Can Cats Detect Seizures? The Feline Scent-Alert Mechanism Explained

Yes, some cats appear to detect seizures before they happen, and the evidence points to scent rather than intuition. Cats carry roughly 200 million olfactory receptors—about 40 times the human count—and can detect some odorants at concentrations near 1 part per trillion. During the pre-ictal phase, which runs from about 10 minutes to over an hour before a seizure, sweating, stress hormones, and metabolic byproducts shift the body's volatile chemistry. A cat positioned near its owner's face or chest is sampling that change continuously. The behavior is real but inconsistent: no cat has been validated as a medical alert animal the way seizure-response dogs have.

The biology behind a cat's seizure sense

A seizure is an electrical storm in the brain, but its chemical footprint leaks outward. Within seconds to minutes of onset, sympathetic nervous system activation drives sweating, especially on the palms, soles, and forehead. Human sweat contains more than 400 volatile organic compounds, and the mix shifts measurably under stress. Adrenaline and cortisol rise, breathing rate climbs, and muscle activity before a convulsion can release trace compounds through the skin.

Cats read that chemistry with equipment humans do not have. Their vomeronasal organ, a paired structure behind the upper incisors, detects pheromones and non-volatile molecules that the main olfactory epithelium misses. When a cat performs the flehmen response—mouth slightly open, lips pulled back—it is drawing air across that organ. The behavior looks like a grimace; functionally it is a chemical assay.

The feline olfactory epithelium itself covers about 20 to 40 square centimeters, compared with roughly 5 square centimeters in humans. The olfactory bulb, the brain region that first processes smell, is proportionally larger in cats and densely wired. Detection thresholds vary by compound, but for some fatty acids and sulfur compounds cats outperform dogs. In plain terms: a cat's nose is a broadband chemical sensor running at a sensitivity humans cannot approach.

Hearing adds a second channel. Cats detect frequencies up to roughly 64,000 hertz, well above the human ceiling of 20,000 hertz. Pre-seizure vocalizations, breathing changes, or the faint sounds of a body tensing may register. Whether cats actually use these cues during a seizure is unproven; the auditory capability exists, but the behavioral evidence is thin.

How often cats actually respond, and how the numbers compare

The honest answer is that nobody knows the base rate. There is no published study that has placed cats with epilepsy patients and measured pre-seizure alerts under controlled conditions. What exists is owner report data, and it is heavily skewed toward people who already believe their cat senses something.

The closest quantitative anchor comes from dog research. A 2019 survey of seizure-alert dogs found that owners reported sensitivity above 70 percent, but when researchers tested dogs under blinded conditions, accuracy fell toward chance. A 2013 study of seizure-response dogs in the journal Neurology found that response behaviors—not prediction—were the reliable finding. Cats have not even reached that stage of study.

The gap between owner perception and measured performance matters. Owners who live with a cat and a seizure disorder are highly motivated to notice coincidences. A cat that happens to sit on the chest 15 minutes before a seizure gets remembered; the dozens of times it sat on the chest without a seizure get forgotten. This is confirmation bias operating at household scale, and it inflates every anecdotal estimate.

Reading the behaviors owners report most

The behaviors attributed to seizure detection cluster into a few recognizable patterns. None is specific to seizures, which is the central problem.

A cat may become unusually clingy in the 10 to 60 minutes before an event, following its owner room to room or refusing to leave the bed. It may stare fixedly at the owner's face, a behavior sometimes called "alerting." Some cats vocalize with a low, repetitive meow. Others pace, sniff the owner's head or hands intently, or position themselves against the body. During a convulsion, many cats stay close and lick the face or hands afterward.

The same behaviors appear in other contexts. Clinginess happens with illness, anxiety, and routine change. Face-staring happens when a cat wants food. Licking happens for grooming and social bonding. A 2015 review of assistance animals noted that no behavioral signature has been shown to distinguish seizure anticipation from general affiliative behavior.

What would constitute real evidence? A prospective study with continuous video and EEG monitoring, where the cat's behavior is timestamped against seizure onset without the owner knowing when events were coming. That study has not been done in cats. Until it is, every claim about feline seizure prediction sits on anecdote.

