Do Seizures Damage the Brain ?—Cumulative Effects of Seizures and Epilepsy: A 2025 Perspective

Abstract

In 1885, William Gowers proposed that epilepsy is a progressive disease, based on clinical evidence before any effective treatments were available. His long-standing hypothesis has been summarized with the statement “seizures beget seizures.” Whether this is the case and related questions about seizure-induced modification and damage of brain circuits are of fundamental importance for neurobiological understanding of epilepsy, development of effective treatment strategies, clinical management, and prognostication. Consensus about progression and seizure-induced damage has remained controversial. Here, we critically review these long-standing questions, incorporating perspectives about perceived inconsistencies in past studies, potential implications of recent longitudinal imaging and cognitive studies, and emphasize experimental and clinical gaps that have proved challenging. Answers to these questions are important for development of management strategies to achieve prompt effective acute control of seizures and prevention of their potential recurrence and long-term comorbidities.

Introduction

“Do seizures damage my brain?” and “Will epilepsy cause dementia?” are common questions posed by patients with new-onset epilepsy, as well as those whose seizures continue in spite of treatment. The questions arise from anxiety that seizures may cause cumulative brain damage and cognitive impairment. In contemporary terms, Gower's hypothesis “seizures beget seizures” ¹ implies that single or repeated seizures may alter brain structure and networks leading to progressive worsening of epilepsy accompanied by comorbidities and neurocognitive impairment. The hypothesis that “seizures beget seizures” has recently been reviewed from historical perspectives. ²

Evidence that status epilepticus (SE) induces brain damage has been accepted as the rationale for increasingly aggressive treatment. The question that remains is whether a single seizure leads to changes contributing to adverse long-term consequences and whether repetitive seizures, other than SE, may cumulatively do so. The neurobiological phenomenon of “kindling” ³ by repeated seizures evoked by various induction methods in multiple species supports the progressive hypothesis. Goddard noted that evoked seizures, not merely stimulation, were required to produce kindling progression. During the 1990s, histological studies provided evidence that repeated brief seizures in kindled rodents cumulatively produce neuronal damage, circuit reorganization, with accompanying comorbidities.^4,5 In 2001, animal and human studies evaluating the question of whether seizures cause neuronal damage were discussed in a workshop and subsequent reviews which concluded, “A firm answer to this question has been surprisingly elusive.” ⁵

During the past 2 decades, new imaging and cognitive studies have added substantial evidence of cumulative chronic effects associated with epilepsy. Here, we critically evaluate the evidence in 2025 with review of the most pertinent past and recent experimental observations and human population, cognitive, neuropathologic, and imaging studies. We conclude with consensus points based on currently available evidence, which may be useful to resolve questions in future studies and provide a rationale for the development of more effective acute interventions to prevent possible consequences of seizures that are so important for patients.

Experimental Studies and Neurobiological Evidence for Epilepsy Progression and Damage—Perspectives in 2025

The potential consequences of seizures, including progression, seizure-induced damage, and neurocognitive effects have been studied for decades in animal epilepsy models ranging from SE to single seizures. The effects of SE and repeated brief seizures developing after SE have been extensively examined in models of SE using kainic acid and pilocarpine. ⁶ Effects of single seizures or seizure threshold models without prior SE have also been examined in models including pentylenetetrazol, maximum electrical shock, 6-Hz corneal-evoked seizures, and after brief kindled seizures. The SE models with recurring seizures offer the opportunity to examine progression and damage after an inciting event, whereas kindling and other models inducing single self-limited seizures offer the opportunity to assess cumulative effects of repeated brief seizures.^6,7 Brief kindled seizures can be evoked by repeated electrical or chemical stimuli in species ranging from amphibians to primates, and with sufficient repetition, progress to increasingly long and eventually spontaneous seizures. The occurrence of kindling in multiple species is an example of seizure-induced brain plasticity evoked by repeated brief seizures producing permanent effects in neural circuits, ie, a phenomenon of circuit plasticity, ⁸ supporting Gower's statement that “seizures beget seizures.” Some have argued that the kindling phenomenon of seizure-induced plasticity as an experimental model does not recapitulate features of typical sporadic epilepsy and may therefore not be relevant in human epilepsy.^9,10 This viewpoint makes an exception for the human species to a phenomenon of circuit plasticity robustly observed across species ranging from amphibians to higher primates. Rather than dismissing the relevance of kindling, a reasonable question is “what aspects of human epilepsy might be influenced by the kindling phenomenon?” One aspect that deserves critical attention is the focus of this review—“do seizures damage the brain?”

