The relationships among metacognitive functions, sleep‐related thought‐control strategies and sleep quality: A mediation analysis

In the context of sleep disturbances, increasing evidence suggests a critical role of sleep‐related dysfunctional metacognitive activity, including metacognitive control of intrusive thoughts in the pre‐sleep period. Although the relationship between sleep‐related thought‐control strategies and poor sleep quality is well recognized, the possible contribution of general metacognitive functioning within this relation is still unclear. In this study, we performed a mediation analysis to examine the role of thought‐control strategies on the relationship between metacognitive abilities and sleep quality in individuals with different self‐reported sleep characteristics. Two‐hundred and forty‐five individuals participated in the study. Participants completed the Pittsburgh Sleep Quality Index, the Thought Control Questionnaire Insomnia‐Revised, and the Metacognition Self‐Assessment Scale to evaluate sleep quality, thought‐control strategies and metacognitive functions, respectively. The results showed that worry strategy in the pre‐sleep period mediates the relationship between metacognitive functions and sleep quality. Particularly, the ability to understand one's mental states and mastery functions could be the two metacognitive domains primarily involved in the dysfunctional metacognitive thought‐control activity responsible for reduced sleep quality. The observed effect suggests that inadequate metacognitive functioning is associated with poor sleep quality in healthy subjects via the mediation of dysfunctional worry strategy. These findings suggest the potential relevance of clinical interventions to enhance specific metacognitive abilities, with the aim to promote more functional strategies for managing cognitive and emotional processes during the pre‐sleep period.


