Hippocampal sclerosis is a pattern of neuronal loss and gliosis affecting the hippocampus, most commonly associated with mesial temporal lobe epilepsy. It is one of the most important structural abnormalities in drug-resistant temporal lobe epilepsy and one of the best-established pathological substrates considered during epilepsy surgery evaluation.
The term is sometimes used loosely alongside mesial temporal sclerosis, but the two are not identical. Hippocampal sclerosis specifically refers to sclerosis within the hippocampal formation. Mesial temporal sclerosis is a broader term that can include changes in nearby mesial temporal structures such as the amygdala and entorhinal cortex.
What the hippocampus normally does
The hippocampus is a paired structure deep in the temporal lobes that is heavily involved in memory formation, spatial processing and the organisation of new experiences. It is also embedded in dense networks linking the temporal lobe, amygdala, entorhinal cortex, thalamus and wider limbic system.
Because of this network position, abnormal hippocampal circuitry can contribute not only to seizures but also to memory and cognitive difficulties.
What “sclerosis” means
In neuropathology, sclerosis does not simply mean that a structure looks small on MRI. Hippocampal sclerosis is defined histologically by selective neuronal cell loss together with astrogliosis within the hippocampal formation.
The damage is not uniform. Different hippocampal subfields can be affected to different degrees, and this variation forms the basis of the International League Against Epilepsy classification of hippocampal sclerosis.
The ILAE classification
The ILAE consensus classification recognises three main histopathological patterns in well-preserved surgical specimens.
- ILAE type 1: the classical pattern, with severe neuronal loss and gliosis particularly involving CA1 and CA4.
- ILAE type 2: a CA1-predominant pattern.
- ILAE type 3: a CA4-predominant pattern, sometimes historically called end-folium sclerosis.
Surgical specimens can also show reactive gliosis without the full neuronal-loss pattern required for hippocampal sclerosis.
These distinctions matter because hippocampal sclerosis is biologically heterogeneous rather than one identical lesion in every patient.
Why hippocampal sclerosis is strongly linked with temporal lobe epilepsy
Mesial temporal structures are capable of sustaining recurrent epileptic activity, and hippocampal sclerosis is the most frequent histopathological abnormality encountered in many surgical series of drug-resistant temporal lobe epilepsy.
The abnormal tissue can contain reorganised excitatory and inhibitory circuits, altered interneuron populations, mossy fibre sprouting and changes in dentate granule cells. These network changes may contribute to hyperexcitability and recurrent seizure generation.
Hippocampal sclerosis is not simply a scar
Calling it a “scar” is understandable, but incomplete. The pathology includes selective cell loss, gliosis and network reorganisation. Surviving neurons can form altered circuits rather than the hippocampus simply becoming inert tissue.
This is important because the epileptogenic effect depends on how the remaining network functions, not only on how much tissue has been lost.
What causes hippocampal sclerosis?
There is no single proven cause in every case. Hippocampal sclerosis may develop through several different pathways and can reflect interactions between early brain injury, prolonged seizures, inflammation, genetic susceptibility, development and recurrent epileptic activity.
Associations have been described with prolonged childhood febrile seizures, status epilepticus, central nervous system infection and other early precipitating injuries, but these associations do not prove that every such event causes later hippocampal sclerosis.
Many children who have febrile seizures never develop hippocampal sclerosis or epilepsy.
The direction of cause and effect is complicated
A long-standing question is whether seizures cause hippocampal sclerosis, hippocampal sclerosis causes seizures, or both occur as parts of a shared disease process.
Evidence suggests that more than one pathway exists. In some people, an early injury may contribute to later hippocampal damage and epileptogenesis. In others, pre-existing vulnerability may make the hippocampus both seizure-prone and susceptible to injury.
Repeated seizures may also contribute to progressive network change in some patients, but hippocampal sclerosis should not be described as proof that every seizure has damaged the brain.
Febrile seizures are an association, not a simple explanation
Historical studies found that some people with hippocampal sclerosis reported prolonged or complicated febrile seizures in childhood. ILAE type 1 hippocampal sclerosis has also been associated more often with early precipitating injuries in some surgical series.
However, most children with febrile seizures do not develop temporal lobe epilepsy. The relationship is likely influenced by duration, severity, underlying susceptibility and other biological factors.
Can hippocampal sclerosis be unilateral or bilateral?
Yes. Many people have clear unilateral hippocampal sclerosis affecting one side, but bilateral hippocampal abnormalities can occur.
Bilateral disease can complicate interpretation because both seizure localisation and memory function may be more widely distributed.
