the closest attention.
The scale of the problem Core
Without continuation treatment, most patients who respond to ECT relapse. Even with it, the risk in the first year is substantial.
- In modern studies of continuation pharmacotherapy, 51% relapsed by 12 months, and most of those — 38% — within the first six months. The six-month relapse rate with continuation ECT was similar, at 37% (Jelovac et al., 2013).
- In randomised trials, antidepressant medication halved the risk of relapse compared with placebo over the first six months (risk ratio 0.49; number needed to treat 3.3) (Jelovac et al., 2013).
- The clearest single demonstration: after a successful course, relapse within 24 weeks was 84% on placebo, 60% on nortriptyline alone, and 39% on nortriptyline plus lithium (Sackeim et al., 2001).
Continuation and maintenance ECT in depression Core
| Study | Design | Result |
|---|---|---|
| Sackeim et al., 2001 | Randomised, continuation pharmacotherapy after ECT | Relapse at 24 weeks: placebo 84%, nortriptyline 60%, nortriptyline plus lithium 39%. |
| CORE: Kellner et al., 2006 | Randomised, C-ECT versus nortriptyline plus lithium | The two strategies performed similarly. Continuation ECT is a legitimate alternative to drugs, not a second-best option. |
| Nordenskjöld et al., 2013 | Randomised, C-ECT plus pharmacotherapy versus pharmacotherapy alone, one year, 56 patients | Relapse at one year: 32% with C-ECT plus drugs versus 61% with drugs alone (hazard ratio 2.32). Cognitive and memory measures were stable in patients who did not relapse. |
| PRIDE phase 2: Kellner et al., 2016b | Randomised, older adults; venlafaxine plus lithium, with or without continuation ECT on a symptom-titrated schedule | Continuation ECT plus medication outperformed medication alone. |
| STABLE: Lisanby et al., 2008 | Algorithm development using CORE data | A symptom-titrated schedule identified every patient who went on to relapse, on average 2.2 weeks beforehand, while exposing only 20% of sustained remitters to extra treatments. The basis for giving C-ECT by clinical state rather than by a fixed calendar. |
Taken together: continuation ECT works, it works at least as well as drug treatment, and it works best added to medication rather than instead of it.
In a ten-year chart review of continuation ECT in major depression, 20 of 22 patients remained on 0.3 ms ultrabrief pulses throughout the continuation phase, and most visits were recorded as "in remission". The median duration was 51 days, treatment intervals widened from about 10 days at one month to about 20 days at four months, and the stimulus dose rose steadily across the phase. Ultrabrief ECT can be used for relapse prevention, not only for the acute course.
Atashnama A, Aly H, Krishnan V, Howari R, Mayur P. Australas Psychiatry 2020;28:286–90. doi:10.1177/1039856220917070
Continuation and maintenance ECT in schizophrenia Core
The evidence here is thinner than in depression but points the same way, and the one controlled trial is striking.
- In treatment-resistant schizophrenia, 51 patients who had remitted on bilateral ECT plus flupenthixol were randomised to continuation with flupenthixol alone, continuation ECT alone, or both. Among completers, relapse over six months was 40% with the combination against 93% in each of the single-treatment arms (Chanpattana et al., 1999). Neither treatment alone held the gain; together they did.
In 46 inpatients with schizophrenia receiving C/M-ECT, improvement in illness severity was preserved for about six months after the acute course, and bitemporal placement was associated with a smaller loss of benefit over time. A narrative review of the case-report literature supports C/M-ECT combined with antipsychotics as a maintenance strategy, with the caveat of likely publication bias.
George R, Krishnan V, Talbot D, et al. Asian J Psychiatr 2024;92:103895. doi:10.1016/j.ajp.2023.103895 · Egan O, Talbot D, Mayur P, Harris A. Australas Psychiatry 2024;32:309–13. doi:10.1177/10398562241256837
Clinical and cognitive monitoring during C/M-ECT Advanced
Long-term ECT is safe to continue only if it is measured. A patient attending monthly for years can drift in either direction — toward relapse, or toward unnecessary treatment — and neither is visible without records.
A monitoring schedule
| At every visit | Every few months | Every 6–12 months |
|---|---|---|
| Depression or illness-severity rating (MADRS, or CGI-S for psychotic illness); time to reorientation; stimulus dose and seizure quality; side effects since the last treatment; the patient's own account of memory | Cognitive screen (MoCA); subjective memory questionnaire and a global memory question (Module 6); review of medication and physical health | Full review of the indication: is continuing still justified? Re-titrate the seizure threshold. Review consent and, where applicable, the legal authority. Consider a planned trial of extending the interval. |
Deciding the interval
Taper by clinical state, not by the calendar alone. A common pattern is weekly for a few treatments, then fortnightly, then every three to four weeks, extending while the patient stays well and contracting at the first sign of return. The STABLE work showed that a symptom-titrated schedule can anticipate relapse by about two weeks while sparing most stable patients extra treatments (Lisanby et al., 2008). The author's chart review shows what this looks like in practice: intervals roughly doubling between the first and fourth month, with the charge rising as the threshold climbs (Atashnama et al., 2020).
What the evidence says about long-term cognition
- In the one-year randomised trial of C-ECT, cognitive and memory measures were stable in patients who did not relapse, in both the ECT and the medication-only arms (Nordenskjöld et al., 2013).
- Across 199 patients with up to 186 lifetime treatments, the number of previous treatments had no effect on cognitive performance. Age, current depression severity and days since the last treatment did (Kirov et al., 2016).
- So a rising treatment count is not in itself a reason to stop. Judge the patient, with measurements.
Interviews with ten patients receiving long-term C/M-ECT found that they described it as life-saving and as what kept them well and out of hospital. Subjective cognitive effects ranged from minimal to pervasive, but were seen as adaptable and an acceptable trade-off. The recurring concerns were practical and social: the logistical and financial barriers to access, the trust built with long-standing staff, family ambivalence about the number of treatments, and stigma leading to selective disclosure.
Mayur P, Du N, Harris A, Ramakrishnan V, Andepalli J, López-Portillo B. J ECT 2026. doi:10.1097/YCT.0000000000001310
Those findings have a direct implication for monitoring: ask about access, transport, cost and family views as routinely as you ask about memory. They are the commonest reasons a working maintenance schedule breaks down.