Multiple studies in rats evaluated effects of kindling-induced focal and secondary generalized tonic-clonic seizures (GTC) ¹¹ on neuronal populations in the hippocampus and other limbic and neocortical regions. Four separate studies using laborious nitrocellulose (celloidin) embedding, which preserves cellular and tissue architecture in detail not afforded by conventional freezing methods, detected neuronal loss using stereological counting techniques.^12–15 These histological techniques, which require >6 months and are rarely applied in epilepsy research or conventional neuropathology, revealed neuronal loss in the dentate gyrus (DG) and hippocampal CA1 after 3 GTC kindled seizures, progressing after 150 seizures to 54% in the hilus of the temporal DG, 82% in CA3, and 81% in CA1, evolving into a pattern resembling classical hippocampal sclerosis. ¹³ Cumulative neuronal loss was associated with progressive spatial memory dysfunction detected after as few as ∼6 focal seizures and 30 GTCs,^14,15 suggesting that repeated seizures may contribute to hippocampal sclerosis and memory deficits in at least some cases.

Studies from multiple laboratories demonstrated that kindled seizures induce cellular alterations including neuronal death, axonal sprouting, astrocytosis, and neurogenesis.^16–20 Apoptosis occurs in the rat DG after a single hippocampal stimulation, with increasing apoptotic cells after 40 kindling stimulations,^16–18 accompanied by neurogenesis in the DG,^16,20 where newborn neurons differentiate into granule cells after amygdala kindling. ²⁰ While the functional impacts of circuits formed by newly born neurons are complex, ²¹ neurogenesis plays a role in hippocampal circuit remodeling potentially influencing the development of epilepsy. ²²

Repeated brief seizures evoked by kindling eventually result in spontaneous seizures.^3,6,7 The emergence of spontaneous seizures in rats after ∼100 evoked seizures ²³ is accompanied by reduced paired pulse inhibition, reduced γ-aminobutyric acid (GABA) currents, and cumulative seizure-induced cell loss of GABA-ergic interneurons labeled by cholecystokinin (CCK) and the neuronal GABA transporter (GAT-1), ²³ which project axo-somatic and axo-axonic inhibitory terminals to granule cells and axon initial segments. Loss of axo-somatic and axo-axonic inhibitory terminals would facilitate spike output from granule cells to CA3 neurons, and is consistent with reduced paired pulse inhibition and GABAergic currents. Seizure-induced loss of axo-somatic and axo-axonic inhibition could contribute to the development of intractable epilepsy. ²³

Methodological differences account for discrepancies among histological studies investigating cumulative seizure-induced neuronal damage, prompting skepticism about the relevance of seizure-induced damage in human epilepsy.^9,10 A recent study examining frozen sections from rats with post-SE seizures and rats with >1000 seizures evoked by kindling did not show cumulative neuronal loss with kindling, ²⁴ prompting editorial comments that kindling does not produce neuronal loss. ²⁵ However, quantitative detection of cell loss in frozen sections is flawed by insensitive preservation of cellular and tissue architecture due to section dehydration and cellular distortion,^26–29 calling into question the reported findings. Subtle but significant apoptotic neuronal loss occurs after individual seizures,^16–18 and even with the most sensitive nitrocellulose embedding techniques neuronal loss becomes visually detectable only after more than 10% loss has occurred.^12–15 Interpretative flaws arising from the use of insensitive histological methods^24,26–29 have confounded understanding of consequences of seizures in neural circuitry,^9,10 and appreciation of the potential damaging effects of seizures and contributions of circuit plasticity to human epilepsy progression. ³⁰ Experimental studies suggest a loss of just a few percent of the neurons could have a substantial functional effect. ³⁰

