SYNGAP1-related disorder

A single gene, and a test that finds it.

SYNGAP1-related disorder is caused by a change in one gene, and it isn’t found at all unless somebody orders the test. “Autism with epilepsy” describes what you see; it doesn’t explain it. Where both are present, a genetic cause is worth asking about.

If any of these is happening, your medical needs may be urgent

These are about seizures. Tick anything true right now.

A seizure lasting more than five minutes.
One seizure after another without properly waking up in between.
Breathing trouble, or lips or face going blue or grey.
A first-ever seizure.
A seizure with a head injury, or in water.
Not returning to their normal self well after it ends.

What we would do, as a friend rather than as your doctor. This is our opinion and we stand behind it. It is not medical advice and we are not examining your child.

Five minutes is the number. A seizure still going at five minutes is a 911 call. You do not wait to see. That is standard advice, not ours.
A first seizure gets seen the same day, even if it stopped.
Two boxes or more — ring 911.
One box, not the five-minute one — ring the triage nurse now.

If you have rescue medication and a plan, follow your plan. It was written for your child and it beats anything on this page.

None of them? Then nothing below is an emergency, and you can read at your own pace.

Also on the list in an earlier edition

Seizures one after another without full recovery between
Breathing difficulty or a blue colour during or after a seizure
A first seizure that has never happened before
Unresponsive, or not returning to normal, well after it ends
One box — concerning. Ring your doctor’s office and ask for the triage nurse today.
Two boxes — ring the triage nurse now, not later today.
Three boxes — go to urgent care.
Four or five boxes — ring 911, and expect that to mean the emergency room.

Not sure how serious this is? Where to call, and when.

Start with a phone call. Your doctor’s office has a triage line and a nurse will talk it through with you. They are genuinely good at this, they do it all day, and it costs nothing. Most practices have a phone line where a nurse takes calls about whether something needs to be seen, and how soon. It is often not advertised — you ring the main number and ask to speak to the triage nurse. There is no charge and you do not need an appointment to use it.

If it feels more serious than that, go to urgent care. You will usually be seen faster than at a hospital, and the care is good.

If it feels life-threatening, call 911 or go to the emergency room. Do not wait until you are sure.

Which one is yours to decide, not ours. But our advice is to assume it is worse than it looks and take the more cautious road. And trust your body — if it is telling you something is wrong, it is probably right. The best possible outcome of a trip to the emergency room is walking out saying well, that was a waste of an evening — but I feel a lot better knowing it is nothing serious.

Before you call, have these ready. They are what the nurse will ask, and the call goes better when you are not working them out on the phone.

How bad is it, one to ten? Pick a number even if it feels arbitrary. They are not testing you — it gives them somewhere to start.
When exactly did it start? Within the hour, six hours, today, yesterday, this week, longer. The number matters less than which side of “today” it falls on.
Which way is it going? Better, worse, or the same — and if worse, over hours or over days. This is often the question that decides things, and it is the one people least often have an answer to.
If there is a fever, the actual number and the time you took it. “I feel hot” and “101.4 at three o’clock” are very different pieces of information.
Anything that affects your immune system, said early. Chemotherapy, steroids, immunosuppressants, a transplant, or a condition that affects it. It changes how they read everything else, so it should not come out at the end.
What you have already tried, and whether it helped. What did not work narrows things down as much as what did.
Anything new alongside it — vomiting and not keeping fluids down, bleeding, trouble passing urine, confusion, breathlessness. Combinations are read differently from single symptoms.
Your medicines, including the ones started or stopped recently.

You are not diagnosing yourself by having this ready. You are handing them the things they would otherwise spend the call extracting — and the decision stays entirely theirs.

Writing to the portal instead of calling? The same list works, in that order, in one message. Put the direction it is going and the immune-system line near the top — portal messages get read quickly, and those two change how the rest is read.

Go deeper

8 layers below. Open any one.

The SYNGAP1 gene carries instructions for a protein used at the junctions between brain cells. When one copy makes too little of it, those junctions don’t settle as they normally would. A combination tends to travel together: developmental delay, difficulty learning, seizures, and autistic features.