Outpatient ECT: practical dos and don'ts Core
Most continuation and maintenance ECT is given as a day procedure. The clinical technique is unchanged; what changes is everything around it.
| Do | Don't | |
|---|---|---|
| Before the day | Give written fasting instructions and confirm them the day before. Specify which regular medicines to take with a sip of water and which to withhold. Confirm an escort is booked. | Don't assume a patient who has attended for months remembers the instructions. Memory is the thing being affected. |
| Escort and transport | Insist on a responsible adult to accompany the patient home and stay with them for the rest of the day. | Don't let a patient leave alone, use public transport alone, or drive themselves on the day of treatment. |
| Driving | State clearly, in writing, that there is no driving on the day of treatment, and advise on the local rules for driving between treatments. | Don't leave driving unaddressed in the notes. Check your state or territory's fitness-to-drive requirements. |
| On the day | Confirm fasting, current medication, physical health changes since the last visit, and that consent still stands. Check the interval since the last treatment and record it. | Don't proceed without reviewing what has changed. Outpatients acquire new illnesses and new medicines between visits. |
| Recovery and discharge | Observe until fully oriented, steady on their feet, able to tolerate oral fluids, and with stable observations. Discharge to the named escort with written instructions. | Don't discharge on the basis of time elapsed alone. Reorientation time varies between treatments. |
| Between treatments | Name who is responsible for monitoring, and how the patient contacts the service if they deteriorate before the next visit. | Don't let the patient fall between the ECT service and the treating team. This is where relapse is missed. |
| Consent and the law | Review capacity and consent at intervals, and keep any tribunal or legal authority current. Document the review. | Don't treat a consent given a year ago as settled, particularly across a long maintenance course. |
| When to admit | Admit for treatment when there is deterioration, significant physical comorbidity, a high anaesthetic risk, no reliable escort, or concern about suicide risk. | Don't keep a deteriorating patient on an outpatient schedule to preserve the arrangement. |
These points are standard day-procedure practice rather than findings from a trial. Check them against your service's own day-surgery and ECT policies, and against your state or territory's driving rules.
Non-response to an acute course
When an adequate course of ECT has not worked, the first task is to confirm that it really was adequate. The second is to know what else there is.
When ECT is not working Core
In depression, consider non-response when the depression score has changed by less than 25% after about six to eight technically adequate treatments. Apply this flexibly: act sooner if the patient is deteriorating, and allow longer in schizophrenia, which needs longer courses (Module 2). Work through the causes in order before abandoning the course. The appendix sets the steps out as an algorithm for shifting gears.
| Step | Questions to ask |
|---|---|
| 1. Diagnosis | Is this the illness ECT treats? Consider personality factors, substance use, medical causes and comorbid anxiety or trauma. |
| 2. Seizure quality | Are seizures adequate on EEG, not just by duration? (Module 5) |
| 3. Dose | Is the dose still the intended multiple of a threshold that may have risen? Has the dose been re-titrated? |
| 4. Medications | Are benzodiazepines, anticonvulsants or the anaesthetic agent blunting seizures? (Module 3) |
| 5. Technique | Switch ultrabrief to brief pulse, 0.5 ms and then 1.0 ms (Module 4), and optimise right unilateral ECT. Use bilateral placement only as a last resort (appendix). |
| 6. Alternatives | Only once the above have been worked through. See below. |
Most apparent ECT non-response is a problem of delivery, not of the treatment. Do not move to step 6 until steps 1 to 5 have been answered and documented.
Alternatives after a failed course: the evidence Advanced
Anand A, Mathew SJ, Sanacora G, et al. Ketamine versus ECT for nonpsychotic treatment-resistant major depression (ELEKT-D). An open-label, randomised non-inferiority trial in 403 patients referred to ECT clinics. The most important recent trial for this decision.
N Engl J Med 2023;388:2315–25. doi:10.1056/NEJMoa2302399 · PMID 37224232
Tricyclic antidepressants
The strongest evidence of the four, though most of it concerns relapse prevention rather than rescue after failure. Tricyclics have the largest evidence base for preventing relapse after ECT (Jelovac et al., 2013), and nortriptyline plus lithium was the most effective continuation regimen tested (Sackeim et al., 2001). Given during a course, nortriptyline improved the efficacy of ECT and reduced its cognitive adverse effects, while venlafaxine gave a smaller benefit and tended to worsen cognition (Sackeim et al., 2009). A patient who has not responded and is not on a tricyclic has an obvious gap to fill.
Lithium augmentation
Lithium added to a tricyclic gave the lowest post-ECT relapse rate in the key trial, 39% against 60% for nortriptyline alone (Sackeim et al., 2001). In treatment-resistant depression more broadly, lithium augmentation is one of the few strategies with consistent placebo-controlled support (Thase, 2004). Note the interaction with ECT itself: lithium is associated with prolonged seizures and post-ictal delirium, so if ECT is continuing, keep the level in the low therapeutic range (Module 1).
Monoamine oxidase inhibitors
A narrative review of treatment-resistant depression concluded that the best-documented strategies — lithium augmentation, switching to an MAOI, and ECT — are among the least used in practice (Thase, 2004). There is no randomised trial of an MAOI specifically after a failed ECT course, so the case rests on the general treatment-resistance literature. The practical obstacles are the dietary and drug restrictions and the washout required before and after. If ECT may be resumed, plan the washout with the anaesthetist.
Ketamine and esketamine
Two randomised trials compared ketamine with ECT directly and reached opposite conclusions:
- ELEKT-D randomised 403 patients with non-psychotic treatment-resistant depression, referred to ECT clinics, to ketamine 0.5 mg/kg twice weekly or ECT three times weekly for three weeks. Ketamine was non-inferior to ECT on response, with better memory outcomes and quality of life (Anand et al., 2023).
- KetECT randomised 186 hospitalised inpatients with unipolar depression to thrice-weekly ketamine or ECT. Remission was 63% with ECT against 46% with ketamine, so ketamine was inferior, though still of value. Serious and long-lasting adverse events, including persisting amnesia, were more common with ECT, while treatment-emergent effects caused more dropouts with ketamine (Ekstrand et al., 2022).
The difference is best explained by the populations: ELEKT-D excluded psychosis and recruited outpatients referred for ECT, while KetECT treated severely ill inpatients. The reasonable reading is that ketamine is a genuine alternative in non-psychotic, less severe treatment resistance, and weaker than ECT in severe or psychotic illness — which is exactly the group in whom ECT has failed. Ketamine is not a substitute for ECT in psychotic depression or catatonia.
Ketamine has also been tried as the anaesthetic for ECT, in the hope of combining the two. It did not speed the antidepressant response (Rasmussen et al., 2014), though it is less anticonvulsant and so useful when seizures are inadequate (Module 3).
Vagus nerve stimulation
VNS is the option designed for exactly this group: chronic, highly treatment-resistant depression in which other options, including ECT, have failed. Its evidence has an unusual shape.
- The acute sham-controlled trial was negative on its primary outcome: response at 10 weeks was 15.2% with active VNS against 10.0% with sham (Rush et al., 2005). VNS does not work quickly.