Human Studies—Perspectives in 2025

Epidemiologic and Population Studies of Recurring Seizures

Epidemiologic studies of single seizures, recurrence rates, and evolution of initial seizures into established epilepsy have demonstrated outcomes ranging from remission to evidence of progression in some patients. As nearly 50% of patients presenting with seizures experience remission, ³¹ epilepsy is clearly not invariably progressive, but evidence of a progressive course is evident in some patients. Rates of seizure recurrence after a first unprovoked seizure in individuals with otherwise low risk for seizures revealed that increasing numbers of seizures increased the risk for additional recurrences, suggesting progressive seizure risk due to recurring seizures. ³²

While many treated patients achieve remission, there is a well-recognized clinical pattern of an inciting event such as febrile SE followed by an unprovoked seizure several years later, evolving into additional seizures occurring with gradually shortening interseizure intervals and eventually into drug-resistant temporal lobe epilepsy (TLE) 10 to 20 years after the initial event. ³³ This pattern in a subset of patients is consistent with a progressive epileptogenic process, either primary or seizure-induced. Little is known about potential structural and etiological factors distinguishing these patients from those who achieve remission. Genetic background differences between remission and progression groups have not been systematically examined.

Cognitive Course in Epilepsy

Concern regarding the cognitive impact of epilepsy has been long-standing. Patients regard cumulative cognitive decline as a form of damage. This voluminous literature^34–36 is touched on very briefly here. Relationships between cognition and duration, course, and complications of epilepsy have been investigated in case studies, cross-sectional and prospective investigations since 1912.^37,38 The cognitive course of epilepsy is best examined by controlled longitudinal designs involving repeated assessment of matched healthy controls and patients with epilepsy including accurate measures of seizure frequency. A recent review identified 19 controlled prospective neuropsychological investigations including adult and pediatric studies. ³⁵ Abnormal declining trajectories in cognitive domains including memory, attention, processing speed, and higher executive functions were noted across nine controlled adult studies. Significant trajectory differences were also noted on tests of intelligence, academic achievement, and executive and motor functions in 10 pediatric studies. Static abnormalities over time were common, with decline occurring in some patients.

That severe intractable epilepsy can exert adverse cognitive effects over time is clear ³⁶ with unambiguous relationships between age-accelerated cognitive change and the interval average annual frequency of GTCs, but also focal (formerly complex partial) seizures. Greater numbers of interval GTC seizures were associated with declining verbal and performance intelligence quotients, verbal learning and memory, naming and semantic fluency over the median 13 year test–retest interval. ³⁴ In another controlled prospective cohort study, patients with TLE showed a worse 4-year cognitive course across measures of memory, executive function, motor speed, language, and intelligence compared to matched education-adjusted, healthy controls. ³⁹

Cognitive declines across specific abilities were most notable in subsets of TLE patients characterized by older chronological age, less cerebral reserve (lower baseline IQ score, fewer years of education), longer duration of epilepsy, and especially the presence, location, and degree of baseline morphometric abnormalities. Cross-sectional studies have shown that the range of neuropsychological morbidity in TLE can be widespread across multiple cognitive domains despite the apparent focal epileptic process. ⁴⁰ Further, increasing epilepsy duration has been associated with greater cognitive morbidity.^41–43

The major concern is how cognition fares over decades of epilepsy. Neuropsychological studies across epilepsy patients over decades demonstrate altered cognitive trajectories not seen in age-matched people without epilepsy. In a large controlled cross-sectional study covering decades, childhood-onset TLE was associated with a neurodevelopmental offset in cognition (memory) that remained largely stable and below controls over decades, with an implied earlier onset of later life age-related cognitive disorders (eg, dementia) compared to controls. ⁴² A recent study showed epilepsy duration exceeding 20 years to be associated with declines in intelligence, verbal learning and memory, and reduction in adult neurogenesis in the DG. ⁴³ Human studies associating cognitive declines with decades of epilepsy, neuronal loss, and reduced neurogenesis, as well as animal studies showing seizure-induced neuronal loss, potentially implicate continuing underestimated seizures as a contributing factor, in addition to primary progressive etiologies and/or effects of other factors such as ASMs. The developmental cognitive impact of seizures is suggested by a recent longitudinal neurodevelopmental study in pediatric patients between 12 and 36 months of age with tuberous sclerosis complex (TSC) identified cumulative seizure burden (defined as days with observed seizures) as the most significant predictor of neurodevelopmental outcome, exceeding the influences of tuber lesion volume or ASM treatment. ⁴⁴ While implying that aggressive treatment might have favorable effects on development and cognitive functions, the potential negative effects of aggressive pharmacological treatment remain important considerations in pediatric populations.