| INTRODUCTION
Sleep has a vital role in physical and mental health across the lifespan (Luyster et al., 2012;Sella et al., 2019).At the same time, poor sleep can be a contributing factor to the initiation and worsening of health problems in general.Indeed, insufficient sleep is widely recognized as a prominent issue in modern society, with a significant impact on public health and society in terms of direct and indirect costs (Léger et al., 2013;Léger & Bayon, 2010).Sleep problems are also associated with feelings of anger, anxiety and depression, and lowered positive affect and life satisfaction (Digdon & Koble, 2011).
Individuals with poor sleep quality (Doos Ali Vand et al., 2014) have features that prompt or exacerbate an excessive intrusive sleeprelated cognitive activity, that is dysfunctional beliefs about sleep, and worry about loss of sleep and its consequences.They are also likely to use dysfunctional behaviours to manage their sleeping difficulties (spending more time in bed, dysregulated circadian rhythms; Ree et al., 2005;Willis et al., 2011).Dysfunctional sleep-related cognitive processes may interfere with arousal regulation, which, in turn, contributes to the onset and perpetuation of sleep disturbances (Harvey, 2002;Morin & Benca, 2012).Literature has provided evidence to support the idea that a more general inappropriate metacognitive activity regarding sleep could be among the factors behind sleep difficulties as insomnia (Ballesio et al., 2019;Harvey, 2003;Ong et al., 2012;Palagini et al., 2014;Sella et al., 2019;Waine et al., 2009).
According to Sella et al. (2019), an individual's metacognitive activity regarding sleep (Ong et al., 2012;Ree et al., 2005;Waine et al., 2009) includes their metacognitive beliefs regarding whether they have difficulty sleeping, and their use of thought-control strategies to manage any intrusive thoughts about their sleep (Ree et al., 2005), and can even give rise to unwanted and unpleasant thoughts.For this reason, metacognitive processes have received increasing interest in the context of cognitive models of sleep disorders (Harvey, 2001;Ong et al., 2012;Ree et al., 2005;Wells & Davies, 1994).In this respect, metacognitive control of intrusive thoughts at bedtime has been shown to significantly influence sleep quality (Ree et al., 2005).At a secondary level, these mechanisms may indeed interact with primary cognitive processes during the pre-sleep period amplifying the cognitive activity directly interfering with sleep, thus potentially modulating the chronicity of sleep disorders (Ong et al., 2012).Palagini et al. (2014) showed that metacognitive beliefs, anxiety and depression levels are higher in patients with primary insomnia.Thielsch et al. (2015) showed the relation among metacognitions, thought-control strategies (worry component) and sleep in patients with generalized anxiety disorder.Anxiety and insomnia are two clinical conditions that can influence the relationship among these variables or even be the cause.It is interesting whether the relationship among metacognitive strategies and sleep quality is also present in the absence of specific clinical conditions, and how they act and influence each other in non-clinical populations.
However, the term metacognition is used not only in reference to mental content, that is, regarding beliefs about one's own beliefs, but also in referring to the abilities of individuals to identify and reflect on their own and others' mental states and the ability to use this knowledge to engage adaptive strategies to manage psychological distress and interpersonal problems.In this context, the term metacognition refers to a multifaceted concept involving higher-order mental activities that serve as a source for interpreting and controlling cognitive events (i.e.thoughts) themselves, and to appraise, monitor, regulate or control cognition.In a functional-focused perspective modelthe Metacognitive Multi-Function Model (MMFM)an important aspect of metacognition refers specifically to a set of mental functions allowing people to identify mental states, reasoning about them, and ascribing them to themselves and others, in order to regulate one's own mental states and interpersonal relationships (Carcione et al., 1997;Semerari et al., 2003).Specifically, these crucial functions are necessary to: (a) identify mental states and ascribe them to oneself and others on the basis of facial expressions, somatic states, behaviour and actions; (b) reflect on and reason about mental states; (c) use information about mental states to make decisions, solve problems or psychological and interpersonal conflicts, and cope with subjective suffering (Carcione et al., 2010).Metacognitive abilities can be therefore operationalized as a set of different functional processes operating on mental states that can be selectively impaired at different levels of regulation abilities (Dimaggio et al., 2009;Semerari et al., 2007).Therefore, metacognitive abilities are necessary to identify one's thoughts and plan appropriate psychological mastery strategies.
In the context of sleep disturbances, relative to good sleepers, self-reported poor sleepers without clinical insomnia (Harvey, 2002;Barclay & Gregory, 2010;Sella & Borella, 2020) may use sleep-related metacognitive "thought-control" strategies (Harvey, 2001;Ree et al., 2005;Wells & Davies, 1994) in an attempt to manage intrusive thoughts in the pre-sleep period.These strategies include aggressive suppression (i.e.critically analysing and judging their thoughts), reappraisal (i.e.reinterpreting their thoughts), worry (i.e.focusing on other problems), social avoidance (i.e.preventing thoughts with the emotional support of someone such as a bed partner, or refusing to discuss them), cognitive and behavioural distraction (i.e.shifting attention to other thoughts, or doing something physically; Ree et al., 2005;Wells & Davies, 1994).With the exception of cognitive distraction, which may predict better sleep quality (Ree et al., 2005;Schmidt et al., 2009), such strategies appear to maintain or exacerbate insomnia symptoms (Ree et al., 2005), and are therefore considered dysfunctional.Particularly, the use of specific thought-control strategies such as reappraisal and worry seems to be one of the crucial aspects accounting for individual differences between self-reported poor and good sleepers (Sella & Borella, 2020).On the whole, maladaptive metacognitive thought-control strategies during the presleep period may generate a vicious cycle responsible for the maintenance of sleep disorders over time.In this respect, although the relationship between sleep-related metacognitive beliefs, thought-control strategies and poor sleep quality is well recognized, the possible contribution of metacognitive processes within this relation is still unclear.Particularly, no study to date has been conducted to explore the relationship between metacognitive functions and sleep quality.
On this basis, this study aimed to investigate the interaction among different metacognitive abilities, sleep-related metacognitive strategies, and sleep quality in healthy individuals.To this purpose, we performed a mediation analysis to examine the role of metacognitive thought-control strategies on the relationship between metacognitive abilities and sleep quality in a sample of healthy subjects with different sleep characteristics.We hypothesized that poor metacognitive functioning could be related to poor sleep quality through the mediation role of dysfunctional thought-control strategies in the pre-sleep period.Particularly, we expected that this relationship would be strongly influenced by specific dysfunctional thought-control strategies such as reappraisal and worry.

