Convulsive Therapy

Abstract

Convulsive therapy is the deliberate induction of a generalised seizure to treat severe psychiatric illness, a somatic treatment whose active ingredient is the seizure itself rather than the agent that provokes it. It began in 1934 when Ladislas Meduna induced seizures chemically on the mistaken premise that epilepsy and schizophrenia were biologically antagonistic, and was transformed in 1938 when Ugo Cerletti and Lucio Bini replaced the convulsant drug with a controlled electrical stimulus, creating electroconvulsive therapy. Modern practice retains the seizure as therapeutic while minimising its cost to memory, because the same treatment that most reliably lifts severe depression also disrupts the consolidation and retrieval of recent memories. Convulsive therapy thus poses a sharp cognitive question: how can a seizure heal mood while transiently harming memory, and can the two effects be separated?

Keywords: convulsive therapy, electroconvulsive therapy, therapeutic seizure, retrograde amnesia, treatment-resistant depression

Convulsive therapy is a somatic psychiatric treatment in which a brief, generalised seizure is induced under controlled conditions to relieve severe mental illness, most often a major depressive episode that has not responded to medication. Its defining and initially counterintuitive claim is that the seizure is the therapeutic event: neither the drug nor the electrical current that triggers it carries the benefit, only the self-limiting burst of synchronised brain activity that follows (Fink, 2001). MeSH classifies convulsive therapy under psychiatric somatic therapies, the family of treatments that act on the brain through physical rather than psychological or purely pharmacological means. For cognitive psychology the procedure is of unusual interest because it is a rare instance in which a single, datable event both lifts mood and, as a side effect, transiently impairs memory — turning the clinic into a natural experiment on the relationship between affect and long-term memory (Sackeim et al., 2007; Semkovska & McLoughlin, 2010).

Types of Convulsive Therapy

In the MeSH tree, convulsive therapy is the parent heading for a small set of narrower somatic techniques, and its immediate descendants are listed in Table 1. The categories are cross-cut by two independent distinctions that are worth keeping apart. One is the trigger used to reach the therapeutic state — a convulsant drug, an electrical current, or a magnetic field — which is a matter of engineering rather than of mechanism. The other is whether a seizure is actually induced at all: the historical members of the family (pharmacoconvulsive and electroconvulsive therapy) depend on a generalised seizure, whereas one indexed descendant, transcranial direct current stimulation, delivers a weak, deliberately subconvulsive current that never provokes a seizure at all. Its placement under a heading named for convulsions is an artefact of how MeSH groups brain-stimulation methods for retrieval, not a claim that it is a convulsive treatment; the MeSH tree is an indexing classification built to organise the literature, not a mechanistic taxonomy, and these entries should be read in that spirit.

Origins: From Camphor to Cerletti

The field was founded on a hypothesis now known to be false. In the early 1930s the Hungarian neuropsychiatrist Ladislas Meduna, impressed by clinical reports that epilepsy and schizophrenia rarely coexisted, proposed a biological antagonism between seizures and psychosis and reasoned that artificially inducing seizures might drive out the illness (Fink, 1984). In 1934 he began injecting camphor, and then the faster-acting convulsant pentylenetetrazol, to provoke seizures in patients with catatonic schizophrenia, reporting striking remissions. This pharmacoconvulsive therapy was the first member of the family, and although Meduna's antagonism premise did not survive scrutiny, his central empirical claim — that induced seizures could relieve severe psychiatric illness — did (Shorter & Healy, 2007).

The chemical method was frightening and hard to control: the latency to seizure was unpredictable and patients experienced a period of dread before losing consciousness. In 1938, in Rome, Ugo Cerletti and Lucio Bini replaced the convulsant drug with a brief electrical stimulus applied to the scalp, which induced a seizure immediately and reproducibly. Electroconvulsive therapy spread rapidly because it was faster, cheaper, and more controllable than its chemical predecessor, and within a decade it had displaced pharmacoconvulsive therapy almost entirely (Fink, 2001; Shorter & Healy, 2007). What the electrical method changed was the trigger, not the principle: the therapeutic agent remained the generalised seizure, a point that would take decades of dose-finding research to establish rigorously.

The Therapeutic Seizure and How It Works

That the seizure is necessary was shown by a clean subtraction. When patients are given the anaesthesia and the electrical stimulus but the current is kept below the seizure threshold, so that no generalised seizure occurs, the antidepressant benefit largely disappears; an adequate seizure must be elicited for the treatment to work (Sackeim et al., 1993). Yet the seizure is necessary without being sufficient: two courses that both produce full-length seizures can differ greatly in efficacy depending on how far the stimulus exceeds the threshold and where the electrodes sit, showing that the details of how the seizure is induced modulate the outcome (Sackeim et al., 2000).