Human Neuropathologic Studies

Neuropathological examination to assess the extent of damage that may be caused by recurring seizures is inherently limited by confounding initial damage or acquired injury, difficulties implementing stereological counting techniques in closely matched control tissue, and unreliability of retrospective estimates of seizure frequency. Not all pathological studies have detected neuronal loss in patients with intractable epilepsy and status epilepticus,^45,46 but the largest study to date of resected pathology of surgically treated, medically intractable patients with TLE and frontal lobe epilepsy (FLE) reported different patterns of neuronal loss and structural changes in various brain regions. ⁴⁷ The frequent observation of mesial temporal sclerosis in patients with TLE contrasts with pathological studies of refractory FLE secondary to focal cortical dysplasia (FCD), which show little or no neuronal loss in adjacent perilesional cortex. Pathological studies of FCDs, including studies using rigorous stereological counting methods, have confirmed reduction of neurons within regions of dysplasia versus perilesional cortex and nonepilepsy controls, but are limited by difficulties in comparing perilesional cortex to normal matched control regions. ⁴⁸ In a study of cortical tissue resected for treatment of refractory epilepsy associated with FCD from patients who experienced “thousands” of seizures, ⁴⁹ perilesional cortex adjacent to dysplasias appeared unaltered in terms of tissue architecture and neuronal density. It was concluded that “this evidence argues against the hypothesis that epileptiform activity per se contributes to focal brain injury, at least in neocortical epilepsies”. ⁴⁹ Neuropathological studies of resected surgical sections have provided quantitative immunohistochemical analyses of neurodegenerative markers such as tau, apolipoprotein epsilon 4 (APOE4), and others that vary with duration of epilepsy and cognitive function.^50–52 Blood and CSF studies have revealed elevations of potential damage-associated biomarkers including neuron specific enolase, neurofilaments, interleukins, tumor-necrosis factor-α, tau/p-tau, and amyloid proteins consistent with damage after acute seizures and SE, but have not been informative regarding questions about cumulative damage.^53,54 Neuropathological studies strongly support the association of seizures with brain damage and degenerative markers, but have not resolved questions about damage from initial etiologies versus effects of ongoing seizures in various epileptic syndromes.

Human Imaging Studies

Imaging studies provide unequivocal evidence that SE can produce brain damage, as directly demonstrated after febrile SE resulting in chronic hippocampal atrophy, malformation, ⁵⁵ and associated cognitive impairment. Not all cases of febrile or nonfebrile SE produce detectable damage, but prolonged complex febrile seizures appear to be associated with greater long-term alterations and risk for chronic epilepsy. Questions remain about the effects of isolated seizures on MRI alterations. Interpretation of progressive changes in serial imaging studies in patients with chronic epilepsy are complicated by potential independent confounding factors: progression of the initial injury, further insults after epilepsy onset (eg, SE or traumatic brain injury), comorbidities, treatment effects from ASMs, surgery, neurostimulation, and the unreliability of historical reporting of seizure frequency.

Case and cross-sectional imaging studies provide evidence that isolated seizures may be followed by cumulative damage. A study comparing patients with newly diagnosed cryptogenic TLE, chronic well-controlled TLE, and drug-resistant TLE detected smaller hippocampal volumes in patients with drug-resistant epilepsy, correlating with estimates of seizure frequency and epilepsy duration. ⁵⁶ A single well-described case of a patient with an initial normal MRI followed by serial MRI scans at 2.75 years after seizure onset and a total of 7 GTC seizures demonstrated evolving bilateral asymmetric volume loss (30% on the left side, 10% on the right side) and increased signal intensity on T2-weighted images, consistent with cumulative damage after isolated focal and GTC seizures. There was no further change 9 months later, after a total of 10 GTC seizures. ⁵⁷

