Schizophrenia Research: Cognition 28 (2022) 100222 Contents lists available at ScienceDirect Schizophrenia Research: Cognition journal homepage: www.elsevier.com/locate/scog Verbal memory performance predicts remission and functional outcome in people at clinical high-risk for psychosis Emily P. Hedges a, *, Hannah Dickson b, Stefania Tognin a, Gemma Modinos a, Mathilde Antoniades c, Mark van der Gaag d, e, Lieuwe de Haan f, g, Patrick McGorry h, Christos Pantelis i, Anita Riecher-Rössler j, Rodrigo Bressan k, Neus Barrantes-Vidal l, Marie-Odile Krebs m, Merete Nordentoft n, Stephan Ruhrmann o, Gabriele Sachs p, Bart P. Rutten q, Jim van Os a, q, EU-GEI High Risk Study1, Lucia R. Valmaggia r, Philip McGuire a, 2, Matthew J. Kempton a, 2 a Department of Psychosis Studies, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, United Kingdom b Department of Forensic and Neurodevelopmental Sciences, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, United Kingdom c Department of Psychiatry, Icahn School of Medicine at Mount Sinai, 1425 Madison Ave, New York, NY 10029, United States d VU University, Faculty of Behavioural and Movement Sciences, Department of Clinical Psychology and EMGO+ Institute for Health and Care Research, van der Boechorststraat 1, 1081, BT, Amsterdam, the Netherlands e Parnassia Psychiatric Institute, Department of Psychosis Research, Zoutkeetsingel 40, 2512, HN, The Hague, the Netherlands f Amsterdam UMC, Department Early Psychosis, Meibergdreef 5, 1105, AZ, Amsterdam, the Netherlands g Arkin Amsterdam, the Netherlands h Centre for Youth Mental Health, University of Melbourne, 35 Poplar Road (Locked Bag 10), Parkville, Victoria 485 3052, Australia i Melbourne Neuropsychiatry Centre, University of Melbourne & Melbourne Health, Carlton South, Vic, Australia j University of Basel, Faculty of Medicine, Basel, Switzerland k LiNC - Lab Interdisciplinar Neurociências Clínicas, Depto Psiquiatria, Escola Paulista de Medicina, Universidade Federal de São Paulo – UNIFESP, Brazil l Departament de Psicologia Clínica i de la Salut (Universitat Autònoma de Barcelona), Fundació Sanitària Sant Pere Claver (Spain), Spanish Mental Health Research Network (CIBERSAM), Spain m University of Paris, GHU Psychiatrie et Neurosciences of Paris, Sainte-Anne, C’JAAD, pôle PEPIT, Inserm 1266, Institut de Psychiatrie (CNRS 3557), Paris, France n Mental Health Center Copenhagen and Center for Clinical Intervention and Neuropsychiatric Schizophrenia Research, CINS, Mental Health Center Glostrup, Mental Health Services in the Capital Region of Copenhagen, University of Copenhagen, Denmark o Department of Psychiatry and Psychotherapy, Faculty of Medicine and University Hospital, University of Cologne, Cologne, Germany p Medical University of Vienna, Department of Psychiatry and Psychotherapy, Austria q Department of Psychiatry and Neuropsychology, School for Mental Health and Neuroscience, Maastricht University Medical Centre, P.O. Box 616, 6200 MD 464 Maastricht, the Netherlands r Department of Psychology, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, United Kingdom A R T I C L E I N F O A B S T R A C T Keywords: Robust deficits in cognitive functioning are present in people with psychosis and are evident in the early stages of Cognition the disorder. Impairments in verbal memory and verbal fluency are reliably seen in individuals at clinical high- Verbal fluency risk for psychosis (CHR) compared to healthy populations. As previous studies have shown a relationship be­ Prodrome tween cognition and longer-term outcomes in schizophrenia, the aim of this paper was to explore whether verbal Transition Early intervention memory and verbal fluency performance predicted outcomes in a large CHR sample recruited as part of the EU- GEI High Risk Study. Participants included 316 CHR individuals, 90.8% of whom were not currently on anti­ psychotic medication, and 60 healthy controls. Verbal memory and verbal fluency performance were measured at baseline. At two-year follow-up, CHR individuals were assessed by three different outcome measures, those who did and did not (1) transition to psychosis, (2) experience burdening impairment or disabilities, or (3) remit clinically from CHR status. Individuals with CHR displayed significant verbal memory and verbal fluency deficits * Corresponding author at: Department of Psychosis Studies, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, 16 De Crespigny Park, SE5 8AF, United Kingdom. E-mail address: emily.p.hedges@kcl.ac.uk (E.P. Hedges). 