Redefining Insomnia: A Comparison of Categorical and Dimensional Approaches

Project Summary Insomnia is the most common sleep disorder and negatively impacts daily functioning, quality of life, and work productivity. Despite the high prevalence of insomnia, particularly in women, and >40 years of extensive research efforts, the field still lacks a solid understanding of how insomnia should be assessed and defined. For example, polysomnography (PSG) has failed to yield a consistent pattern of objective impairment and, in fact, often fails to distinguish individuals with insomnia from good sleepers. Clinical diagnosis of Insomnia Disorder is therefore based solely on subjective report of difficulty initiating or maintaining sleep (DIMS) associated with perceived distress or impairment. Earlier diagnostic editions differentiated multiple subtypes based on patterns of clinical symptoms that failed to demonstrate reliability and validity and have since been abandoned. Insomnia is now operationalized as a monolithic subjective entity despite the likelihood that it is instead a heterogeneous disorder with multifactorial etiology and presentation with complexity akin to depression. Heterogeneity significantly reduces the likelihood of understanding the pathophysiology of insomnia or realizing a goal of personalized insomnia treatment. We propose that the time is ripe for a reconceptualization that combines knowledge about subjective and objective DIMS with mechanistic factors known to be associated with insomnia: hyperarousal and circadian misalignment (i.e. sleeping out of phase with endogenous rhythms). To that end, this proposal seeks to revisit the operationalization of insomnia to derive a classification system with high reliability and validity. We will recruit n=400 individuals with subjective DIMS, defined as sleep onset latency, wakefulness after sleep onset, and/or early morning awakenings >30 minutes. Following home sleep apnea screening, in depth phenotyping will assess four constructs of insomnia: 1) subjective DIMS based on sleep diaries and retrospective self-report; 2) objective DIMS with actigraphy and home PSG; 3) physiologic (heart rate variability), cognitive (self-report) and cortical (beta EEG) hyperarousal; and 4) circadian misalignment defined as the phase angle between dim light melatonin onset and midpoint of sleep along with subjective and actigraphic measures of rhythmicity. Structural equation modeling and latent profile analysis will be used to both derive latent factors underlying each dimension (dimensional model) and to categorize individuals into distinct subgroups based on these inputs (categorical model). Reliability will be examined in a subset (n=200) by conducting a second assessment one month later to assess for temporal stability of categories and dimensions. Concurrent validity of each dimension and category will be assessed by examining associations with validators of subjective distress, mental health symptoms, quality of life, neurocognitive performance and peripheral inflammation.