واکنش پذیری به محرک های عاطفی روانی فیزیولوژیایی مربوط به خواب در بی خوابی اولیه

Abstract The present study examined psychophysiological reactivity to emotional stimuli related and non-related to sleep in people with primary insomnia (PPI) and in good sleepers (GS). Twenty-one PPI and 18 GS were presented with five blocks of neutral, negative, positive, sleep-related negative and sleep-related positive pictures. During the presentation of the pictures, facial electromyography (EMG) of the corrugator and the zygomatic muscles, heart rate (HR) and cardiac vagal tone (CVT) were recorded. Subjective ratings of the stimuli were also collected. We found that only PPI exhibited greater inhibition of the corrugator activity in response to sleep-related positive stimuli compared to the other blocks of stimuli. Furthermore, PPI rated the sleep-related negative stimuli as more unpleasant and arousing and showed higher CVT in response to all stimuli as compared to GS. Results were interpreted as indicating that PPI exhibit craving for sleep-related positive stimuli, and also hyper-arousability in response to sleep-related negative stimuli, as compared to GS. Our results suggest that psychological treatment of insomnia could benefit by the inclusion of strategies dealing with emotional processes linked with sleep processes.

Introduction Insomnia as a disorder is defined as difficulties in initiating/maintaining sleep and/or non-restorative sleep accompanied by decreased daytime functioning persisting for at least four weeks (American Academy of Sleep Medicine, AASM, 2005). Aetiological theories (Espie, 2002, Espie et al., 2006, Harvey, 2002, Lundh and Broman, 2000, Morin, 1993, Perlis et al., 1997 and Riemann et al., 2010) consider heightened levels of autonomic, cortical, cognitive, and emotional arousal as relevant maintaining factors of insomnia. Although widely recognized as important (e.g. Espie, 2002, Lundh and Broman, 2000, Morin, 1993 and Riemann et al., 2010), investigations of the aetiological role of emotional arousal in insomnia have been classically based on indirect evidence. These have involved studies showing that neurotic temperamental traits are associated with insomnia (e.g. Espie, 1991 and LeBlanc et al., 2007), studies investigating the content of intrusive thoughts experienced by people with insomnia during the pre-sleep period (e.g. Van Egeren et al., 1983 and Wicklow and Espie, 2000), and studies demonstrating significant comorbidity with clinical anxiety or depression (e.g. Chang et al., 1997, Ford and Kamerow, 1989 and Riemann, 2007). Considering the dimensional approach to the study of emotions (e.g. Bradley, 2000 and Lang, 2002, chap. 14), the emotional experience is described as referring to two main dimensions: the valence (positive vs negative) and the arousal (from low to high). Based on this, the majority of research considering insomnia has largely focussed on the dimension of arousal, but not on the dimension of valence. A small number of studies have investigated the relationship between positive (e.g. excited, enthusiastic) and negative (e.g. hostile, upset) affective states and poor sleep quality using self-report questionnaires (e.g. McCrae et al., 2008, Norlander et al., 2005 and Scott and Judge, 2006). These studies showed that people with insomnia report heightened negative and diminished positive emotional states as compared to good sleepers. Three studies have examined the relationship between emotions and sleep measuring psychophysiological responses to visual stimuli validated both for the dimensions of arousal and valence. In two studies, Franzen, Buysse, Dahl, Thompson, and Siegle (2009) and Franzen, Siegle, and Buysse (2008) measured pupillary responses to high arousing positive and negative emotional pictures and low arousing neutral pictures in sleep-deprived participants as compared to control groups. Main results showed that responses to negative pictures were larger in sleep-deprived participants as compared to controls. In a functional magnetic resonance imaging (fMRI) design study of sleep deprivation, Yoo, Gujarm, Hu, Jolesz, and Walker (2007) presented pictures ranging from emotionally neutral (neutral valence, low arousal) to increasingly aversive quality (negative valence, high arousal). The sleep deprivation group displayed enhanced activity in the amygdala and reduced functional connectivity between the