Sleep architecture and subjective sleep complaints are also affected by aging 22, 23. Deep learning methods that utilize all relevant information in PSGs may provide additional useful clinical insights such as important health outcomes.Īge is one of the strongest predictors of morbidity and mortality. Further, they merely imitate human scoring without attempting to capture all the rich incipient information contained in a full night PSG study discussed above. However, these new methods have mostly been confined to replicating a scoring practice that is limited by arbitrary definitions 1 that may not capture all relevant information available in the data. These algorithms provide added information such as higher resolution sleep stages and probabilistic measures, in contrast to manual scoring that only offers categorical classification. Recently, promising deep learning methods have been developed that efficiently and objectively assist PSG analyses 19, 20, 21. Finally, specific abnormalities such as REM sleep behavior disorder (RBD) and loss of sleep-stage specific autonomic regulation during sleep are well established early precursors of synucleinopathies 16, 17, 18. Similarly, decreased slow-wave sleep and low SE have been associated with hypertension incidence and a variety of cardiovascular outcomes among participants in the Sleep Heart Health Study (SHHS) 14, 15. All-cause mortality has been associated with an increase in arousal burden 11 (a measure of sleep fragmentation), decreased sleep efficiency (SE) 12 and decreased rapid eye movement (REM) sleep amounts 13. Sleep apnea has also been shown to be associated with increased mortality risk independent of obesity, age, and sex 10.Īlthough sleep apnea measures are currently the main rationale for conducting clinical sleep studies, there is evidence that other aspects of objective sleep influence mortality and health outcomes. Of particular clinical importance are measures of sleep disordered breathing events such as the AHI or associated hypoxic burden, which has been associated with daytime sleepiness 3, cognitive impairment, and increased risk of cardiovascular disease such as development of high blood pressure and stroke in multiple studies independent of age, sex and obesity 4, 5, 6, 7, 8, 9. This scoring is time-consuming and prone to inter- and intra-rater variability 2. Scoring is done manually by trained technicians and supervised by medical doctors, according to American Academy of Sleep Medicine (AASM) guidelines 1. Sadly, the millions of PSGs collected every year are primarily used clinically to visually extract simple metrics such as sleep latency, proportion of time in various sleep stages, rates of sleep apnea events (apnea-hypopnea index, AHI), periodic leg movement (PLM), and arousals (arousal index, ArI). It thus contains a wealth of information on the normal physiology of a given individual (notably brain physiology). The PSG provides recording of multiple physiological measures during sleep, at a time when the individual is mostly immobile and uncontaminated by sensory inputs. The gold standard diagnostic test for this evaluation is nocturnal polysomnography (PSG), a test comprised of multiple physiological signals, i.e., electroencephalogram (EEG), electrocardiogram (ECG), electrooculogram (EOG), chin and leg electromyogram (EMG), breathing effort and airflow, all of which are recorded overnight. Sleep clinics throughout the world evaluate millions of patients every year.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |