The Association between Sleep Problems and Neuropsychological Deficits in Medication-naïve Children with ADHD

ABSTRACT

Background

Children with ADHD are reported to have sleep problems and neuropsychological deficits, but studies examining a potential association between the two are scarce and the use of varying methodology can complicate conclusions.

Participants

A clinical sample of 59 medication-naïve children with ADHD between the ages of 6 and 14 years (71% male).

Methods

Children underwent polysomnography and multiple sleep latency test, and parent rated sleep habits on the Children’s Sleep Habits Questionnaire. Children also completed an extensive neuropsychological battery of executive function and delay aversion tasks, and parents and teachers rated executive function behavior on the Behavior Rating Inventory of Executive Function. Linear regression analyses were conducted with each of the neuropsychological outcomes included as the outcome variable and the sleep parameters as the predictor variables.

Results

The correlations between sleep and neuropsychological outcomes were generally modest, but some sleep parameters (primarily sleep stages and sleep latencies) were associated with objectively and subjectively measured executive function and delay aversion.

Conclusions

Using objective and subjective gold standard assessment procedures this study supports a (modest) association between sleep and neuropsychological function in children with ADHD.

Disclosure statement

In the last three years, Per Hove Thomsen has received speakers fee from Medice and Shire, and Edmund Sonuga-Barke has received speaker fees from Shire & Medice, consultancy from Neurotech Solutions, and research support from QBTech. None of the remaining authors report disclosures.

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website.

Notes

¹ A post-hoc power analysis for a multiple regression model (deviation from zero) with parameters derived from the overall prediction model (with eight sleep parameters and age and gender) was conducted. The effect size was f²=1.78 (equivalent to R² = .64), alpha was .05, sample size was 59, with 10 predictors. Based on these parameters GPOWER (Erdfelder et al., 1996) estimated the power at 1.00. This would indicate that the primary analysis was adequately powered.

Additional information

Funding

This work was supported by the TRYGFoundation under Grants 7–10-0098 and 101183 and the Lundbeck Foundation under Grant R67-A6449.