Abstracts

Rationale: Dravet Syndrome (DS) is an epileptic encephalopathy caused primarily (~80-90% of cases) by haploinsufficiency of the SCN1A gene. This SCN1A gene deficiency disproportionately affects GABAergic inhibitory interneurons, of which fast-spiking parvalbumin-positive (PV+) and somatostatin-positive (Sst+) interneurons, rely on proper SCN1A function for the generation-transmission of action potentials and long-term potentiation, a key cellular mechanism underlying learning and memory. We hypothesized that the resultant deficiency in interneuron signaling should reflect in network oscillations recorded with scalp EEG. Published work using metrics of Spindle activity (distinctive ‘bursts’ of ~9-16 Hz activity lasting ~.5-2 seconds during stage 2 NREM sleep) in DS are lacking. We explored spindle metrics as potential biomarkers for drug target engagement for future therapeutic intervention, particularly those targeted at improving intellectual function and sensory information integration.


Methods: EEG data of patients with DS (N = 15) and age-matched neuro-typical control participants (N = 20), were analyzed. Stage 2 NREM sleep were extracted and analyzed for spindle density (per minute) and spindle duration. EEG were band-pass filtered in the sigma range from 9-16 Hz and separated into slow (9-12.5 Hz) and fast (12.5-16 Hz) spindle phenomena and assessed for spindle density and duration. These metrics were further analyzed for age-specific correlations along the continuum of pediatric development from 0-12 years.

Results: During stage 2 NREM sleep, in frontal and central locations, participants with DS had significantly fewer slow spindles per minute (i.e. less density), and significantly shorter slow spindle duration, than controls. Further, only frontal locations showed a strong positive correlation of increasing spindle density with age (ρ=.78, p< .000). Interestingly, fast spindle density in central locations showed the opposite trend, with DS producing more significantly fast spindles per minute then controls, throughout N2 sleep.


Conclusions: Lower slow spindle density and shorter spindle duration in DS during N2 sleep may reflect deficits in inhibitory signaling among key networks essential for synchronizing and maintaining network firing, giving rise to rhythmic spindles (and allowing for full information processing during sleep). A full manifestation of complete spindle activity throughout the night is tied to optimal higher-order information integration of events occurring throughout the day. Anything less than optimal spindle activity is tied to intellectual/psychiatric disability, among others. Abnormalities in fast spindles in central scalp EEG, may uniquely reflect a lack of inhibition possibly stemming from the inhibitory-interneuron-rich thalamic reticular nucleus (TRN), as these types of spindles may more heavily rely on signaling from the TRN to produce them. We therefore proposed that sleep spindle metrics may be useful biomarkers for interventions targeted at improving intellectual function (slow spindles), and motor or sensory integration (fast spindles), as they may each rely on shared, and unique, mechanisms.


Funding: This work was funded by Encoded Therapeutics.