It is present from birth. It is not caused by anything a parent did or didn’t do.

The delay is visible, and the seizures are visible. Each can be managed as a problem in its own right for years, without anyone asking whether one cause lies beneath both.

Evidence: Grade C. Our source describes SYNGAP1 as one of the more common single-gene causes of intellectual disability, with far more people affected than identified. We haven’t re-checked those figures yet. The pattern is the point: many affected, few found.

Ask whether an epilepsy gene panel or exome sequencing has been done. A panel checks a set list of genes. An exome looks much more widely. Which one fits is a clinical decision, but anyone can ask whether either has been done.

A result changes practical things: which seizure medicines are more likely to suit, what to watch for, and a way into families and research focused on this exact condition rather than on delay in general.

Think of a smoke alarm that goes off a year after the kitchen started to smell of smoke. The smell was the first sign; the alarm only made it impossible to ignore. In SYNGAP1-related disorder, slow development is usually the smell and the seizures are the alarm.

In the largest study of the seizures, delay in development came before the first seizure in 54 of the 56 children whose early history was known. The first seizure came at a median age of about two years. Good evidence (Vlaskamp DRM et al., Neurology 2019, PMID 30541864.) A smaller Korean study that followed 13 children over years found the same order: delay from early infancy, seizures later, at a median of about two and a half years. Good evidence (Kim HJ et al., American Journal of Medical Genetics A 2024, PMID 38563110.)

That gap of a year or more is a window. A child with unexplained delay can have genetic testing before the seizures start. Since 2021 the American College of Medical Genetics has recommended exome or genome sequencing as a first- or second-tier test for children with developmental delay or intellectual disability. Exome sequencing reads the instructions in nearly all of a person’s genes at once. Strong evidence (Manickam K et al., Genetics in Medicine 2021 (ACMG guideline), PMID 34211152.) The National Society of Genetic Counselors recommends genetic testing for anyone with epilepsy that has no other explanation, whatever their age. Strong evidence (Smith L et al., Journal of Genetic Counseling 2023 (NSGC guideline), PMID 36281494.)

The Korean study also noticed something else. Language and learning tended to level off, or plateau, between about ages two and five, and the authors think the seizures may contribute to that. Early evidence (Thirteen children; Kim HJ et al., American Journal of Medical Genetics A 2024, PMID 38563110.) If that holds, the years when the plateau happens are also the years when good seizure control matters most.

What you might notice: slow sitting, walking or talking, months before anything that looks like a seizure.
What it might mean: one of many genetic causes of delay, SYNGAP1 among them. Most delay has other causes.
The test that would clarify it: exome sequencing (or a broad epilepsy and delay gene panel), with a chromosomal microarray.

A child bangs a knee hard on the table edge and keeps playing. Another has a burning ear infection and shows it only by hitting their head, or by not sleeping. In SYNGAP1-related disorder this is common, and it is not toughness.

In the 57-child seizure study, 72% of families reported a high pain threshold, meaning it takes much more than usual for a child to show pain. Good evidence (Vlaskamp DRM et al., Neurology 2019, PMID 30541864.) Research in people with SYNGAP1 and in mice found that the brain’s touch-sensing area receives weaker signals from the skin than usual. It is like a doorbell wired to a speaker that is turned down. Good evidence (Mouse experiments plus human touch testing; Michaelson SD et al., Nature Neuroscience 2018, PMID 30455457.)

In practice, the usual alarms can go quiet: crying, limping, holding the sore ear. Pain may show up instead as a change in behaviour, in sleep, or in eating, or as more seizures. Our reading is that a sudden change in behaviour in a child with SYNGAP1 calls for a body check before anyone decides it is “behavioural.” That means ears, teeth, throat, skin, bones and bowels. Early evidence (This is our inference from the two findings above. No trial has tested it.)

What you might notice: new hitting, biting, refusing food, or bad nights, with no obvious cause.
What it might mean: hidden pain, such as an ear infection, a tooth, constipation or a missed fracture, as often as anything about the brain.
The test that would clarify it: an ordinary physical exam with this in mind. The primary care doctor or pediatrician who has known the child over time is best placed to see what is different this week.