1 Members of the EU-GEI High Risk Study group author are listed at the end of the paper. 2 These authors contributed equally to this paper. https://doi.org/10.1016/j.scog.2021.100222 Received 23 July 2021; Received in revised form 11 October 2021; Accepted 11 October 2021 Available online 28 October 2021 2215-0013/© 2021 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). E.P. Hedges et al. Schizophrenia Research: Cognition 28 (2022) 100222 at baseline compared to healthy controls (Hedges' g effect size = 0.24 to 0.66). There were no significant dif­ ferences in cognitive performance of those who did and did not transition to psychosis. However, impaired immediate verbal recall predicted both functional disability and non-remission from the CHR state. Results remained significant when analyses were restricted to only include antipsychotic-free CHR participants. These findings may inform the development of early interventions designed to improve cognitive deficits in the early stages of psychosis. 1. Introduction representative sample, findings from EU-GEI have the potential to pro­ vide a better understanding of the role of cognition and its effect on Cognitive impairment is a core feature of schizophrenia (Sheffield subsequent outcomes in CHR. Furthermore, EU-GEI consists of a pre­ et al., 2018) and is an important predictor of poor functional outcome dominantly antipsychotic-free CHR cohort; 90.8% of individuals who (Bowie and Harvey, 2005; Green et al., 2004). Deficits in verbal memory completed cognitive testing were not on antipsychotic medication, and verbal fluency are consistently reported among individuals with therefore helping to provide robust findings on neurocognition in CHR schizophrenia (Mesholam-Gately et al., 2009; Schaefer et al., 2013), individuals. those experiencing first-episode psychosis (Aas et al., 2014; Mesholam- Gately et al., 2009) and individuals at clinical high-risk for psychosis 2. Methods (CHR) (Bora et al., 2014; Catalan et al., 2021). Although cognitive impairment in CHR is less pronounced than individuals with first- 2.1. Participants episode psychosis relative to healthy individuals (HC), certain do­ mains, and particularly memory, may be impaired to a comparable de­ The EU-GEI High Risk study includes 344 CHR individuals and 67 gree as first-episode psychosis (Becker et al., 2010; Sheffield et al., HC. CHR participants were recruited from 11 early detection centres 2018). In such instance, verbal memory and verbal fluency performance (London, Amsterdam, The Hague, Basel, Cologne, Melbourne, Vienna, may be useful in predicting psychosis and functional outcome, as well as Copenhagen, Paris, Barcelona, and Sao Paolo), having been referred by targeting early interventions to improve performance (Catalan et al., their local mental health service. HC were recruited at four centres; 2021; Hauser et al., 2017). London including GP lists, national postal address file and Gumtree Prospective studies of CHR cohorts, where individuals are clinically website; Melbourne, by online advertisement; and Amsterdam and The followed-up after 1–3 years, allow for comparisons of cognitive perfor­ Hague, by Proefbunny website. Ethical approval for EU-GEI study was mance at baseline between CHR individuals who do and do not subse­ obtained locally at each site and participants gave written informed quently (1) transition to psychosis, (2) experience burdening consent. impairment or disabilities, or (3) remit from CHR status. Among in­ For inclusion criteria of CHR participants, the Comprehensive dividuals who transition to psychosis (CHR-T) compared to those who Assessment of At-Risk Mental States (CAARMS) (Yung et al., 2005) was do not (CHR-NT), existing meta-analytic evidence is inconsistent used to determine whether individuals met at least one of CHR criteria: reporting both significant differences (Bora et al., 2014; Catalan et al., Attenuated Psychosis Group, Vulnerability Group or Brief Limited 2021; Fusar-Poli et al., 2012; Hauser et al., 2017) and no differences Intermittent Psychotic Symptoms Group. Exclusion criteria for all par­ (Catalan et al., 2021; De Herdt et al., 2013) in verbal memory and verbal ticipants were: (1) past/present diagnosis of psychotic disorder, deter­ fluency performance. In the largest CHR sample of North American mined by CAARMS and Structural Clinical Interview for DSM Disorders