amygdala and the medial prefrontal cortex (MPFC). These results were interpreted as an increased neurobiological response to emotional stimuli and a reduced inhibitory influence of the MPFC on emotional reactivity after sleep deprivation. Taken together, these studies show that sleep deprivation alters emotional responses to negative stimuli, thus, sleep seems to be important for maintaining adaptive emotional processes. However, as these studies focussed on sleep deprivation, no data about emotional reactivity in insomnia are available up to date. In fact, insomnia is a condition which differs from sleep deprivation because it implies chronic sleep difficulties (more than four weeks), primarily subjective complaints, impairments in sleep quality and not necessarily in sleep quantity, and adaptation processes. Moreover, the studies aforementioned did not specifically considered the valence dimension. In fact, Yoo et al. (2007) did not use positive stimuli, and Franzen et al., 2009 and Franzen et al., 2008 used pupillary responses which are indices of more general emotional information processing. Bradley (2000) has indicated that facial electromyography (EMG) measures specifically the emotional valence. The activity of the corrugator muscle (knits the eye brow into a frown) is both potentiated by unpleasant pictures and inhibited by pleasant pictures when compared to neutral stimuli. In contrast, the activity of the zygomatic muscle (pulls the corners of the mouth back and up into a smile) increases in response to pleasant pictures, while unpleasant stimuli elicit no inhibition (e.g. Larsen, Norris, & Cacioppo, 2003). The aim of the present study was to evaluate facial EMG responses to positive and negative emotional stimuli related and not related to sleep in people with primary insomnia and good sleepers. We assumed that the presentation of symptom-relevant emotional stimuli would enhance responses in the group with insomnia due to a greater attention directed to this material and due to personal relevance. The following predictions were considered. As compared to neutral stimuli and as compared to good sleepers, people with primary insomnia should exhibit enhanced corrugator activity in response to sleep-related negative stimuli. Additionally, they may show diminished corrugator activity and heightened zygomatic activity in response to sleep-related positive stimuli. With respect to stimuli not related to sleep, people with primary insomnia should present enhanced responses in the corrugator activity to negative stimuli. Additionally, two alternative predictions were considered with respect to positive stimuli. Diminished responses to positive stimuli in the zygomatic muscle have been reported for individuals with dysphoria compared to healthy controls (Sloan, Bradley, Dimoulas, & Lang, 2002). As insomnia is a condition highly linked with depression and with a waking complaint of negative mood changes, we may see a reduced response in the zygomatic muscle to positive stimuli in individuals with primary insomnia. Alternatively, this reduced response may be unique to depression as primary insomnia is an independent diagnostic entity from mood disorders (e.g. Perlis et al., 2006 and Riemann and Voderholzer, 2003). In addition to these hypotheses, we also expected that people with primary insomnia would respond with higher autonomic arousal to all stimuli. Thus, numerous studies have demonstrated increased levels of arousal in people with primary insomnia when compared to good sleepers (for a review see Riemann et al., 2010). Additionally, Lundh and Broman (2000) suggested that people with insomnia are characterized by an increase in the levels of arousal in response to sudden, new or emotional stimuli (hyper-arousability). In the current experiment, we also recorded heart rate (HR) and cardiac vagal tone (CVT) during the presentation of the visual stimuli. As an increase of the heart rate is linked to heightened arousal stimulation ( Witvliet & Vrana, 1995) and higher CVT has been associated with greater reactivity to stimuli ( Movius & Allen, 2005), we predicted alterations in these measures. Finally, subjective ratings of the valence and the arousal of the pictorial stimuli were also collected. We predicted that people with primary insomnia would demonstrate an altered valence and arousal ratings of sleep-related stimuli as compared to good sleepers.