A medical record is searched the way a library is: by its labels. The right words help the next clinician, the next school and the next research study find the child.

SYNGAP1-related disorder (sometimes “SYNGAP1-related intellectual disability”). This is the name of the condition. The protein the gene makes is called SynGAP.
A diagnosis code of its own. In the US, SYNGAP1-related intellectual disability has had its own billing and diagnosis code since 1 October 2021 (F78.A1). Before that it was filed under general headings. Strong evidence (ICD-10-CM 2022 code set (US), effective 1 October 2021; confirmed in the published code tables.) If the diagnosis is on the problem list under its own code, anyone who opens the chart sees it.
Haploinsufficiency. One of the two copies of the gene does not work, so the body makes about half the usual amount of protein. Think of a two-engine plane flying on one engine. Strong evidence (Holder JL, Hamdan FF, Michaud JL, GeneReviews (NCBI Bookshelf), PMID 30789692.)
De novo. The change is new in the child, and neither parent carries it in their blood. That is true for almost every family tested so far. Strong evidence (Holder JL, Hamdan FF, Michaud JL, GeneReviews (NCBI Bookshelf), PMID 30789692; also the original report, Hamdan FF et al., New England Journal of Medicine 2009, PMID 19196676.)
Germline mosaicism. A parent can carry the change in some egg or sperm cells and not in their blood. A parent passing it on has been reported in one family. This is why the chance of a brother or sister being affected is low but not zero. Strong evidence (Holder JL, Hamdan FF, Michaud JL, GeneReviews (NCBI Bookshelf), PMID 30789692.)
Developmental and epileptic encephalopathy (DEE). A condition in which both the underlying cause and the seizures affect development. The 2019 seizure study described SYNGAP1 as a “generalized DEE” with its own recognizable pattern. Good evidence (Vlaskamp DRM et al., Neurology 2019, PMID 30541864.)
Eyelid myoclonia with absences. Quick, repeated fluttering of the eyelids, often with the eyes drifting up, during a few seconds of blankness. It is easy to mistake for a tic or for not paying attention.
Pathogenic, likely pathogenic, variant of uncertain significance (VUS). These are the three verdicts a lab can give a gene change. Only the first two count as a diagnosis. A VUS means “we can’t tell yet,” and labs can reclassify it as more families are studied. It is reasonable to ask for a VUS to be looked at again after a year or two.

Each figure below gives its source and the size of the group behind it. A percentage from 57 children is a sketch. It can tell you what to watch for, but it can’t tell you what your child will be like.

From the 57-child seizure study (Vlaskamp DRM et al., Neurology 2019, PMID 30541864). Good evidence

Epilepsy: 56 of 57.
Eyelid myoclonia with absences: about two in three (65%). Seizures triggered by eating: about one in four (25%).
Intellectual disability: 55 of 57, moderate to severe in 50.
Behaviour problems 73%; high pain threshold 72%; eating problems, including refusing food in the mouth, 68%; low muscle tone 67%; sleep problems 62%; autism 54%; unsteady walking or balance 51%.
Development stalled or slipped back around the time of seizures in nearly every child.

From the GeneReviews summary (Holder JL, Hamdan FF, Michaud JL, GeneReviews (NCBI Bookshelf), PMID 30789692). Strong evidence About half of children’s seizures respond to the first medicine tried; in the rest they are hard to control. About 9 in 10 cases are a change in the gene’s spelling, and about 1 in 10 is a missing piece of chromosome 6 that includes the gene.

From health records and insurance claims for about 400 people (McKee JL et al., Genetics in Medicine 2025, PMID 40119723). Good evidence Autistic behaviour was much more likely to appear between about 27 and 30 months of age. Generalized seizures became more common after age three. Valproate and lamotrigine did better than other seizure medicines at reducing seizures or keeping a child seizure-free. That comes from records, not from a trial.

Age at diagnosis. In a registry of 65 children, the median age at diagnosis was 3 years, and the range ran from 1 to 17. Good evidence (Scott MR et al., Orphanet Journal of Rare Diseases 2025, PMID 40717090.) In a clinic series of 15, the average was about five and a half years. Good evidence (Jimenez-Gomez A et al., Journal of Neurodevelopmental Disorders 2019, PMID 31395010.)

