Sleep EEG slow-wave activity in medicated and unmedicated children and adolescents with attention-deficit/hyperactivity disorder

0
68

The main result of this study is that sleep EEG SWA is reduced across the scalp in children and adolescents with ADHD compared to healthy controls. Interestingly, we found that the SWA reduction was mostly driven by patients without stimulant medication at the time of their participation in the study. Both of these findings are in agreement with the well-established anatomical differences in ADHD patients. Specifically, several studies found that children with ADHD exhibit widespread reductions in gray matter volume and cortical thickness3,4,27,28. Moreover, psychostimulant medication has been shown to normalize both gray matter3,18 and cortical thickness29. This parallelism between sleep SWA and cortical gray matter might not be surprising given that several studies have established a direct relationship between the two10,30.

Our knowledge about how slow waves are generated supports a close relationship between SWA and gray matter. Slow waves on the cortical surface reflect slow oscillations of cortical neurons31. These slow oscillations of cortical neurons during deep sleep are the result of an unique alternating activity pattern between active “on” states and completely silent “off” states. The more neurons are involved and the more synchronized they display these “on”/“off” states, the larger the amplitude of the surface slow wave32. Thus, higher synaptic connectivity and stronger synaptic connections result in larger and highly synchronized networks of neurons involved in slow oscillations and in more SWA recorded on the cortical surface.

This relationship might explain why particularly during development, when major reorganization processes take place, sleep SWA parallels major maturational changes in cortical gray matter, i.e. (1) the inverted U-shape time course during the first two decades of life10,12,14,33, (2) the regional maturation along the posterior-anterior axis13,15,34,35, and (3) experience-dependent changes in synaptic plasticity, which are favored by processes of brain maturation36. These close relationships seem not to be exclusive for gray matter, but specific aspects of slow waves may also relate to the maturation of white matter microstructure37.

Sleep EEG SWA represents a promising readout of age-dependent changes in functional neuronal connectivity and may thus complement the anatomical markers of brain maturation36. As such, this electrophysiological readout potentially contributes to the understanding of why ADHD patients show reduced cortical gray matter for the following reasons. It was proposed that a delay in cortical maturation underlies the brain anatomical differences7. Indeed, in our previous work published in 2013 by Ringli and coworkers, we concluded, based on the topographical differences in SWA, that children with ADHD show a less mature pattern than healthy control participants. Our novel analysis with a larger study population, allowing a comparison of absolute SWA, reveals a global reduction of SWA in patients with ADHD. This observation does not exclude that additional local changes are present. The directionality of the relationship between sleep SWA and a maturational delay is unknown. In other words, whether decreased SWA supports a cortical developmental delay, or alternatively whether a maturational delay results in decreased SWA remains to be established with longitudinal studies. Two observations, which both support an active role of SWA in cortical gray matter maturation, encourage the exploration in more detail: (1) cortical maturation assessed by SWA precedes gray matter maturation measured by means of MRI38 and (2) good evidence exists that SWA plays a critical role in synaptic plasticity39. Thus, reduced SWA in ADHD patients may actively contribute to the delay of brain development.

Finally, we would like to discuss major confounders and potential limitations of the study.

First, the attenuated effect on SWA in the medicated group may be caused by a withdrawal effect rather than by medication itself, because participants were asked to refrain from taking medication on the day of EEG measurement. This, however, is unlikely because patients who refrained from taking their medication on the day of measurement did not significantly differ from patients who continued taking psychostimulants on the day of measurement. On the other hand, psychostimulants may have an acute effect on SWA, since only patients who continued taking their medication on the day of EEG assessment, but not those who withdrew from it, differed significantly from unmedicated patients. Additionally, in agreement with previous studies40,41, we observed an acute medication effect on sleep onset latency (Supplementary Table S2), supposedly due to the arousal-enhancing effects of psychostimulants. Thus, based on our data we cannot conclude if the main effect of stimulant medication is mediated by a long-term or an immediate effect.

Second, symptom severity could mediate some of the group differences we observe. Unfortunately, across the extended time window in which measurements were taking place, no standard procedures for the assessment of symptom severity was conducted. Consistent across all patients is that they had a diagnosed ADHD and fulfilled the same exclusion and inclusion criteria. However, our data do not allow to explore the relationship between present findings and symptom severity. Furthermore, the proportion of children recruited by one or the other institution varies across medication subgroups. While the majority of unmedicated patients was recruited by one institution, most medicated patients who refrained from medication intake on the day of measurement were recruited by the other (Table 1).

Third, the level of SWA depends on experiences and activities performed during the day36,42,43. Therefore, it is possible that alterations in SWA are not only a result of different brain anatomy between children with and without ADHD, but also of differing daytime activities. To account for differences in time awake prior to the EEG assessment and potential effects on SWA26, we incorporated the evaluation of sleep diaries, which showed that sleep–wake history did not differ between ADHD and control participants. It is thus very unlikely that the decrease in SWA was caused by lower sleep pressure in ADHD patients.

Fourth, one might speculate that global SWA changes in ADHD patients result from disturbed sleep in this patient group. By looking at the architecture of the first NREM sleep hour we could show that durations of wake and sleep stage N1 did not differ between ADHD and control subjects. A more direct measure for sleep fragmentation is the number of awakenings. Thus, we compared awakening counts during the first NREM sleep hour of ADHD patients (n = 50) with those of healthy controls (n = 86) and found that ADHD patients tended to wake up even fewer times than healthy controls (p = 0.04). When performing a medication-subgroup analysis, we counted significantly more awakenings for the ADHD group “med in past” as compared to the ADHD group “med day before” (p = 0.04, Tukey-Kramer post hoc test), which is consistent with different durations of wake and N1 in the same subgroups (see above). Thus, apart from the “ADHD-med in past” group, decreased SWA in ADHD patients cannot be explained by more sleep fragmentation.

Fifth, undetected sleep problems in the ADHD group might have influenced our results. Sleep disorders, such as obstructive sleep apnea and periodic limb movements, were shown to affect a high percentage of ADHD patients44. In the present study, we excluded patients who reported to suffer from sleep problems and/or a diagnosed sleep disorder, but did not perform a polysomnography allowing a clinical assessment of the participant’s sleep. It is, however, unlikely that the ADHD group differed from the control group in terms of sleep problems, since we did not detect differences in the amount of awakenings, the length of wake after sleep onset, nor in sleep–wake history.

Finally, also power in other frequency ranges may show differences between ADHD and control participants. Relatedly, recent studies also reported lower EEG power in the sigma-frequency range (12–16 Hz) in children with ADHD45,46, however, without focusing on topographical differences. We limited our hypothesis driven approach to the SWA frequency range to avoid a reduction in statistical power because of multiple comparisons.

Having these limitations in mind, our results show that children and adolescents diagnosed with ADHD experience a reduction in EEG SWA during NREM sleep. Stimulant medication seems to have normalizing effects on SWA. These findings are in line with neuroimaging studies showing decreased levels of gray matter, possibly caused by a maturational delay in ADHD patients. This parallelism between ADHD-related gray matter changes and SWA changes needs to be further investigated by assessing both MRI and sleep EEG in a large patient sample.

Source link

LEAVE A REPLY

Please enter your comment!
Please enter your name here