Results Data preparation and analyses Statistical analyses were conducted using the STATISTICA software (version 6, www.statsoft.com). 1) Description of the sample. Groups were compared using one-way ANOVAs (with respect to age, sleepiness, anxiety and severity of the insomnia symptoms) and chi-squares (with respect to GENDER). 2) Facial electromyography (EMG). The Spike-2 software, version 5 (www.ced.co.uk), was used for preparing the EMG data. Raw data were high-pass filtered for eliminating the DC signal. Data were then rectified, smoothed and transformed into logarithm to normalise the distribution. According to the “Guidelines for human electromyographic research” (Fridlund & Cacioppo, 1986), most psychophysiological research using EMG has focussed on some variations of EMG signal amplitude as the dependent measure. Commonly, reflexes are often quantified using peak-to-peak EMG amplitude and onset latency (phasic responses). However, the guidelines also underline that it has been suggested that maintained EMG activity represent muscular contraction more accurately than a simple amplitude. Research has used, thus, also the mean of the rectified, smoothed, and filtered EMG signal (tonic responses). In this study, we have considered both phasic and tonic responses to analyse the EMG responses to the stimuli. Phasic responses: Changes in the phasic responses were obtained as the difference between the peak-to-peak amplitude of the EMG recorded during the presentation of each block of pictures and the peak-to-peak amplitude of the EMG recorded during the second before the stimulus onset. Two mixed design factorial ANOVAs were computed using the GROUP as between subjects factor, the VALENCE and the SIDE OF THE FACE (left vs right) as within subjects factors, and the mean changes of activity recorded in the corrugator and the zygomatic muscles respectively as dependent variables. Tonic responses: Tonic EMG activity was obtained as the mean values recorded during each block of pictures (10 neutral, 10 negative not related to sleep, 10 negative related to sleep, 10 positive, 10 positive related to sleep). Two mixed design factorial ANOVAs were computed using the GROUP as between subjects factor, the VALENCE and the SIDE OF THE FACE as within subjects factors, and the mean activity recorded over the corrugator and the zygomatic muscles respectively during each block as dependent variables. 3) HR and CVT. Mean heart rate (HR) and mean cardiac vagal tone (CVT) were calculated for each stimulus block. With skewness and kurtosis < 1, the distributions of the means were considered normal. Data were analysed within two mixed design factorial ANOVAs using the GROUP as between subjects factor, the VALENCE as within subject factor and respectively the mean HR and CVT as dependent variables. 4) Subjective measures. Two mixed design factorial ANOVAs GROUP × VALENCE were computed using the GROUP (PPI vs GS) as between subjects factor, the VALENCE (neutral vs negative not related to sleep vs negative related to sleep vs positive not related to sleep vs positive related to sleep) as within subjects factor, and the subjective valence ratings (pleasant vs unpleasant) and arousal ratings (low vs high) respectively as dependent variables. 5) Cultural differences. As our sample was recruited half in Scotland and half in Italy, we consider possible cultural differences in the emotional responses. Descriptive analyses and the same mixed design factorial ANOVAs analyses aforementioned were repeated considering as between subjects factor the PLACE (Glasgow vs Rome). When a significant principal effect or interaction was found, specific hypothesis were tested through planned comparisons. Findings Means ± standard deviations are reported in the text in brackets. 1) Description of the sample. A summary of the descriptive results is provided in Table 1. The PPI group included a marginally significant higher number of females, reported higher scores on the STAI-T and had higher scores on the ISI compared to the GS group. There was no difference between the two groups due to the time of day of recording (X2(2,N=39) = 1.48, p = 0.48). Nine PPI and 8 GS were recorded in the morning (between 9:00 and 12:30), 9 PPI and 5 GS at lunch time (between 12:30 and 15:30) and 3 PPI and 5 GS in the afternoon (between 15:30 and 18:00). Two good sleepers and 4 people with insomnia reported scores higher than 7 on the KSS, however no difference between the groups was found, as shown in Table 1. 