Metacognition is generally described as “thinking about one’s own thinking” [1
]. Metacognitive beliefs and dysfunctions have gained research focus in the past two decades. Their potential relevance with respect to the maintenance and even induction of symptoms in different psychiatric diseases such as psychotic disorders, anxiety disorders and depression [2
] is being increasingly discussed. It is speculated that the appraisal of anomalous experiences could play a critical role in the development of psychosis [4
]. As a matter of fact, metacognitive beliefs have been shown to be significantly altered in patients with psychotic disorders compared to healthy controls [6
]. Furthermore, a recent meta-analysis has provided evidence for significant differences in metacognitive beliefs in individuals in an at-risk mental state for psychosis (ARMS) [7
]. The ARMS concept identifies young people at markedly increased risk for psychosis [8
]. It is defined as experiencing at least one of the three ultra-high risk (UHR) criteria: (i) genetic risk and deterioration syndrome, i.e. genetic risk for psychosis in a first-degree relative or schizotypal disorder in the individual and relevant drop in functioning; (ii) attenuated psychotic syndrome, and (iii) brief limited intermittent psychotic symptoms.
Sex aspects play an important role in individuals in an ARMS. Similarly to findings in patients with established diagnosis of schizophrenia [9
], sex differences in UHR individuals with respect to poorer mentalizing abilities, more serious negative symptoms, higher substance abuse comorbidity and poorer social functioning have been described, with men showing more pronounced alterations compared to women [10
]. However, sex differences in metacognitive beliefs have not been previously described in the ARMS population.
Most studies on metacognition examined metacognitive beliefs, a specific aspect of metacognition, designating beliefs about the perceived importance of directing or controlling one’s own cognitive processes [7
]. Metacognitive beliefs have been shown to lead to attention and cognitive processing biases [11
]. The most commonly used instrument for the assessment of metacognitive beliefs to date is the Metacognitions Questionnaire (MCQ), which was developed on the basis of the Self-Regulatory Executive Function model to measure dysfunctional metacognitive beliefs [12
]. The original version of the MCQ consists of 65 items; however a shorter version (MCQ-30) with 30 items is also available [13
]. Both versions have 5 subscales, each assessing one dimension of dysfunctional metacognitive beliefs: (1) positive beliefs about worry (e.g. ‘Worrying helps me to get things sorted out in my mind’), (2) negative beliefs about uncontrollability of thoughts and corresponding danger (e.g. ‘Worrying is dangerous for me’ and ‘I cannot ignore my worrying thoughts’), (3) cognitive confidence (e.g. ‘I have a poor memory’), (4) negative beliefs about thoughts in general, including themes of responsibility, punishment and superstition (e.g. ‘If I did not control a worrying thought, and then it happened, it would be my fault’) and (5) cognitive self-consciousness (e.g. ‘I pay close attention to the way my mind works’).
The original MCQ was developed in 1997 and its validation study showed significantly lower scores for healthy women in the subscales ‘negative beliefs about thoughts in general’ and ‘cognitive self-consciousness’ compared to healthy men [12
]. The primary validation study for MCQ-30 revealed non-significantly higher scores for healthy males in subscales ‘cognitive confidence’ and ‘cognitive self-consciousness’ and in overall scores compared to healthy females [13
]. The validation study of the Greek translation of the MCQ-30 found that women scored significantly higher than men in the overall MCQ-30 and the ‘negative beliefs about uncontrollability of thoughts and corresponding danger’ (2) and ‘negative beliefs about thoughts in general’ (4) subscales [14
]. In the validation study of the Spanish version of the MCQ-30 on the other hand men had significantly higher results in the subsections ‘positive beliefs about worry’ (1) and ‘negative beliefs about thoughts in general’ (4) [15
Metacognitive beliefs are, however, not the only aspect of metacognitive functioning investigated in the ultrahigh risk for psychosis population. As a matter of fact, global metalevel of performance monitoring the correctness of cognitive tasks has been used as a proxy for metacognitive functioning [16
]. Furthermore, Buchy et al. [17
] depict four components of metacognition (self-reflectivity, understanding others’ minds, decentration and mastery) using the Metacognition Assessment Scale (MAS-A) [18
]. Eisenacher and her colleagues [19
] examined metamemory functioning by investigating performance in memory monitoring in individuals with ARMS.
