Diazepam induced sleep spindle increase correlates with cognitive recovery in a child with epileptic encephalopathy

Abstract

Background: Continuous spike and wave of sleep with encephalopathy (CSWS) is a rare and severe developmental electroclinical epileptic encephalopathy characterized by seizures, abundant sleep activated interictal epileptiform discharges, and cognitive regression or deceleration of expected cognitive growth. The cause of the cognitive symptoms is unknown, and efforts to link epileptiform activity to cognitive function have been unrevealing. Converging lines of evidence implicate thalamocortical circuits in these disorders. Sleep spindles are generated and propagated by the same thalamocortical circuits that can generate spikes and, in healthy sleep, support memory consolidation. As such, sleep spindle deficits may provide a physiologically relevant mechanistic biomarker for cognitive dysfunction in epileptic encephalopathies.

Case presentation: We describe the longitudinal course of a child with CSWS with initial cognitive regression followed by dramatic cognitive improvement after treatment. Using validated automated detection algorithms, we analyzed electroencephalograms for epileptiform discharges and sleep spindles alongside contemporaneous neuropsychological evaluations over the course of the patient's disease. We found that sleep spindles increased dramatically with high-dose diazepam treatment, corresponding with marked improvements in cognitive performance. We also found that the sleep spindle rate was anticorrelated to spike rate, consistent with a competitively shared underlying thalamocortical circuitry.

Conclusions: Epileptic encephalopathies are challenging electroclinical syndromes characterized by combined seizures and a deceleration or regression in cognitive skills over childhood. This report identifies thalamocortical circuit dysfunction in a case of epileptic encephalopathy and motivates future investigations of sleep spindles as a biomarker of cognitive function and a potential therapeutic target in this challenging disease.

Keywords: CSWS; Case report; Continuous spike and wave of sleep with encephalopathy; Epileptic encephalopathy; High-dose diazepam.

Fig. 1

Longitudinal EEGs, neuropsychological (NP) evaluations, and medication changes over 6 years of follow-up. Neuroactive medications are indicated in green circles. LEV:Levetiracetam; CLO:Clobazam; LTG:Lamotrigine; DZP: Diazepam, DEX: Dexmethylphenidate

Fig. 2

Longitudinal IIS rate during N2 sleep. Topographical map of the spike rate across ages shows a dramatic decrease on the night of high-dose diazepam treatment that was sustained with treatment. Spikes were detected with Persyst and were restricted to HD-EEG recordings

Fig. 3

Course of neuropsychological testing scores. The thick black line is the mean across all cognitive scores shown. A consistent decline across cognitive domains is apparent from age 6.4 years to age 9.6 years after which improvements were observed following high-dose diazepam treatment

Fig. 4

Example spindle detections at age 9.6 years pre- and post- high dose diazepam treatment. Data is referenced the common average and spindle detections at each channel are highlighted. For detection, spindles are required to have regular oscillations that stand out from the background activity in the sigma band (9–15 Hz), not have an increase in theta activity (4-8 Hz) which can occur with epileptiform spikes, and last at least 0.5 s in duration [27]

Fig. 5

Longitudinal spindle rate and duration during N2 sleep. Topographical map of the a spindle rate and b duration across ages shows a dramatic increase on the night of high-dose diazepam treatment (red arrow) that is sustained. A transient increase in spindles at ages 6.2–7.0 years coincides with the start of dexmethylphenidate. c Spindle rate and d Spindle duration for 56 control children using the same detector [32] Open black circles indicate individual children, horizontal red line indicates the median value. The case values are shown in blue (pre-treatment) and orange (post treatment). Here, the higher orange value corresponds to the night of high-dose valium treatment (1 mg/kg) and the lower value to a night several months later on a decreased dose (0.2 mg/kg)

Fig. 6

Power Spectral Density curves before and after treatment with high dose diazepam in wake (a) and N2 sleep (b). During post-treatment sleep there is a discrete bump in the sleep spindle range (10-15 Hz) that is much smaller in pre-treatment sleep. (Inset) magnification of the sleep spindle frequency band. Both post-treatment wake and sleep EEGs show the expected increase in beta band frequencies (~ 20-25 Hz) after treatment with a benzodiazepine

Fig. 7

Relationship between detected spike rate and spindle rate. Data obtained using HDEEG recordings are shown here for consistency