2) Facial electromyography (EMG). Results are summarized in Table 3. Table 2. Results related to facial EMG data. SOURCE OF VARIATION Corrugator Zygomatic df Sum of squares df error Mean square F p Observed Power df Sum of squares df error Mean square F p Observed power Phasic responses GROUP 1 5.179 17 5.179 0.520 0.48 0.11 1 34.626 21 14.876 2.327 0.14 0.142 SIDE 1 0.016 17 0.016 0.003 0.96 0.05 1 13.712 21 13.712 1.469 0.24 0.212 VALENCE 4 9.548 68 2.387 1.521 0.21 0.45 4 22.359 84 5.590 4.197 0.004 0.909 GROUP × SIDE 1 2.505 17 2.505 0.490 0.49 0.10 1 0.905 21 0.905 0.097 0.76 0.060 GROUP × VALENCE 4 5.550 68 1.387 0.884 0.48 0.27 4 4.808 84 1.202 0.903 0.76 0.275 SIDE × VALENCE 4 3.131 68 0.783 1.479 0.22 0.44 4 3.426 84 0.856 1.539 0.20 0.457 GROUP × SIDE × VALENCE 4 2.382 68 0.596 1.125 0.35 0.34 4 1.114 84 0.278 0.500 0.74 0.164 Tonic responses GROUP 1 1.551 36 1.551 0.117 0.73 0.06 1 51.143 36 51.143 3.758 0.06 0.471 SIDE 1 2.703 36 2.703 0.590 0.45 0.12 1 21.330 36 21.330 3.388 0.74 0.433 VALENCE 4 0.287 144 0.072 0.631 0.64 0.20 4 3.976 144 0.994 7.562 <0.001 0.996 GROUP × SIDE 1 0.014 36 0.014 0.003 0.96 0.05 1 0.116 36 0.116 0.018 0.89 0.052 GROUP × VALENCE 4 1.137 144 0.284 2.502 0.04 0.70 4 0.297 144 0.074 0.565 0.69 0.185 SIDE × VALENCE 4 0.272 144 0.068 1.027 0.40 0.32 4 0.009 144 0.002 0.048 0.10 0.059 GROUP × SIDE × VALENCE 4 0.376 144 0.094 1.422 0.23 0.43 4 0.355 144 0.089 1.828 0.13 0.545 Table options Table 3. HR and CVT. SOURCE OF VARIATION HR CVT df Sum of squares df error Mean square F p Observed power df Sum of squares df error Mean square F p Observed power GROUP 1 82.087 21 82.087 0.274 0.60 0.08 1 228.49 20 228.49 7.538 0.01 0.74 VALENCE 4 16.464 84 4.116 1.011 0.41 0.31 4 5.864 80 1.466 2.255 0.07 0.64 GROUP × VALENCE 4 19.265 84 4.816 1.182 0.32 0.36 4 0.608 80 0.152 0.234 0.91 0.10 Table options No significant effects were evidenced for phasic responses of the corrugator muscle. Phasic responses of the zygomatic muscle showed a significant main effect for the factor VALENCE. Planned comparisons evidenced that both groups presented heightened activity in response to positive stimuli (related and not related to sleep) compared to negative stimuli (related and not related to sleep) (positive not related to sleep: −1.17 ± 1.77; sleep-related positive: −1.61 ± 2.03; negative not related to sleep: −2.01 ± 1.49; sleep-related negative: −1.98, p = 0.03; F(1,21) = 12.17, p = 0.002). Additionally all participants presented increased zygomatic activity in response to positive stimuli related to sleep compared to negative stimuli related to sleep (F(1,21) = 7.60, p = 0.01). Finally, all participants displayed different responses to neutral stimuli (−1.88 ± 2.03) as compared with all the emotional stimuli (F(1,21) = 5.53, p = 0.03). When tonic responses of the corrugator muscle were analysed, a significant interaction GROUP × VALENCE was found. Planned comparisons showed that GS presented similar responses to all types of stimuli (neutral: 0.03 ± 1.39; positive not related to sleep: −0.09 ± 1.38; negative not related to sleep: 0.02 ± 1.36; sleep-related positive: 0.10 ± 1.37; sleep-related negative: 0.0005 ± 1.36; F(1,36) = 2.30 p = 0.14). In contrast, corrugator activity in PPI decreases significantly in response to positive stimuli related to sleep (−0.239 ± 1.325), compared to the activity produced in response to all the other types of stimuli (neutral: −0.0685 ± 1.34; positive not related to sleep: −0.099 ± 1.40; negative not related to sleep: −0.0645 ± 1.37; sleep-related negative: −0.1205 ± 1.345; F(1,36) = 5.426, p = 0.03). The interaction is shown in Fig. 1. Tonic responses for corrugator muscle: interaction GROUP × VALENCE. Fig. 1. Tonic responses for corrugator muscle: interaction GROUP × VALENCE. Figure options When tonic responses of the zygomatic muscle were analysed, a significant main effect of the VALENCE was found. Planned comparisons showed only that both groups displayed heightened activity in response to positive stimuli (related and not related to sleep) compared to negative stimuli (related and not related to sleep) (positive not related to sleep: 0.66 ± 1.42; sleep-related positive: 0.55 ± 1.48; negative not related to sleep: 0.41 ± 1.40; sleep-related negative: 0.37 ± 1.37; F(1,36) = 16.34, p < 0.001). Additionally, a marginally significant main effect of the GROUP has to be noted, showing that PPI display enhanced reactivity to all stimuli as compared to GS. 3) HR and CVT. Results from HR and CVT responses are presented in Table 4. Table 4. Subjective responses. SOURCE OF VARIATION Valence ratings Arousal ratings df Sum of squares df error Mean square F p Observed power df Sum of squares df error Mean square F p Observed power GROUP 1 7.329 37 7.329 1.962 0.17 0.28 1 139.531 37 139.531 22.251 <0.001 1.00 VALENCE 4 593.784 148 148.446 86.256 <0.001 1.00 4 143.838 148 35.960 16.716 <0.001 1.00 GROUP × VALENCE 4 22.748 148 5.687 3.305 0.01 0.83 4 19.961 148 4.990 2.320 0.06 0.66 Table options No significant effects were evidenced for HR responses. When CVT was analysed, a significant main effect of the GROUP was found. People with primary insomnia (8.12 ± 2.93) displayed increased CVT in response to all stimuli compared to GS (5.027 ± 1.315). 