Epileptogenesis is the process through which brain tissue develops an enduring tendency to generate spontaneous seizures. It is not the seizure itself, and it is not simply another name for a low seizure threshold. It describes the longer-term biological changes that can transform a previously non-epileptic network into one capable of producing recurrent unprovoked seizures.
Modern research treats epileptogenesis as a complex, evolving process rather than one single event. It can involve molecular signalling, inflammation, altered ion channels, changes in inhibition and excitation, cell loss, gliosis, synaptic reorganisation, network rewiring and changes in gene expression. The importance of each mechanism depends on the underlying cause.
Epileptogenesis is not ictogenesis
Ictogenesis refers to the immediate transition into a seizure state. Epileptogenesis refers to the longer biological process that creates or strengthens the enduring predisposition to seizures. The two are related but distinct.
A person may have an epileptogenic network for years without having a seizure every day. Likewise, a provoked seizure caused by severe hypoglycaemia or alcohol withdrawal can occur without the person developing epilepsy. This distinction is central to understanding what epilepsy actually is.
What can start epileptogenesis?
Epileptogenesis can follow several types of brain insult, including traumatic brain injury, stroke, central nervous system infection, prolonged seizures, hypoxic–ischaemic injury and some developmental or structural abnormalities. In genetic epilepsies, the epileptogenic process may begin during brain development because the underlying molecular or circuit abnormality is present from early life.
Not everyone exposed to the same insult develops epilepsy. That variability is one of the major scientific questions in the field. The severity and location of injury, age, genetics, inflammation, network structure and many other factors influence whether epileptogenesis occurs.
The latent period
After some acquired brain injuries there can be a period of weeks, months or even years before the first unprovoked seizure. This interval is often called the latent period. It was once imagined as a quiet waiting phase, but research now suggests that substantial biological change can occur during this time.
Inflammation, changes in receptors and ion channels, altered gene expression, axonal sprouting, cell loss and glial responses may all evolve during the latent period. In some people, these changes eventually create a network capable of generating spontaneous seizures.
Inflammation can reshape excitability
Brain injury activates immune and inflammatory pathways. Microglia and astrocytes release signalling molecules such as cytokines, and the blood–brain barrier can become disrupted. These changes can alter neurotransmission, ion homeostasis and neuronal excitability.
Inflammation is not automatically harmful. It is part of normal tissue repair. The problem arises when inflammatory signalling becomes excessive, prolonged or maladaptive and starts to support the development of hyperexcitable networks.
The blood–brain barrier
The blood–brain barrier normally tightly regulates the movement of substances between the bloodstream and brain tissue. After trauma, stroke, infection or prolonged seizures, barrier integrity can be disrupted.
Experimental work suggests that blood-derived proteins entering brain tissue can activate signalling pathways in astrocytes and other cells, altering potassium regulation, inflammation and synaptic function. Blood–brain barrier dysfunction is therefore being studied as both a marker and potential contributor to epileptogenesis.
Neuron loss can change circuit balance
Some acquired epilepsies are associated with selective neuronal loss. Loss of inhibitory interneurons can reduce local restraint, while loss of principal neurons can trigger compensatory rewiring in surviving cells. The effect depends on which cells are lost and how the surviving network reorganises.
Importantly, neuron loss alone does not automatically produce epilepsy. Some experimental and human data show that circuit reorganisation is often more important than simple cell counts.
Axonal sprouting and synaptic reorganisation
After injury, surviving neurons can grow new axonal branches and form new synaptic connections. In temporal lobe epilepsy, mossy-fibre sprouting in the dentate gyrus is one of the most studied examples.
New recurrent excitatory connections can make a circuit more capable of sustaining seizure activity. But sprouting is not uniformly epileptogenic: some new connections may be inhibitory or compensatory. The functional effect depends on where the new synapses form and which cell types are connected.
Changes in GABAergic inhibition
Epileptogenesis can alter inhibitory signalling in several ways: interneuron loss, receptor redistribution, changes in GABA synthesis and release, altered chloride transport and changes in GABAA-receptor subunits have all been described in experimental and human epilepsy.
These changes can make inhibition less effective, spatially misplaced or altered in timing. However, epilepsy is not explained by a simple universal loss of GABA. In some networks, inhibitory circuits remain active or even become excessively synchronising.
Ion-channel remodelling
Neurons can change which ion channels they express after brain injury or repeated seizures. Sodium, potassium, calcium, chloride and HCN channel systems can all be altered.
These changes can affect resting membrane potential, firing threshold, burst firing, adaptation and rhythmicity. In genetic epilepsies, abnormal channel function may be present from the outset; in acquired epilepsies, channel expression can change as part of the epileptogenic process.
Gene expression changes
Epileptogenesis is accompanied by widespread changes in transcription. Genes involved in inflammation, synaptic plasticity, ion channels, growth factors, extracellular matrix and cell survival can become upregulated or downregulated.
These transcriptional changes are not merely passive markers. Some are thought to drive network remodelling. Others may be compensatory attempts by the brain to stabilise itself. Separating harmful from protective changes is one of the main challenges in developing disease-modifying treatments.
BDNF and TrkB signalling
Brain-derived neurotrophic factor (BDNF) and its receptor TrkB are heavily studied in acquired epileptogenesis. These molecules normally support neuronal survival and synaptic plasticity, but after injury or prolonged seizures, excessive or poorly regulated TrkB signalling can promote circuit changes associated with epileptogenesis in experimental models.
This is a useful example of why epileptogenesis is difficult to treat: pathways that are beneficial in normal brain development and repair can become maladaptive under pathological conditions.
