Authors :
Presenting Author: Darion Toutant, – University of Manitoba
Marcus Ng, BMSc, MD, FRCPC, CSCN(EEG), FAES – Associate Professor of Neurology / Epileptologist, University of Manitoba
Rationale: Interictal epileptiform discharges (IEDs) exhibit distinct characteristic differences throughout the sleep-wake states. This study aims to create a gold-standard database of IEDs and investigate their characteristics during sleep-wake states. Understanding these variations in IED features across wakefulness, NREM sleep stages 1-3, and REM could hold crucial information for why REM expresses a suppressive ability towards IED propagation.
Methods: 10 EMU patients with multiple nights have been professionally marked for sleep states and IED validity using lab-developed software to streamline the process. With a gold-standard dataset, a computational pipeline of nine defining IED morphological features (Figure 1) is preprocessed using filters and extracted from three montages (Laplacian, Raw, Referential). Because data was non-normal, we tested differences of each feature in each sleep-wake state using the Kruskal-Wallis test, with Dunn’s post hoc individual state vs. state testing. We used Bonferroni correction for multiple comparisons within similar features based on grouping into slope-related, amplitude, duration, and area types. Corrected statistical significance was set at p< 4.167e-4 for slope-related, p< 5.56e-4 for amplitude, and p< 0.00167 for duration/area features. Table 1 shows all median value comparisons within each cell followed by the p-value from Dunn’s test.