Prospective longitudinal studies have provided suggestive, but hardly conclusive, evidence for seizure-induced progression. A study of serial MRI volumetrics in 24 patients identified progressive ipsilateral hippocampal volume decrease of 9% over ∼3.5 years, with volume loss correlated to generalized seizure count between scans. ⁵⁸ Another prospective study compared 179 patients with TLE, extratemporal partial epilepsy, and generalized epilepsy to 90 control participants with MRI scans 3.5 years apart, concluding that while significant atrophy developed in individual patients, overt structural cerebral damage was not an inevitable outcome of epileptic seizures. ⁵⁹ In a meta-analysis of 42 neuroimaging studies, ipsilateral hippocampal atrophy correlated with epilepsy duration (r = 0.42), reported seizure frequency (r = 0.35), and epilepsy duration-related progression of atrophy in extratemporal cortical regions, subcortical regions, and whole brain. ⁶⁰ The review concluded that there was a low to moderate level of evidence for progressive atrophy across studies.

A world-wide multicenter MRI analysis (ENIGMA) identified patterns of grey matter reduction across epilepsy syndromes with distinctive differences between syndromes, implying more widespread structural compromise than previously appreciated. ⁶¹ A cross-sectional MRI study in patients with TLE and idiopathic generalized epilepsy employing artificial intelligence-based tools and modeling identified distinct spatiotemporal MRI patterns that were consistent with epilepsy-related progressive brain atrophy, prompting the authors to suggest that both focal and generalized epilepsies are dynamic disorders associated with brain structure deterioration and features shared with classical neurodegenerative disorders. ⁶²

Longitudinal MRI studies have demonstrated striking evidence of progressive cortical thinning in the brains of patients with epilepsy. In a longitudinal case-control study 3T MRI scans obtained 6 months to >3 years apart in patients with focal epilepsy were compared to age-matched healthy control participants. ⁶³ Widespread cortical thinning was seen in 77% of patients with epilepsy with an annualized rate of global cortical thinning twice that of the thinning associated with aging in the control group. In patients with TLE or FLE, progressive atrophy was extensive beyond the focal onset area and commonly occurred in the contralateral hemisphere. In patients with TLE, cortical thinning was noted within structures connected with the ipsilateral hippocampus, and was most evident in adults aged >55 years and during the first 5 years after seizure onset. However, progression was not associated with historical seizure frequency, generalized seizure type, or medication burden, and no difference was observed between patients with or without ongoing seizures. ⁶³ The latter observations would appear to support progression from an underlying disease process rather than effects of individual seizures.

After epilepsy surgery in patients with TLE, patients who were seizure free had rates of progressive cortical thinning comparable to healthy control groups during normal aging, whereas those with postoperative seizures had small areas of continued accelerated thinning after surgery. These latter observations support a relationship of cortical thinning to seizures.^64,65 Another recent imaging study of TLE with unilateral hippocampal sclerosis revealed gray matter hypertrophy in the contralateral amygdala in addition to thinning in contralateral network areas, suggesting network plasticity that could be seizure-induced. ⁶⁶

The increasing documentation of heterogeneous and progressive structural alterations across the epilepsies may reflect both variable etiologies and secondary processes including seizure-induced damage and plasticity. The cellular basis for progressive thinning and other evolving structural alterations occurring as a function of epilepsy duration is uncertain, but the potential influence of seizure-induced processes is made more compelling by recordings from indwelling and responsive neurostimulation (RNS) electrodes demonstrating that seizure frequency is greater than suggested by clinical reporting.^67,68 Subjective counts and RNS devices have revealed that subclinical seizures are underestimated, undermining interpretations about lack of relationships between historical seizure frequency, cortical thinning, and other cumulative structural alterations.