| Participants
The 245 healthy students were recruited from the University of L'Aquila, Italy.To evaluate subjective sleep quality, the participants completed the Pittsburgh Sleep Quality Index (PSQI; Buysse et al., 1989;Curcio et al., 2013).To have a heterogeneous distribution of scores reflecting the different quality of sleep in the general population, the PSQI score was purposely not used as an inclusion criterion.
The exclusion criteria were subjects with a history of medical, neurological or psychiatric disorders and drug/alcohol abuse.In addition, the presence of clinically relevant insomnia, anxiety and depression symptoms was an exclusion criterion.Specifically, all participants completed the State-Trait Anxiety Inventory (STAI Form Y-2; Pedrabissi & Santinello, 1989;Spielberger et al., 1983), the Beck Depression Inventory-II (BDI-II; Beck et al., 1996;Sica & Ghisi, 2007) and the Insomnia Severity Index (ISI; Castronovo et al., 2016;Morin et al., 2011) to exclude subjects with clinically relevant conditions.
The STAI Y-2 measures trait (T) anxiety, which is assumed to be a predisposition to perceive situations as potentially threatening, possibly leading to increased state anxiety (Spielberger et al., 1983;Pedrabissi & Santinello, 1989).The trait section of this questionnaire comprises 20 questions, with a total score ranging from 0 to 60.Both scales have good psychometric qualities (Cronbach α: STAI-T 0.91).
The BDI-II is a widely used 21-item self-report inventory assessing depressive symptoms in adults (Beck et al., 1996;Sica & Ghisi, 2007), with a cut-off score for clinical depression usually set above 14.
Demographic and clinical characteristics of the participants are reported in Table 1.
Ethical approval was obtained from the Internal Review Board of Department of Biotechnological and Applied Clinical Sciences, University of L'Aquila, L'Aquila, Italy (number of protocol 23,406).

| Pittsburgh Sleep Quality Index (PSQI)
The PSQI is a nine-item measure of global sleep quality over the past month (Curcio et al., 2013)  Items are rated on a 0 (never) to 3 (three or more times per week) point scale and can be added together to create the total score, where a score of > 5 indicates poor sleep quality (Buysse et al., 1989;Curcio et al., 2013).more minor things").The dependent variable was the total score for all items.

| The Metacognition Self-Assessment Scale (MSAS)
The MSAS (Pedone et al., 2017) is an 18-item self-report measure specifically developed for the assessment of MMFM sub-functions.The regression analysis and subsequently the mediation models were performed according to the correlation results (see below).

| Regression analysis
A linear regression analysis was performed to evaluate the relations between a dependent variable and one (simple linear regression) or more (multiple linear regression) explanatory variables.In this study, the variables for the regression analysis were the TCQI-R factors, the PSQI and the MSAS components.

| Mediation analysis
The mediation analysis has long been of interest to psychologists (Kenny et al., 2003;Mazza et al., 2017;Peretti et al., 2019), and it is used to understand the psychological processes by which the independent variables affect the dependent variables.According to Baron and Kenny's (1986) study, the mediation model shows how variable X's causal effect can be apportioned into its indirect effect on Y through the mediator (M) and its direct effect on Y (path c, Figure 1a).
Path a represents the effect of X on the proposed M, whereas path b is the effect of M on Y partialling out the effect of X.All of these paths would typically be quantified with unstandardized regression coefficients.The indirect effect of X on Y through M can then be quantified as the product of a and b (i.e.ab; Rucker et al., 2011).
Therefore, in the mediation process, the relationship between the independent variable (X) and the dependent variable (Y) is hypothesized to be an indirect effect (path c') that exists due to the influence of a third variable (M; Figure 1b).We used the Sobel test (Sobel, 1982) to explore the significance of the mediation effects.Specifically, the Sobel test was performed to evaluate whether the relation between metacognition and insomnia is mediated by thought-