Why a seizure should relieve depression at all remains incompletely understood, and several mechanisms are under active consideration rather than settled (Bolwig, 2011). Candidate accounts include an anticonvulsant effect, in which the brain's own seizure-suppressing systems are recruited and raise the seizure threshold over a course of treatment; changes in monoaminergic and neuroendocrine signalling; and the induction of synaptic and structural plasticity, including neurogenesis in the hippocampus. Whatever the final account, the efficacy itself is not in doubt: pooled analyses and a large systematic review establish electroconvulsive therapy as more effective than pharmacological treatment for severe depression, and as the most effective acute antidepressant treatment available (UK ECT Review Group, 2003; Pagnin et al., 2004; Lisanby, 2007). Response is not uniform, and clinical features such as older age and the presence of psychotic depression predict a stronger response, helping target the treatment to those most likely to benefit (van Diermen et al., 2018).

Electrode Placement and Stimulus Dose

The single most consequential decision in modern practice, and the one that matters most to cognition, is how the seizure is elicited. Two variables dominate. Electrode placement sets the path of current through the brain: bitemporal (bilateral) placement drives current through both temporal lobes and the medial temporal memory structures beneath them, whereas right unilateral placement confines the current largely to the non-dominant hemisphere, sparing the left-hemisphere systems most involved in verbal memory. Stimulus dose, the electrical charge delivered, is expressed relative to each patient's individually measured seizure threshold, because that threshold varies severalfold across people (Sackeim et al., 1993).

These two variables interact in a way that reframes dosing as a cognitive problem. Bitemporal treatment is reliably effective even at low multiples of the seizure threshold, but exacts the greatest cost to memory. Right unilateral treatment is gentler on memory, but only becomes comparably effective when dosed well above threshold — at roughly six times threshold it matches bitemporal efficacy while producing markedly fewer cognitive side effects (Sackeim et al., 2000). The advent of the ultrabrief pulse, an electrical pulse shorter than the classic waveform and closer to the minimum needed to depolarise a neuron, reduced cognitive side effects further by wasting less charge outside the moment of neural firing (Weiner & Reti, 2017). Placement and dose are therefore not merely clinical settings but the levers by which the therapeutic seizure is separated, as far as possible, from its amnestic shadow.

The Cognitive Cost: Amnesia

The characteristic side effect of convulsive therapy is amnesia, and its structure closely parallels the amnesia produced by medial temporal lobe injury, which is unsurprising given the anatomy the current traverses. Two components combine. An anterograde deficit impairs the formation of new memories during and shortly after a course of treatment, so that events around the treatment period are poorly retained. A retrograde deficit impairs access to memories formed before treatment, and — like the retrograde amnesia of organic injury — it is temporally graded, falling most heavily on memories from the weeks and months before treatment while sparing remote and well-consolidated ones (Sackeim et al., 2007). The graded pattern implicates the consolidation of recent episodic memory: the seizure appears to disrupt memories that are still labile and hippocampus-dependent, leaving those already stabilised in cortex comparatively intact.

The clinically decisive question is how much of this loss persists, and here a large meta-analysis reshaped the field. Pooling objective cognitive tests across dozens of studies and thousands of patients, the analysis found that measurable deficits are concentrated in the first three days after treatment, that performance in most domains returns to baseline within about fifteen days, and that beyond that window many cognitive measures actually rise above pre-treatment levels — because lifting the depression itself removes the concentration and effort deficits that depression imposes (Semkovska & McLoughlin, 2010). The anterograde component thus recovers substantially. What can persist is a patch of retrograde amnesia for the period immediately surrounding treatment, most pronounced after bitemporal, high-dose courses and reduced by right unilateral and ultrabrief techniques (Sackeim et al., 2007). The cognitive profile is therefore not a global dementing effect, as lay belief often holds, but a bounded and largely reversible disruption of recent-memory systems.

Placement, Path, and the Memory Trade-off

Because electrode placement determines which memory systems the current reaches, it is the clearest illustration of the treatment's central trade-off. The medial temporal lobe structures that convulsive therapy most affects are the same ones whose bilateral damage produces the classic amnesic syndrome, so a current path that spares them ought to spare memory — and it does. Confining the current to the non-dominant hemisphere, as right unilateral placement does, leaves the verbal-memory machinery of the language-dominant hemisphere comparatively untouched, which is why right unilateral courses produce less impairment of verbal recall and recall of autobiographical detail than bitemporal courses of equivalent antidepressant effect (Sackeim et al., 2000; Sackeim et al., 2007).