How many there are, and how many have been found. A genetics-based model estimates about 6 new cases per 100,000 births, roughly 1 in 16,000. Good evidence (A model, not a head count; López-Rivera JA et al., Brain 2020, PMID 32168371.) Families known to advocacy groups worldwide numbered 1,806 on 1 July 2026. Good evidence (SYNGAP1 Census, Q2 2026 update, Syngap Global Network via curesyngap1.org.) Our own arithmetic: at 1 in 16,000, a world of about eight billion people would hold several hundred thousand. On that basis, well under one in a hundred has been found. Early evidence (This rests on the model holding across all ages and countries. Use it to show the size of the gap, not to quote.) An older figure of “more than a million” appears in some of our earlier drafts. We could not trace it to a source, and we have not used it.

The “window” question has evidence on both sides. A 2012 study in mice found that the brain connections matured too early. Repairing the gene in adult mice did not improve their behaviour or memory. Good evidence (Mouse work; Clement JP et al., Cell 2012, PMID 23141534.) A 2019 study by an overlapping group, using a different method, restored the protein in adult mice and did improve memory and seizure measures. It also removed abnormal electrical bursts that got worse during sleep. Good evidence (Mouse work; Creson TK et al., eLife 2019, PMID 31025938.) Neither study was done in people. Taken together, they suggest that some problems may be fixed early in development while others stay changeable for life. That is a reason to keep teaching and keep treating at every age. It is not a promise. Not yet tested

The studies count the children who were found. The 57-child study recruited partly through social media and through specialists’ own practices. Strong evidence (Stated in its methods; Vlaskamp DRM et al., Neurology 2019, PMID 30541864.) Children with milder forms, or who were never tested, are missing from every percentage. The real picture is probably wider, and gentler at the mild end.

Small numbers, and early treatment reports. One sleep study used wrist activity monitors on five children with SYNGAP1 and found broken-up sleep and early-morning waking. It is useful, but it is five children. Early evidence (Doucoure A et al., Sleep Medicine 2026, PMID 41887121.) A low dose of perampanel, a seizure medicine that blocks one type of signal receptor, was reported to help one toddler’s development and sleep. It did not stop the seizures. Early evidence (One child; Gupta S et al., Frontiers in Neurology 2023, PMID 37662032.) Reports like these suggest questions for a specialist. They are not a basis for changing treatment on their own.

No treatment yet targets the cause. At the start of 2024 there was no approved treatment for the cause of SYNGAP1-related disorder. Several cause-targeted treatments were in development, many of them helped along by family organizations. Good evidence (A review written by the main patient organization; Graglia JM, Harding AJ, Helde KA, Therapeutic Advances in Rare Disease 2025, PMID 39807402.) Any trial result announced after this page was written needs checking against its own published report.

A single gene, and a test that finds it. SYNGAP1-related disorder is caused by a change in one gene — and it is not found at all unless somebody orders the test. “Autism with epilepsy” is a description, not a cause; where both are present, a genetic cause is worth asking about.

What it is

One gene, one protein, one junction

The SYNGAP1 gene carries instructions for a protein used at the junctions between brain cells. Where one copy makes too little of it, those junctions do not settle as they normally would — and a combination tends to travel together: developmental delay, difficulty learning, seizures, and autistic features. It is present from birth, and not caused by anything a parent did or did not do.

Why it goes unfound

Nothing says “order a genetic test”

The delay is visible, the seizures are visible, and both can be managed as findings in their own right for years without the question of a single underlying cause ever being put. Early evidence The source records this as among the more common single-gene causes of intellectual disability, with far more affected than identified. The numbers are carried from that document and not re-checked here — the shape (many affected, few found) is the point; treat the figures as needing confirmation.

Getting the test

Panel or exome

Ask specifically whether an epilepsy gene panel or exome sequencing has been done — a panel looks at a defined list, exome looks much more widely. Which is appropriate is a clinical decision; whether either has been done is a question anyone can ask. A result changes practical things: which seizure medicines are more likely to suit, what to watch for, and access to a community of families and to research about this condition rather than about delay in general.