Epilepsy and the Alzheimer's Disease Spectrum

Relationships among neuronal loss, functional deficits, and epilepsy have been increasingly recognized in patients with Alzheimer's disease (AD) and epilepsy.^69,70 Epilepsy is more common in patients with AD, and AD is more common in patients with epilepsy. Patients with late-onset epilepsy (LOE, >55 years old) of unknown etiology have 2–4 times higher risk for developing mild cognitive impairment and AD compared to individuals without epilepsy. Epilepsy duration and seizure burden predict dementia. Cognitive decline was present in more than half of patients with LOE compared to nonepileptic controls, and further declined over 12 months. In patients with mild AD and epilepsy, seizures may exacerbate cognitive deterioration with 4-fold faster decline at 1 year. There is high prevalence of subclinical epileptiform activity in AD, with higher spike numbers correlating with 2.3-fold faster cognitive decline at 1 year.^71,72 A recent population-based prospective study demonstrated that patients with childhood-onset epilepsy have a trajectory of increased amyloid accumulation measured by [11C]PIB imaging compared to matched nonepileptic controls. ⁷³ In 5xFAD transgenic mice which exhibit Aβ amyloid deposition, increased kindling susceptibility, and model Alzheimer's disease, 40 Hz auditory-visual stimulation reduced both Aβ accumulation and kindling susceptibility. ⁷⁴ Intriguing associations between epilepsy and AD underscore questions about bidirectional influences between seizures and neuronal loss.

Perspectives About Seizure-Induced Progression and Cumulative Damage in 2025

Are perspectives in 2025 potentially advantageous to resolve questions about progression posed by Gower's 1885 hypothesis and the 2002 review conclusion that “a firm answer to this question has been surprisingly elusive”? Noteworthy in 2025 is the new evidence of progressive cortical thinning occurring in association with seizures, the observation of progressive declines in cognitive function evolving with epilepsy duration over decades in association with diminished neurogenesis in resected human brain tissue, and evidence from continuous scalp EEG and depth EEG recording that electrographic seizures occur far more frequently than clinically recognized seizures, implying that seizure frequency may be significantly underestimated in many patients and studies.

Compared to 2002, damaging effects of SE have been further documented, demonstrating that longer focal seizures have a greater impact than shorter seizures. ⁷⁵ The dividing line in terms of seizure duration and damage in humans has become less clearly defined and may be a continuum. Experimental studies demonstrate that specific brain regions are particularly prone to seizure-induced neuronal death. Effects of extratemporal seizures have been less studied, but neuronal death can be widespread and variable. The regional variability of damage introduces complexity into investigations of possible progression and cumulative damage in both human and animal studies that remain a challenge today.

As significant numbers of patients achieve seizure control and remission, and with many new effective ASMs, widespread use of effective surgical resection, and growing evidence that device and stimulation therapy can improve control in some refractory patients, Gower's question in 2025 should be “which, if any of the epilepsies, are progressive?” Which patients demonstrate “progression” based on progressive etiologies versus effects of continuing seizures? Imaging and cognitive studies in epilepsies previously considered self-limited or benign (eg, benign epilepsy with centrotemporal spikes, benign Rolandic epilepsy, juvenile myoclonic epilepsy and absence epilepsy) have less association with structural abnormalities than focal seizures, but are revealing impairments such as subsets of absence patients demonstrating long-term cognitive and functional deficits. ⁷⁶

In 2025, the increasing number of new transgenic animal models manifesting seizures and epilepsy and expanded awareness of molecular, cellular, and circuit abnormalities in conventional animal models have not simplified the questions. In many transgenic models, similar to human disorders, it may be difficult to dissect contributions of abnormalities reflecting the primary genetic mutation versus secondary consequences seizures, as in mouse models with SCN1A and K(V)1.1 mutations. Conventional models with controlled seizure induction methods offer possibilities to distinguish etiologic cause from secondary seizure-induced changes. The cumulative effects of individual seizures in chronic epilepsy may be dependent on seizure type, duration, locus and spread, clustered versus isolated seizures, intervals between seizures, number and frequency of seizures, and are likely influenced by genetic background and differential effects of interseizure intervals on seizure-induced gene expression. Patterns of gene expression influenced by the above temporal and spatial dynamics of individual seizures may contribute to different structural and functional effects of a single seizure and seizure-induced circuit plasticity, but offer potential avenues for future tailored treatments.

