Quality of sleep and P300 amplitude in primary insomniacs

The hyperarousal hypothesis of primary insomnia (e.g. [1]) yelded controversial results (e.g. [2]) which could depend on the weakness of the assumption which consider insomniac’s high arousal level as a stable individual characteristic. Since there is a wide variability in bad and good nights of sleep, a low arousal level could be present following the good nights of sleep. This view is consistent with results of a recent study that showed higher P300 amplitude in insomniacs only following a bad night [3]. The aim of the present study is to replicate previous results and to extend the finding, recording also a presleep measure of cortical arousal. Methods: seven primary insomniacs and seven healthy age matched controls took part in the study for 1 week, during which they were actigraphically monitored. At their homes, each day, before and after sleep, subjects filled in a questionnaire diary and their P300 ERP (oddball paradigm) was recorded through portable equipment (Evo-Quick, Micromed, Italy). For each subject diary and actigraphic data were used for identifying the worse (N–) and the best (N+) sleep quality night in the week. N– was defined as the night with the lowest Sleep Efficiency Index and the N+ the one with the highest SEI. The mean amplitudes of the P300 from FZ, CZ and PZ were submitted to a MANOVA considering NIGHT (N– vs. N+), GROUP (insomniacs vs. controls) and TIME of day (presleep vs. postsleep) as factors. As a control, means of sleep parameters (i.e. SOL, WASO and SEI) of the selected nights were submitted to different factorial ANOVAs with NIGHT and GROUP as factors. Results: MANOVA On P300 amplitude shows a significant GROUP X NIGHT interaction [R Rao (3.10) = 4.56; <0.0293]. Post hoc LSD Test indicates that in insomniacs P300 amplitude recorded from FZ is significantly higher (P < 0.05) compared to controls, but this difference is shown only with the occurrence of the N–, independently by the TIME of day. Control ANOVAs on sleep parameters showed significant NIGHT effects indicating that IES is significantly lower (F1.10 = 6.79; P < 0.05) and WASO is significantly higher (F1.10 = 12.12; P < 0.01) in the N– compared to the N+. Conclusion: Insomniacs showed a higher cortical arousal with respect to controls but only with the occurrence of a bad night of sleep. These results agree with the previous study and are consistent with the hyperarousal hypothesis but confirm that the level of cortical arousal is also related to the quality of sleep. Hyperarousal is not a stable trait in primary insoniacs but varies as a function of sleep quality.