Metacognitive skills training and computerized cognitive remediation among individuals with first-episode psychosis: influence on social cognition

ABSTRACT Background Metacognitive skills training (MST) is often integrated into cognitive remediation programs for psychosis. Social cognition – the mental processes underlying social perception and behavior – is robustly related to outcomes in psychosis and is increasingly addressed with targeted treatments. Though metacognition and social cognition are related constructs, little is known about how MST may influence social cognition among individuals with psychosis participating in broad-based, non-social cognitive remediation. Methods Individuals with first-episode psychosis who completed six months of metacognitive remediation (MCR; n = 12) were compared to a historical control group who received six months of computerized cognition remediation (CCR; n = 10) alone (ClinicalTrials.gov Identifier NCT01570972). Results Though individuals receiving MCR experienced gains in emotion processing and theory of mind, these changes were not significantly different when compared to individuals receiving CCR. MST did not contribute to social cognitive change in the context of CCR. Discussion Though MST may be relevant to facilitating social cognitive gains within broader cognitive remediation programs for first-episode psychosis, these benefits are limited and may not exceed those conferred by standard cognitive remediation. Opportunities for investigation of other potential mechanisms of social cognitive response to interventions remain.


Introduction
Impaired cognition is a hallmark feature of psychotic disorders that is present by the first episode of psychosis (Mesholam-Gately et al., 2009) and is closely tied to functional outcomes (Green et al., 2000).Psychosocial interventions aimed at enhancing cognition have been developed and implemented for individuals with psychosis, with demonstrated success in improving cognitive performance (Wykes et al., 2011) and aspects of functioning (Medalia & Saperstein, 2013).In addition to addressing the primary cognitive domains impaired among people with psychosis (e.g.attention, memory, executive functioning; Barch et al., 2009), cognitive-enhancing therapies are thought to be most effective when they include both opportunities to develop awareness of the problem-solving strategies used during cognitive training tasks and deliberate efforts to facilitate transfer of these strategies to real-world activities (Bowie et al., 2020).Metacognition, which describes the capacity to "think about thinking" and to leverage control over the cognitive processes demanded by learning (Livingston, 2003), is one such area for skill development relevant to the goals of cognitive enhancing therapies.Notably, research suggests that programs that include metacognitive skills training lead to enhanced cognitive gains compared to interventions that do not include this component (Breitborde et al., 2017).
Social cognition -the mental processes underlying social perception and social interactions -has also been shown to contribute significantly to functional outcome (Fett et al., 2011), even beyond the influence of non-social neurocognition (Halverson et al., 2019).Though an increasing number of interventions targeting social cognition have been developed and studied for individuals with psychosis (Horan & Green, 2019), evidence suggests that improvements do not translate consistently to real-world functioning (Bowie et al., 2020;Fiszdon & Reddy, 2012).With regard to cognitiveenhancing interventions, some preliminary evidence suggests that though cognitive remediation alone does not improve social perception (Wright et al., 2021), social cognitive performance is improved when cognitive remediation is combined with socially focused training (Gaudelus et al., 2016;Lindenmayer et al., 2018;Nahum et al., 2014).At the same time, participation in a multicomponent treatment program for first-episode psychosis (FEP) may lead to improvements in social cognition even when no social-cognition targeted interventions are present (Breitborde et al., 2018).Additional research clarifying the impact of cognitive-enhancing interventions -particularly those paired with additional skills training -on social cognition among individuals with psychosis is thus warranted.
While aspects of metacognition and social cognition are distinct among people with psychosis (Hasson-Ohayon et al., 2015), they also share some conceptual underpinnings (e.g. the ability to become a "spectator" of one's own thoughts as well as the possible internal experiences of others; Zawidzki, 2019).However, whether metacognitive skills training may facilitate social cognition within a broad-based cognitive enhancing intervention is unknown.The goal of this pilot study was thus to assess pre-and post-social cognitive performance among young people with FEP participating in a cognitive-enhancing intervention.To determine whether the inclusion of metacognitive skills training contributed to social cognitive improvements, participants who completed a standard, broadbased computerized cognitive remediation program (CCR) were compared to participants who completed a program designed to target metacognitive skill development specifically within the context of cognitive remediation (Metacognitive Remediation; MCR; Breitborde & Moe, 2016).

Participants
Twenty-two individuals with first-episode psychosis were recruited from the Early Psychosis Intervention Center (EPICENTER; Breitborde et al., 2015) for the study.These individuals, as part of their eligibility for EPICENTER, met the following criteria: (i) diagnosis of a schizophrenia-spectrum or affective disorder with psychotic features as confirmed by the Structured Clinical Interview for DSM-IV-TR (M.B. First et al., 2002), (ii) between the ages of 15-35, (iii) initial onset of psychosis within the past 5 years, and (iv) estimated pre-morbid IQ > 70.

