0420 Characterization of Brain Age in Patients with Prolonged Sleep Duration. (25th May 2022)

• Record Type: Journal Article
• Authors: Crawford, Bredon; Pawar, Rahul; Sun, Haoqi; Westover, Michael; Thomas, Robert; Blattner, Margaret
• Is Part Of: Sleep. Volume 45(2022)Supplement 1
• Page Start: A187
• Page End: A188
• Publication Date: 2022-05-25
• Subjects: Sleep -- Physiological aspects -- Periodicals; Sleep disorders -- Periodicals; Sommeil -- Aspect physiologique -- Périodiques; Sommeil, Troubles du -- Périodiques; Sleep disorders; Sleep -- Physiological aspects; Sleep -- physiological aspects; Sleep Wake Disorders; Psychophysiology; Electronic journals; Periodicals; 616.8498
• DOI: 10.1093/sleep/zsac079.417
• Languages: English
• ISSNs: 0161-8105

Abstract

Abstract: Introduction: We identified 35 consecutive extended sleep studies for patients with 10 or more hours of total sleep time, and applied the BAI model to these studies. The BAI model was trained using sleep studies from relatively healthy participants. The sleep EEG features were extracted from both the spectral domain and the waveform. For each sleep stage, we extracted 96 features, and each of the features was averaged across the sleep stages. The resulting features were concatenated to form 480 features to represent the entire recording of sleep. These features are fed into a linear regression model and then adjusted to reduce age-dependent bias. Methods: We identified 35 consecutive extended sleep studies for patients with 10 or more hours of total sleep time, and applied the BAI model to these studies. The BAI model was trained using sleep studies from relatively healthy participants. The sleep EEG features were extracted from both the spectral domain and the waveform. For each sleep stage, we extracted 96 features, and each of the features was averaged across the sleep stages. The resulting features were concatenated to form 480 features to represent the entire recording of sleep. These features are fed into a linear regression model and then adjusted to reduce age-dependent bias Results: 35 extended polysomnograms were reviewed for patients undergoing evaluation for hypersomnia, with a median age of 27 years (range 16-76), and female predominance (28/35, 80%). This hypersomnia cohort forms two clusters by Gaussian mixture modeling. Patients in the first cluster exhibit a brain age that correlates well to chronological age (mean BAI = 4.4yr), while patients in the second cluster exhibit a brain age 12.4 years younger than chronological age (mean BAI = -12.4yr). The EEG spectrogram in the low BAI cluster shows high spindle band power, high delta band power, and high peak frequency of posterior dominant rhythm when compared to their age norms. Conclusion: A subset of hypersomnia patients demonstrates a combination of EEG features associated with lower chronological age. We anticipate that the findings here will lead to further insights in IH physiology and phenotypes. Support (If Any): NIH R01NS102190, R01NS102574, R01NS107291, RF1AG064312, R01AG062989, R01AG073410