The Alzheimer's evidence guide
Seizures are a documented part of Alzheimer's disease. Some cause convulsions. Others may briefly interrupt memory, speech, awareness, or behavior. Researchers have also recorded seizures with no obvious outward signs.
These events can be easy to miss or mistake for the dementia itself. This guide explains how seizures can appear, what brain recordings have found, why testing has limits, and what researchers still need to learn.
Start here
A seizure may briefly interrupt awareness, speech, memory, or behavior without causing shaking. In Alzheimer's disease, that change may be mistaken for part of the dementia unless someone notices its sudden start, repeated pattern, and recovery. [4]
Researchers have recorded actual seizures when no clear outward change was noticed. They have also recorded brief electrical discharges linked with seizure risk. A brief discharge is not itself a seizure, and neither finding is the same as general slowing on a brain recording. [7] [8] [9]
Better testing may reveal activity that would otherwise remain hidden. Researchers must still determine which findings affect memory or daily life and whether detecting and treating them leads to better care. [15]
How often seizures are reported
Researchers have reported different percentages because they studied different groups and time periods. One study counted active seizures near the time of dementia diagnosis. Another estimated the chance of developing a seizure after symptoms began. A third reviewed seizures documented across the full course of the disease. Each percentage answers a different question.
In a 2026 study, 174 of 3,040 people whose Alzheimer's disease was confirmed by examining brain tissue after death had clinically active seizures. The researchers counted seizures documented within three years before or after the dementia diagnosis. They excluded people with a clinical history of stroke or traumatic brain injury. This percentage covers a defined period around diagnosis rather than lifetime seizure risk, and routine assessments may underrecord seizures. [1]
Researchers reviewed medical records from 1,320 people treated at a memory center. They estimated that 13.4% would have a seizure after Alzheimer's symptoms began. The rate was nearly ten times that of a comparison group of the same age. This percentage is a calculation based on when seizures occurred, rather than a count of everyone in the study with a recorded seizure. [2]
Researchers reviewed the medical records of 144 people whose Alzheimer's disease was confirmed by examining brain tissue after death. Seizures were recorded in 31.3% at some point between the first symptoms and death. The result applies to this selected group and is not an estimate for the wider community. [3]
Studies found particularly high seizure risk among people whose Alzheimer's symptoms began at younger ages. Seizures also became more common as the disease progressed. These patterns help explain why estimates vary among studies and among patients. [2] [5]
Together, the studies show that seizures affect a meaningful share of people with Alzheimer's disease. Each study examined a different group and period of time. Reading the percentages in their individual contexts provides a more accurate picture than combining them into one estimate.
How seizures can appear
Some seizures cause shaking. Others briefly interrupt awareness, speech, memory, or behavior. Memory-clinic studies found that these less obvious events were sometimes attributed to dementia. A sudden start, repeated pattern, and clear recovery can make an event important to investigate. [2] [4]
An activity stops abruptly. The person stares, does not respond as usual, or appears briefly disconnected. Details such as a sudden start, repeated pattern, and clear recovery can help distinguish the event from ordinary inattention.
Speech may stop or become unexpectedly difficult. The person may not recall a short interval. Usual language and memory problems can make the change harder to recognize or describe.
Some seizures involve an unusual sensation, sudden fear, or a repeated movement. A person with dementia may have difficulty explaining the experience, so observations made during the event can be especially helpful.
Some seizures cause stiffening, shaking, or repeated movement in one part of the body. Confusion, tiredness, or reduced response afterward can also be part of the event. The recovery may be easier for an observer to notice than the seizure itself. [2] [4]
Example of a useful event description
“Stopped speaking during lunch and stared ahead. Did not answer two questions for about 40 seconds. Then resumed speaking but did not recall the pause. A similar sequence was noticed yesterday.”
This note tells a clinician what changed, how long it lasted, how the person responded, and whether it happened before. Those details can support further evaluation without assigning a cause.
Memory loss, word-finding difficulty, and changes in attention may already be familiar. An observer may describe a brief event simply as “more confused.” An incomplete description, an unwitnessed event, or difficulty recalling it can make the pattern harder to recognize. [4]
Fainting, sleep-related events, medication effects, and sudden illness can resemble seizures. A sudden or lasting change may also signal another urgent problem. The person's usual pattern and the full event history help guide evaluation.
