Abstracts

Rationale:
Sleep-wake states (SWS) can greatly affect localization of the presumed epileptic source with implications for resective surgery, but this effect has not been studied in detail among all 5 canonical SWS: wakefulness, rapid eye movement sleep (REM), non-REM1 (N1), N2, N3. We systematically applied electrical source localization (ESL) to interictal epileptiform activity (“spikes”) of all 5 canonical SWS to quantitatively compare the spatial extent, spatial overlap agreement (“concordance”), and spatial non-overlap disagreement (“discordance”) of epileptic source localizations between SWS.
Method:
50 epilepsy monitoring unit patients were recruited prospectively and sequentially at the University of Manitoba. We analyzed 1132 spikes from 295 day-night recordings. Spikes underwent ESL by standardized low resolution electromagnetic tomography analysis to yield probabilistic source space of spikes arising in each SWS for each patient. ESL results were mapped to cortical grey matter (CGM) in individualized 3-D brain models using recent MRI when possible (n=40). Results reported as mean (range). Statistical significance for one-way ANOVA and linear regression was p< 0.05 with adjustment for multiple comparisons where appropriate.
Results:
Inter-individual source space sizes yielded by 5 SWS did not differ significantly between SWS (p=0.929); but intra-individual source spaces significantly differed (p=0.0089) with REM producing greatest differences. On a per-person basis out of all SWS, REM’s source space was most often largest (38%), and most often largest-or-smallest (64.9%). The 5 SWS convergently overlapped (concurred) to a shared source space of 20.3% CGM, with each SWS contribution not significantly different (p=0.982). 18.7% (0-81.5%) of the REM source space reveals novel source localizations untouched by any other SWS, uniquely activating 9.3% (0-45.6%) of a given patient’s CGM, far more than any other SWS (p=0.0292). Amount of unique localization in REM correlates positively to REM’s source space overall size (r=0.55, p=0.0011); with 3.1% unique source space gained for every 10% enlargement (% CGM).
Conclusion:
We show that REM consistently reveals new epileptic source activations not given by any other canonical SWS. REM’s novelty does not diminish its contribution to the 5-way shared concordant space; rather, REM reveals new activations by enlarging the source space to reach beyond other SWS and into truly unique grey matter. REM also shows bipolarizing tendency to enlarge or shrink the source space, which may relate to its high degree of cortical asynchrony wherein destructive interference may temper spike propagation, and constructive interference may propagate spikes into new regions by drawing on unique cortico-cortical connections known to occur in REM. Overall, one may conceive of a dynamic epileptic source space changing over SWS as a malleable “source cloud”, with REM having the most impact.
Funding:
:Health Sciences Centre Foundation (Winnipeg), University of Manitoba Rady Faculty of Health Sciences & Department of Internal Medicine, Mathematics of Information Technology and Complex Systems