Questions, as worded then

Has my child had genetic testing — a panel, or exome sequencing?
Both delay and seizures — has a single underlying cause been looked for rather than treating the two separately?
If the first test was a normal panel, would a broader test find something it would not?
If a cause were found, would it change which medicines we try?

Figures marked Grade C are carried from the source guide and not re-verified here. Nothing comes from any individual’s medical record.

Questions to bring

Copy these, or read them out. They fit in a short visit.

Has my child had genetic testing: a gene panel, or exome sequencing?
With both delay and seizures, has anyone looked for one cause behind both?
If the first test was a normal panel, would a broader test find something it missed?
If a cause were found, would it change which medicines we try?
Can SYNGAP1-related disorder go on the problem list under its own diagnosis code, so every clinician who opens the chart sees it?
Because many children with SYNGAP1 show pain less than other children, what should we check (ears, teeth, bowels, bones) when behaviour, sleep or seizures suddenly change?
Should we as parents be tested too, so we know the real chance for a future pregnancy, given that a parent can carry it in egg or sperm cells only?

The short version

Getting ahead of it: Where developmental delay and seizures occur together, ask for genetic testing rather than treating each as its own thing.
Your testing regime: An epilepsy gene panel or exome sequencing; if a panel came back normal, ask whether a broader test would find more.
What they don’t tell you: Far more people are affected than are identified, and a diagnosis changes which seizure medicines are likely to suit.
What it can spawn: As the underlying cause it shapes the seizures and development; unrecognised, it means years of unguided management.
What it’s confused with: “Autism,” “global developmental delay,” and “epilepsy of unknown cause.”

The short version, from the 2 October draft

Getting ahead of it: Getting ahead of it: Delay usually comes a year or more before the first seizure. Asking about exome sequencing when the delay is first noticed, as the 2021 medical genetics guideline supports, can give a family the name before the seizures start rather than years after (Manickam K et al., Genetics in Medicine 2021 (ACMG guideline), PMID 34211152; Grade A).
Your testing regime: Your testing regime: After the diagnosis, the core checks are seizure type and control (video of spells, and video EEG when something new appears), development and school needs, behaviour, feeding and growth, and sleep (GeneReviews; Grade A for the list). One more is worth adding: a body check for hidden pain whenever behaviour changes (our inference; Grade C). The primary care doctor holds the single list and watches the trend.
What they don’t tell you: What they don’t tell you: About seven in ten children feel or show pain less than usual (Vlaskamp 2019; Grade B). The chance of a brother or sister being affected is low but not zero (GeneReviews; Grade A). The worldwide count of 1,806 is only the families known to advocacy groups (Census Q2 2026; Grade B). And the mouse evidence on whether adults can still improve points both ways (Clement 2012 against Creson 2019; Grade B each).
What it can spawn: What it can spawn: Seizures that do not settle on medicines in about half of children (GeneReviews; Grade A). Eating problems and refusing food in about two in three, which can mean weight loss and feeding therapy (Vlaskamp 2019; Grade B). Injuries and infections that go unnoticed because pain is muted (Grade C). Broken sleep that wears down the whole household (Grade B). And when behaviour medicines are used, the routine monitoring each of those medicines needs.
What it’s confused with: What it’s confused with: Autism on its own; myoclonic-atonic epilepsy (Doose); eyelid myoclonia with absences (Jeavons); Lennox–Gastaut syndrome; Angelman and Rett syndromes; and “epileptic encephalopathy, cause unknown.” Several of these names describe a pattern of seizures rather than a cause, so a child can correctly carry one of them and SYNGAP1 underneath it. Genetic testing tells them apart (Grade A that testing separates them; Grade C for any single bedside feature offered as a tie-breaker).

Figures marked Grade C come from our source guide and have not yet been re-verified. Nothing on this page comes from any individual’s medical record.

The longer chapter

From the 11 September 2026 edition (Vermont Synergy Initiative site): “Nobody has found a cause for my child’s delays and seizures”. Every section below is that edition’s own words; open any one.