Compare the event with the person's usual pattern. What changed? How suddenly did it begin? How long did it last? Did the person respond? What was recovery like? Has the same sequence happened before?
What brain recordings can show
An electroencephalogram, usually called an EEG, records the brain's electrical activity. It may show an actual seizure, a brief electrical discharge linked with seizure risk, general slowing, or another change. These findings have different meanings.
Check what a study means by “silent.” The word may refer to an actual seizure recorded without outward signs or to a brief electrical discharge between seizures. This atlas states which finding each study recorded.
The studies below measured brief seizure-related electrical discharges. They used different participants, equipment, recording lengths, and definitions of an abnormal signal. Those differences help explain the wide range of results.
| Study group | How the recording was done | What researchers found | How to understand the result |
|---|---|---|---|
| Vossel, 2016 33 with Alzheimer's; 19 controls [7] | Overnight EEG plus magnetoencephalography, which measures brain magnetic signals. | 42.4% of the Alzheimer's group; 10.5% of controls. | Small, relatively young, selected sample. Sleep contributed to detection. These are not clinical seizure prevalence figures. |
| Lam, 2020 41 with Alzheimer's without epilepsy; 43 controls [9] | Twenty-four-hour ambulatory EEG, with expert consensus review. | 22% of this Alzheimer's group; 4.7% of controls. | A separate Alzheimer's-with-epilepsy group had more abnormalities. Discharge features differed, so a positive finding is not one uniform category. |
| Szabo, 2026 30 with Alzheimer's; 30 controls [12] | Enhanced overnight video EEG with sleep recording, additional temporal electrodes, and structured review. | 10% of the Alzheimer's group; 3.3% of controls. | The group difference was not statistically significant. Participant selection and interpretation criteria matter alongside recording duration. |
A 2026 review combined results from 13 studies involving 1,373 people without known epilepsy. The results varied widely. A single combined percentage cannot explain differences in who was studied, how the recordings were done, and what each research team counted as a discharge. [16]
Limits of EEG detection
Alzheimer's studies have recorded brain activity in several ways. Some used overnight video EEG, 24-hour ambulatory scalp EEG, or 24-hour inpatient EEG. One used small electrodes worn in the ears for two days at a time. Others recorded from the scalp while also placing electrodes closer to deep areas involved in memory. Longer recordings give researchers more time to capture activity that comes and goes, especially during sleep. Deep recordings have also found activity that did not appear on scalp EEG. The type and length of recording affect what a study can find.
Brief activity may not occur during a short recording. Sleep is especially important in this research. Repeated or longer recordings provide more opportunities to capture activity that comes and goes. [7] [11]
Electrical activity that begins deep in the brain may produce little or no clear signal on scalp EEG. Small Alzheimer's studies used electrodes placed close to deep memory areas and recorded discharges and seizures that did not appear on the scalp recording. [8] [10]
Movement, normal brain patterns, electrode placement, and the rules used to identify a discharge can all affect the result. A useful method must separate meaningful electrical activity from recording interference and false findings. [11] [12]
A normal scalp EEG may leave an event unexplained. Its meaning depends on what happened, when the recording took place, how long it lasted, where the electrodes were placed, and how the signal was interpreted.
A detection percentage tells us how many people had a particular finding under one study's conditions. Test sensitivity means how often a test finds activity that is truly present. Measuring sensitivity requires another dependable method for confirming whether the activity occurred.
Very small studies using electrodes close to deep memory areas found that at least 95% of the brief electrical discharges had no matching signal on scalp EEG. The activity counted in that percentage included brief discharges as well as seizures. It does not mean that ambulatory EEG fails to detect 95% of seizures without outward signs in all people with Alzheimer's disease. [8] [10]
Why this matters for care
Studies have linked seizures and brief seizure-related electrical discharges with faster changes in memory and everyday function. Researchers are also studying their possible connections with sleep and the brain changes caused by Alzheimer's disease. These findings show why the field matters, although they do not yet establish cause and effect.
