If the PDF viewer does not appear initially, please reload the page to refresh your browser.
We analyzed the EEG signal from the C3-M1 electrode derivation of a participant in unrelated research who underwent two consecutive polysomnography (PSG) studies. The multitaper spectral analysis of sleep EEG could depict the nonstationary oscillatory structure of sleeping brain states instantaneously (Fig 1, middle column), which was not only closely paralleling the traditional hypnogram (Fig 1, upper column) but also with much richness of objective neurophysiological information, including the SO range. The relative SO power was extracted from the multitaper spectral analysis as an objective, continuous-valued quantitative metric estimating sleep depth during the night (Fig 1, lower column). The high intra-individual variability of temporal sleep evolution pattern across nights is inevitable because sleep is an active and adaptive process reflecting homeostatic sleep pressure and circadian rhythm of daily life.
To represent personal intrinsic brain activities during sleep with sturdy intra-individual stability as less affected by the nightly different temporal sleep evolution pattern, we computed the average spectral power (0.5-30 Hz) corresponding to the SO-power to obtain the SO-spectrogram (Fig2, upper column). The probability of conventional wake and NREM stages being observed showed sequential peaks from wake to N3 state. The peak of REM sleep probability was similar to the N1 stage’s peak (Fig2, middle column). The number of SO-power bins used in the average calculation was different across nights (Fig2, lower column); however, the SO-spectrogram of night 1 and night 2 showed highly similar patterns.
We validated the similarity of SO-spectrogram across nights by reconstructing the whole night’s spectral power and the SO-power using the SO-spectrogram as a template (Fig3/Table). When we compared the reconstructed spectral power with the original multitaper spectrogram, the model quite accurately predicted the actual data. However, as certain areas differ, we need further exploration to find the parameter (ex. Minimal number of SO-power bin for average) to minimize the difference.
10:00 – 11:30 AM ET
HMS DSM Annual Faculty Meeting
10:00 – 11:30 AM ET
Mary A. Carskadon, PhD Introductory Meeting with HMS DSM Trainees
12:00 – 1:15 PM ET
Division of Sleep Medicine Annual Prize Lecture by Mary A. Carskadon, PhD
1:15 – 1:30 PM ET
Awarding of 2020 Harvard Medical School Division of Sleep Medicine Prize to Mary A. Carskadon, PhD
3:00 – 4:30 PM ET
4:30 – 5:30 PM ET
6:00 – 7:00 PM ET
Evening Public Lecture by Mary A. Carskadon, PhD
“Changes in Sleep Biology Create a Perfect Storm Affecting Teen Health and Well-Being”