Acquired Epilepsy & Epileptogenesis
Can we stop epilepsy before it starts?
After a stroke, head injury, tumour or brain infection, epilepsy can emerge months, even years, later. That delay is one of neurology’s rare open doors, a chance to see epilepsy coming, and perhaps close it before it begins. Our programme turns that window into a research strategy.
We treat the silent window between injury and epilepsy as a target, from predicting who is at risk to finding, and testing, ways to prevent it.
A window that opens in advance
Unlike most epilepsies, acquired epilepsy announces itself, a brain insult, a silent latent period, then seizures. Stroke alone is the leading cause of epilepsy in older adults. If we can see it coming, epileptogenesis becomes a target, not just a diagnosis.
A global consortium
We built and co-lead the SeLECT Consortium, the largest research collaboration dedicated to seizures after stroke, spanning more than 20 cohorts across four continents on an open-science framework that turns scattered datasets into a shared engine for discovery.
Gold-standard prediction
Our SeLECT and SeLECT 2.0 models are validated internationally and give individual, absolute risk at the bedside, and adding early EEG (SeLECT-EEG) sharpens them further. predictepilepsy.com brings our models, and others, to clinicians everywhere.
Rethinking the 7-day rule
For over a century, a single 7-day cutoff has separated acute symptomatic from remote symptomatic seizures, and dictated how they are treated. Our data suggest that line may sit in the wrong place, with direct consequences for treatment and for everyday questions like driving and work (the COSY framework, the Chance Of a Seizure in the next Year).
The true burden of acquired epilepsy
Disease-burden frameworks usually fold epilepsy into the parent condition, hiding its real cost. Using one of the world’s largest federated health datasets across many brain insults, we are building the case that acquired epilepsy is a distinct contributor to morbidity and mortality, not a bystander, and that early treatment may improve outcomes.
Causal inference for antiepileptogenesis
There are still no treatments that prevent epilepsy after brain injury. We mine large real-world datasets with causal inference methods such as target-trial emulation to surface existing drugs that might interrupt epileptogenesis, prioritising the strongest candidates for future prevention trials. This work is ongoing.
Help us predict, and prevent, epilepsy after stroke.
PRe-SeLECT (Prospective Risk Evaluation: Seizures after Stroke) is our prospective, biomarker-rich, trial-ready cohort, the foundation for the next generation of prevention trials. We’re inviting sites to join.
Our work on this topic, in the literature.
Continuous versus short EEG after ischaemic stroke
Clarified what continuous EEG adds over a short recording for detecting abnormalities and predicting who develops epilepsy after stroke.
Driving in people with seizures and epilepsy
Set out how individual seizure risk, expressed as the Chance Of a Seizure in the next Year (COSY), can inform fairer, evidence-based driving rules.
Epilepsy as a dynamic disease, toward individualized seizure risk prediction
Argued that seizure risk is not fixed but changes over time, and that individualized, continually updated prediction should guide real-world decisions.
The role of EEG in predicting post-stroke seizures (SeLECT-EEG)
Showed that adding early EEG to the SeLECT model sharpens prediction, especially in patients without acute symptomatic seizures.
Timing and type of acute symptomatic seizures, post-stroke epilepsy and mortality
Detailed how the timing and type of acute symptomatic seizures shape later epilepsy and death, refining who needs close follow-up.
Implications for driving based on the risk of seizures after ischaemic stroke
Translated post-stroke seizure risk into concrete, individualized guidance on fitness to drive after stroke.
Mortality and epilepsy risk by type of acute symptomatic seizure (SeLECT 2.0)
Revealed that different acute symptomatic seizure types carry very different epilepsy and mortality risks, and updated the model to SeLECT 2.0.
Seizures after ischaemic stroke, a matched multicentre study
Quantified the risk and timing of seizures after stroke in a large matched cohort, the evidence base behind modern risk stratification.
Prediction of late seizures after ischaemic stroke (the SeLECT score)
Introduced SeLECT, the first validated model to predict late seizures after ischaemic stroke from five routine clinical variables.
