Authors :
Presenting Author: Jay Pathmanathan, MD, PhD – Beacon Biosignals
Franz Furbass, PhD – Beacon Biosignals
Anna Marszalek, BA – Beacon Biosignals
Sherry Santarina, RN – Beacon Biosignals
Michelle Fogerson, PhD – Beacon Biosignals
Tanja Hellier, PhD – UCB
Jan Vandenneucker, MSc – UCB
Matthew Alkaitis, MD, PhD – UCB
Jacob Donoghue, MD, PhD – Beacon Biosignals
Rationale:
Sleep dysfunction is a common comorbidity in Lennox-Gastaut syndrome (LGS) and is closely tied to epileptic abnormalities. Sleep abnormalities have been suggested to contribute to seizures and cognitive impairment but are poorly characterized in this population. There is no longitudinal data on brain-derived sleep staging in LGS patients in the home setting. Here we report preliminary findings from the HEADFIRST study.
Methods:
The HEADFIRST study examined limited montage night-time EEG as recorded on the Dreem 3S EEG Headband. The Dreem 3S records frontotemporal and occipital EEG and is coupled to a machine learning sleep classifier that provides automated sleep staging. The system is FDA cleared to assess disturbed nighttime sleep in adults and is operable without technologist oversight. The Dreem headband was deployed to assess sleep differences between participants with LGS and healthy siblings (controls). Participants were asked to attempt recordings for 10 total nights over 2 weeks. At the time of this report, 10 participants with LGS and 9 neurotypical sibling controls had completed the study. Welch’s t-test was used to assess mean differences between groups for selected sleep metrics.Results:
For the 19 participants, 149 total nights were recorded and 116 (55 LGS and 61 control nights) were deemed high enough quality for analyses. All but one participant (with LGS) had ≥3 nights of good quality recording spanning ≥4 hours. Sleep disturbances and pathological sleep organization were universally present in participants with LGS. Normal sleep cycling patterns were markedly disrupted, with control and LGS participants having 4.6±0.3 and 2.6±1.5 sleep cycles, respectively (P< 0.05, Figure 1). While total sleep time and sleep onset latency were similar between groups, wake after sleep onset (WASO) and total REM duration were significantly reduced in participants with LGS vs controls (WASO, 52±24 vs 21±10 minutes, P< 0.05; total REM, 57±56 vs 104±15 minutes, P< 0.05, for LGS and control, respectively). Within subject variability was greater for LGS participants than controls (total sleep time