Prodrome Longitudinal Study (NAPLS-2), authors reported differences (First et al., 2002); (2) relevant symptoms explained by neurological in immediate verbal memory and semantic fluency performance of CHR- disorder or drug/alcohol dependency; (3) contraindications to MRI T and CHR-NT, but these did not survive Bonferroni-corrections for scanning or unwillingness to provide blood/saliva sample, and (3) IQ multiple comparisons (Seidman et al., 2016). Yet, in the latest meta- estimate < 60. HC participants did not meet CHR criteria. CHR and HC analysis, impaired verbal memory was the strongest predictor of tran­ individuals were included in the present study if they had completed at sition to psychosis (Catalan et al., 2021). In addition, few longitudinal least one measure of verbal memory or verbal fluency at baseline. studies have explored individual associations of verbal memory and Typical age of participants was 18–35 years but not restricted to due to verbal fluency with CHR functional and remission outcomes. Among variation between sites in the age at which persons are accepted by them, there have been conflicting findings for verbal fluency and verbal clinical services. memory for predicting social and occupational functioning (Bolt et al., 2019; Lin et al., 2011; Niendam et al., 2007) and remission outcomes 2.2. Procedure (Addington et al., 2019b; Glenthøj et al., 2021; Lee et al., 2014; Simon et al., 2012). Here, the lack of consensus is likely attributable to limited Using a naturalistic, prospective design, the EU-GEI study collected sample sizes, attrition during follow-up and phenotypic heterogeneity. multi-modal data at baseline from July 2010 to August 2015. Partici­ Although underexplored, it is important to determine cognitive pre­ pants were invited for follow-up assessments at 12 months and 24 dictors of functional disability alongside clinical outcomes. Long-term months. For CHR-T participants, further assessments were conducted as functional difficulties are highly prevalent within CHR individuals, soon as possible after transition, one-year and two-years later. Face-to- irrespective of transition, and they are not addressed by psychological face assessments were carried out by trained researchers. Researchers interventions (van der Gaag et al., 2019) so represent an equally had passed the online training course which involved rating CAARMS important target for prevention (Carrión et al., 2013). and Global Assessment of Functioning (GAF) training videos. Inter-rater The aims of the present study were to examine (1) differences in reliability across the EU-GEI centres was assessed from the training verbal memory and verbal fluency in CHR and HC, as well as (2) in CHR- videos and scores greater than 0.7 were deemed acceptable (eTable 1). NT and CHR-T, and (3) the relationship between verbal memory and verbal fluency at baseline and two-year functional and remission out­ comes in a large CHR sample. The “EU Network of National Schizo­ 2.3. Baseline demographics and cognitive measures phrenia Networks study gene-environment interactions” (EU-GEI) is to- date the largest multi-national study of 344 CHR individuals (van Os Participants were assessed at each visit on a wide range of social, et al., 2014). As a well-characterised and uniquely globally- cognitive, clinical, imaging and blood-based measures (van Os et al., 2014). Instruments were translated into the language local to each site 2 E.P. Hedges et al. Schizophrenia Research: Cognition 28 (2022) 100222 and subsequently translated back for accuracy. The following de­ baseline cognitive performance in CHR-T and CHR-NT. scriptions focus only on the measures that are relevant to this paper. In relation to clinical outcomes, general linear model univariate Data on demographic characteristics of participants were collected at analyses were conducted to determine the effect of baseline cognitive baseline (e.g., age, gender) using the Medical Research Council socio- performance in CHR on GAF disability impairment at two-year follow- demographic schedule (Mallett, 1997). Socio-economic status (SES) up. Binary logistic regression was carried out to examine whether was defined by father's social class at participant's birth. SES was cat­ baseline cognitive performance in CHR predicts remission at follow-up. egorised into a three-class model: salariat, intermediate and working Analyses were repeated for each parameter of verbal memory (im­ class. Fathers who were long-term unemployed were reclassified ac­ mediate and delayed recall) and verbal fluency (phonemic and semantic cording to their last main paid job and the never worked/full-time stu­ fluency). These were chosen as the most commonly reported in primary dents were excluded (n = 2) (Harrison and Rose, 2006). Intelligence (IQ) studies for measures of verbal memory and verbal fluency (Fusar-Poli was measured using a short version of the Wechsler Adult Intelligence et al., 2012). Statistical significance was set at p < 0.05 (two-tailed). In Scale-III, including the four subtests of Block Design, Arithmetic, Digit Model 1, analyses accounted for age, site, and gender (Menghini-Müller Symbol, and Information (Blyler et al., 2000), which have been short­ et al., 2020). For any significant finding, analyses were repeated to ened, and demonstrates good reliability and predictive validity (Velth­ include only CHR participants who were not on antipsychotic medica­ orst et al., 2013). tion at baseline cognitive assessment (Model 2). SES was included as an This paper focuses on two widely-used instruments as measures of additional covariate in Model 3 as this measure differed between groups. verbal memory and verbal fluency. Verbal memory was assessed with It was not included in Model 1 because 14.8% of SES data was missing the Rey Auditory Verbal Learning Test (RAVLT) (Delaney et al., 1992), and may have reduced statistical power; conversely there was no which uses a fixed-order 15-word list to measure immediate recall (total missing data for the other covariates. Confounds were selected apriori number of words correctly recalled for Trials I-V) and delayed recall based on previous research to account for additional variance in the (total words correctly recalled for Trial VI). Verbal fluency was evalu­ model and increase sensitivity of the analyses. IQ was not added as a ated using the Verbal Fluency Test (Henry and Crawford, 2005) during covariate. As low IQ has been reported as a risk factor for later devel­ which participants are asked to generate as many words as possible in opment of schizophrenia (Dickson et al., 2012) and IQ is correlated with 60 s for a given letter or category. Semantic fluency was defined by the verbal memory fluency tests, including IQ in statistical analyses may total number of words produced in the animal names subtest; and result in an underestimation of impairments in these neurocognitive phonemic fluency (or letter fluency), by the total words produced in the domains (MacCabe et al., 2012). letter subtest. Letter subtests differed between sites to match first letter frequencies within each language where possible (e.g., FAS was utilised 3. Results for London/Cologne and SNA for Amsterdam/The Hague). Raw scores were converted into standardised z-scores based on the HC group (CHR 3.1. Sample characteristics mean minus HC mean, divided by HC standard deviation), so that the HC group had a mean of 0 and standard deviation of 1. The final sample consisted of 316 CHR participants and 60 HC. Table 1 provides information on baseline characteristics of participants. 2.4. Clinical measures at follow-up There were significant differences in IQ, years in education, SES and GAF scores of CHR and HC. Most CHR participants (90.8%) were not on CHR participants were first categorised into CHR-Ts and CHR-NTs. antipsychotic medication at the baseline cognitive assessment. A total of Transition to psychosis was defined as the development of full 156 CHR participants returned at two-year follow-up. CHR participants threshold psychotic disorder using the CAARMS (Yung et al., 2005). who did and did not complete follow-up measures differed significantly Available clinical records were used to determine any diagnosis of a at baseline in age and years in education but not in any cognitive or psychotic disorder when participants did not return for follow-up as­ other demographic measure (eTable 2). sessments. CHR participants were further grouped into remitters (CHR- R) and non-remitters (CHR-NR). CHR-R participants were those whose symptoms remitted at two-year follow-up (i.e., no longer met CHR Table 1 criteria). Participants who still met CHR criteria at follow-up or had Characteristics of CHR participants and healthy controls at baseline. transitioned to psychosis were categorised as CHR-NRs. Global func­ HC CHR HC vs CHR tioning was also assessed using GAF (Hall, 1995) which was