The short answer, before anything else.

SYNGAP1 is a gene. When one copy does not work, there is not enough of the protein it makes. That single fact explains most of what follows: the seizures, the delay, the trouble with noise and crowds, the sleep.
→ where this is explained: What this page is about

Autism is a description, not a cause. Being given that description does not rule out a genetic cause underneath it, and looking for one changes what can be done.
→ where this is explained: What this page is about

Some seizure types point at it specifically — much more than others. If you have a phone video of one, bring it. It is often worth more than the description.
→ where this is explained: The seizures that point at it specifically

The one thing to do next: Worth asking: “He has autism features and seizures. Has anybody looked for a genetic cause that would explain both?” and “Which genes were on that panel, and was a microarray done as well as sequencing?” Sequencing alone can miss it.

Everything below explains each of those, in whatever order suits you.

SYNGAP1 is a gene. Everyone has two copies. It makes a protein that acts as a brake on brain signalling — it helps the brain filter, time, and settle down after being stimulated.

When one copy does not work, there is not enough of that protein. The brain is not damaged. It is under-braked. Signals arrive without being filtered or timed properly.

That single fact explains most of what follows: the seizures, the delay, the trouble with noise and crowds, the sleep.

Across children identified with it:

Developmental delay — 100%. It is always there.
Epilepsy — 84%.
Autism features — 68%.
Sleep disturbance — common, and often the thing families find hardest.

GRADE B These figures are carried from this project’s own SYNGAP1 work rather than re-checked against the primary papers today. They are good enough to tell you what to ask. They are not good enough to quote.

And one caution that matters for every number above. They describe children who were found to have it. Children with milder versions may never be tested, so the real spread is probably wider and gentler at one end than these figures suggest.

Some seizure types are much more suggestive than others. If any of these sound familiar, that is worth saying out loud to a neurologist:

Brief staring spells with rapid eyelid fluttering. Easily mistaken for daydreaming or not listening.
Sudden drops — the body goes limp for an instant and the child falls.
Seizures that happen while eating, or just after. This one is unusual enough that it is worth mentioning even if nobody has asked.

If you have a phone video, bring it. Thirty seconds of footage settles questions that an hour of description cannot. Neurologists ask for this and are glad to get it.

Pick whichever sounds like your family. There is no wrong door and no right order. Each of these feeds the next, and the last feeds the first.

“We were told it was autism.”

Autism features are present in about two thirds of children with SYNGAP1, so the description fits. It is not a wrong observation.

But autism is a description, not a cause. It says what is happening. It does not say why.

The combination that should prompt a genetic test: autism features plus seizures plus intellectual disability. Autism alone, in a child with no seizures and typical development elsewhere, is a different situation.

GRADE A That developmental delay with epilepsy warrants genetic testing is standard practice, not a fringe view.

“He has the autism diagnosis and he has seizures. Has anybody looked for a genetic cause that would explain both?”

“We had genetic testing and it was normal.” the most useful thing here

“Genetic testing” is not one test, and a normal result may mean the test did not look.

Two things worth checking, and both are reasonable questions:

Was SYNGAP1 on the panel? Older panels do not all include it. A panel from several years ago may simply not have been looking.
Was a microarray done as well as sequencing? These find different things. Sequencing reads the letters of a gene. A microarray finds whole missing chunks. Roughly one in nine SYNGAP1 cases is a deletion — and sequencing alone can miss those.

GRADE B The deletion proportion is carried from our own work. The principle beneath it — that sequencing and microarray detect different classes of change — is Grade A and not disputed.

panel did not include it → or sequencing missed a deletion → “normal” → the search stops → years pass

“Which genes were on that panel, and was a microarray done as well as sequencing?”

Testing has moved fast. A test from a few years ago is not the test available now, and repeating it is not a criticism of anyone.

“They named a syndrome after the seizures.”

Some diagnoses describe a pattern of seizures rather than a cause. Doose syndrome is one; Jeavons is another.

These are not wrong, and they are not the end of the road. They are descriptions of what the seizures do. SYNGAP1 is one of the things that can cause that pattern.