- A five-year registry of 795 patients who had failed four or more treatments including ECT found a cumulative response rate of 67.6% with adjunctive VNS against 40.9% with treatment as usual, with higher remission rates (Aaronson et al., 2017). The benefit emerges over months to years, not weeks.
- The RECOVER trial randomised 493 patients with marked treatment resistance to active or sham VNS. A history of ECT, whether lifetime or in the current episode, predicted a poorer outcome overall — but those same patients showed greater benefit from active VNS than from sham (Aaronson et al., 2025). A failed ECT course therefore marks out a patient who does badly in general yet may be among those most likely to gain from VNS.
Practical points: VNS requires surgical implantation, the response is slow, and access and funding in Australia are limited. Discuss it early with a service that offers it rather than as a last resort after years have passed.
Putting it together
| Option | Strength of evidence after failed ECT | Speed |
|---|---|---|
| Tricyclic, with lithium | Good for relapse prevention; indirect for rescue | Weeks |
| Lithium augmentation | Good in treatment resistance generally | Weeks |
| MAOI | Good in treatment resistance generally; no post-ECT trial | Weeks |
| Ketamine or esketamine | Two head-to-head trials against ECT, disagreeing; neither in post-ECT failure | Days |
| VNS | Negative acutely; positive long-term, with a signal favouring patients who have had ECT | Months to years |
A reasonable sequence for a patient who has genuinely failed an optimised course is: ensure an adequate tricyclic trial with lithium augmentation; consider an MAOI; use ketamine where the illness is non-psychotic and a rapid effect is needed; and refer for VNS where the illness is chronic and the patient has failed several modalities including ECT.
The sequence above is a teaching synthesis of the evidence cited, not a guideline recommendation.
Special situations
ECT in patients with medical comorbidity or implanted devices, in pregnancy and the puerperium, in children and adolescents, and in older people. In each the question is not whether ECT can be given, but how.
Tess AV, Smetana GW. Medical evaluation of patients undergoing electroconvulsive therapy. The standard reference for the pre-ECT medical work-up.
N Engl J Med 2009;360:1437–44. doi:10.1056/NEJMra0707755 · PMID 19339723
Elias A, Das S, Kirkland J, Loyal S, Thomas N. Safety of electroconvulsive therapy in the context of physiological and medical complexity: a state-of-the-art review. A recent Australian review that gathers the scattered literature on medical complexity into one place.
PCN Rep 2025;4:e70051. doi:10.1002/pcn5.70051 · PMID 39759458
How to think about risk Core
There are no absolute contraindications to ECT (Module 1). Every situation in this module is a question of modifying risk rather than accepting or refusing the treatment.
The recent review of medical complexity concluded that mortality directly attributable to ECT is extremely rare. Deaths, including fetal deaths, have been reported after ECT in the presence of recent myocardial infarction, deep vein thrombosis, intracranial aneurysm and tumours, phaeochromocytoma, sepsis and pregnancy — but in no case could the outcome be attributed to ECT with confidence, because of the time lag between treatment and death and the presence of confounding factors. The authors concluded that risk can be mitigated by safety protocols, clear communication with everyone involved, collaboration with anaesthetists and specialists, consultation-liaison input, and education about ECT (Elias et al., 2025).
| Step | What it means in practice |
|---|---|
| 1. Weigh the alternative | Compare the risk of ECT against the risk of the untreated illness, not against no risk at all. Severe depression in a frail patient is itself lethal, and in cardiac disease the alternative of a tricyclic may carry more risk than ECT (Zielinski et al., 1993). |
| 2. Optimise the condition first | Treat the heart failure, control the blood pressure, stabilise the diabetes. Most risk is reduced before the first treatment, not during it. |
| 3. Get the right people in the room | Anaesthetist, and the relevant specialist: cardiologist, neurologist, obstetrician, paediatrician, geriatrician. Agree the plan in writing before the first treatment. |
| 4. Modify the technique | Right unilateral placement, ultrabrief pulses, a less anticonvulsant anaesthetic, a beta-blocker for the surge, twice-weekly treatment. Each is a lever. |
| 5. Document the reasoning | Record the risk–benefit discussion, who was consulted, and what was modified. This is both good practice and the medicolegal record. |
Comorbid medical disorders Core
| System | The concern | How to manage it |
|---|---|---|
| Cardiac | The vagal then sympathetic surge (Module 1) stresses the myocardium. Risk is highest with recent infarction, unstable angina, decompensated failure, severe valvular disease and significant arrhythmia. | Cardiology review and optimisation. Continue beta-blockers and most cardiac medication. Consider a short-acting beta-blocker at induction. In 40 patients with left ventricular impairment, arrhythmia or conduction delay, cardiac complications were commoner than in matched controls but were mostly transitory, and 38 of 40 completed the course — while 11 of 21 had previously had to stop a tricyclic because of cardiovascular effects (Zielinski et al., 1993). |
| Hypertension | Marked rises in blood pressure during the seizure. | Control before starting. Give the morning antihypertensives with a sip of water. A short-acting beta-blocker blunts the surge (Module 3). |
| Intracranial | ECT transiently raises intracranial pressure and cerebral blood flow. Space-occupying lesions, raised pressure, recent stroke, aneurysm and vascular malformation all raise concern. | Neurology or neurosurgery review and imaging. Small, non-oedematous, incidental lesions have been treated safely. Control blood pressure tightly. Avoid where there is raised pressure or significant mass effect. |
| Epilepsy | Interaction with anticonvulsants, and theoretical concern about seizure control. | Continue anticonvulsants prescribed for epilepsy — do not withhold them as you would a mood stabiliser. Expect a higher seizure threshold. ECT itself raises the threshold and is not contraindicated. |
| Respiratory | Hypoxia during apnoea; bronchospasm. Severe COPD and poorly controlled asthma raise risk. | Optimise before starting; bronchodilators on the day. Minimise theophylline, which is linked to prolonged seizures. Consider high-flow nasal oxygen for patients who desaturate (Module 3). |
| Diabetes | Fasting with usual hypoglycaemic treatment; variable glycaemic control across a course. | Schedule early in the list. Withhold or reduce the morning insulin or oral agent per an agreed plan, check glucose before and after, and have glucose available. |
| Gastro-oesophageal reflux | Aspiration during bag-and-mask ventilation. | Confirm fasting. Consider a proton-pump inhibitor and a prokinetic, and discuss airway protection with the anaesthetist. |
| Musculoskeletal | Osteoporosis, recent fracture or unstable joints, with risk from the convulsion. | Use a fuller relaxant dose, around 1 mg/kg or above (Module 3), and position carefully. |
| Thromboembolic | Deep vein thrombosis has featured among reported deaths (Elias et al., 2025). | Assess and treat before starting; use prophylaxis in the immobile or catatonic patient. |
| Endocrine | Phaeochromocytoma is the one condition where the catecholamine surge may be dangerous. Hyperthyroidism also raises risk. | Exclude phaeochromocytoma where suspected and treat it before ECT. Stabilise thyroid function. |
| Ophthalmic / dental | Raised intraocular pressure; glaucoma; retinal detachment; loose or fragile teeth. | Ophthalmology advice for retinal disease. Dental assessment, a well-fitted bite block, and a custom guard where needed. |
This table summarises the general approach from the sources cited. Individual management is a matter for the treating team and the anaesthetist.