Interventions
Individuals in each group completed computerized cognitive training.For individuals in the computerized cognitive remediation (CCR; n = 10) group, sessions involved completion of PSSCogRehab (Bracy, 2012) cognitive training exercises that were administered by a therapist.For the metacognitive remediation (MCR; n = 12) group, sessions also involved the same PSSCogRehab cognitive training exercises along with metacognitive skills training and discussion completed with a therapist.Specific content of these discussions, the development of the MCR intervention, and the impact of MCR on general cognition is described in greater detail elsewhere (Breitborde et al., 2017).In each intervention group, readiness to "pass" from one exercise to the next was determined by a-priori mastery criteria developed and manualized at EPICENTER (Breitborde & Moe, 2016).Each session of either the CCR or MCR intervention was completed individually with a therapist (i.e.not in a group setting).All participants completed two, 45-60 minute sessions of training per week over the 6month interval between enrollment and follow-up.
Allocation to intervention for study participants was not random as the CCR group was a historical control (see, Breitborde et al., 2017).However, the CCR and MCR groups did not have statistically significant differences with regard to baseline social cognition scores, baseline neurocognition, diagnosis, age, illness duration, or sex.Participant demographics appear in Table 1.
This study was approved by the University of Arizona Institutional Review Board.All participants completed written informed consent.

Social cognition battery
Participants completed five measures assessing different domains of social cognitive functioning at baseline and again after 6 months of either CCR or MCR.The Hinting Task (Corcoran et al., 1995) was used to assess theory of mind.Emotion processing was assessed via the Facial Emotion Identification and Facial Emotion Discrimination Test (FEIT/FEDT; Kerr & Neale, 1993).The Social Cue Recognition Task (SCRT; Corrigan & Green, 1993a) was used to assess social processing with regard to both concrete and abstract social cue recognition in low emotion and high emotion social interactions.The Situational Feature Recognition Test (SRFT; Corrigan & Green, 1993b) was used to measure social knowledge of goals and actions associated with both familiar and unfamiliar activities.Per scoring guidelines for the SCRT and SRFT, the A' sensitivity index was used for these measures.Finally, the Internal, Personal, and Situational Attributions Questionnaire (IPSAQ; Kinderman & Bentall, 1996) was used to assess externalizing and personalizing attributional biases.

Statistical analyses
Within-group changes in variable scores from baseline to 6-month assessment were examined using dependent sample t-tests.Between-group changes in social cognition in the CCR compared to the MCR group were examined using a separate mixed-model ANOVA to examine the influence of group (MCR or CCR) and time (baseline versus 6-months) for each social cognitive measure.Missing data were addressed using multiple imputation.The fraction of missing information (λ) for our analyses was trivial (i.e.λ ≤ .17;Bodner, 2008).

Results
Individuals in the CCR group did not demonstrate within-subject changes in any of the social cognitive measures.Individuals in the MCR group demonstrated increases in performance on theory of mind (t = 3.13, p < .01,d = .9)and facial emotion identification (t = 2.46, p = .01,d = .7).
Comparison of changes in social cognition between the CCR and MCR groups revealed no significant group x time interactions for social cognition scores (all Fs < 1).

Discussion
In summary, our results suggest that individuals with FEP participating in MCR experience improvements in theory of mind and aspects of emotional processing, but that these improvements are not significantly greater than those experienced by individuals with FEP who participate in CCR without metacognitive skills development.Furthermore, participants within both the CCR and MCR groups did not demonstrate significant improvements on the majority of the social cognitive tasks included in the current study, suggesting that broad-based cognitive remediation does not translate to general social cognitive improvements.Though metacognition includes a broad set of mental activities that involve thinking about thinking, it can also encompass the integration of social information and perceptions about the thoughts and/or intentions of others into larger complex representations of one's self and people around them (Lysaker & Dimaggio, 2014).However, our results suggest that inclusion of metacognitive skills training in the context of broad-based cognitive enhancing interventions is likely inadequate to facilitate gains in social cognition, despite its conceptual overlap with social cognitive processes.
The present study has significant limitations.Notably, we had a small sample size and nonrandom assignment to intervention.Our use of a historical control allowed this novel investigation but lacks the scientific rigor of a full comparative trial.Though we used multiple social cognitive measures, our measurement of social cognition was not exhaustive and thus current results should be interpreted with appropriate caution.Relatedly, these tasks were selected prior to publication of findings from the Social Cognition Psychometric Evaluation (SCOPE) in early psychosis study (Ludwig et al., 2017), and thus poor psychometric properties may have contributed to our null findings.MCR was also provided in a 1:1 format to FEP individuals and thus how these findings may extend to group interventions or to individuals with chronic psychosis is not known.We do not have data on the number of eligible individuals in the program who were not enrolled in this study, and thus cannot make inferences about representativity of our current sample.Finally, we do not have measures of participant response to or satisfaction with the CCR or MCR intervention and thus how differences in these variables contributed to our findings is unknown.
Though we did not observe social cognitive benefits from a broad-based metacognitive skills training individuals with FEP in our study, opportunities for additional studies examining other potential mechanisms of social cognitive change in response to interventions remain.Though emerging evidence suggests that interventions targeting social cognition lead to social cognitive and functional improvements among young adults with psychosis (Bartholomeusz et al., 2013;Nahum et al., 2014), studies to date have not investigated factors that may influence or drive response to treatment.Thus, research focused on elucidating the mechanisms of social cognitive interventions for psychosis may facilitate greater precision with regard to treatment targets and subsequently more consistency with regard to translation of social cognitive gains to real-world functioning.

Table 1 .
Group demographics and clinical characteristics.