We are pleased to be able to share the latest work from our team: ‘Estimating the likelihood of epilepsy from clinically non-contributory electroencephalograms using computational analysis: A retrospective, multisite case–control study’. 

This work has just appeared in Epilepsia, a leading, authoritative journal for innovative clinical and basic science research for all aspects of epilepsy and seizures.  

The work presents a retrospective multi-site analysis from the largest (to date) database of clinically non-contributory EEG recordings (i.e. free of diagnostic abnormalities such as interictal epileptiform discharges). To minimise the risk of overfitting and bias, the markers informing the analysis were pre-determined (and based on their performance in the equivalence of Phase I studies). We were able to demonstrate predictive power (68% balanced accuracy, 61% sensitivity, 4.64 diagnostic odds ratio) in otherwise clinically inconclusive EEG. This demonstrates the potential to accelerate and improve clinical diagnosis of epilepsy without additional burden to the patient or the clinician. 

To read the complete manuscript, please visit here.
 
Congratulations to our contributing authors: 

Scientific Director, Wessel Woldman 

Managing Director, John Terry 

Chief Technology Officer, Ben Howes 

Clinical Scientist, Kay Meiklejohn 

A huge thanks to all the other authors who made this work possible, especially Professor Rohit Shankar MBE.