Implanted medical devices Advanced
The number of patients presenting for ECT with an implanted device is rising. The two questions in each case are whether the ECT current can damage or trigger the device, and whether the seizure or the haemodynamic response can.
| Device | Concern | Approach |
|---|---|---|
| Cardiac pacemaker | Electrical interference, and the risk of the device misreading the stimulus. Modern pacemakers are well shielded. | ECT is not contraindicated. Obtain a device check and the manufacturer's advice. Consider programming to an asynchronous mode, have a magnet and external pacing available, and insulate the patient from earthed metal. Successful courses are reported, including in older patients (Mędrala et al., 2018). |
| Implantable cardioverter defibrillator (ICD) | The tachycardia of the seizure, or the stimulus artefact, may be misread as a shockable rhythm. | Deactivate before treatment and reactivate afterwards, with cardiology involvement, and have external defibrillation available throughout. |
| Deep brain stimulation (DBS) | Current could in principle be conducted along the leads and heat the tissue around the electrodes. | A systematic review found 25 reported cases of ECT in patients with DBS implants and ECT appeared effective and safe. The two most important recommendations are to set the DBS system to 0 V and turn it off before ECT, and to avoid electrode sites near the DBS leads (Bukowski et al., 2022). Involve the DBS team. |
| Vagus nerve stimulator (VNS) | The same conduction and heating concerns, in a patient group who may need both treatments (Module 7). | Turn the device off before treatment and on afterwards. Discuss with the implanting service. ECT and VNS are both recognised neurostimulation options with separate evidence bases (Milev et al., 2016). |
| Cochlear implant | Current near the device and its electrode array. | Seek the manufacturer's advice. Remove the external processor. Consider electrode placement away from the implant. |
| Insulin or intrathecal pump | Device malfunction; and for insulin pumps, glycaemic control during fasting. | Device check and an agreed plan for the fasting period. Monitor accordingly. |
| Aneurysm clips, coils and shunts | These are a reason for caution about the underlying lesion rather than the hardware itself. | Neurosurgical review. Treated and secured aneurysms have been managed with tight blood-pressure control. |
Pregnancy and the puerperium Core
ECT has a particular place in pregnancy because the alternative is not no treatment but either untreated severe illness or medication exposure, and because it works quickly in exactly the presentations that threaten mother and baby.
What the evidence shows
- A review of 339 cases from 1941 to 2007 found at least partial remission in 78% of cases where efficacy data were available. There were 25 fetal or neonatal complications, of which 11 were judged likely to be related to ECT, including two deaths, and 20 maternal complications, of which 18 were judged likely related. The authors concluded that ECT is effective for severe mental illness in pregnancy and that the risks to mother and fetus are low (Anderson and Reti, 2009).
- A systematic review reached the same conclusion, with total or at least partial remission in almost all patients and no reported maternal deaths, and argued that ECT is probably under-used in pregnancy because of stigma among both patients and professionals (Pompili et al., 2014).
- Be careful how you quote these numbers. A meta-review of the four systematic reviews found that they varied widely in method, in the studies and cases they included — from 16 studies and 300 cases to 67 studies and 169 cases — and in how they interpreted adverse outcomes. One review treated every adverse outcome as potentially ECT-related, while others gave reasons for ruling ECT out. The apparent disagreement in the literature is largely a disagreement about attribution (Sinha et al., 2017).
What changes in pregnancy
| Issue | Why it matters | What to do |
|---|---|---|
| Aspiration risk | Raised intra-abdominal pressure and lower sphincter tone increase the risk during bag-and-mask ventilation (Rose et al., 2020). | Strict fasting; consider a proton-pump inhibitor and antacid; discuss airway protection with the anaesthetist. |
| Aortocaval compression | In the second half of pregnancy, lying supine can compress the vena cava and reduce placental perfusion (Rose et al., 2020). | Position with left lateral tilt or a wedge. |
| Fetal monitoring | Fetal heart rate changes have been associated with prolonged seizure activity (Rose et al., 2020). | Fetal heart rate monitoring before and after each treatment; continuous monitoring and an obstetric presence later in pregnancy, by local protocol. |
| Preterm labour | Reported among the complications in the case literature. | Obstetric review; tocodynamometry where indicated; have a plan for the onset of labour. |
| The team | No single specialty can cover this safely. | A multidisciplinary plan agreed in advance between psychiatry, obstetrics, maternal–fetal medicine and anaesthesia (Rose et al., 2020). |
The puerperium
Postpartum psychosis and severe postnatal depression are among the clearest indications for ECT. The illness is typically sudden, severe and dangerous to both mother and infant; the response to ECT is often rapid; and a quick recovery protects the developing relationship with the baby. Practical points: there is no bar to breastfeeding, and feeds can be timed around the anaesthetic with advice on the drugs used; arrange care for the infant during treatment and recovery; and keep mother and baby together where the service allows.
Children and adolescents Advanced
ECT is used rarely in young people and is tightly regulated. The evidence base is small, but what there is points in the same direction as the adult literature.
- The classic review — by an Australian group — identified 60 reports describing ECT in 396 patients aged 18 or under. Most were single case reports and there were no controlled trials. Rates of improvement were 63% for depression, 80% for mania, 42% for schizophrenia and 80% for catatonia. Serious complications were very rare, while minor transient side effects were common. The authors concluded that ECT in the young appears similar in effectiveness and side effects to ECT in adults, while stressing the lack of systematic evidence and calling for registers and controlled trials (Rey and Walter, 1997).
- Catatonia is the indication with the strongest signal, and it is frequently missed. It occurs in children and adolescents with autism spectrum disorder and with intellectual disability, where benzodiazepines and ECT are the only known effective treatments, and where timely recognition and treatment can be life-saving (Ghaziuddin et al., 2020).
- Cognition is the parents' first question. It has now been examined directly in depressed adolescents treated with ECT (Ghaziuddin et al., 2024), and the framework in Module 6 applies with more rather than less rigour in this age group.