So a child can correctly have both labels: the pattern name, and the genetic cause underneath it. Having the first does not mean anybody looked for the second.

“That name describes his seizure pattern. Do we know what is causing the pattern?”

“The sleep is the thing that is breaking us.” the loop

Sleep is often treated as the side issue, to be dealt with once the seizures are sorted. It may be the most reachable part of the whole picture.

Here is the loop, and every arrow in it is ordinary neurology:

poor sleep → seizures come more easily → more seizures → tiredness and poor concentration → less learning → more frustration and harder behaviour → more stress → worse sleep

Read it round again and you end where you started. There is no first cause to find, which is why arguing about which came first goes nowhere.

What matters is which part you can reach. The gene cannot be changed. Sleep sometimes can. And because it is a loop, improving sleep does not only mean a better night — it pushes on the seizure threshold, and on the learning, and on the behaviour.

GRADE B The individual links — sleep and seizure threshold, seizures and cognition — are established. Presenting them as one self-reinforcing loop is this project’s reading, and we are saying so.

“Can we treat the sleep as a real target rather than something to sort out later?”

“Which medicines, and which ones make it worse?”

About half of children respond well to anti-seizure medication. About half do not, and that is called drug-resistant epilepsy. It is a description of what happened, not a verdict on your child or on anybody’s effort.

Seizure type changes which medicine helps. Some drugs that work well for one type can make another type worse. That is exactly why naming the seizure type — and having a genetic cause — changes treatment rather than just satisfying curiosity.

The ketogenic diet is a recognised option where medication has not controlled seizures. It is demanding, it is properly supervised, and it is not a fringe idea.

GRADE A That some anti-seizure drugs worsen particular seizure types, and that the ketogenic diet is used in drug-resistant epilepsy, are both standard.

“Given his seizure types, which medicines are most likely to help — and are any of them likely to make things worse?”

“Is it too late? Has the window closed?” read this one carefully

Parents are often told, gently and with real kindness, that the early years are the window and that after them the picture is fixed.

That assumption is being questioned, and it is worth knowing why.

In mice — and we are saying in mice because it matters — restoring SynGAP function in adult animals improved cognition. Not in newborns. In adults, after the supposed window.

GRADE U This is animal work and it has not been shown in people. Mouse findings often do not carry across, and nobody should change what they are doing on the strength of it. We are not telling you there is a treatment. There is not one yet.

What it does mean is that “the window has closed” is a hypothesis rather than a fact — and one with evidence against it. If somebody sets the expectation low, the effort follows the expectation. That is a reason to keep teaching, keep working, and keep expecting.

“Is there any reason to stop working on new skills at his age?”

Read them in any order and you end up where you started. That is not a weakness in the explanation. It is what the thing is.

You are the observer here, and you have watched longer and closer than anyone. These are the questions a good history asks. Answer the ones that matter and skip the rest. What you write stays on your own machine.

When did you first think something was different? Not when somebody agreed with you — when you first thought it. The gap between those two dates is often years, and it is worth saying.

Which way is it going? Over months, not days. Gaining, holding, or losing ground.

What do the episodes actually look like? Describe them as you would to a neighbour. If you have video, that is better than any description.

When do they happen? Time of day, around meals, after poor nights, during illness.

What changed around the same time? Any medicine started, stopped or increased. Any illness. Any change at school.

What have you tried, and what happened? Including what made things worse. That is information, not failure.

What has he stopped doing — and what have you stopped trying? Both halves. The second one is the question nobody asks, and it often says the most.

What worries you most? Not the medical question. The real one.

“He has autism features and seizures. Has anybody looked for a genetic cause that would explain both?”
“Given his seizure types, which medicines help — and which could make it worse?”
“Can we treat the sleep as a target in its own right?”
“If we did find a genetic cause, what would change about his care?”
“Is there a registry or a study he could join?”

Number five is the one worth pressing. A genetic answer is not just a label — it changes which medicines are chosen, and it opens research that is closed to children whose cause is unknown.

Not a diagnosis. Nothing here can give you one.