No previous work has systematically reviewed and analysed whether alterations in metacognitive beliefs differ between male and female UHR individuals.
Our meta-analysis of the available data on metacognitive beliefs in UHR individuals revealed no statistically significant differences in overall scores between males and females. To our knowledge, this is the first systematic review on sex differences in metacognition and metacognitive beliefs measured by different instruments.
Although the overall finding shows no differences between women and men at UHR for psychosis in metacognitive beliefs, a closer look at the individual data is worthy. The subsection on ‘positive beliefs about worry’ yielded significant results in one study [28
], with higher scores in UHR women. However, upon addition of the remaining data, this difference diminished. The same effect was found in the subscale ‘cognitive self-consciousness’ with higher scores in females in the study by Barbato et al. [28
], but lower scores in women in the study by Welsh et al. [29
]. These effects might be due to differences in study samples and methodology of the included studies. As a matter of fact, the data made available derives from three heterogeneous studies with samples differing in several basic characteristics (see the Quantitative results
section Figs. 2
showing demographic data). Mean age was quite disparate, i.e. Brett et al. [27
] included adults only, whereas Welsh et al. [29
] assessed only youths between the age of 12 and 17. While these two latter studies had smaller sample sizes (around 30 participants each), three-quarters of the data we analysed was derived from Barbato et al. [28
] with a sample size of 171 ARMS individuals. Accordingly, the latter will be expected to most significantly influence the result of our meta-analysis. Indeed, the participants in the study by Barbato et al. [28
] were older than those in the study by Welsh et al. [29
] (mean age of 19.7 years versus 15.3 years). With respect to metacognitive beliefs in younger age, a validation study for the children and adolescent version of the MCQ has shown the following: adolescents reported greater ‘cognitive self-consciousness’ (subscale 5) than children, and adolescent girls scored higher in overall MCQ than adolescent boys [30
]. In the overall development in healthy individuals, metacognitive abilities increase over adolescence, peaking in adulthood, with healthy women showing better abilities than men [31
]. Thus, the comparison of UHR populations at disparate ages might have biased our results. It is also noteworthy that recently the validity of the attenuated psychosis concept has been described as decreasing with younger age [32
], possibly adding to the heterogeneity of the sample.
Though adequate efforts were made, we were not able to access all published MCQ data in UHR samples. Therefore, our analysis included MCQ data of around 60% of the over 400 ARMS individuals, whose data have been published.
With respect to operationalisation of ARMS status, the Barbato et al. study [28
] is the only one that used the Structured Interview of Prodromal Symptoms (SIPS) as the standardized assessment, whereas the other studies used the Comprehensive Assessment of “At-Risk Mental State” (CAARMS); however we do not expect a major bias from this since definitions of UHR status are maintained in all studies.
While there has been great effort in the past 20 years to understand the metacognitive alterations in prodromal stages of psychosis [6
], only limited research emphasized sex-specific differences in metacognition. That is even in spite of the fact that validation studies in various languages recognized differences in metacognitive beliefs even in healthy participants [12
]. In the last few years alterations in metacognition have not only been interpreted as characteristic markers for psychotic prodromes, but more generally as an important factor in inducing and aggravating psychiatric illness [33
]. As previously discussed, there are increasing indications pointing towards the importance of metacognition in a broader aspect of psychologic functioning as well as psychiatric disease development and maintenance [33
]. Alterations in metacognition are possibly decisive cognitive factors not only in psychotic episodes, but also in mood disorders and anxiety disorders [33
]. On the other hand metacognitive biases could also be just a broad symptom of various psychiatric diseases and may also decline with remission of the respective episode.
Neurocognitive abilities have been shown to be a necessity for the functioning of social cognition and metacognition, but not solely [34
]. The association between cognition and metacognitive performance in patients with schizophrenia spectrum disorder was found to grow weaker with increasing conceptual disorganization as measured in the Positive and Negative Syndrome Scale (PANSS); however this effect was not shown between neuro- and social cognition [35
]. The studies included in this meta-analysis adjusted neither for neurocognitive nor for social cognition parameters, as these were not their primary markers.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.