Practical differences
| Issue | Approach |
|---|---|
| Seizure threshold | Usually lower than in adults. Start titration at the lowest step and watch for prolonged seizures. Have a benzodiazepine drawn up. |
| Placement and pulse width | Ultrabrief right unilateral ECT is the preferred starting technique to protect cognition in the developing brain (Module 4). |
| Consent and the law | Extra safeguards apply, and they are stricter the younger the patient. The APA task force report (3rd ed, 2024) specifies that for a child under 13 years of age, two consultant child psychiatrists must each provide an opinion supporting ECT before treatment proceeds. Beyond this, most Australian jurisdictions place their own restrictions on ECT for minors, often including a lower age limit and mandatory tribunal or court approval regardless of parental consent; in New South Wales additional provisions apply to patients under 16. Check the current legislation in your state or territory before proposing treatment. |
| Who should treat | A child and adolescent psychiatrist, an anaesthetist experienced with children, and an ECT service familiar with the age group. Get a second opinion as a matter of routine. |
| Family and school | Involve the family throughout. Plan for schooling, and for the cognitive effects during the course. |
The legal position for minors differs between Australian states and territories and changes over time. Nothing on this page substitutes for the current Act and your local policy.
Older patients Core
Older people are the largest group receiving ECT and, contrary to the common assumption, among the most likely to respond well.
- In a prospective multisite study of 268 patients with unipolar major depression, response rates were 54% in adults under 60, 73% in the young-old (60–74) and 67% in the old-old (75 and over). The two older groups carried a greater burden of physical illness and more baseline cognitive impairment, and had higher seizure thresholds, yet tolerated ECT in a manner similar to younger patients and showed similar or better acute response (Tew et al., 1999).
- In the PRIDE study, 240 older adults received ultrabrief right unilateral ECT at six times threshold, three times a week, with venlafaxine. Remission was 62%, reached after a mean of 7.3 treatments, with good safety and tolerability (Kellner et al., 2016a). Phase 2 showed that continuation ECT added to medication outperformed medication alone (Module 7).
- Age also appears to protect autobiographical memory: in the current meta-analysis, higher age was associated with a smaller effect of ECT on autobiographical memory (Module 6).
What to do differently
| Issue | Approach |
|---|---|
| Higher seizure threshold | Expect it, and titrate rather than relying on a formula. Check that the target dose is within the device maximum before starting (dose calculator). |
| Medical comorbidity | Assume it is present. Work through the first two tables in this module. |
| Polypharmacy | Review every medicine for seizure threshold, anticholinergic burden and anaesthetic interaction. Anticholinergics worsen post-ictal confusion. |
| Post-ictal delirium | Commoner than in younger patients. Keep the environment quiet and oriented, minimise anticholinergics, review lithium, and record reorientation time at every treatment. |
| Cognition | Prefer ultrabrief right unilateral ECT. Baseline cognitive screening is essential, since poorer baseline cognition predicts greater retrograde amnesia (Module 6). |
| Falls and frailty | Assess falls risk, and supervise mobilisation in recovery. Consider bone health before a strong convulsion. |
| Pace of the course | Twice weekly is a reasonable default with brief-pulse ECT if confusion accumulates (Module 2). |
Consent, capacity and the law
No other psychiatric treatment is hedged with as much law as ECT. This module covers what valid consent requires, how capacity is assessed, how much to disclose about cognitive effects, and the hardest case of all — the patient who cannot consent and will not agree.
Leong GB, Eth S. Legal and ethical issues in electroconvulsive therapy. The classic statement of the problem: obtaining valid informed consent is the treating psychiatrist's duty, and failure to obtain it is the commonest area of professional liability in ECT.
Psychiatr Clin North Am 1991;14:1007–20. PMID 1771147
McSherry B. Electroconvulsive therapy without consent: the influence of human rights law. Recent, and written for Australian practice: the UN Special Rapporteur's call for a ban on non-consensual ECT, and the low capacity threshold set by Bell J in the Supreme Court of Victoria.
J Law Med 2019;26:732–6. PMID 31682352
What valid consent requires Core
Consent is not a signature on a form. It is a process with four elements, and all four must be present.
| Element | What it means for ECT |
|---|---|
| Capacity | The patient can understand, retain, weigh and communicate a decision about this treatment. Depression itself can impair it — through hopelessness, nihilistic delusions, poor concentration or psychomotor retardation. |
| Information | What ECT is, why it is proposed, what the alternatives are (including no treatment), what the likely benefit is, and what the adverse effects are — including the cognitive effects, discussed below. |
| Voluntariness | Free of coercion, express or implied. The involuntary patient on a treatment order, and the patient whose family is pressing hard, both need particular attention here. |
| Documentation | Record what was discussed, with whom, what was asked and answered, the capacity assessment and its reasoning, and the patient's decision. The note is the evidence that the process happened. |
Assessing capacity Core
Capacity is decision-specific and time-specific. A patient may lack capacity to manage finances and retain capacity to consent to ECT, or the reverse. The four functional abilities to test and record are:
- Understanding — can the patient give back, in their own words, what ECT involves and why it is being offered?
- Retention — can they hold the information long enough to use it? Brief retention is enough; indefinite recall is not required.
- Appreciation and weighing — can they apply the information to their own situation and weigh the options? This is where delusions bite: a patient who accepts every fact but believes they are already dead is not weighing.
- Communicating a choice — can they express a decision, by any means?
Two cautions. First, a refusal is not evidence of incapacity. A patient who understands ECT, weighs it and declines it has exercised capacity, not lost it. Second, the threshold is low as a matter of law: in the Victorian Supreme Court decision discussed by McSherry (2019), Bell J set a low threshold for the capacity to consent to or refuse ECT, and there has been real divergence in how tribunals have understood what capacity means in this context.
How much to disclose about cognitive effects Core
This is the most contested part of the ECT consent conversation, and the part most often done badly. The evidence is set out in Module 6; what follows is how to translate it into what you actually say.
Why the old formula will not do
For years patients were told, in substance, that memory effects are temporary and resolve within weeks. Rose and colleagues' systematic review of patients' own accounts — 26 clinician-led studies and nine studies conducted by or with patients — found that at least one third of patients reported persistent memory loss, and that the level of perceived benefit varied with the method used to ask. Their conclusion was blunt: the Royal College of Psychiatrists' then statement that over 80% of patients are satisfied and that memory loss is not clinically important was unfounded (Rose et al., 2003).
You do not have to accept every inference drawn from that review to accept its implication for consent. If a third of patients say they have lasting memory difficulty, then telling a patient that memory effects always resolve is not a reassurance — it is a statement that a substantial minority of patients will later experience as having been misled. That is a consent failure whatever the mechanism of their complaint turns out to be.
A defensible way to put it
| Topic | What to say |
|---|---|
| Confusion after each treatment | "You will be confused for a while after each treatment — usually minutes to an hour. Some people are more confused than others, and it can build up over a course." |
| Difficulty forming new memories | "During the course and for a few weeks afterwards it is common to have trouble taking in and holding new information — appointments, conversations, what you read. This part usually settles." |
| Loss of memories from around the treatment | "You may lose memories of events from the weeks or months around the course, and sometimes further back. Memories closest to the treatment are the most vulnerable. Some of this comes back and some may not." |
| Persistent autobiographical gaps | "A minority of people describe lasting patchiness in memory for parts of their own life. Studies of this disagree about how common it is and how much is due to ECT as opposed to the depression itself, but it is real enough that I should tell you about it rather than promise you it will not happen." |
| What reduces the risk | "We can reduce the risk by the electrode position we use, the pulse width, the dose and how often we treat — and we will monitor your memory through the course and change the technique if we need to." |
| What we do not know | "Long-term cognitive research in ECT is weaker than it should be. I will not pretend to a precision the evidence does not support." |
The suggested wording above is the author's, offered as a worked model rather than as approved or validated text. Use your service's information materials and your own words.
Treatment without consent Advanced
Here the ethics are genuinely hard, and the literature cuts against the intuition that the cautious course is always to withhold.
- A survey of all German psychiatric hospitals performing ECT identified, over 12 months, just 15 cases of ECT in patients who lacked capacity and were refusing. A marked clinical response was seen in 14 of the 15, and capacity to consent was restored in 11. Alongside this very favourable benefit-to-risk ratio, the authors found widespread inexperience, reservation and uncertainty among psychiatrists, and concluded that a general rejection of involuntary ECT does not seem ethically appropriate (Besse et al., 2019).
- In a chart review of 168 patients treated for depressive episodes, 34 (20%) received ECT on relatives' consent. They were older, had shorter episodes and more often had psychotic, melancholic or catatonic features; the main indication was the need for rapid recovery. Remission (62%) and response (82%) rates and readmission rates did not differ from those who consented themselves. The authors argued that rejecting this practice out of concern about consent may itself be unethical, measured against the principles of beneficence and non-maleficence (Takamiya et al., 2021).
- Against this sits a serious human-rights argument. The UN Special Rapporteur on Torture has called for a ban on the non-consensual use of ECT, and the Convention on the Rights of Persons with Disabilities is read by many as incompatible with substitute decision-making in psychiatry altogether (McSherry, 2019). This is not a fringe position and should not be taught as one.
The rest of the team Advanced
Consent is usually framed as a problem for the treating psychiatrist, but the nurse in the ECT suite is the person most often asked "do I have to do this?". A critical realist review of the qualitative evidence found four themes running through mental health nurses' ethical dilemmas in ECT: the nurse's conflicting role as advocate and participant, consent, questioned efficacy and uncertain mechanism, and side effects and legal issues. The author argued that nurses need an ethical footing broader than the four biomedical principles, including an understanding of the social and political factors surrounding ECT (Sweetmore, 2021).
Practically: anyone who may be asked that question in the suite should know what the patient was told, what the legal status of the treatment is, and who to escalate to. If a nurse in your service cannot answer those three things for the patient in front of them, the consent process has a gap in it.
The Australian legal framework Core
ECT is more tightly regulated in Australia than almost anywhere, and the detail is state and territory legislation that changes. The structure is broadly consistent:
- Informed consent in a prescribed form is required for a voluntary patient with capacity, with specified information given and a specified period before treatment begins.
- A tribunal or board authorises ECT where the patient cannot or does not consent, with the patient entitled to representation. In New South Wales this is the Mental Health Review Tribunal.
- Additional safeguards for minors apply, and are stricter the younger the patient (Module 8).
- Emergency provisions exist in some jurisdictions for genuinely life-threatening situations, with mandatory notification and review afterwards.
- Reporting and audit obligations attach to the service, not only to the treating doctor.
Lived experience and public image
Every patient arrives having already formed a view of ECT, usually from a film. This module covers what patients themselves report, how ECT is portrayed and why, and the history that gives the fear its purchase.
Rose D, Fleischmann P, Wykes T, Leese M, Bindman J. Patients' perspectives on electroconvulsive therapy: systematic review. The paper that changed what services can honestly claim about satisfaction and memory.
BMJ 2003;326:1363. doi:10.1136/bmj.326.7403.1363 · PMID 12816822
Sienaert P. Based on a true story? The portrayal of ECT in international movies and television programs. A systematic count of what the public has actually been shown.
Brain Stimul 2016;9:882–91. doi:10.1016/j.brs.2016.07.005 · PMID 27522170
Lived experience
What patients themselves report about ECT, and what they say they want done differently.
What patients report Core
- How you ask changes what you hear. Rose and colleagues found that perceived benefit tracked the method used to elicit it — clinician-administered questionnaires in clinical settings produced higher satisfaction than studies designed or conducted by patients themselves. At least a third reported persistent memory loss (Rose et al., 2003). This is the same methodological point made about subjective memory in Module 6, and it applies to satisfaction just as much as to side effects.
- Experience is genuinely divided, not merely mixed. A meta-synthesis of 12 qualitative studies published between 2010 and 2024 — titled Between Hope and Trauma at the Last Resort — identified nine subthemes under three primary themes and concluded that patients' experiences of ECT are conflicting (Shin et al., 2025). The same treatment is described as rescue by some patients and as harm by others, and both accounts are first-hand.
- What patients say they want is mostly about information and support. In a consumer-led Australian study, 17 people who had received ECT were interviewed by consumer researchers who had also received it. Four categories emerged: information matters; preparation and decisions before ECT; the experience of ECT; and suggestions for improvement. Most participants wanted more information, offered repeatedly rather than once, greater involvement of family and friends, the chance to hear from others who had received ECT, and more support for managing memory and cognitive effects (Wells et al., 2018).
- Consent can be formally complete and substantively thin. In a qualitative study of 52 patients and their relatives in Vellore, more than half of recipients did not know the details of ECT even at the end of the course, though they were not unhappy to have received it. Relatives generally felt well informed — and also admitted to perceiving coercion. Patient and relative perceptions diverged significantly, as did those of voluntarily and involuntarily admitted patients (Rajkumar et al., 2006).
- Families carry their own burden. A recent qualitative study of parents deciding about ECT for adolescents with severe depression found decision-making anxiety, fear of ECT, a sense of "deciding in a fog", and the weight of considering the adolescent's own feelings — with media representations named among the influences on them (Zhan et al., 2025).