This is why dosing decisions are, at bottom, decisions about how much memory disruption is acceptable for a given gain in speed or certainty of response. A severely ill, acutely suicidal patient may warrant bitemporal treatment for its reliability; a patient for whom cognitive preservation is paramount may be treated with high-dose right unilateral or ultrabrief stimulation, accepting a possibly slower response for a lighter cognitive footprint (Kellner et al., 2012). The existence of this dial — a single treatment whose cognitive cost can be turned up or down by changing the current path — is what makes convulsive therapy a standing case study in the neuroscience of memory as well as a psychiatric treatment.

Worked Example

Consider an idealised patient tested on a standardised autobiographical memory interview that samples 50 personal facts and events from the year before treatment. Suppose the patient consistently recalls 45 of the 50 items at baseline, then 30 of 50 one week after a course of bitemporal electroconvulsive therapy, and 42 of 50 at a six-month follow-up. The acute retrograde loss is (45 − 30) / 45 = 15 / 45 = 0.333, a 33.3% drop from the patient's own baseline. By six months, recall has recovered to 42 / 45 = 0.933, or 93.3% of baseline, so the fraction of the acute loss that has resolved is (42 − 30) / (45 − 30) = 12 / 15 = 0.80 — four-fifths of the deficit recovered — leaving a persistent retrograde gap of (45 − 42) / 45 = 3 / 45 = 0.067, about 6.7% (Sackeim et al., 2007; Semkovska & McLoughlin, 2010).

Now place this cognitive cost against the therapeutic gain. Suppose the same patient's depression score falls from 32 to 9 on a standard 52-point rating scale over the course. The symptom reduction is (32 − 9) / 32 = 23 / 32 = 0.719, a 71.9% improvement, and a final score of 9 is below the conventional remission threshold of about 10, so the patient has remitted. The example makes the trade-off numerically explicit: a 71.9% reduction in depressive symptoms is bought at the price of a retrograde amnesia that is 33.3% at its acute worst but resolves to a residual 6.7%. Choosing right unilateral placement instead would be expected to shrink both the acute and the residual memory loss while, if dosed to roughly six times the seizure threshold, preserving most of the antidepressant gain (Sackeim et al., 2000) — the quantitative shape of the placement decision described above.

Discussion

Convulsive therapy occupies a peculiar and instructive place in the sciences of the mind. Clinically it is the most effective acute treatment for severe depression, a status established across pooled trials and unchallenged by any pharmacological alternative (UK ECT Review Group, 2003; Pagnin et al., 2004). The qualifier acute is doing real work: the benefit of a course is powerful but not self-sustaining, and without continuation treatment a large fraction of remitted patients relapse within six months — a landmark randomised trial found that relapse after successful electroconvulsive therapy is common on placebo and is roughly halved by continuation pharmacotherapy, which is why a completed course is now routinely followed by maintenance medication or maintenance treatment rather than treated as a cure (Sackeim et al., 2001). Historically it is a reminder that a treatment can be discovered from a false theory and remain valid once the theory is discarded: Meduna's antagonism hypothesis was wrong, but the seizures it motivated worked (Fink, 2001). And for cognitive psychology it is a naturalistic model of retrograde amnesia with a known onset, a gradient over recent memory, and a largely predictable time course of recovery — a controlled disruption of the very consolidation processes that the study of organic amnesia had inferred indirectly (Sackeim et al., 2007; Semkovska & McLoughlin, 2010).

The unresolved problems are correspondingly deep. The mechanism by which a seizure lifts mood is still not known, which is a remarkable gap for a treatment nearly a century old (Bolwig, 2011). More pointedly for cognition, it is not yet established whether the antidepressant effect and the amnestic effect are inseparable consequences of one process or two dissociable effects that better technique could fully pull apart. The steady reduction of cognitive side effects through right unilateral placement, ultrabrief pulses, and careful dosing suggests the latter is at least partly true, and the search for a form of the treatment that keeps the mood benefit while shedding the memory cost is the practical expression of that hope (Weiner & Reti, 2017; Kellner et al., 2012).

