Simulating stimulus dose titration to seizure threshold in electroconvulsive therapy — from empirical first-treatment dosing to individualized maintenance across a treatment course
Electroconvulsive therapy (ECT) remains the most effective acute treatment for severe, treatment-resistant, or life-threatening episodes of major depression, catatonia, and certain other psychiatric conditions. Before any electrical stimulus is delivered, a full pre-anesthetic and psychiatric baseline is established, and the very first treatment in a course is dosed using an estimation method rather than a value already known to be therapeutic for that individual.
Before the first treatment, patients undergo psychiatric, medical, cardiac, and cognitive baseline assessment (history, physical exam, ECG, basic labs, cognitive screening such as MMSE/MoCA). An anesthesiologist evaluates airway and cardiac risk, since ECT is delivered under brief general anesthesia.
During the procedure itself, monitoring includes: • Continuous single- or multi-channel EEG — the definitive record of generalized seizure activity • A blood-pressure cuff placed on one limb before muscle relaxant is given ("cuff method"), isolating that limb from paralysis so the motor seizure can still be observed and timed • Continuous ECG and pulse oximetry throughout anesthesia • Bispectral index or clinical anesthesia depth monitoring in some centers
Oxygen is given by mask, an anesthetic agent (methohexital, etomidate, or occasionally propofol/ketamine) is administered at a sub-anticonvulsant induction dose, followed by succinylcholine to minimize motor convulsion and the risk of musculoskeletal injury.
Because an individual patient's seizure threshold is unknown before the first treatment, clinicians rely on one of two established estimation strategies rather than guessing:
• Age-based (half-age) method: an empirical rule of thumb — dose (as % of device maximum output) is initially set close to half the patient's age in years, reflecting the well-documented rise in seizure threshold with age, sex, and electrode placement.
• Formula-based method (e.g., benchmarking algorithms incorporating age, sex, and electrode placement): produces a starting estimate meant to fall near, but not necessarily at, that individual's actual threshold.
Neither method identifies the true threshold with precision — both exist to select a reasonable, safe starting point for the empirical titration that follows in Stage 2. Device output is delivered as a brief-pulse or ultra-brief-pulse constant-current square-wave stimulus, characterized by pulse width, frequency, current, and train duration, and expressed clinically as a percentage of the device's maximum available charge.
The dose titration method treats the first ECT session as a measurement procedure: rather than assuming a single estimated dose will produce an adequate seizure, the clinician delivers a low initial stimulus and, if it fails to produce a generalized seizure, re-stimulates at a higher output after a brief pause — repeating until an adequate seizure occurs. The lowest dose that produces an adequate generalized seizure is that session's empirical seizure threshold.
Titration proceeds as a controlled dose-escalation sequence within a single anesthetic episode:
1. Deliver an initial low stimulus (well below the anticipated threshold, e.g. ~5% of device output). 2. Observe the motor cuff limb and the EEG trace for 20–50 seconds. 3. If no generalized motor and EEG seizure occurs (a "missed" or subconvulsive stimulation — brief afterdischarge only, no sustained rhythmic activity), wait roughly 20–60 seconds for anesthetic and physiologic stability, then re-stimulate at a higher dose. 4. Repeat, typically for up to 3–4 total stimulations in one session, until an adequate generalized seizure is obtained. 5. The lowest dose producing an adequate seizure is recorded as the empirical seizure threshold for that session.
Because each re-stimulation still requires anesthesia and carries cumulative risk (prolonged seizure, cardiac stress, apnea time), the number of attempts per session is capped in practice — if threshold is not reached within the safe number of re-stimulations, titration is deferred to the next session rather than pursued indefinitely.
Seizure threshold itself is not the treatment dose — it is the calibration point. Once threshold is known, subsequent treatments in the course are dosed as a multiple above that threshold, because dosing relative to threshold (rather than as a fixed device percentage) is what predicts both efficacy and cognitive burden:
• Bilateral (BL) electrode placement: typically dosed at roughly 1.5–2.5× the measured seizure threshold • Right unilateral (RUL) placement: requires substantially higher relative dosing, commonly 5–6× threshold, to achieve comparable antidepressant efficacy — a landmark finding from Sackeim and colleagues' dose-response studies
This relative-dosing principle is precisely why threshold titration matters clinically: two patients receiving the identical absolute device percentage can be at very different points relative to their own threshold, with correspondingly different odds of an adequate seizure and different side-effect burden.
Titration is only performed at treatment initiation (and periodically for re-titration, see Stage 4) — not before every session. Once the working dose is established, most subsequent treatments are delivered directly at the calculated maintenance multiple.
Not every electrical stimulus that reaches the brain produces a therapeutically adequate seizure. Clinicians confirm adequacy using both directly observed motor convulsion (in the cuffed limb) and the EEG record, which is the more reliable indicator because succinylcholine can markedly attenuate visible motor activity while the electroencephalographic seizure proceeds normally.
An adequate ECT seizure on EEG has a recognizable morphology:
• Onset: an abrupt rise in amplitude and frequency, often with brief low-voltage fast activity immediately after the stimulus • Ictal phase: high-amplitude, rhythmic polyspike-and-wave or spike-and-slow-wave activity, becoming more synchronized and slowing in frequency as the seizure progresses • Termination: an abrupt drop in amplitude • Post-ictal suppression: a period of markedly flattened, low-amplitude EEG immediately following seizure termination — considered a favorable sign of a well-organized generalized seizure
The motor convulsion, observed in the cuffed limb, typically appears as tonic then clonic movement but frequently runs shorter than the EEG seizure duration because succinylcholine dampens peripheral muscular expression without altering central seizure activity — this is precisely why EEG, not motor observation alone, is the primary adequacy criterion.