What to do with this in practice Core
| Finding | What it changes |
|---|---|
| Patients want information more than once | Give the information at referral, at consent, at the first treatment and mid-course. Assume it has not been retained — which, given what ECT does to new learning, is a clinical fact rather than a criticism of the patient. |
| Satisfaction depends on who asks | Do not rely on your own post-course question to the patient in the suite. Build in a way of hearing from patients that is not administered by the treating team. |
| Experiences are divided | Do not present a uniformly positive account. A patient who has been told that most people do well, and who then does badly, loses trust in everything else they were told. |
| Peer contact is wanted | Where your service can offer contact with someone who has had ECT, offer it. Patients asked for this specifically. |
| Families perceive coercion even when consent is documented | Ask the family directly whether they feel they had a real choice. A signature does not answer that question. |
| Memory support is asked for and rarely given | Practical strategies — written notes, a diary, a support person at appointments, deferring major decisions — belong in the treatment plan, not in an afterthought. |
Media, activism and the history behind the fear
How ECT is portrayed, the scholarly critique, and the historical abuses that give the fear its purchase.
What the public has been shown Core
Sienaert's review counted every traceable depiction from 1948 to 2016: 52 movies (57 scenes), 21 television programmes (23 scenes) and 2 animated sitcoms. The findings are worth knowing precisely, because they explain the specific fears patients arrive with:
| Feature of the depiction | Proportion |
|---|---|
| Given without consent | 57% of scenes (47/82) |
| Given without anaesthesia | 72% of scenes (59/82) |
| Bilateral electrode placement | 89% of scenes (73/82) |
| Main indication in movies: behavioural control or torture | 29.8% (17/57) |
| Main indication in TV: erasing memories | 28% (7/25) |
| Negative and inaccurate overall | 81% of movies, 72% of TV programmes |
Sienaert's conclusion was that ECT functions on screen as a metaphor for repression and mind control, shown as memory-erasing, painful and damaging. McDonald and Walter traced the same trajectory in American film across 22 titles from The Snake Pit (1948) to Requiem for a Dream (2000): ECT begins its screen career as a severe but helpful remedy and becomes progressively more brutal, until the depiction has little in common with the treatment as practised. Film-makers, they observed, now appear influenced more by One Flew Over the Cuckoo's Nest than by any evidence about ECT (McDonald and Walter, 2001; 2009).
Stigma, rather than evidence, remains the principal barrier to public acceptance (Dowman et al., 2005).
The historical antecedents: ECT, coercion and political abuse Core
The screen image did not come from nowhere. Three distinct historical strands feed it, and they need separating, because conflating them is how the argument is usually lost.
1. Unmodified ECT and institutional behaviour control
ECT was introduced by Cerletti and Bini in 1938 and spread quickly, partly because it was cheap and simple to administer (Patriarca et al., 2021; Berrios, 1997). For its first two decades it was given without anaesthesia or muscle relaxant — the patient was conscious at the moment of stimulation and sustained a full tonic–clonic convulsion, with fractures a recognised complication. In the large institutions of the 1940s to 1960s it was also, in places, administered in ways that are not defensible as treatment: in high frequency to produce "regression", to disturbed or merely troublesome patients, and with an understood disciplinary function. The memory of that practice is accurate. Many of the people who hold it, or whose parents held it, are still alive.
2. Political abuse of psychiatry
The wider fear that psychiatry can be turned into an instrument of the state is also well founded. Van Voren's historical overview records that in the Soviet Union approximately one third of political prisoners were held in psychiatric hospitals, that the issue split the World Psychiatric Association so deeply that the Soviet society was forced to withdraw in 1983 and returned only conditionally in 1989, that systematic abuse also occurred in other socialist states and on a systematic scale in Romania, and that comparable systematic political abuse became evident in the People's Republic of China in the 2000s (van Voren, 2010; van Voren, 2002).
ECT is not prominent in that catalogue. It was not the characteristic instrument of Soviet political psychiatry, and the standard summaries of these sources do not list it among the punitive treatments. But the widespread impression that it was is easy to understand and is not simply an error. Several of the documented methods were themselves shock or convulsive treatments — insulin shock, and sulfazin, which induces convulsions. English-language accounts sometimes refer to "electric shocks" without distinguishing punitive electrocution from ECT. And "electroshock" is the word that circulates in this literature generally. A reader of this material may reasonably come away with that impression.
Why the distinction is still worth making. It cuts both ways. Telling a patient that ECT was the Soviet torture instrument is not supported by the documentary record. Dismissing their fear is worse, because the abuse of psychiatric treatment for political control is established historical fact, and ECT's screen image (Part B above) has become the symbol of that abuse precisely because it looks like what people imagine coercion to look like. The patient is afraid of something real. The accurate answer is not "that never happened" but "that happened, here is what was actually done, and here is what we are proposing instead".
3. Coercive use within living memory, and the live legal question
The question is not only historical. Non-consensual ECT is lawful in Australia under tribunal authorisation today, the UN Special Rapporteur on Torture has called for it to be banned, and the Convention on the Rights of Persons with Disabilities is read by many advocates as prohibiting it (Module 9). A patient or family member who says "they can give it to you against your will" is not repeating a film plot. They are describing the law correctly. What they usually do not know is the scale — the German national survey found 15 such cases in a year across every ECT-providing hospital in the country (Besse et al., 2019) — or the safeguards.
The three-strand framing in this section, and the characterisation of institutional practice in the mid-twentieth century, are the author's synthesis of the historical sources cited rather than claims drawn from a single paper. The statement that ECT is not prominent in the documented Soviet catalogue rests on the standard secondary summaries of Bloch and Reddaway and on the 1991 US delegation report, not on a reading of the primary Russian-language material; isolated instances cannot be excluded.
The scholarly critique, and the exchange Advanced
Distinct from popular demonisation, there is a sustained academic argument against ECT. Trainees should be able to state it accurately, because dismissing it unread is both bad scholarship and bad persuasion.
| The critique | The response |
|---|---|
| Read and Bentall's review of ECT versus simulated ECT concluded that the placebo-controlled studies show minimal support for effectiveness in depression or schizophrenia during treatment, no evidence of benefit beyond the treatment period, no placebo-controlled evidence that ECT prevents suicide, and — set against evidence of persistent amnesia and a small increase in mortality risk — a cost–benefit balance so poor that its use "cannot be scientifically justified" (Read and Bentall, 2010). The argument was extended to antidepressants and ECT together, framing depression as a meaningful response to life circumstances rather than a biological deficit (Read and Moncrieff, 2022). | Kirov and colleagues set out the evidence-based case that modern ECT is the most effective treatment for severe, psychotic and treatment-resistant depression, and is safe (Kirov et al., 2021). Meechan and colleagues argued specifically that the critique rests on narrative reviews that contradict existing systematic reviews and meta-analyses, and set out the interpretive errors that narrative reviewing invites (Meechan et al., 2022). The dispute was also run as a formal head-to-head in the BMJ — "Should we stop using electroconvulsive therapy?" — with Read and Cunliffe arguing yes and Jauhar and McLoughlin arguing no (2019); it is the most efficient single place to read both sides. |
Talking to a frightened patient or family Core
- Ask what they have seen or heard first. You cannot correct a picture you have not been shown. "What do you already know about ECT?" and "where did that come from?" take thirty seconds.