Current Directions

The most active research front applies modern neuroimaging to the old question of mechanism. Large multi-site collaborations pooling structural magnetic resonance scans have shown that a course of electroconvulsive therapy reliably increases the volume of the hippocampus, and that the change is dose-dependent and related to the current path — though, tellingly, the volume increase does not straightforwardly track clinical improvement, complicating any simple neurogenesis account of efficacy (Oltedal et al., 2018). Broader reviews of this imaging work map a widespread pattern of grey-matter and connectivity changes and frame the central puzzle of dissociating the changes that carry the antidepressant benefit from those that accompany the cognitive cost (Ousdal et al., 2022).

A second front seeks to keep the therapeutic seizure while further narrowing its cognitive footprint. Magnetic seizure therapy induces the seizure with a focused magnetic field rather than an electrical current, allowing more precise control of where the seizure begins and shallower spread into medial temporal memory structures; early controlled comparisons found it feasible and safe with a more favourable cognitive profile than conventional electroconvulsive therapy (Lisanby et al., 2003). Alongside these mechanistic and technical advances, large propensity-matched cohort studies have addressed the treatment's medical safety directly, finding no increase in serious medical events attributable to electroconvulsive therapy relative to matched controls, which counters a persistent public perception of danger (Kaster et al., 2021). Together these lines are converging on a treatment that is better understood, more precisely targeted, and increasingly able to separate its healing effect from its cost to memory.

Common Misconceptions

Convulsive therapy causes permanent, global memory loss.

The deficit is selective and largely reversible: an anterograde impairment that recovers within about two weeks and a temporally graded retrograde loss for the period around treatment, not a global erasure of memory or intellect (Semkovska & McLoughlin, 2010).

The electric shock itself is what treats the illness.

The current is only the trigger; the therapeutic agent is the generalised seizure. A subthreshold stimulus that produces no seizure lacks the antidepressant effect (Sackeim et al., 1993).

It is an outdated treatment with no evidence behind it.

It remains the most effective acute treatment for severe depression, supported by systematic review and meta-analysis, and is in routine modern use with refined dosing (UK ECT Review Group, 2003; Pagnin et al., 2004).

Glossary

Anterograde amnesia.

Impaired formation of new memories during and shortly after a course of convulsive therapy; the component that recovers most fully.

Bitemporal placement.

A bilateral electrode montage driving current through both temporal lobes; reliably effective but with the greatest cost to memory.

Convulsive therapy.

The deliberate induction of a generalised seizure to treat severe psychiatric illness; the seizure, not its trigger, is the therapeutic event.

Electroconvulsive therapy.

Convulsive therapy in which the seizure is induced by a controlled electrical stimulus through scalp electrodes; the dominant modern form.

Magnetic seizure therapy.

A newer technique inducing the therapeutic seizure with a focused magnetic field, giving more precise control of seizure onset and a lighter cognitive profile.

Meduna's principle.

The founding, now-discarded premise of a biological antagonism between seizures and psychosis that motivated the first convulsive treatments.

Pharmacoconvulsive therapy.

The original form of convulsive therapy, in which seizures were induced chemically with camphor and then pentylenetetrazol.

Retrograde amnesia.

Loss of access to memories formed before treatment, temporally graded so that recent memories are more vulnerable than remote ones.

Right unilateral placement.

An electrode montage confining current largely to the non-dominant hemisphere, sparing verbal memory; comparably effective only at high stimulus doses.

Seizure threshold.

The minimum electrical charge needed to elicit a generalised seizure in a given patient; it varies severalfold and anchors individualised dosing.

Stimulus dose.

The electrical charge delivered, expressed as a multiple of the patient's seizure threshold; a key determinant of both efficacy and cognitive cost.

Subconvulsive stimulation.

Brain stimulation delivered below the seizure threshold so that no seizure occurs, as in transcranial direct current stimulation.

Therapeutic seizure.

The generalised seizure that carries the treatment's benefit, distinguished from the drug, current, or field that triggers it.

Treatment-resistant depression.

A major depressive episode that has failed to respond to adequate trials of medication; the leading modern indication for electroconvulsive therapy.

Ultrabrief pulse.

A very short electrical pulse closer to the minimum needed to fire a neuron, which reduces cognitive side effects by wasting less charge.

