Polysomnographic (PSG) recordings are the gold-standard of sleep research in humans. Yet, PSG recordings are not always analysed to their full potential. This is particularly striking in the case of insomnia, a disorder for which PSG recordings are not necessary to establish a diagnosis. I will show here how PSG recordings carry very rich and reliable information about one's sleep, in particular in the case of insomnia. In fact, the paradox of SSM (presence of subjective symptoms of insomnia without objective impairments of sleep) is apparent only when focusing on superficial, large-scale metrics of sleep but dissolves when examining the finer dynamics of brain activity. The limitations outlined in the past and leading to the exclusion of PSG recordings from the diagnosis of insomnia are no longer warranted and clinicians should revisit the benefits of PSG recordings. This reassessment is timely in the light of two ongoing revolutions in the field of Sleep Medicine: (1) the emergence of consumer-based PSG devices will bring PSG recordings to the masses, (2) the translation of computational tools from the field of Artificial Intelligence to Sleep Medicine allows the rapid, automated and massive analysis of large datasets. I will illustrate the advantages of these new methods in the field of Sleep Research by showing how we can even move beyond the current classification of sleep stages using unsupervised clustering.
Google meet: https://meet.google.com/uaw-keoy-xbj