Scalp EEG and the Question of Antipsychotic Response

A Psychiatric Times article has put a straightforward question to its clinical readership: can EEG help predict which patients will respond to antipsychotics? The headline — "Predicting Antipsychotic Response: Could EEG Help?" — signals a field still at the hypothesis-testing stage, not a device awaiting a clearance letter.

The framing deserves attention on its own terms. Psychiatric Times addresses prescribers, not electroneurodiagnostic technologists. When a psychiatry outlet raises EEG, the practical question arrives quickly: what would a predictor of antipsychotic response actually change at the bedside?

The answer, at minimum, is timing. Treatment trials in psychosis run for weeks before efficacy becomes clear. A biomarker drawn from a routine scalp recording — no sedation, no radiation, no specialized implant — would let clinicians abandon failing regimens earlier. That is the prize. Whether EEG can deliver it is what the article examines.

What kind of EEG are we discussing?

The source headline does not specify a montage, cohort, or modality, and this report will not invent one. The relevant distinction still matters for readers of this site.

Quantitative EEG work in psychiatry typically rests on scalp recordings, analyzed offline for spectral power, coherence, or event-related potentials. That is a different animal from the visual-read routine EEG of the epilepsy monitoring unit. If a response-predictor emerges from this literature, it will almost certainly live in the quantitative pipeline — dense-array acquisition, processed metrics, and a validated classifier — rather than in a visual pattern any reader can call from a page.

Nothing in the source indicates an FDA-cleared predictive device. Readers should treat this as investigational research territory until a sponsor, a trial identifier, and a clearance or approval appear.

The precedent problem

EEG-based prediction is not a new ambition. The field has seen quantitative EEG algorithms proposed for distinguishing dementia subtypes, forecasting antidepressant response, and stratifying patients across several drug classes. Commercial ventures — the now-familiar story of EEG-based diagnostics startups — have promised more than the evidence sustained.

So restraint is warranted twice over. First, psychiatric response prediction is multi-factorial; a single-modality predictor faces a high bar against clinical judgment and pharmacogenomic panels. Second, replication in psychiatry has been historically fragile.

None of this makes the question illegitimate. Antipsychotic non-response is common, costly, and clinically demoralizing for patients and families. A cheap, repeatable physiological measure that improves first-choice prescribing even modestly would carry real value.

What does this change for the reading room?

For now, nothing operationally. No guideline asks electroencephalographers to produce a response prognosis alongside routine reports. The story's significance is directional: it marks continued interest in moving scalp EEG from a seizure-detection tool toward a pharmacodynamic assay.

Technologists and lab managers may see the downstream effects first — requests for standardized quantitative acquisition protocols, or research collaborations seeking archival scalp EEG datasets with linked medication outcomes. Well-annotated archives will be the raw material for any classifier that follows.

Where the burden now sits

The Psychiatric Times piece frames a question rather than announcing a result. The burden sits with prospective, blinded, multi-site studies: preregistered outcomes, defined montages, and a stated classifier locked before data collection. Until those elements appear in print, the honest summary is that EEG remains a candidate, not a tool.

Watch the trial registries. The next durable fact in this story will arrive there first.