| Regression analysis
Based on the results obtained from the correlation analysis, we performed linear regression analyses.These were further used to create the mediation model.We reported only the significant linear regression and mediation models analyses.
Thus, we obtained two significant regressions and respective mediation models using the Worry factor of TCQI-R.Based on the results of the correlation analysis, no other significant mediation models emerged.
The significant regressions were described below.
i.We carried out a first regression analysis using the Worry factor of TCQI-R as the dependent variable and the MSAS_A component of MSAS as the predictor (R 2 = 0.04; SE = 3.32; p = 0.0001).This was followed by a second regression analysis where the Worry factor of TCQI-R became the predictor together with MSAS_A component, while the PSQI was the dependent variable (R 2 = 0.13, SE = 2.98; p = 0.0001).
ii.A first regression analysis was performed using the Worry factor of TCQI-R as the dependent variable and the MSAS_D component of MSAS as the predictor (R 2 = 0.02; SE = 3.35; p = 0.003).
This was followed by a second regression analysis where the Worry factor of TCQI-R became the predictor together with MSAS_D component, while the PSQI was the dependent variable (R 2 = 0.13, SE = 0.012; p = 0.0001).

| Mediation analysis
The significant mediation models were described below.
i.The relationship between the MSAS_A component of MSAS (X) and PSQI (Y), using Worry factor of TCQI-R as a mediator (M), was explored.The Sobel test showed that this model was significant (Z = 3.69; SE = 0.013; p = 0.0002; Figure 2).
ii.We explored the relationship between the MSAS_D component of MSAS (X) and PSQI (Y), using Worry factor of TCQI-R as a mediator (M).The Sobel test showed that this model was significant (Z = 2.66; SE = 0.019; p = 0.007; Figure 3).