As in 2002, the major challenge in assessment of the consequences of human epilepsy continues to be the uncertainty and inaccuracy of reported historical seizure frequency. Continuous scalp EEG and depth EEG and RNS recordings demonstrate that subclinical seizures are underestimated by patient and family historical reports,^66,67 limiting interpretation of studies of relationships among seizure frequency, cumulative damage, and progression. These challenges also apply to many experimental studies with the exception of seizure induction methods that can be directly controlled, as in kindling, where seizure number and cumulative damage have been directly correlated when sufficiently sensitive histologic methods are employed.

Consensus Points in 2025?

The long-standing controversies about progressive brain damage in chronic epilepsy remain difficult to conclusively resolve, but recent longitudinal human studies demonstrating progressive cortical thinning associated with continuing seizures and cognitive studies revealing progressive declines over decades associated with neuronal loss and decreased neurogenesis provide new context to critically reevaluate Gower's progression hypothesis and the significance of seizure-induced neuronal loss observed in the most sensitive experimental studies. We conclude with a summary of potential consensus points across experimental and clinical perspectives in 2025:

Epilepsy is not progressive in all patients

The >60% of patients achieving seizure control do not have progressive courses or cognitive decline, but a subset of patients with TLE and other focal epilepsies experience progressive decline without apparent progressive initial insults. The decline may include increasing frequency of seizures over time, with or without functional or cognitive declines. Typical absence seizures and other “benign” childhood epilepsies do not invariably lead to progressive decline, but in a subset of absence and “benign” epilepsy patients, attention, cognitive, behavioral, and psychiatric issues may emerge in adolescence after seizures abate, perhaps reflecting initial etiology or a consequence of seizures.

Damage is greater with longer versus shorter seizures

Damage in SE has increased urgency for rapid treatment, with intervention thresholds shifting from 30 min of unremitting seizures to 5 min. Whether a single seizure or repeated brief seizures lead to significant structural or functional changes in the human brain remains controversial, but the most sensitive histologic methods in experimental animals demonstrate subtle, but cumulative, neuronal death with repeated brief seizures accompanied by functional deficits.

Progressive cortical thinning occurs with recurring seizures in human longitudinal imaging studies

Cortical thinning increases with duration of epilepsy, appears to be related to seizures, and progresses in postsurgical patients whose seizures continue compared to postsurgical patients who are seizure free. These observations are consistent with seizure-induced cumulative cortical damage.

Neuropsychological studies across epilepsy patients over decades demonstrate a declining cognitive trajectory not seen in age-matched people without epilepsy

Recent human studies associating cognitive declines with decades of epilepsy, neuronal loss, and reduced neurogenesis, as well as animal studies showing seizure-induced neuronal loss, potentially implicate continuing underestimated seizures as a contributing factor, but primary progressive etiological and/or effects of other factors such as ASMs cannot be excluded.

Studies of progression in humans are significantly limited by the inaccuracy of seizure counts over long periods

Continuous scalp EEG recordings reveal seizures not detected clinically, and depth EEG recordings demonstrate that electrographic seizures are more frequent than detected in scalp recordings or clinical records, and may not be recognized as a clinically apparent seizure. Patient, family, and medical historical records are inadequate for resolution of questions about relationships of seizure burden to imaging patterns of damage and cumulative cognitive deficits. The insensitivity and inaccuracy of human seizure counts is a major impediment to resolving questions about progressive impact of uncontrolled seizures, but experimental studies offer the potential to determine seizure counts more accurately.

Across both experimental and human domains, it remains difficult to separate progression caused by underlying primary etiologies from progression secondary to recurring seizures

Experimental studies that induce focal seizures without initial pathologic alterations offer the potential to separate these confounding processes. Transgenic animals with epilepsy secondary to genetic or epigenetic defects do not typically provide this opportunity, or resolve the question of initial etiological or secondary seizure-induced contributions to progressive damage.

These points derived from recent clinical, imaging, cognitive studies, and previous experimental observations may serve as guidelines and provide potential directions for investigations to address cumulative adverse consequences of recurring seizures. Concerns about progressive damage remain compelling for the ∼35% of patients who do not achieve control with current therapies and for those who lack access to adequate treatment. The available data about progressive adverse long-term effects of epilepsy provides a compelling need for more effective acute interventions to prevent recurring seizures and their potential consequences.