Astrocytes and potassium regulation
Astrocytes normally help remove extracellular glutamate, buffer potassium, regulate water balance and support the blood–brain barrier. Reactive gliosis can alter these functions after injury.
If potassium buffering or glutamate uptake becomes impaired, neuronal networks can become more excitable. Astrocytes are therefore increasingly viewed as active participants in epileptogenesis rather than passive scar-forming cells.
The extracellular matrix and perineuronal nets
The extracellular matrix helps stabilise synapses and regulates structural plasticity. After brain injury or seizures, matrix proteins and enzymes can change. Perineuronal nets surrounding some inhibitory interneurons can also be disrupted.
These changes may affect how easily circuits reorganise and how stable inhibitory connections remain. This is an emerging area of epileptogenesis research rather than a fully settled clinical mechanism.
Adult neurogenesis and abnormal integration
In the hippocampus, new neurons can continue to be generated in adulthood. After prolonged seizures or injury, the rate and pattern of neurogenesis can change. Some newly generated granule cells may migrate abnormally, develop unusual dendrites or integrate into circuits in atypical ways.
Experimental work suggests that abnormal integration of new neurons may contribute to temporal lobe epileptogenesis, although its importance in human epilepsy remains under investigation.
Network reorganisation beyond the original injury
Epileptogenesis can involve a network larger than the original lesion. Changes in connectivity, synchrony and excitability can spread across connected regions, creating a distributed epileptic network rather than one isolated abnormal spot.
This helps explain why some people continue to have seizures even when a structural lesion appears small, and why presurgical assessment often has to map a wider network rather than simply identify one abnormal MRI finding.
Why some brain injuries cause epilepsy and others do not
The answer is not one factor. Risk depends on injury severity, location, age, genetics, inflammation, the extent of haemorrhage or tissue damage, seizure activity during the acute phase, and how the network repairs itself.
This question has its own dedicated page because it deals with individual risk after injury. This page focuses instead on the biological transformation itself.
Epileptogenesis may continue after the first seizure
Older models sometimes treated epileptogenesis as a process that ended once spontaneous seizures began. Newer models are more dynamic. In some epilepsies, disease-related changes continue after diagnosis, and the epileptic network can evolve over time.
This does not mean every epilepsy is inevitably progressive. Some epilepsies remain stable, remit, or are strongly age-dependent. The concept is that epileptogenesis can include both the initial development of epilepsy and later modification of the epileptic network in some conditions.
Can seizures themselves promote further epileptogenesis?
Seizures can trigger inflammation, gene-expression changes, synaptic plasticity and ionic stress. In some experimental settings, repeated severe seizures can strengthen epileptogenic mechanisms.
However, it is too simplistic to say that every seizure automatically makes epilepsy worse. Human epilepsy is highly heterogeneous, and many people do not show progressive worsening despite recurrent seizures.
Biomarkers of epileptogenesis
One major research goal is to identify biomarkers that show when epileptogenesis is occurring before the first unprovoked seizure. Candidate biomarkers include EEG abnormalities, high-frequency oscillations, imaging changes, blood or cerebrospinal-fluid markers, inflammatory signatures and network-connectivity measures.
No single biomarker is currently reliable enough to predict with certainty who will develop epilepsy after brain injury. This is a major limitation because disease-modifying treatment would ideally be given during the epileptogenic period, before recurrent seizures become established.
Antiseizure is not the same as antiepileptogenic
Most current antiseizure medicines suppress the probability of seizures while the medicine is being taken. That is different from preventing the underlying development of epilepsy.
An antiepileptogenic treatment would prevent or permanently reduce the development of an epileptic network after a high-risk brain insult. A disease-modifying treatment would alter the long-term course of established epilepsy rather than simply suppressing seizures temporarily.
At present, preventing post-traumatic or post-stroke epilepsy remains an important unmet need. Clinical trials have not yet established a broadly effective antiepileptogenic drug for routine use.
Why animal models are necessary — and limited
Much of what is known about epileptogenesis comes from animal models of status epilepticus, traumatic brain injury, infection and genetic epilepsy. These models allow researchers to examine brain tissue repeatedly across the latent period and test specific molecular pathways.
But no single model reproduces the full diversity of human epilepsy. Findings from one species, one injury model or one brain region cannot automatically be assumed to apply to every person.
What science still does not know
Researchers still cannot explain exactly why two people with apparently similar brain injuries have different outcomes, which molecular changes are essential rather than incidental, or which combination of biomarkers will identify an individual in time for preventive treatment.
Epileptogenesis is best understood as a systems problem: many processes interact over time, and the same diagnosis can be reached through different biological routes.
What this page does — and does not — explain
This page explains the long-term transformation that creates an enduring seizure-prone network. It does not repeat the immediate biology of a seizure, the concept of seizure threshold, or the question of why one particular injury causes epilepsy while another does not. Those subjects each have their own Science & Discovery pages.
Sources
Jasper’s Basic Mechanisms of the Epilepsies, 5th ed. — Epileptogenesis: Molecular Mechanisms and Treatments
Jasper’s Basic Mechanisms of the Epilepsies, 5th ed. — EEG Biomarkers of Epileptogenesis
Jasper’s Basic Mechanisms of the Epilepsies, 5th ed. — BDNF/TrkB Signaling and Epileptogenesis
Jasper’s Basic Mechanisms of the Epilepsies, 5th ed. — Transcriptomic Alterations in Epileptogenesis
Jasper’s Basic Mechanisms of the Epilepsies, 5th ed. — full reference work
Evidence checked: September 2026.