VEPs Tested as Biomarker for Visual Hallucinations in Late-Stage Alzheimer's

By Elena Vasquez

A study in the September 2026 issue of Clinical EEG and Neuroscience asks whether visual evoked potentials can serve as an electrophysiological biomarker for visual hallucination pathway integrity in late-stage Alzheimer's disease.

The article appears in Volume 57, Issue 5, pages 458-463. Its abstract frames VEPs as noninvasive electrophysiological measures that reflect the functional integrity of the visual conduction pathway, then sets out to test that hypothesis against the visual hallucination phenotype of late-stage AD.

Why does visual hallucination matter in late-stage AD?

Visual hallucinations complicate the differential diagnosis against Lewy body dementia. They accelerate caregiver burnout. They frequently trigger antipsychotic prescriptions whose own morbidity is well documented. The clinical question is whether the underlying pathway breakdown can be measured objectively, before bedside observation accumulates enough events to make the diagnosis a self-fulfilling one.

The economic question is simpler still. A biomarker that distinguishes functional from organic visual disturbance in a non-cooperative patient is worth a great deal to the memory clinic, the nursing home, and the antipsychotic-prescribing primary care physician.

What do VEPs bring to the EEG lab?

Pattern-reversal VEPs are among the cheapest, fastest studies a clinical neurophysiology laboratory can offer. The principal output is the P100 peak latency and amplitude recorded over occipital midline electrodes (Oz, with O1 and O2 lateral), reflecting conduction through the optic radiations and arrival at the primary visual cortex. Flash VEPs substitute for pattern reversal when fixation fails, a non-trivial consideration in late-stage dementia, where gaze cannot be reliably controlled.

The clinical neurophysiology community has run VEPs for decades to localize anterior visual pathway lesions, distinguish functional from organic visual loss, and track demyelinating disease activity. Extending that tradition into neuropsychiatric phenotyping is the conceptual move the September study makes.

What does the abstract actually say?

The published snippet states only the study's framing objective. Sample size, stimulation protocol, electrode montage, comparison group, and effect sizes are not given in the available excerpt. Any verdict on clinical utility must wait for the full text.

For now, the practical reading is straightforward. If P100 latency prolongation or amplitude reduction correlates with current or subsequent visual hallucination burden in late-stage AD, then a 20-minute chair-side test could flag patients warranting antipsychotic caution earlier than bedside observation alone. If the correlation proves weak or absent, VEPs join a long list of physiological measures neuropsychiatry has tried and shelved.

What does this change for the monitoring unit and the memory clinic?

Probably nothing this cycle. Clinical neurophysiology laboratories already run VEPs on ophthalmology referrals, and the September paper does not propose a new protocol or cut-off. What it does propose is a hypothesis worth testing in larger, prospective cohorts that pair standardized VEP recording with structured psychiatric phenotyping.

The forward-looking question is whether memory clinics will begin ordering VEPs as part of the late-stage AD workup, or whether the modality will remain confined to ophthalmology and multiple sclerosis referrals. The answer arrives when the next cohort study reads out.