| DISCUSSION
In this study, we evaluated for the first time the role of metacognitive functioning as conceptualized in the MMFM (Carcione et al., 1997;Pedone et al., 2017;Semerari et al., 2003)   previous evidence indicating this specific metacognitive strategy as a predictor of better sleep quality (Ree et al., 2005;Schmidt et al., 2009).However, the reappraisal strategy showed no association with sleep quality in this study.Thus, contrary to some previous investigations (Gellis & Park, 2013;Schmidt et al., 2010Schmidt et al., , 2011;;Sella & Borella, 2020), this specific strategy could be implicated to a lesser extent in the genesis and maintenance of sleep difficulties, at least in individuals without clinical insomnia.Taken together, these results are consistent with the literature showing direct associations between the use of metacognitive thought-control strategies in the pre-sleep period and self-reported poor sleep quality.
Further, our results suggest an association between some sleeprelated thought-control strategies and specific aspects of metacognitive functioning.Particularly, significant negative correlations were found between the domain of understanding one's own mind and social avoidance and worry strategies, thus suggesting that abilities in these metacognitive domains are related to less use of these dysfunctional metacognitive thought-control strategies at bedtime; further, understanding one's own mind was associated with cognitive distraction, a sleep-related strategy that may help reduce sleeping difficulties.A similar pattern of relationships emerged for the mastery domains, with better mastery abilities related to less use of dysfunctional worry and more cognitive distraction.Finally, both understanding one's own mind and mastery metacognitive domains showed an association with the reappraisal strategy, suggesting that these crucial metacognitive dimensions may be directly associated also with a thought-control strategy that is considered particularly dysfunctional in the pre-sleep period.However, the reappraisal thought-control strategy showed no association with sleep quality in our sample.
In line with our hypothesis, our results revealed significant relationships between metacognitive functioning and sleep quality, particularly for understanding one's mental states and mastery dimensions.
Therefore, the ability to understand one's mental states and mastery functions could be the two metacognitive domains primarily involved in the dysfunctional metacognitive thought-control activity responsible for reduced sleep quality.
The major implication of these two dimensions is not surprising when considering the self-oriented nature of the specific set of abilities they include.In a functional-focused perspective model (Carcione et al., 1997;Semerari et al., 2003) In order to evaluate the impact of sleep-related metacognitive strategies on the relationship between metacognitive functioning and sleep quality, we used a mediation analysis approach.We obtained two significant mediation models with the worry strategy as a mediator and the two specific metacognitive domains, understanding one's own mind and mastery, as independent variables.Specifically, results showed that the relationship between understanding one's own mind and sleep quality is sustained by worry; similarly, worry mediated the relationship between mastery abilities and sleep quality.These results agree with previous investigations indicating a particularly dysfunctional role of worry strategy to cope with intrusive thoughts at bedtime (Gellis & Park, 2013;Schmidt et al., 2010Schmidt et al., , 2011;;Sella & Borella, 2020).Furthermore, the involvement of the worry strategy as a mediator between metacognition and sleep quality could be explained in light of the centrality of this strategy in metacognitive functioning.Worry is indeed conceptualized as a perseverative style of mental processing that may give rise to psychological distress (Wells & Matthews, 1994).In a content-oriented perspective (Wells, 2000;Wells & Matthews, 1994), metacognition is considered a set of beliefs about one's mental content aimed at regulating attentional processes that, in some cases, might promote the maintenance of dysfunctional attentional processes such as worry.According to the literature, metacognitive beliefs and worry are therefore related and both implicated in the development of emotional distress such as anxiety (Ryum et al., 2017).Potentially, the interactions between metacognitive functioning, metacognitive beliefs and worry might be critically involved also in the pre-sleep arousal state that characterizes individuals with sleep difficulties.Among the different metacognitive functions, these results indicate that the ability to understand one's own mind and mastery strategies could be essential to avoid maladaptive worry in the pre-sleep period.
This study has some limitations.First, the assessment of sleep stable representations of self and other (Dimaggio et al., 2007).Poor metacognition is considered a core element of the general pathology of personality, with abundant evidence supporting this notion (Carcione et al., 2019;Semerari et al., 2014).Alterations in metacognitive functioning have the potential to affect a wide number of clinical variables negatively; at various levels, such dysfunctions have been associated with several clinical conditions, including emotional disturbances (Wells, 2000) and difficulties in emotional regulation (Harder & Folke, 2012).As indicated by the current results, the impact of poor metacognitive functioning may also extend to variables such as sleep quality.Therefore, these findings suggest the potential relevance of clinical interventions to enhance specific metacognitive abilities, with the aim to promote more functional strategies for managing cognitive and emotional processes during the pre-sleep period.
that measures seven domains: sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleeping medications, and daytime dysfunction.
Questionnaire Insomnia-Revised (TCQI-R)This scale is written byRee et al. (2005) and depends on self-report; it consists of 35 items and is used to evaluate strategies used by individuals to control their thoughts while suffering from insomnia.The Italian version of TCQI-R(Sella et al., 2016) contains 35 items that are rated on a four-point Likert scale ranging from 1 (almost never) to 4 (almost always).These items are designed to capture five different mental control strategies that people may rely on when being kept awake by thoughts at night: (i) aggressive suppression (e.g."I get angry at myself for having the thought"); (ii) behavioural distraction/suppression (e.g."I try to block them out by reading, watching TV, or listening to the radio"); (iii) cognitive distraction/ suppression (e.g."I think pleasant thoughts instead"); (iv) reappraisal (e.g."I try to reinterpret the thought"); (v) worry (e.g."I worry about T A B L E 1 Demographic and clinical data of the participants The MSAS is scored using a five-point Likert scale (1 = never, 2 = rarely, 3 = sometimes, 4 = frequently, 5 = almost always), which yields a raw score range of 18-90.High scores on the MSAS indicate better self-evaluation of metacognitive abilities than low scores.The MSAS is designed to measure five abilities of metacognition: (1) monitoring; (2) differentiation; (3) integration; (4) decentration; and (5) mastery, which are evaluated across four sections (i.e.understanding one's own mind, MSAS-A; understanding other's mind, MSAS-B; decentering, MSAS-C; mastery, MSAS-D).Scores from the five subscales are summed to give a total score that represents the individual's overall level of metacognitive functioning.2.3 | Data analysis 2.3.1 | Correlations Pearson's correlations were performed to explore the relationship among the TCQI-R factors (aggressive suppression, cognitive distraction, reappraisal, social avoidance, behavioural distraction, and worry), the PSQI and the MSAS components (MSAS_A, MSAS_B, MSAS_C and MSAS-D).
in the relationship between maladaptive thought-control strategies and subjective sleep quality in a sample of healthy individuals.Specifically, we performed a mediation model to test the hypothesis that metacognitive abilities could be related to poor sleep quality via the mediation role of dysfunctional thought-control strategies in the pre-sleep period.Overall, resultsshowed that metacognition and sleep quality are related, and this relationship is mediated by worry, a dysfunctional thought-control strategy in the pre-sleep period.The observed effect suggests that inadequate metacognitive functioning, particularly in understanding one's own mind and mastery domains, is associated with poor sleep quality in healthy subjects via the mediation of dysfunctional worry strategy at bedtime.