split into (N = 60) (N = 316) (p-value) disability and symptoms subscales. GAF disability subscale resembles Age in years, M(SD) 23.68 (4.15) 23.05 (4.96) 0.357 the established Social and Occupational Functioning Assessment Scale Gender female, n(%) 28 (46.67) 148 (46.84) 0.981 (Goldman et al., 1992). GAF scores range from 0 to 100; higher scores Years in educationa, M(SD) 16.25 (2.77) 14.42 (3.11) <0.001** IQb, M(SD) 112.08 98.37 indicate superior functioning or fewer symptoms within the last month. <0.001** (18.01) (16.98) Due to natural variation in follow-up visits, analyses for the present SESc, n(%) 0.005** paper were restricted to include two-year data that was collected be­ Salariat 27 (54.00) 95 (34.17) tween 1.5 and 2.5 years after baseline. Intermediate 18 (36.00) 101 (36.33) Working class 5 (10.00) 82 (29.50) GAF symptomsd, M(SD) 86.80 (11.03) 55.06 2.5. Statistical analyses <0.001** (10.20) e GAF symptoms at follow-up , M – 62.49 Statistical analyses were performed using IBM SPSS Statistics 26 for (SD) (12.69) Windows. CHR individuals and HC were compared on baseline de­ GAF disabilityf, M(SD) 85.37 (9.11) 55.76 <0.001** (12.46) mographic characteristics using independent t-tests for continuous GAF disability at follow-upg, M – 63.16 dependent variables and chi-square tests for categorical dependent (SD) (14.69) variables. Baseline characteristics were also compared between CHR Antipsychotic useh, n(%) – 28 (9.20) who had clinical data at follow-up and CHR who did not complete Current cannabis usei, n(%) 17 (43.59) 82 (34.89) 0.295 clinical follow-up measures. Data was missing for: 24 CHR; 1 HC and 18 CHR; 10 HC and 38 CHR; d1 HC a b c For case-control comparisons, univariate general linear models were and 16 CHR; e 203 CHR; f1 HC and 8 CHR; g194 CHR; h10 CHR; i21 HC and 81 implemented to determine any differences in verbal memory and verbal CHR. fluency performance. The same model was then utilised to compare ** p-value <0.01. 3 E.P. Hedges et al. Schizophrenia Research: Cognition 28 (2022) 100222 3.2. Baseline cognition group comparisons CHR-NT groups. Compared to HC, the magnitude of impairment of CHR- T group was similar to that of the overall CHR sample. Therefore, any At baseline, the CHR group performed significantly worse than HC potential decline in verbal memory and fluency performance was not on tests of immediate recall [F(1,359) = 14.43, p = 0.00017], delayed detectable early on in the prodromal phase of psychosis. Inconsistencies recall [F(1,349) = 6.47, p = 0.011], phonemic fluency [F(1,367) = in results between studies may reflect methodological differences 19.12, p = 0.00002], and semantic fluency [F(1,366) = 7.34, p = 0.007], (Addington et al., 2019a) such as the use of different memory tasks, adjusting for gender, age, and site (Table 2). All findings remained sig­ which may employ different encoding strategies (Grimes et al., 2017). A nificant in analyses restricted to antipsychotic-free participants and further contributory factor may be differences in the nature of CHR additionally controlling for SES (eTable 3). Fig. 1 displays the mean z- samples across centres (Catalan et al., 2021), which vary widely in the scores of performance on each cognitive test for CHR and HC groups. way that subjects are ascertained, their average age, and the incidence of Of 316 CHR participants, 60 participants transitioned to psychosis transition. during follow-up. There were no significant differences seen in imme­ However, poor immediate verbal memory was a predictor of both diate recall, delayed recall, phonemic fluency, or semantic fluency non-remission from CHR state and greater functional disability at two- scores at baseline between CHR-NT and CHR-T groups (Table 2 and eFig. year follow-up. To date, few longitudinal studies have explored the as­ 1). sociation between verbal memory and these two outcomes (Addington et al., 2019b; Bolt et al., 2019; Glenthøj et al., 2021; Lee et al., 2014; Lin 3.3. Cognitive functioning and clinical outcome et al., 2011; Niendam et al., 2007; Simon et al., 2012) and the findings have been inconsistent. As with transition to psychosis, this lack of At follow-up, clinical data was available for 156 CHR participants. consensus may be attributable to the use of different methods for testing Among CHR, immediate recall significantly predicted global disability verbal memory and defining outcome measures, the use of small sample score [F(1,119) = 5.39, p = 0.022] and non-remission [OR = 0.54, 95% sizes and variable durations of follow-up. These issues may be addressed CI, 0.34–0.85, p = 0.008] at two-year follow-up, controlling for age, by standardising methods across studies and the use of large sample sizes gender, and site (Tables 3 and 4, respectively). Fig. 2 shows that higher (Addington and Barbato, 2012). Cognition, and particularly verbal immediate recall performance at baseline was associated with greater memory, is generally viewed as an important contributor of functional functioning at follow-up. The results did not change in Model 2 or Model impairment in schizophrenia (Green et al., 2000; Kahn and Keefe, 2013) 3 (eTable 4). and functional difficulties are highly prevalent in CHR populations The findings indicated that delayed recall, phonemic and semantic (Carrión et al., 2013). Preliminary evidence suggests that cognitive fluency were not statistically significant predictors of global disability or remediation therapies are an effective clinical intervention during non-remission at follow-up (Tables 3 and 4). clinical high-risk stage (Glenthøj et al., 2017) and are associated with improvements in cognition that enhance functioning in CHR samples 4. Discussion (Fiszdon et al., 2016). However, research into other forms of cognitive treatment is also needed (Sheffield et al., 2018). Findings from the Our first main finding was that CHR individuals displayed deficits in present study indicate that verbal memory may be a potentially prom­ verbal memory and verbal fluency compared to HC. Our second major ising target for early interventions as greater performance in this domain finding was that, within the CHR sample, impairments in immediate was predictive of remission from CHR state and higher functioning. In verbal recall at baseline were associated with both non-remission and addition, cognitive remediation in schizophrenia has the largest effect functional disability at two-year follow-up. The results remained un­ on improvements in verbal memory as well as in community and work changed following sensitivity analyses with participants who were not functioning (Lejeune et al., 2021). Future research should examine on antipsychotic medication. whether improving verbal memory in CHR reduces the probability of an In line with previous meta-analytic findings, CHR individuals dis­ unfavourable outcome. played small-to-medium effect size differences in verbal memory and The present study has several strengths. First, EU-GEI study collected verbal fluency compared to HC (Bora et al., 2014; Catalan et al., 2021; comprehensive cognitive, clinical, and functional data from a large and Giuliano et al., 2012). Also, in line with earlier studies, the magnitude of globally-represented sample of CHR individuals. Second, the study in­ these impairments was smaller than those seen in first-episode psychosis volves a predominantly antipsychotic-free sample. Therefore, we could patients relative to controls (standardised mean differences of − 1.30 to explore the role of verbal memory and fluency without the influence of − 0.69) (Mesholam-Gately et al., 2009), indicating that there may be a antipsychotic use acting as a confounding variable (Bora and Murray, subsequent decline in performance as the illness progresses. Although 2014). Sensitivity analyses confirmed that findings remained significant such decline is inferred from CHR and first-episode psychosis perfor­ when the minority (9.2%) of participants on antipsychotic medication mance in the same cognitive tasks, the comparison is cross-sectional in were excluded, suggesting that cognitive deficits in CHR may be mini­ nature and so should be interpreted cautiously. Contrary to the results of mally influenced by current antipsychotic use (Catalan et al., 2021; meta-analyses (Fusar-Poli et al., 2012), we did not observe differences in Pukrop et al., 2006; Seidman et al., 2016). A limitation is that, although verbal memory or verbal fluency performance between the CHR-T and typical of similar studies, follow-up lasted for only two years so it is Table 2 Mean (standard deviation) raw scores and between-group comparisons of cognitive performance at baseline. HC (N = 60) CHR (N = 316) HC vs CHR (p-value) Effect size (Hedges' g) CHR-NT (N = 256) CHR-T (N = 60) NT vs T (p-value) Verbal learning, M(SD) Immediate recalla 56.27 (8.50) 51.31 (9.97) <0.001** 0.51 51.26 (9.62) 51.51 (11.49) 0.327 Delayed recallb 11.45 (3.48) 10.57 (3.05) 0.011* 0.28 10.60 (3.04) 10.45 (3.11) 0.313 Verbal fluency, M(SD) Phonemic fluencyc 43.97 (14.43) 35.52 (12.52) <0.001** 0.66 34.73 (12.37) 38.95 (12.67) 0.196 Semantic fluencyd 22.83 (8.00) 21.32 (6.03) 0.007** 0.24 21.33 (5.97) 21.31 (6.37) 0.789 Model 1 for each analysis adjusting for age, gender, and site. Data was missing for: a12 CHR-NT and 5 CHR-T; b4 HC, 18 CHR-NT and 5 CHR-T; c2 HC, 5 CHR-NT and 2 CHR-T; d2 HC, 6 CHR-NT and 2 CHR-T. * p-value <0.05. ** p-value <0.01. 