A large national database study found greater problems with memory, thinking, and daily function among people with Alzheimer's dementia who had a history of seizures. The researchers accounted for age and how long each person had lived with the disease. The study found a relationship, but it could not show that seizures caused the decline. [5]
About 7.5 months later, 38 of 54 people with available follow-up were still classified as having active seizures. “Active” included a seizure during the previous year or continued medical management for seizures. Many participants had no follow-up information, and the study did not analyze the effects of medication. The 70.4% figure applies only to this small follow-up group. [5]
A study used 24-hour EEG in 52 people with Alzheimer's disease and included 38 in its three-year follow-up analysis. People with brief seizure-related electrical activity declined about 1.5 times faster on the study's overall measure of memory and thinking. People who developed seizures during follow-up were excluded from this analysis. [6]
Another small study found a similar link. Together, the findings support further research. They cannot tell us whether the electrical activity contributes to decline, signals a more aggressive form of Alzheimer's disease, or reflects both processes. [7]
The hippocampus is a deep brain area that is important for memory and affected by Alzheimer's disease. Deep recordings have found abnormal electrical activity in and around this area, often during sleep. Researchers use the term “network hyperexcitability” for a tendency toward excessive electrical activation. The term does not mean that every person with Alzheimer's disease has epilepsy. [8] [10]
A recent study of people whose diagnosis was confirmed after death also linked seizures with advanced tau changes and amyloid buildup in the walls of brain blood vessels, known as cerebral amyloid angiopathy. These findings offer clues about possible biological connections. An Alzheimer's diagnosis alone cannot show whether one person is having seizures. [1]
A trial involving 34 people tested levetiracetam, an antiseizure medicine. The medicine did not improve the study's main measure of memory and thinking across the full group. A small subgroup with seizure-related electrical activity improved on selected tasks, but those results were exploratory. [15]
The results support more focused trials. Current evidence does not show that every EEG discharge should be treated or that antiseizure medicine slows Alzheimer's disease.
New ways to detect hard-to-find activity
Some technologies record for longer. Others use different sensors or new ways of interpreting a scalp signal. Several have only been tested in people with epilepsy. Alzheimer's-specific studies are needed before their performance or clinical value can be applied to dementia care.
Tested in Alzheimer's disease | Early research
Musaeus and colleagues, 2023
Read the study ↗What researchers tested: Twenty-four people with Alzheimer's disease and 15 people without it used small electrodes worn in the ears. The Alzheimer's group completed as many as three two-day recordings over six months. During the first recording, brief seizure-related discharges appeared in 75% of the Alzheimer's group and 46.7% of the comparison group. The difference in the number of people with discharges could have occurred by chance, although the discharges were more frequent in the Alzheimer's group. [11]
What remains unknown: Similar findings in many people without Alzheimer's disease, limited electrode coverage, recording interference, and discomfort make the results harder to interpret. The study did not record ear EEG and scalp EEG at the same time, and it did not find a clear relationship with memory or thinking. The authors disclosed funding and relationships with the device company.
Seagull's perspective: Repeated ear recordings could make it easier to study activity over time. Researchers must first confirm what the signals mean and how reliably the method separates meaningful activity from false findings.
Tested in temporal lobe epilepsy | Not yet tested in dementia
Abou Jaoude and colleagues, 2022
Read the study ↗What researchers tested: A machine-learning system called HEAnet analyzed scalp EEG for patterns linked with seizure-related activity in the hippocampus, a deep brain area involved in memory. The study included 141 people with temporal lobe epilepsy or in comparison groups. Recordings taken from the scalp and from deeper electrodes at the same time helped researchers develop the system. [13]
What remains unknown: The study did not include people with Alzheimer's disease. Research in dementia must show whether the system can accurately identify brief electrical discharges and actual seizures while limiting false alarms.
Seagull's perspective: Scalp recordings may contain useful information that ordinary visual review does not reveal. The next step is careful testing in people with dementia, using a dependable comparison method.