F
I G U R E 2 The Sobel test significance levels, for the relationship between understanding one's own mind measured by MSAS_A and quality of sleep (PSQI) as mediated by thought-control strategies focusing on problems (Worry factor of TCQI-R).MSAS, Metacognition Self-Assessment Scale; PSQI, Pittsburgh Sleep Quality Index; TCQI-R, Thought Control Questionnaire Insomnia-Revised F I G U R E 3 The Sobel test significance levels, for the relationship between Mastery measured by MSAS_D and quality of sleep (PSQI) as mediated by thought-control strategies focusing on problems (Worry factor of TCQI-R).MSAS, Metacognition Self-Assessment Scale; PSQI, Pittsburgh Sleep Quality Index; TCQI-R, Thought Control Questionnaire Insomnia-Revised In the last decades, research has pointed to sleep-related dysfunctional metacognitive processes as crucially implicated in the genesis and maintenance of sleep difficulties in individuals with clinical insomnia and subjects with poor sleep quality (Palagini et al., 2014).Poor sleepers may use such thought-control strategies to reduce intrusive thoughts about their sleep, resulting in an exacerbation of sleep difficulties.In line with the current literature, our correlational analysis revealed significant relationships among the TCQI-R factors aggressive suppression, behavioural distraction, social avoidance, cognitive distraction and worry, and subjective sleep quality in healthy individuals, thus suggesting a critical implication of such dysfunctional metacognitive processes in the pre-sleep period.Moreover, a significant negative relationship emerged between the thoughtcontrol strategy cognitive distraction and PSQI score, thus confirming , understanding one's own mind dimension (MSAS-A) includes specific sub-functions such as the ability to monitor one's own inner state (i.e.Self-monitoring), the ability to recognize one's own representations and emotions (i.e.Identification), the ability to integrate different aspects of a mental state with their evolution over time and form a coherent narrative (i.e.Integration), and the ability to differentiate between inner classes of representations and between representations and reality (i.e.Differentiation).Mastery (MSAS-D) refers to a set of regulation and control strategies to solve psychological and interpersonal problems.Therefore, it is reasonable to hypothesize that the two MSAS domains reflecting abilities crucial to the identification and regulation of one's own mental states are those most involved in the inner regulation of dysfunctional activity in the pre-sleep period when compared with understanding other's mind (MSAS-B) and decentring (MSAS-C) that are both oriented toward understanding others' mental states.
quality and metacognitive functioning was based entirely on self-rated measures that may return different ratings compared with clinical or objective evaluations.In addition, the study was conducted on a sample of people without clinically relevant sleep disorders; therefore, the generalizability of these results to people with insomnia or other sleep-related conditions is limited.Similarly, the interaction between metacognitive functioning, worry and sleep quality remains to be further clarified in individuals with mental health conditions such as depression or anxiety disorders, particularly considering the welldocumented association between anxiety and worry regulation strategy (Coles & Heimberg, 2005; Wells & Carter, 2009).Future investigations are needed to confirm the current results, possibly including different clinical groups, such as individuals with insomnia and anxiety disorders.Nevertheless, the current study enriches the literature on the interactions between metacognitive processes and sleep quality.In this respect, previous studies have highlighted the relationship between metacognitive activity and sleep quality by investigating specific, sleep-related metacognitive processes, such as thought-control strategies or metacognitive beliefs about sleeping difficulties.This study extends the focus of the investigation to general metacognitive functioning in its distinct sub-functions, evaluating the possible link between different metacognitive abilities and sleep quality.Our results indicate for the first time an association between metacognitive functions and sleep quality sustained by the use of specific dysfunctional metacognitive processes (i.e.worry strategy during the presleep period).These results have also important clinical implications toward the development of improved interventions for individuals with sleep disturbances.Increasing evidence suggests that metacognitive dysfunction plays a key role in several psychopathological dimensions due to deficits in understanding one's own and others' internal experience, with implications for the development of integrated and