Presurgical teams therefore look beyond the MRI image and compare structural findings with EEG, seizure semiology, neuropsychology and sometimes PET, SPECT, MEG or intracranial EEG.
What MRI can show
Typical MRI features include reduced hippocampal volume, increased T2 or FLAIR signal and loss or distortion of normal internal architecture. Coronal imaging oriented appropriately through the hippocampus improves assessment.
The hippocampus may appear visibly smaller on one side, but volume alone is not sufficient for diagnosis.
Why a dedicated epilepsy MRI matters
Subtle hippocampal abnormalities can be missed on a routine brain scan. Epilepsy protocols use high-resolution sequences and planes designed to assess the hippocampus and other structures associated with epilepsy.
A report of “normal MRI” therefore does not always mean that subtle hippocampal sclerosis has been excluded.
This is especially relevant where the scan was performed on an older scanner, with thick slices, motion artefact or without epilepsy-specific sequences.
Whole-hippocampal volume can miss subfield disease
Recent radiological work has emphasised that some pathological subtypes affect selected hippocampal subfields without producing dramatic whole-hippocampal shrinkage.
This helps explain why automated whole-structure volumetry can miss disease that an expert reader identifies through morphology, signal and internal architecture.
Advanced MRI methods increasingly aim to characterise hippocampal subfields rather than treating the structure as a single uniform volume.
A normal-looking hippocampus can still be abnormal
Some patients with temporal lobe epilepsy have no obvious MRI evidence of hippocampal sclerosis but later show subtle pathology, functional abnormalities or microscopic changes.
This does not mean every MRI-negative temporal lobe epilepsy hides hippocampal sclerosis. It means that conventional imaging has limits.
Dual pathology can occur
Hippocampal sclerosis may coexist with another potentially epileptogenic lesion. Examples include focal cortical dysplasia, tumour, vascular malformation, encephalitis-related injury or another glial scar.
The ILAE classification specifically recognises that hippocampal sclerosis can occur alongside another lesion rather than always existing in isolation.
When two abnormalities coexist, presurgical teams need to determine which structures are involved in the seizure network.
Focal cortical dysplasia can coexist with hippocampal sclerosis
When focal cortical dysplasia is associated with hippocampal sclerosis, the combination can be described within the ILAE focal cortical dysplasia framework as a type IIIa pattern in appropriate pathological circumstances.
This is one reason surgery planning cannot simply target whichever abnormality appears most obvious on MRI.
What seizures can look like
Hippocampal sclerosis is commonly associated with focal temporal lobe seizures. Possible symptoms include déjà vu, rising epigastric sensations, sudden fear, altered awareness, behavioural arrest, automatisms and postictal confusion.
Some seizures spread to bilateral tonic-clonic activity.
However, semiology varies and no single symptom proves hippocampal sclerosis.
Memory problems can occur even between seizures
The hippocampus is central to memory formation, so people with mesial temporal lobe epilepsy and hippocampal sclerosis can experience difficulties with learning and recall.
The pattern depends partly on the affected side, language dominance, seizure burden, medication effects and the integrity of the opposite hippocampus.
Memory assessment is therefore an important part of both routine care and presurgical evaluation.
Left and right hippocampal disease can affect memory differently
In many people, the language-dominant temporal lobe is more strongly involved in verbal learning and memory, while the non-dominant side may contribute more to visual or spatial memory.
These are broad patterns, not absolute rules. Individual organisation varies, and formal neuropsychological testing is used when the distinction matters clinically.
EEG may point to the same temporal lobe — but not always
Scalp EEG may show temporal epileptiform discharges or seizures arising over the side of the sclerotic hippocampus.
But deep mesial temporal activity can be difficult to record accurately from the scalp. Sometimes the scalp EEG reflects seizure spread rather than the exact site of onset.
Discordance between MRI and EEG does not automatically invalidate either test. It signals the need for more careful evaluation.
When SEEG becomes relevant
Stereoelectroencephalography may be considered when non-invasive tests do not provide enough certainty about seizure onset, when bilateral temporal involvement is suspected or when more than one structural abnormality is present.
SEEG records directly from implanted depth electrodes and can clarify the relationship between the hippocampus and wider seizure network.
Why hippocampal sclerosis is associated with drug resistance
Hippocampal sclerosis is strongly represented among people with drug-resistant temporal lobe epilepsy. That does not mean medication never works, but seizure freedom may remain difficult to achieve in a substantial subgroup.
Once appropriate medication trials have failed, specialist referral becomes important because continued cycling through medicines may offer diminishing chances of seizure freedom.