Why the pre-ictal window is so hard to test

Seizure prediction is difficult because the pre-ictal state is not sharply defined. The window is commonly cited as 10 minutes to 1 hour, but some people report aura symptoms up to 24 hours before an event, and others report none at all. A cat that alerts "an hour before" has a very wide target to hit by chance.

The natural variability of seizures compounds the problem. A person with epilepsy may have clusters of events separated by weeks, then a run of several in a day. Chance-level behavior will look impressive during the clusters and unremarkable during quiet periods. Owners remember the clusters.

There is also a measurement artifact specific to this question. Seizures themselves are undercounted even with EEG, because many events are subtle and unwitnessed. If the ground truth is noisy, any animal's "accuracy" is partly measuring the monitoring, not the animal. In plain terms: you cannot grade a detector when the events it is supposed to detect are themselves poorly recorded.

The 2019 dog survey illustrates the ceiling. It relied on owner questionnaires, and even so, the authors cautioned that blinding changed outcomes. Cats, with no equivalent survey at all, sit below that already-limited evidence base.

Separating real alerts from confirmation bias and folklore

Two misconceptions dominate this topic, and both need correcting.

The first is that cats sense seizures through some form of psychic or energetic perception. There is no mechanism for this, and none is needed: scent, hearing, and close physical proximity explain everything that has been plausibly reported. Owners should be skeptical of any source that invokes vibration, "energy fields," or intuition without chemistry behind it.

The second is the opposite error—dismissing all cat behavior around seizures as meaningless. The response behaviors during and after a seizure (staying close, licking, vocalizing) are real and are not prediction claims. A cat that lies against its owner during a convulsion and grooms the face afterward is showing a genuine behavioral response, not a paranormal one.

A harder caveat concerns harm. A cat that lies across a person's chest during a convulsion can obstruct breathing or be injured. A cat that scratches or bites during a seizure poses a wound risk. Some epilepsy patients report their cat's alert gives them enough warning to sit down safely; others report the cat simply gets in the way. The distinction matters for safety planning.

Practical steps if you want to evaluate your cat

If you live with epilepsy and a cat, you can generate better evidence than anecdote—for yourself, not for science.

• Log every seizure with the exact time, type, and duration, plus where the cat was and what it was doing in the prior hour.
• Record 2 to 4 weeks of baseline cat behavior without seizures, so you know how often the "alerting" behavior happens on quiet days.
• Compare alert frequency before seizures against alert frequency in matched control windows, such as the same hour on seizure-free days.
• Keep the cat off the chest and away from the head during a convulsion, and move it to the floor if it clings.
• Never delay rescue medication or a safety plan on the basis of a cat's behavior.

The baseline step is the one most owners skip and the one that matters most. If your cat stares at your face five times a week and you have a seizure once a month, the stare is not an alert. It is a cat.

The comparison is not close. A log costs nothing and produces more honest data than any amount of remembered incidents.

What the timeline looks like during a seizure event

The pre-ictal and post-ictal phases are where cat behavior is most often reported. The table below lays out the typical course and what has actually been documented in cats at each stage.

The numbers overlap because the phases themselves overlap. An aura that lasts 60 minutes overlaps an ictal event that starts at minute 55. This is why the pre-ictal window is often quoted as a range, not a fixed value.

In every phase, the reported cat behavior is identical to ordinary affiliative behavior. That is the core problem, and no amount of owner confidence solves it.

Limits of the evidence and what would change the picture

The current state of knowledge can be stated in one sentence: cats plausibly detect the chemical and behavioral changes around seizures, but no controlled study has demonstrated predictive accuracy, and the base rate of coincidental behavior is unknown.

Fixing that would require three things. First, a blinded prospective cohort of epilepsy patients living with cats, with continuous EEG and video. Second, chemical analysis of pre-ictal sweat and breath to identify which compounds shift and whether cats could detect them. Third, a behavioral scoring system validated against the chemical data. None of these exists for cats. The dog literature has moved partway down this road; the cat literature has not started.

What would falsify the claim? A blinded study showing cat alert frequency before seizures equals alert frequency in control windows. That is the test, and it is simple enough that its absence after decades of anecdote is itself informative.

For now, treat a cat's alertness as a possible supplement to, never a replacement for, a written seizure diary, rescue medication, and a safety plan. The cat may be reading your chemistry. It is also reading your habits, and those two readings look identical from the outside.