Authors :
Presenting Author: Sarah Gascoigne, BSc, PGDip – Newcastle University
Nathan Evans, BSc – Newcastle University; Gerard Hall, PhD – Newcastle University; Csaba Kozma, BSc – Newcastle University; Mariella Panagiotopoulou, BSc, PGDip – Newcastle University; Gabrielle Schroeder, PhD – Newcastle University; Callum Simpson, BSc, MSc – Newcastle University; Christopher Thornton, PhD – Newcastle University; Frances Turner, PhD – Newcastle University; Heather Woodhouse, BSc, PGDip – Newcastle University; Leonard Waldmann, BS – Technical University Munich; Jessica Blickwedel, PhD – Newcastle University; Fahmida Chowdhury, MD – University College London Hospitals; Beate Deihl, MD – University College London Hospitals; John Duncan, MD – University College London Hospitals; Ryan Faulder, PhD – Newcastle University; Rhys Thomas, MD – Newcastle University; Kevin Wilson, PhD – Newcastle University; Peter Taylor, PhD – Newcastle University; Yujiang Wang, PhD – Newcastle University
Rationale:
Estimation of the seizure onset zone is often based on a "typical" seizure for the individual. However, this approach may ignore relevant regions involved in epileptogenesis. Further, in introducing novel algorithms for identifying epileptogenic tissue, performance is often assessed through comparison with the clinically labelled seizure onset zone (e.g., Pellegrino et al., 2016). Little work has been dedicated to understanding how seizure onsets vary within individuals.Methods:
We analyzed iEEG recordings of 367 focal seizures across 23 subjects. Seizure onsets were labelled by clinical teams and algorithmically. Channel-wise automatic seizure onset detection was performed using a seizure "imprint" algorithm (see Gascoigne et al., 2023). Onsets were then converted to regions of interest (ROIs) using the Lausanne 120 atlas. Using clinically labelled onsets, we computed the proportion of onset regions resected. Using automatically detected seizure onsets, for each ROI we computed the proportion of seizures in which the ROI was included in the onset on a within-subject basis. In this work, the term "onset region" is used to describe one or more ROIs included in the onset of ictal activity. Results:
Our results show that clinically labelled seizure onsets tended to be resected (median percentage of regions resected: 80%) irrespective of post-surgical outcomes. However, on average 67% of automatically detected seizure onset ROIs were subsequently resected. Further, our results show that seizure onsets vary on a within-subject basis. Consistent onset regions, defined as those present in at least 75% of seizures were seen in only 13 subjects (57%). Interestingly, inconsistent onset regions, defined as those present in no more than half of seizures, were seen in all subjects. There was no relationship found with post-surgical outcomes.Conclusions:
The significance of this work is that we have demonstrated that seizure onsets are not static in refractory focal epilepsy. We envisage that this information could be applied in future work to better understand seizure onsets and apply this knowledge to tailored treatment plans.
References:
Gascoigne, S. J., Waldmann, L., Schroeder, G. M., Panagiotopoulou, M., Blickwedel, J., Chowdhury, F., ... & Wang, Y. (2023). A library of quantitative markers of seizure severity. Epilepsia, 64(4), 1074–1086.
Pellegrino, G., Hedrich, T., Chowdhury, R., Hall, J. A., Lina, J. M., Dubeau, F., ... & Grova, C. (2016). Source localization of the seizure onset zone from ictal EEG/MEG data. Human brain mapping, 37(7), 2528-2546.
Funding: Engineering and Physical Sciences Research Council (EP/L015358/1) and ADLINK
Epilepsy Research UK (
OSR/0550/ERUK/NE02) UKRI Future Leaders Fellowships (MR/T04294X/1, MR/V026569/1)
Wellcome Trust Innovation grant 218380
NIHR UCLH/UCL Biomedical Research Centre