Authors :
Presenting Author: Emory Peng, BS – Massachusetts General Hospital and Harvard Medical School
Giovanna Aiello, PhD – Massachusetts General Hospital and Harvard Medical School
Peter Hadar, MD, MS – Massachusetts General Hospital; Harvard Medical School
Pariya Salami, Ph.D. – Massachusetts General Hospital
Sydney Cash, MD, PhD – Massachusetts General Hospital
Rina Zelmann, M.Eng. Ph.D. – Mass General Research Institute
Rationale:
Sleep and epilepsy are closely intertwined, as epilepsy has been shown to disturb patients’ sleep and interfere with sleep spindles. Usually, epilepsy patients remain on a stable anti-epilepsy medication regimen; however, when they are admitted for a Phase II stereo-electroencephalography (sEEG) EMU stay, their medications will be tapered down as a means of inducing seizure activity to pinpoint the seizure onset zone. Based on this, we hypothesize that patients admitted for a Phase II sEEG EMU stay would experience changes in the abundance and frequency of sleep spindles that are correlated with the medication changes.Methods:
We performed a preliminary retrospective case study using NREM data collected from patients who were admitted to the MGH EMU for Phase II sEEG (N=5). Patients’ seizure history, anti-epilepsy medications, and details regarding each sEEG channel were collected through chart review in the EHR.
The raw EEG data was downsampled from 1024Hz to 200Hz, and the epileptic channels excluded. An automated spindle detector was applied to the remaining channels, and the spindle rate was calculated from the results. Each patient had at least 30 minutes of NREM sleep for each day at least 2 hours apart from any seizure was analyzed. Each patient took between 1-4 anti-epilepsy medications, which included Lamotrigine, Topiramate, Cenobamate, Clobazam, Oxcarbazepine, Briviact, Carbamazepine, Lacosamide, and Zonisamide. Sleep spindle counts were computed for each channel and then averaged over the included channels for each night.
The patients’ prescribed daily anti-epilepsy medication dosage was used as the baseline daily dose measurement, and each day’s medication dosages were then calculated as a percent of the defined daily dose for each patient (Table 1). Spearman’s correlation between spindle rate and medication dosage was computed.
Results:
An average of 124.8 channels (range: 99-152) and 7.4 nights (range 5-9) per patient were analyzed. When considering the average spindle rate across all non-epileptic channels, one patient showed a significant correlation where an increase in medication corresponded to a higher spindle rate (rho = 0.97, p = 0.033). A non-significant trend was observed for another patient. Similar results for each patient were maintained when considering only the contacts in the thalamus (including ANT, CM, PUL, and Ventral).
Conclusions:
This preliminary analysis into muti-medication effects on sleep spindles during a multi-night EMU stay with intracranial EEG, suggests that there could be a relationship between anti-epilepsy medication dosages and sleep spindle rates. If confirmed in a larger cohort, the observed increase in sleep spindle rate with higher medication doses might suggest a positive effect of medication on sleep. This could be an indirect consequence of the decrease in epileptiform activity, which might permit the augmentation of sleep spindles even outside the seizure onset zone.
Funding: NIH NIA (1R01AG090302), AES Junior Investigator Research Award