Abstracts

Rationale: Epilepsy is recognised as a dynamic disease, where susceptibility to seizures and seizure characteristics change over time. Specifically, we recently found variable seizure EEG dynamics within individual patients (Schroeder et al., PNAS, 2020). Additionally, the variability appeared to follow subject-specific circadian or longer timescale modulations. However, whether signatures of these modulations over different timescales can be captured on continuously recorded EEG remains unclear.

Methods: In this work, we analyse continuous interictal intracranial electroencephalographic (iEEG) recordings from video-telemetry units and find fluctuations in iEEG band power over different timescales ranging from minutes up to twelve days. Then, using a linear regression framework, we evaluate how well these band power fluctuations can explain differences in EEG dynamics of epileptic seizures within each patient.

We conducted a quantitative analysis using long-term monitoring iEEG recordings from 18 drug-resistant epilepsy patients with a total of 2656 hours of recordings. We calculated the band power of the 5 main frequency bands in 30 s windows over the entire continuous recording in each patient. Dimensionality reduction was performed and we subsequently investigated fluctuations of the band power data on different timescales using Multivariate Empirical Mode Decomposition (MEMD) (Rehman et al., Proc. R. Soc. A, 466(2117):1291-1302, 2010), a data-adaptive and empirical method suitable for non-stationary data.

Results: As expected from previous literature, and as a validation, we found that all subjects show a circadian fluctuation in their EEG band power. Furthermore, we also observed many other fluctuations on subject-specific timescales, such as ultradian and infradian rhythms, in the band power data. Importantly, we found that a combination of fluctuations on different timescales can explain changes in seizure EEG dynamics in most subjects above chance level.

Conclusions: These results suggest that subject-specific fluctuations in iEEG band power over timescales of minutes to days are associated with how seizures are modulated over time. Future work is needed to link the detected fluctuations to the exact biological time-varying processes. Understanding seizure modulating factors enables development of novel treatment strategies that minimise the seizure spread, duration, or severity and therefore clinical impact of seizures.

Funding: Please list any funding that was received in support of this abstract.: MP was supported by the Engineering and Physical Sciences Research Council, Centre for Doctoral Training in Cloud Computing for Big Data (grant number EP/L015358/1). PNT was supported by the Wellcome Trust (105617/Z/14/Z and (210109/Z/18/Z). YW was supported by the Wellcome Trust (208940/Z/17/Z). We thank the members of the CNNP lab (\url{www.cnnp‐lab.com}) for discussions.