Seizure duration is a necessary but imperfect proxy for seizure quality and therapeutic adequacy, and both ends of the spectrum carry meaning:
• Too short (<15–20 s) / subconvulsive: the stimulus was likely below or only marginally at threshold; a brief afterdischarge without sustained generalized rhythmic activity is not considered therapeutic and should prompt re-stimulation at a higher dose (within the same session, per Stage 2) or dose escalation for the next session.
• Within target (roughly 20–50 s): generally taken as an adequate seizure supporting continued treatment at the current dose.
• Prolonged (>~120–180 s): raises concern for evolving status epilepticus and typically prompts active pharmacologic termination (e.g., additional benzodiazepine or anesthetic agent) and dose reduction at the next session.
Duration alone does not guarantee clinical response — some adequate-appearing seizures are still followed by non-response, and clinicians increasingly also weigh seizure morphology (postictal suppression, interhemispheric coherence) alongside raw duration.
Duration targets are guidelines, not rigid cutoffs — a 19-second seizure with robust post-ictal suppression may still be judged adequate by an experienced clinician, while a 45-second seizure with poor morphology may prompt dose reconsideration.
ECT is, paradoxically, its own anticonvulsant. Repeated induced seizures produce durable neuroadaptive changes — increased GABAergic inhibitory tone, altered receptor sensitivity — that raise the seizure threshold as the treatment course progresses. A dose that produced a comfortably adequate seizure at treatment 1 can become subthreshold by treatment 8 or 9 if never re-evaluated.
Each induced generalized seizure appears to leave behind a lasting anticonvulsant trace — analogous, in reverse, to kindling. Proposed mechanisms include upregulation of inhibitory GABA-A receptor signaling, changes in endogenous neuropeptide and neurosteroid anticonvulsant systems, and altered ion channel excitability following repeated seizure induction.
The practical consequence is that seizure threshold is not a fixed patient trait to be measured once — it is a moving target that typically increases as more treatments are delivered, most notably across the first several sessions of a course, though it continues to drift more gradually thereafter. This is distinct from (and layered on top of) baseline threshold differences driven by age, sex, medication effects (notably anticonvulsant and benzodiazepine co-administration, which independently raise threshold), and electrode placement.
Because threshold is not re-measured before every treatment, drift is usually inferred indirectly:
• Seizure duration trending shorter at a fixed dose over successive sessions • A treatment producing a frankly subconvulsive response at a dose that had reliably worked previously • Progressive weakening of seizure morphology (reduced amplitude, less clear post-ictal suppression) even when duration nominally remains in range
When drift is suspected, clinicians typically respond by empirically increasing the stimulus dose at the next session (often in defined increments, e.g., 5–15 percentage points of device output) and monitoring the resulting seizure, rather than repeating a full multi-stimulation titration procedure. Formal re-titration (repeating the Stage 2 stepwise procedure) is reserved for larger or unexplained shifts, or when a treatment produces an unequivocally subconvulsive response.
A course that never adjusts for threshold drift risks silently under-dosing later treatments — seizures that look "present" on cursory motor observation may in fact be subconvulsive on EEG, reducing the antidepressant benefit of the remaining sessions without the treatment team realizing it.
The entire titration process exists to solve one balancing problem: deliver enough suprathreshold stimulus to reliably produce a therapeutic seizure, while minimizing the dose- and placement-dependent cognitive side effects — anterograde and retrograde amnesia — that are ECT's principal drawback. Dosing decisions made across a course are ultimately individualized trade-offs, not a single universal number.
Efficacy is strongly related to dosing relative to threshold rather than to absolute device output. Below threshold, no generalized seizure — and no antidepressant benefit — occurs. Just above threshold, seizures may be adequate in duration but weaker in therapeutic effect, particularly with right unilateral placement, which requires markedly higher relative dosing (roughly 5–6× threshold) to match the efficacy of bilateral placement at only 1.5–2.5× threshold.
Beyond a certain suprathreshold point, further dose increases yield diminishing incremental efficacy — the dose-response curve for antidepressant benefit tends to plateau, while the dose-response curve for cognitive side effects continues to climb. This divergence is the crux of individualized dosing: identify the point that captures most of the efficacy benefit without paying an unnecessary cognitive price.
Cognitive side effects of ECT — most commonly transient post-treatment disorientation, anterograde memory difficulty during the course, and retrograde amnesia for events around the treatment period — scale with both electrode placement and relative stimulus dose:
• Bilateral placement produces broader cognitive effects than unilateral placement at comparable relative doses • Right unilateral placement, especially with ultra-brief pulse width, substantially reduces cognitive burden while requiring higher relative dosing to preserve efficacy • Higher suprathreshold multiples within a given placement increase cognitive side effects incrementally
Managing this trade-off individually means: selecting an electrode placement and pulse width matched to the patient's efficacy needs and cognitive vulnerability, dosing at the point on the relative-dose curve that yields adequate seizures without excess margin, tracking drift (Stage 4) so dose is not needlessly higher than required, and monitoring cognition serially through the course so the strategy can be adjusted — including switching placement — if side effects become limiting.
There is no single "correct" dose in ECT. The individualized strategy is a moving target across the whole course: track seizure adequacy every session, expect threshold to rise, and titrate dose to the minimum that reliably produces an adequate seizure for the chosen electrode placement — not the maximum the device can deliver.