Tested in epilepsy | Not yet tested in dementia
Halliday and colleagues, UMPIRE study, 2025
Read the study ↗What researchers tested: Twenty-six adults with epilepsy received a recording system implanted beneath the scalp. The main safety assessment lasted six months. In eight participants, comparison scalp EEG captured 25 seizures, and the implanted system recorded all 25. [14]
What remains unknown: The comparison involved seizures already visible on scalp EEG, so the study cannot show how well the system detects deep seizures with no scalp signal. The study did not include people with Alzheimer's disease. Surgery, comfort, and usefulness in dementia require separate evaluation. The authors disclosed relationships with the manufacturer.
Seagull's perspective: Long-term recording could reduce uncertainty about events that happen only occasionally. Researchers must still determine whether the device records the type of activity that is difficult to find in Alzheimer's disease and whether that information improves care.
How Seagull evaluates a technology claim
Inclusion in this guide is research commentary. It is not a product endorsement or a statement that the technology is available for dementia care.
Research insights
One study recorded electrical activity deep in the brain that scalp EEG did not show. Another reported a much lower detection rate after applying structured definitions and review. Together, they show how the recording method, the people studied, and the definition of an abnormal signal can change the result.
Direct recording evidence | Five selected participants
Devulder and colleagues, Brain, 2025
Read the study ↗Five people with early or diagnosed Alzheimer's disease underwent scalp EEG while thin electrodes were placed close to deep areas involved in memory. At least 95% of the brief seizure-related electrical discharges recorded by the deep electrodes produced no matching signal on scalp EEG. Two seizures in one participant also produced no matching scalp signal. The study explored possible relationships with sleep and tau changes in the brain. [10]
The participants had previously shown possible seizures or seizure-related electrical activity. This small, selected group cannot tell us how often hidden activity occurs across the wider Alzheimer's population.
Seagull's perspective: These five cases confirm that deep electrical activity can remain unseen on scalp EEG. They cannot show how often this happens across the wider Alzheimer's population. The research priority is to find safer and more practical ways to investigate this detection gap.
Controlled recording study | Few findings
Szabo and colleagues, 2026
Read the study ↗Researchers used enhanced overnight recording and a structured expert review. They identified brief seizure-related electrical activity in three of 30 participants with Alzheimer's disease and one of 30 participants without it. The difference was small enough that it could have occurred by chance. [12]
The finding helps explain why study estimates differ. Who was selected, which electrodes were used, whether sleep was captured, and what counted as abnormal all affected the result.
Seagull's perspective: Researchers need shared definitions of seizure-related activity and dependable ways to confirm it. A higher count is useful only when the findings are accurate and clinically meaningful.
What research needs to answer next
Researchers need to separate uncertain findings from electrical activity linked with symptoms, future seizures, or meaningful changes in memory and daily function.
Age when symptoms began, stage of the disease, repeated event patterns, and other health information may help identify the right groups to study. An Alzheimer's diagnosis alone cannot determine which test would be useful for one person.
Trials should show whether better detection and treatment improve memory, daily function, safety, sleep, or quality of life. A lower number of electrical discharges on a recording is only one possible result.
New methods must be tested across stages of dementia, combinations of brain disease, care settings, and diverse populations. Comfort, burden on families and caregivers, access, and false alarms all matter.
Sources and editorial approach
Seagull Health prepared this focused review using recent Alzheimer's research and selected studies of emerging detection technology. The evidence was reviewed on . This is an educational review rather than a clinical guideline or complete systematic review.
Study findings are separated from Seagull's perspective. Every percentage stays connected to the people studied, what researchers measured, and the period of time covered. The page identifies differences between studies that find relationships, small studies using deep electrodes, early technology research, and treatment trials.
The review includes core Alzheimer's research and selected studies of new detection technology. Priority is given to research published during the past ten years. The September 2016 Vossel study falls within this review period. Older research is included only when it is foundational or important for understanding later work.
Ting et al. (2026). Neuropathologic Correlates of Seizures in Patients With Alzheimer Disease and Dementia With Lewy Bodies. Neurology.
Study of people whose dementia diagnosis was confirmed after death. Provides a seizure estimate near diagnosis and examines links with brain changes. It does not measure lifetime seizure risk.
Beagle et al. (2017). Relative Incidence of Seizures and Myoclonus in Alzheimer's Disease, Dementia with Lewy Bodies, and Frontotemporal Dementia. Journal of Alzheimer's Disease.