- Name the film. If it is Cuckoo's Nest, say so. Explaining that the scene shows ECT given without anaesthetic, without consent, as a punishment, and that all three are the opposite of what is proposed, addresses the actual fear rather than a general one.
- Concede what is true. It was once given without anaesthesia. It has been used coercively. It can still be given under tribunal authorisation. Memory effects are real and in a minority may persist. Every concession spends credibility that then buys the rest.
- Do not oversell. "Most people do very well" is defensible; "you will be fine" is not, and will be remembered if they are not.
- Describe what they will actually experience. Walking in, the cannula, the oxygen mask, falling asleep, waking in recovery with a headache and some confusion. Concrete description displaces imagined imagery better than statistics do.
- Offer peer contact and a second conversation. Both were asked for directly by patients (Wells et al., 2018).
Shifting gears in non-response
An algorithm for the acute treatment of depression: start with ultrabrief right unilateral ECT under the right conditions, measure response, and step up the treatment in order when there is non-response. Bilateral placement is the last resort.
Ultrabrief right unilateral ECT
- Pulse width 0.25–0.3 ms, right unilateral (d'Elia) placement
- Dose 6 × titrated seizure threshold
- Three times a week
- Record a baseline depression score (e.g. MADRS) and a cognitive screen
If the situation is life-threatening (refusal of food or fluids, high suicide risk), consider starting in second or third gear, with brief-pulse right unilateral ECT. Bitemporal ECT acted faster in one trial (Kellner et al., 2010); Module 2 discusses this.
After about six to eight treatments: has the depression score changed by 25% or more?
The timing is flexible. Review sooner if the patient is deteriorating.
Stay in gear
Continue to remission, or until improvement has plateaued over two or three treatments. Then plan relapse prevention (Module 7).
Check the delivery first
- Diagnosis: is this an illness ECT treats?
- Seizure quality on EEG, not just duration (Module 5)
- Dose: still 6 × a threshold that may have risen? Re-titrate.
- Drugs: benzodiazepines, anticonvulsants
- Anaesthesia: agent, dose, anaesthetic–ECT interval (Module 3)
- Troubleshooting for poor seizures, including a trial of hyperventilation
Shift to 0.5 ms brief pulse, right unilateral
Widen the pulse from 0.3 ms to 0.5 ms, the preferred first step. Keep right unilateral placement and three treatments a week. Review response again.
Dose. Re-titrating at the new pulse width is the ideal. Whether it is needed in practice depends on the device:
- Pulse width, frequency and train duration set individually: with the other settings unchanged, the charge rises automatically with the pulse width, to about 1.7 times the ultrabrief dose. Re-titration may be unnecessary.
- Dose set as a percentage, as on a Thymatron: the charge can stay the same when the pulse width changes. The absolute dose needs to go up, so re-titrating at the new pulse width is more appropriate.
Shift to 1.0 ms brief pulse, right unilateral
Widen the pulse to 1.0 ms. With the other settings unchanged, the charge is now about 3.3 times the original ultrabrief dose, so re-titration may again be unnecessary. Review response again.
On a Thymatron, re-titrate again at 1.0 ms, as in second gear.
Optimise right unilateral ECT further
- Change the anaesthetic to a less anticonvulsant agent (etomidate or ketamine with propofol)
- Lengthen the anaesthetic–ECT interval
- Review concurrent medication and the diagnosis again
Bilateral placement
Bitemporal or bifrontal brief-pulse ECT at 1.5–2.5 × threshold. Use it only when right unilateral ECT has been optimised and has failed. Expect more cognitive effects; monitor reorientation time and cognition closely.
Consider alternatives
Ketamine or esketamine, an MAOI, lithium augmentation or VNS. Module 7 sets out the evidence for each.
Shifting down
The gears also work in reverse. If the patient is responding but cognitive effects are building up, shift down: from bilateral to right unilateral placement, from brief to ultrabrief pulses, or from three to two treatments a week with brief-pulse ECT.
This algorithm applies to the acute treatment of depression. Schizophrenia needs longer courses before response is judged (Module 2). It is a teaching guide and does not replace clinical judgement or local policy.
Dosing calculators
Three teaching tools to make the arithmetic of ECT dosing concrete. They are for learning, not for prescribing.
Charge from stimulus settings
Charge (mC) = current (A) × pulse width (ms) × number of pulses. With bidirectional pulses, the number of pulses is 2 × frequency (Hz) × train duration (s).
Treatment dose from titrated threshold
Enter the threshold found at titration and the planned placement. The tool shows the usual target range.
Age-based estimate vs titration
The half-age method sets the Thymatron percentage at half the patient's age. Compare it with a titrated dose to see how far the two can diverge.
Age explained only 15% of the variance in threshold, and formula dosing would have over-treated most patients. The more the two figures above differ, the stronger the case for titration.
Girish K, Mayur PM, et al. J ECT 2000;16:258–62. doi:10.1097/00124509-200009000-00006
Case vignettes
Work through each case step by step. Choose an answer to see feedback and unlock the next decision.
Assessment answers
The answer and a short explanation for each module assessment question. Attempt the assessment at the end of each module before reading these.
References and further reading
Key guidelines and trials cited in the modules. The author's own studies are listed separately on the Author's research page.
Textbooks and guidelines
- Royal Australian and New Zealand College of Psychiatrists. Professional practice guidelines for the administration of ECT (Weiss et al., 2019). Listed above.
- Royal College of Psychiatrists. The ECT Handbook, 4th edition. Cambridge University Press, 2019.
- American Psychiatric Association. The Practice of Electroconvulsive Therapy: Recommendations for Treatment, Training and Privileging. Task Force report, 3rd edition. APA, 2024. The source of the anaesthetic doses and anticonvulsant ratings in Module 3.
- Your state or territory mental health legislation and local health district ECT policy.
Author's research
Citations retrieved from PubMed. Links open the publisher's page via DOI, or the PubMed record where no DOI is listed.