What you have is a specific question with a specific test behind it, and the words to ask it with. This is missed for an ordinary reason: the description fits something commoner, the description gets recorded, and the search stops. You are the only person holding the whole picture, because you are the only one who is there for all of it.

Take one question. Not six.

Grade A here: that delay with epilepsy warrants genetic testing; that sequencing and microarray find different things; that seizure type changes drug choice; that the ketogenic diet is used in drug-resistant epilepsy.

Grade B: the percentages, the deletion proportion, and the sleep loop as one connected circle. Carried from our own work, not re-verified today.

Grade U, and said plainly: the adult-restoration finding is mouse work. It is a reason for hope and a reason to keep expecting. It is not a treatment, and we will not dress it as one.

Why this page exists. It was promised in January 2026, alongside another room that was built. This one was written and then left in a folder. It is here now because somebody asked what else had already been written and never used.

From an earlier edition

From the first edition of the Book (8–11 September 2026): “My child is delayed, has seizures, and nobody can say why”. Every section below is that edition’s own words; open any one.

Rather listen?

Takes about 7 minutes, read aloud. It is the same words as the page below, so nothing is missing if you would rather read. Download it to keep on your phone — it will play with no signal, in a waiting room or a car.

The short answer

A single gene, and a test that finds it. SYNGAP1-related disorder is caused by a change in one gene. It is found by genetic testing — and it is not found at all unless somebody orders that test.
→ where this is explained: What it is

Most of the people who have it do not know. The gap between how many are thought to be affected and how many have been identified is very large.
→ where this is explained: Why it goes unfound

“Autism with epilepsy” is a description, not a cause. Where both are present together, a genetic cause is worth asking about.
→ where this is explained: What it gets taken for

Chapter Four · SYNGAP1-related disorder

Years of appointments, a diagnosis of “developmental delay” which is a description rather than a cause, and no name for what is happening.

A gene called SYNGAP1 carries the instructions for a protein used at the junctions between brain cells. Where one copy of that gene is altered so it makes too little of the protein, those junctions do not settle the way they normally would. The result is a combination that tends to travel together: delay in development, difficulty with learning, seizures, and autistic features.

It is present from birth. It is not caused by anything a parent did or did not do.

Because nothing about the child says “order a genetic test” unless somebody is thinking of one. The delay is visible, the seizures are visible, and both can be managed as findings in their own right for years without the question of a single underlying cause ever being put.

Grade C The source this chapter was built from records that this is among the more common single-gene causes of intellectual disability, and that the number of people identified is a small fraction of the number thought to be affected. Those figures are carried from that document and have not been re-checked against the primary literature for this page. They are here because the shape of the claim — many affected, few found — is the point, and the shape does not depend on the exact number. Treat the numbers themselves as needing confirmation.

Autism, on its own. The features overlap. The distinguishing question is whether epilepsy is also present, and whether anyone has looked for a single cause behind both.

“Global developmental delay”. A true description of what is happening and not an explanation of why, in exactly the way that chronic kidney disease is in Chapter One.

Epilepsy of unknown cause. Where seizures and developmental difficulty occur in the same child, a genetic cause is a reasonable thing to ask about.

The test is genetic. Ask specifically whether an epilepsy gene panel or exome sequencing has been done — a panel looks at a defined list of genes, and exome sequencing looks much more widely. Which is appropriate is a clinical decision, but whether either has been done is a question anyone can ask.

A result changes practical things: which seizure medicines are more likely to suit, what to watch for, and access to a community of families and to research that is specifically about this condition rather than about delay in general.

“Has my child had genetic testing? If so, which test — a panel, or exome sequencing?”

“My child has both developmental delay and seizures. Has a single underlying cause been looked for, rather than treating the two separately?”

“If the first test was normal, was it a panel? Would a broader test find something a panel would not?”

Next chapter: Itching and skin changes nobody has looked at properly

Built from SYNGAP1-Disease-Room.html, 4 May 2026, recovered from the archive on 8 September 2026 and rewritten for this site. Figures marked Grade C are carried from that document and have not been re-verified here. Nothing on this page comes from any individual’s medical record.