4) Subjective measures. The results from the subjective ratings of VALENCE and AROUSAL are presented in Table 2. Valence: Results showed a significant main effect of the VALENCE and a significant interaction GROUP × VALENCE. Planned comparisons showed that both groups rated positive stimuli (related and not related to sleep) as more pleasant as compared with negative stimuli (related and not related to sleep) (positive not related to sleep: 7.83 ± 1.35, sleep-related positive: 7.15 ± 1.48, negative stimuli: 3.13 ± 1.47, sleep-related negative: 4.44 ± 1.65; F(1,37) = 181.03, p < 0.001). Moreover, both groups rated positive stimuli related to sleep as more pleasant compared to negative stimuli related to sleep (F(1,37) = 79.37, p < 0.001). It has to be noted that neutral stimuli (6.42 ± 1.50) were differently rated by all participants as compared with all the emotional stimuli (F(1,37) = 12.17, p = 0.001). With respect to the interaction GROUP × VALENCE, planned comparison evidenced that GS rated negative stimuli not related to sleep (3.17 ± 1.42) as more unpleasant compared to negative stimuli related to sleep (5.11 ± 1.23; F(1,37) = 24.30, p < 0.001). In contrast, PPI rated negative pictures related to sleep (3.76 ± 1.73) as unpleasant as negative pictures not related to sleep (3.10 ± 1.55; F(1,37) = 3.33, p = 0.08). Arousal: The mixed design factorial ANOVA showed a main effect of the GROUP, and a main effect of the VALENCE. Additionally, it has to be noted that a marginally significant interaction GROUP × VALENCE was detected. People with primary insomnia (4.49 ± 1.93) rated all stimuli as more arousing compared to GS (2.788 ± 1.40) regardless to the valence. Planned comparisons evidenced that both groups rated negative stimuli (related and not related to sleep) as more arousing than positive stimuli (related and not related to sleep) (negative not related to sleep: 4.949 ± 1.89; sleep-related negative: 4.462 ± 2.08, positive not related to sleep: 3.359 ± 2.12, sleep-related positive: 2.564 ± 1.54; F(1,37) = 34.09, p < 0.001). Moreover, both groups rated negative stimuli related to sleep as more arousing compared to positive stimuli related to sleep (F(1,37) = 30.33, p < 0.001). Finally neutral stimuli (3.10 ± 1.99) were differently rated by both PPI and GS as compared with all the emotional stimuli (F(1,37) = 10.88, p = 0.002). With respect to the interaction GROUP × VALENCE, planned comparison showed that GS rated negative stimuli not related to sleep as more arousing as compared with negative stimuli related to sleep (F(1,37) = 4.04, p = 0.05). They also rated as equally arousing negative and positive sleep-related stimuli (F(1,37) = 3.40, p = 0.08). In contrast, PPI rated negative stimuli not related to sleep as arousing as negative stimuli related to sleep (F(1,37) = 0.01, p = 0.92). Moreover, PPI rated negative stimuli related to sleep as more arousing than positive stimuli relative to sleep (F(1,37) = 37.61, p < 0.001). 5) Cultural differences. We found no statistical significant differences between the group recruited in Rome and the group recruited in Glasgow with respect to gender (X2(1,N=39) = 0.51, p = 0.38) and with respect to age (Rome: 23.05 ± 3.14; Glasgow: 21.68 ± 2.67; F(1,37) = 2.14, p = 0.15). The results from the mixed design factorial ANOVAs applied considering the PLACE (Glasgow vs Rome) as between subjects factor evidenced the following: - Main effect of the PLACE: The sample recruited in Rome, as compared to the sample recruited in Glasgow, presented enhanced tonic EMG activity in response to all the stimuli both in the corrugator (Rome: 0.30 ± 1.33; Glasgow: −0.52 ± 0.99; F(1,37) = 5.56, p = 0.02) and in the zygomatic (Rome: 1.08 ± 1.42; Glasgow: −0.09 ± 1.02; F(1,37) = 11.78, p = 0.001) muscles. - Interaction PLACE × VALENCE: A significant interaction was detected with respect to the subjective ratings of the arousal level of the stimuli (F(4,48) = 2.60, p = 0.04). Planned comparisons showed that the sample recruited in Rome (4.10 ± 2.25), compared to the sample recruited in Glasgow (2.58 ± 1.71), rated the positive stimuli not related to sleep as more arousing (F(1,37) = 5.62, p = 0.02). No other statistically significant differences were observed.