Review of records from a memory center. Provides an estimate of seizure risk over time and examines differences related to age.
Vöglein et al. (2022). Seizure prevalence in neurodegenerative diseases: a study of autopsy-proven cases. European Journal of Neurology.
Review of medical records from a selected group whose diagnoses were confirmed after death. Reports seizures recorded across the course of the disease.
Baker et al. (2019). The prevalence and clinical features of epileptic seizures in a memory clinic population. Seizure.
Memory-clinic study describing subtle seizures and why they can be difficult to recognize. EEG did not confirm every possible or probable epilepsy diagnosis.
Vöglein et al. (2020). Seizures in Alzheimer's disease are highly recurrent and associated with a poor disease course. Journal of Neurology.
Large database study with a smaller group followed for seizure recurrence. Shows relationships between seizures and decline but cannot prove that seizures caused the decline.
Horvath et al. (2021). Subclinical epileptiform activity accelerates the progression of Alzheimer's disease: A long-term EEG study. Clinical Neurophysiology.
Study following people over time. Links brief seizure-related electrical activity with faster decline, but the study design cannot prove that the activity caused it.
Vossel et al. (2016). Incidence and impact of subclinical epileptiform activity in Alzheimer's disease. Annals of Neurology.
Small study using overnight video EEG and a test that measures magnetic signals from the brain. Shows the importance of sleep and reports a link with faster decline.
Lam et al. (2017). Silent hippocampal seizures and spikes identified by foramen ovale electrodes in Alzheimer's disease. Nature Medicine.
Two selected cases studied with electrodes placed close to deep memory areas. Demonstrates activity that scalp EEG did not show. It cannot estimate how common this activity is.
Lam et al. (2020). Association of epileptiform abnormalities and seizures in Alzheimer disease. Neurology.
Study using 24-hour ambulatory scalp EEG and review by several experts. Examines how often different electrical patterns appeared in each group.
Devulder et al. (2025; online 2024). Epileptic activity on foramen ovale electrodes is associated with sleep and tau pathology in Alzheimer's disease. Brain.
Five selected people with early or diagnosed Alzheimer's disease. Shows limits of scalp EEG and examines possible links with sleep and tau changes.
Musaeus et al. (2023). Detection of subclinical epileptiform discharges in Alzheimer's disease using long-term outpatient EEG monitoring. Neurobiology of Disease.
Early study using ear EEG in 24 people with Alzheimer's disease and 15 comparison participants. Examines whether longer ear recordings are practical and what they detect. It does not establish diagnostic accuracy.
Szabo et al. (2026). Refining Detection of Subclinical Epileptiform Activity in Alzheimer's Disease: A Case-Control Study and Call for a Consensus. Annals of Neurology.
Thirty people with Alzheimer's disease and 30 comparison participants underwent enhanced overnight recording and structured review. Shows why definitions and recording methods affect detection rates.
Abou Jaoude et al. (2022). Noninvasive Detection of Hippocampal Epileptiform Activity on Scalp Electroencephalogram. JAMA Neurology.
Machine-learning study involving people with temporal lobe epilepsy and comparison groups. Relevant to detecting deep electrical activity, but it did not test people with Alzheimer's disease.
Halliday et al. (2025). The UMPIRE study: A first-in-human multicenter trial of bilateral subscalp monitoring for epileptic seizure detection. Epilepsia.
First study of an implanted recording system in 26 adults with epilepsy. Examines safety and longer recording. It did not test people with dementia.
Vossel et al. (2021). Effect of Levetiracetam on Cognition in Patients With Alzheimer Disease With and Without Epileptiform Activity: A Randomized Clinical Trial. JAMA Neurology.
Small treatment trial in which participants received both levetiracetam and placebo at different times. The main measure of memory and thinking did not improve. Results from a small subgroup support further research.
Yeh et al. (2026). Subclinical Epileptiform Discharge in Patients With Alzheimer Dementia: A Systematic Review and Meta-Analysis. Journal of Clinical Neurophysiology.
Review combining 13 studies with 1,373 people who had no known epilepsy. Wide differences among the studies limit the usefulness of one combined detection estimate.