Key Researchers

Lucio Bini (1908-1964). Italian neuropsychiatrist at the University of Rome La Sapienza; with Cerletti he devised the electrical method of inducing the therapeutic seizure in 1938, converting Meduna's chemical convulsive therapy into the controllable procedure that became electroconvulsive therapy. Wikipedia - Wikidata

Ugo Cerletti (1877-1963). Italian neurologist at the Sapienza University of Rome; he led the 1938 introduction of electrically induced seizures for psychiatric treatment, the founding event of electroconvulsive therapy. Wikipedia - Wikidata

Max Fink (1923-2025). Psychiatrist at Stony Brook University and the field's leading modern historian and advocate; his analyses established the therapeutic seizure as the active ingredient and documented the origins and scientific basis of convulsive therapy across more than half a century. Faculty Page - Google Scholar - Wikipedia - Wikidata

Charles H. Kellner (contemporary). Psychiatrist at the Icahn School of Medicine at Mount Sinai; a leading clinical investigator of electroconvulsive therapy in treatment-resistant depression and of the dosing and placement strategies that balance efficacy against cognitive side effects. Faculty Page - ORCID - Wikidata

Sarah H. Lisanby (contemporary). Psychiatrist at the National Institute of Mental Health and professor emeritus at Duke University; she developed magnetic seizure therapy as a more focal form of convulsive treatment and led its first controlled comparisons with electroconvulsive therapy. Faculty Page - ORCID - Google Scholar - Wikipedia

Declan M. McLoughlin (contemporary). Psychiatrist at Trinity College Dublin and St Patrick's Mental Health Services; his meta-analysis of objective cognitive testing reshaped understanding of the time course and reversibility of the cognitive side effects of electroconvulsive therapy. Faculty Page - ORCID

Ladislas J. Meduna (1896-1964). Hungarian neuropsychiatrist, later at the University of Illinois; he founded convulsive therapy in 1934 by inducing seizures chemically to treat schizophrenia, establishing the empirical claim that induced seizures relieve severe psychiatric illness. Wikipedia - Wikidata

Leif Oltedal (contemporary). Physician-researcher at the University of Bergen and Haukeland University Hospital; he coordinates large multi-site neuroimaging collaborations that have mapped the structural brain changes produced by electroconvulsive therapy, including dose-dependent hippocampal volume increase. Faculty Page - ORCID - Google Scholar

Harold A. Sackeim (contemporary). Psychologist at Columbia University; his controlled trials of electrode placement and stimulus dosing defined the trade-off between antidepressant efficacy and memory impairment and established the cognitive profile of the treatment. ORCID - Google Scholar - Wikipedia

Frequently Asked Questions

What is convulsive therapy?
It is a somatic psychiatric treatment in which a brief, generalised seizure is deliberately induced under controlled conditions to relieve severe mental illness, most often major depression. The seizure itself is the therapeutic event, not the drug or current used to trigger it (Fink, 2001).

How is convulsive therapy different from electroconvulsive therapy?
Convulsive therapy is the general category of inducing a therapeutic seizure; electroconvulsive therapy is the specific, now-dominant form that uses a controlled electrical stimulus. The original form induced seizures chemically, and electroconvulsive therapy replaced that trigger in 1938 (Shorter & Healy, 2007).

Does convulsive therapy cause permanent memory loss?
Mostly no. The anterograde deficit typically resolves within about two weeks, and overall cognition often exceeds pre-treatment levels once the depression lifts; a patch of retrograde amnesia for the period around treatment can persist (Semkovska & McLoughlin, 2010).

Why does electrode placement matter for memory?
Placement sets the path of current through the brain. Bitemporal placement drives current through the medial temporal memory structures and impairs memory more, while right unilateral placement spares the language-dominant hemisphere and causes less verbal memory loss (Sackeim et al., 2000).

Is convulsive therapy effective?
It is the most effective acute treatment for severe depression, outperforming medication in systematic review and meta-analysis, with especially strong response in older and psychotically depressed patients (UK ECT Review Group, 2003; van Diermen et al., 2018).

How does inducing a seizure treat depression?
The mechanism is not fully known. Leading candidates include an anticonvulsant effect that raises the seizure threshold over a course, changes in neuroendocrine and monoamine signalling, and seizure-induced synaptic and structural plasticity in the hippocampus (Bolwig, 2011).

Is convulsive therapy dangerous?
Modern practice under anaesthesia has a strong safety record. A large propensity-matched study found no increase in serious medical events attributable to electroconvulsive therapy relative to matched controls, countering its fearsome public image (Kaster et al., 2021).

What newer techniques are being developed?
Magnetic seizure therapy induces the seizure with a focused magnetic field for a lighter cognitive profile, and neuroimaging collaborations are mapping the brain changes that carry the benefit versus the cognitive cost (Lisanby et al., 2003; Oltedal et al., 2018).

References