Authors :
Presenting Author: Hunki Kwon, PhD – Massachusetts General Hospital/Harvard Medical School
Dhinakaran Chinappen, PhD, MBA, MEng – Massachusetts General Hospital & Harvard Medical School
Anirudh Wodeyar, PhD – Massachusetts General Hospital, Harvard Medical School
Elizabeth Kinard, BA – Massachusetts General Hospital
Skyler Goodman, BS – Massachusetts General Hospital
Wen Shi, PhD – Mass General Hospital / Harvard Medical School
Katherine Walsh, BS – Massachusetts General Hospital/Harvard Medical School
Bryan Baxter, PhD – Psychiatry, MGH
Dara Manoach, PhD – Massachusetts General Hospital/Harvard Medical School
Mark Kramer, PhD – Boston University
Catherine Chu, MD – Massachusetts General Hospital/Harvard Medical School
Rationale: Cognitive deficits can be predicted by sleep spindle rate in Rolandic epilepsy. Auditory stimulation during non-rapid eye movement (NREM) sleep enhances both slow oscillations (SO, 0.5-2 Hz) and sleep spindles (9-15 Hz) in healthy adults. We evaluated the impact of auditory stimulation on SO, sleep spindles, and memory consolidation in children with RE and age-matched controls.
Methods: We prospectively recruited 9 children with RE and 10 age-matched controls. Each subject participated in two sessions, where each session included training on a motor sequence typing (MST) task, a 90-minute nap opportunity during high-density EEG recording, and testing on the MST task after sleep
(Figure 1). One session included auditory stimulation with randomly timed 50 ms bursts of pink noise during NREM sleep and one session included no auditory stimulation (sham).
For the MST, subjects typed a 5-digit sequence for twelve 30 s trials as quickly and accurately as they could with each hand separated by a 10 min break. Sleep-dependent memory consolidation was measured as the percentage difference between performance on the last three trials before sleep and the first three trials after awakening. Offline SOs were detected at frontal channels referenced to the mastoid following bandpass filtering (0.5-4 Hz). SOs were detected if two consecutive positive-to-negative zero crossings occurred within a range of 0.5 to 2 s, and the negative peak during that time window was below -40 μV. Sleep spindles were detected at frontal channels referenced to the average across channels using a validated, automated spindle detector. SO-spindle coupling was identified if a spindle occurred within +1s of the negative peak of a SO. To evaluate the impact of auditory stimulation on SO or SO-spindle coupling in RE or controls and the relationship between sleep oscillations and sleep-dependent memory consolidation, linear mixed-effects models were used with age and visit order as covariates and a subject specific intercept to account for multiple observations per subject.
Results: In patients with RE, auditory stimulation during NREM sleep increased frontal SO rate (mean increase 10.6 SO/min, p< 0.0001) and SO-spindle coupling rate (mean increase 0.53 SO-spindle/min p=0.03;