When surgery is considered
Surgery is considered when seizures remain drug-resistant and evidence indicates that the epileptogenic zone can be treated without unacceptable neurological risk.
For classic unilateral mesial temporal lobe epilepsy with concordant hippocampal sclerosis, surgery has some of the strongest outcome evidence in epilepsy treatment.
Historical and modern reviews report long-term seizure freedom in a substantial majority of appropriately selected patients. Earlier reviews commonly reported around 60–80%, while a 2026 systematic review found many contemporary series clustering around roughly 50–75%, with anterior temporal lobectomy around 70% at one to two years in selected cohorts.
Concordance matters more than MRI alone
The strongest surgical candidates are generally those in whom multiple lines of evidence point to the same epileptogenic region.
A 2026 systematic review found that MRI-visible hippocampal sclerosis, unilateral EEG onset and concordant PET abnormalities together predicted better seizure outcomes than structural measures alone. Bilateral or discordant abnormalities were associated with less favourable outcomes.
This is why surgery is a network decision, not an MRI decision.
What operations may be used
Operations can include anterior temporal lobectomy, selective amygdalohippocampectomy and, in selected centres and patients, minimally invasive approaches such as laser ablation.
The choice depends on anatomy, seizure localisation, cognitive risk, centre expertise and patient-specific factors.
No single operation is automatically best for every patient with hippocampal sclerosis.
Seizure freedom is not the only surgical outcome
Teams also consider memory, language, mood, visual fields, medication burden, quality of life and the risk of postoperative deficits.
A technically successful operation can still have important cognitive consequences, particularly when the dominant temporal lobe is involved.
Presurgical counselling therefore includes both the possibility of seizure benefit and the possibility of cognitive change.
Memory outcome after surgery is complex
Removing or disconnecting epileptogenic mesial temporal tissue can improve seizure control, but memory change depends on preoperative function, side of surgery, language dominance, the condition of the opposite hippocampus and the surgical approach.
Recent systematic review work confirms that hippocampal sclerosis subtype, seizure outcome and postoperative memory do not follow one simple rule.
Can hippocampal sclerosis progress?
Some longitudinal MRI studies suggest that hippocampal volume and wider temporal networks can change over time in subsets of people with chronic temporal lobe epilepsy.
However, hippocampal sclerosis should not automatically be described as a relentlessly progressive disease.
Different patients follow different trajectories, and apparent progression may reflect seizure burden, ageing, comorbid disease and methodological differences between scans.
Hippocampal sclerosis can be found outside epilepsy
Selective hippocampal neuronal loss can also occur in ageing, neurodegenerative disease, ischaemia and other conditions.
The pattern and clinical context matter. The ILAE classification discussed here applies specifically to hippocampal sclerosis in temporal lobe epilepsy.
Why pathology after surgery still matters
When hippocampal tissue is removed, neuropathological examination can confirm the presence and subtype of hippocampal sclerosis and identify additional pathology.
This can refine understanding of why the epilepsy developed and improve clinicopathological research.
Pathology also reminds us that MRI is an indirect view of tissue. The microscopic disease can be more complex than the scan suggests.
Hippocampal sclerosis is a network disorder as well as a local lesion
Modern epilepsy research increasingly treats mesial temporal lobe epilepsy as a network disorder rather than a problem confined to one small structure.
Functional and structural changes can extend into the amygdala, entorhinal cortex, temporal neocortex, thalamus and connected networks.
This helps explain why some people continue to have seizures after technically adequate removal of visibly abnormal hippocampal tissue.
What this page does — and does not — explain
This page focuses on hippocampal sclerosis as a structural epilepsy substrate. The normal role of the hippocampus is covered on The Hippocampus: Memory, Seizures and Epilepsy. Broader temporal seizure patterns are covered on Temporal Lobe Epilepsy. MRI technique is covered on MRI and Neuroimaging in Epilepsy and Why a Dedicated Epilepsy MRI Is Different. Surgical candidacy is covered on Epilepsy Surgery and Presurgical Assessment.
Sources
International League Against Epilepsy — International consensus classification of hippocampal sclerosis
Blümcke et al. — ILAE Task Force consensus classification of hippocampal sclerosis
Thom — Hippocampal sclerosis in epilepsy: neuropathology review
Hippocampal sclerosis — origins and imaging
Radiologic Classification of Hippocampal Sclerosis in Epilepsy
Preoperative predictors of seizure outcomes after epilepsy surgery for mesial temporal sclerosis — 2026 systematic review
Postoperative seizure and memory outcome of temporal lobe epilepsy with hippocampal sclerosis
Evidence checked: September 2026.