The Prospects for Enhancing Sleep Across the Lifespan

IT IS INCREASINGLY RECOGNIZED THAT SLEEP VARIABLES ARE TRAIT-LIKE.¹ A RECENT STUDY OF HEALTHY YOUNG MALE TWIN PAIRS, FOUND THAT spectral sleep variables are among the most heritable traits in humans, constituting a signature so stable and individual they likened it to a “fingerprint.”² A number of studies have corroborated that sleep variables are trait-like in healthy young adults, but the degree to which many phenotypes are trait-like, or heritable, differs across the lifespan. Yet to date there have been no investigations of the heritability of objectively measured sleep traits in relation to the development of important neurobehavioral functions. The study by Geiger and colleagues³ in this issue of SLEEP is the first study to do so, investigating trait-like features of the sleep EEG power spectrum in children, and relating these to aspects of intelligence.

The study involved two different recordings, which allowed the authors to demonstrate that the inter-individual differences of the spectral sleep EEG are reliable and correlated with IQ. The authors further make the claim that they observed trait-like characteristics of sleep. Yet, replicability is necessary but not sufficient for establishing variables as traits. That is, establishing reliability of these measures in children is an important first step, but to fully establish these individual differences as “trait-like,” robustness to experimental manipulation and assessment of magnitude are also required. A relevant discussion of this topic can be found in Tucker et al.¹

A further challenge to establishing trait differences is the small sample size studied by Geiger et al. (i.e., N = 14 subjects 9–12 years of age), which makes it difficult to isolate the trait-like (between-subject) variance from age and developmental sources of between-subject variance. An even greater problem is that stable environmental sources of variance between children were not controlled. This is important because although spectral sleep variables, such as sigma power, are highly trait-like, they additionally show significant changes in response to environmental differences.¹ For example, experimental manipulation of learning experience significantly increases spindle activity during the night.⁴ Thus, in the children studied by Geiger et al., some of the differences in sleep variables may reflect differential exposure to learning opportunities within their respective environments (i.e., stable environmental differences, as opposed to trait variance).

By housing participants in the laboratory for the day before the night in which sleep was recorded, this environmental noise could be reduced. Further, an experimental design involving repeated measurements before and after multiple exposures to manipulated learning experience would have allowed for the simultaneous assessment of all three requirements for establishing sleep EEG spectra as a trait (i.e., robustness, replicability, and magnitude).¹ Since the study by Geiger et al.³ only assessed one of these factors, and involved limited control over environmental variance, the full criteria for trait-like variance in the children's sleep EEG spectra was not established.

Additionally, it is important but often forgotten that even rigorous experimental designs only establish the degree to which variance is trait-like for the given sample and the segment of the population it represents.⁵ That is, trait-like variance differs across segments of a given population. As one concrete example, trait differences in IQ have been shown to differ as a function of SES.^6,7 Specifically, as SES increases, children are more likely to receive the resources necessary to achieve their full genetic intellectual potential; accordingly, at high SES levels it has been estimated that 60% of the variance in the IQ of high SES children is due to trait factors and almost none to environmental factors; in low SES children this appears to be reversed.⁷ Thus, any statements of the degree to which variance is trait-like needs to be couched in terms of the specific segments of the population studied.

The above concerns notwithstanding, the authors present the tantalizing finding that sleep EEG spectral variables may be trait-like in children and further may serve as a marker for intellectual ability.³ This study paves the way for additional investigations of the degree to which objectively measured sleep variables are fixed as opposed to fluid at different stages of development, and how these fixed and fluid sources of variance in sleep measures relate to the fixed and fluid sources of variance in cognitive functions, such as I.Q. The degree to which sleep is susceptible to environmental influences has implications for the potential efficacy of interventions designed to enhance aspects of sleep to support learning and intelligence across the lifespan.

DISCLOSURE STATEMENT

Dr. Tucker has indicated no financial conflicts of interest.