4 E.P. Hedges et al. Schizophrenia Research: Cognition 28 (2022) 100222 0.300 Healthy controls CHR parcipants 0.200 Mean (SE) z-scores at baseline assessment 0.100 0.000 -0.100 -0.200 -0.300 -0.400 -0.500 -0.600 -0.700 Immediate Delayed Phonemic Semanc Recall Recall Fluency Fluency Fig. 1. Mean (SE) standardised scores of cognitive performance for CHR participants and healthy controls at baseline. complete follow-up measures reported no differences between the Table 3 groups. Effect of baseline cognition on GAF disability score in 122 CHR at two-year In conclusion, psychosis is characterised by dysfunctions in several follow-up. cognitive domains and impairments are reliably reported in verbal F df Effect size (Cohen's f) p-value memory and verbal fluency (Sheffield et al., 2018). Our findings showed Verbal memory that verbal memory and fluency were associated with vulnerability to Immediate recall 5.39 1119 0.22 0.022* psychosis but were not implicated in the subsequent development of the Delayed recall 0.24 1118 0.05 0.626 disorder. However, we identified that verbal memory deficits were Verbal fluency Phonemic fluency 2.70 1121 0.16 0.104 prominent characteristics of the CHR state and contributed to functional Semantic fluency 0.62 1121 0.08 0.433 impairment. Therefore, verbal memory may represent an important target for early interventions for cognitive impairment in CHR pop­ Model 1 for each analysis adjusting for age, gender, and site. * p-value <0.05. ulations, regardless of who goes on to develop psychosis. CRediT authorship contribution statement Table 4 Effect of baseline cognition on non-remission status at two-year follow-up in Emily P Hedges: Conceptualization, Methodology, Software, Vali­ CHR participants. dation, Formal analysis, Data curation, Writing – original draft, Writing CHR-R CHR-NR Effect size 95% CI p-value – review & editing, Visualization. (N = 49) (N = 96) (odds ratio) Hannah Dickson: Conceptualization, Methodology, Writing – re­ Verbal memory (M, SD) view & editing. Immediate 51.90 49.83 0.54 0.34, 0.008** Stefania Tognin: Investigation, Data curation, Writing – review & recalla (8.62) (10.74) 0.85 editing. Delayed 10.42 10.38 0.86 0.46, 0.621 Gemma Modinos: Investigation, Data curation, Writing – review & b recall (2.42) (2.95) 1.58 Verbal fluency (M, SD) editing. Phonemic 34.16 36.21 0.68 0.40, 0.152 Mathilde Antoniades: Investigation, Data curation, Writing – re­ fluencyc (12.11) (12.86) 1.15 view & editing. Semantic 22.43 21.46 0.90 0.50, 0.722 Mark van der Gaag: Resources, Data curation, Writing – review & fluencyd (6.06) (6.34) 1.63 editing, Supervision, Project administration, Funding acquisition. Model 1 for each analysis adjusting for age, gender, and site. Lieuwe de Haan: Resources, Data curation, Writing – review & Data was missing for: a6 CHR-NR and 1 CHR-R; b7 CHR-NR and 1 CHR-R; c2 editing, Supervision, Project administration, Funding acquisition. CHR-NR; d2 CHR-NR. Patrick McGorry: Resources, Data curation, Writing – review & ** p-value <0.01. editing, Supervision, Project administration, Funding acquisition. Christos Pantelis: Resources, Data curation, Writing – review & possible that the outcome of CHR individuals may subsequently change editing, Supervision, Project administration, Funding acquisition. (Addington et al., 2019b; Salazar de Pablo et al., 2021). Second, as re­ Anita Richer-Rössler: Resources, Data curation, Writing – review & searchers rated several different clinical scales at follow-up, they were editing, Supervision, Project administration, Funding acquisition. not blinded to transition or remission status. A further limitation is that Rodrigo Bressan: Resources, Data curation, Writing – review & the study may have been subject to attrition bias. However, baseline editing, Supervision, Project administration, Funding acquisition. cognitive and clinical comparisons of those who did and did not Neus Barrantes-Vidal: Resources, Data curation, Writing – review 5 E.P. Hedges et al. Schizophrenia Research: Cognition 28 (2022) 100222 100 90 80 GAF disability score at follow-up 70 60 50 40 30 20 10 0 0 10 20 30 40 50 60 70 80 Immediate recall at baseline assessment Fig. 2. Scatterplot of immediate recall at baseline against GAF disability scores at two-year follow-up in CHR participants. & editing, Supervision, Project administration, Funding acquisition. NHMRC Program Grant (ID:1150083). Marie-Odile Krebs: Resources, Data curation, Writing – review & editing, Supervision, Project administration, Funding acquisition. Group Author Merete Nordentoft: Resources, Data curation, Writing – review & editing, Supervision, Project administration, Funding acquisition. EU-GEI High Risk Study group authors: Philip McGuire1, Sara Stephan Ruhrmann: Resources, Data curation, Writing – review & Pisani1, Alessia Avila1, Maria Calem1, Stefania Tognin1, Gemma Mod­ editing, Supervision, Project administration, Funding acquisition. inos1, Lieuwe de Haan2,3, Mark van der Gaag4,5, Eva Velthorst2,6, Tamar Gabriele Sachs: Resources, Data curation, Writing – review & C. Kraan2, Daniella S. van Dam2, Nadine Burger5, Barnaby Nelson7, editing, Supervision, Project administration, Funding acquisition. Patrick McGorry7, G. Paul Amminger7, Christos Pantelis8, Athena Bart P Rutten: Conceptualization, Resources, Data curation, Writing Politis7, Joanne Goodall7, Anita Riecher-Rössler9, Stefan Borgwardt9, – review & editing, Supervision, Project administration, Funding Erich Studerus9, Bruno Sini Scarpato10, Ary Gadelha10, Elisa Brietzke11, acquisition. Graccielle Asevedo10, Elson Asevedo10, Andre Zugman10, Neus Bar­ Jim van Os: Conceptualization, Resources, Data curation, Writing – rantes-Vidal12, Tecelli Domínguez-Martínez13, Anna Racioppi14, review & editing, Supervision, Project administration, Funding Thomas R. Kwapil15, Manel Monsonet14, Lídia Hinojosa14, Mathilde acquisition. Kazes16, Claire Daban16, Julie Bourgin16, Olivier Gay16, Célia Mam-Lam- EU-GEI High Risk Study: Investigation, Data curation. Fook16, Marie-Odile Krebs17, Dorte Nordholm18, Lasse Randers18, Kris­ Lucia R Valmaggia: Conceptualization, Resources, Data curation, tine Krakauer18, Louise Glenthøj18, Birte Glenthøj19, Merete Norden­ Writing – review & editing, Supervision, Project administration, Fund­ toft18, Stephan Ruhrmann20, Dominika Gebhard20, Julia Arnhold21, ing acquisition. Joachim Klosterkötter20, Gabriele Sachs22, Harald Aschauer22, Iris Philip McGuire: Conceptualization, Resources, Data curation, Lasser22, Philippe A. Delespaul23,24, Bart P. Rutten23, Jim van Os1,23 Writing – review & editing, Supervision, Project administration, Fund­ Affiliations: ing acquisition. Matthew J Kempton: Conceptualization, Methodology, Resources, 1 Department of Psychosis Studies, Institute of Psychiatry, Psy­ Data curation, Writing – review & editing, Supervision, Project admin­ chology & Neuroscience, King’s College London, De Crespigny istration, Funding acquisition. Park, Denmark 458 Hill, London, United Kingdom SE5 8AF. 2 Amsterdam UMC, Department Early Psychosis, Meibergdreef 5, 1105 AZ Amsterdam, The Netherlands Declaration of competing interest 3 Arkin Amsterdam. 4 VU University, Faculty of Behavioural and Movement Sciences, The authors declare no conflict of interest. Department of Clinical Psychology and EMGO+ Institute for Health and Care Research, van der Boechorststraat 1, 1081 BT Acknowledgements Amsterdam, The Netherlands. 5 Parnassia Psychiatric Institute, Department of Psychosis The European Network of National Schizophrenia Networks Study­ Research, Zoutkeetsingel 40, 2512 HN The Hague, The ing Gene-Environment Interactions (EU-GEI) Project is funded by grant Netherlands. agreement HEALTH-F2-2010-241909 (Project EU-GEI) from the Euro­ 6 Department of Psychiatry, Icahn School of Medicine at Mount pean Community's Seventh Framework Programme. Additional support Sinai, 1425 Madison Ave, New York, NY 10029, United States. was provided by a Medical Research Council Fellowship to M Kempton 7 Centre for Youth Mental Health, University of Melbourne, 35 (grant MR/J008915/1) and by the Ministerio de Ciencia, Innovación e Poplar Road (Locked Bag 10), Parkville, Victoria 485 3052, Universidades (PSI2017-87512-C2-1-R) and the Generalitat de Cata­ Australia. lunya (2017SGR1612) to N Barrantes-Vidal. C Pantelis was supported by a NHMRC Senior Principal Research Fellowship (628386 & 1105825), a 6 E.P. Hedges et al. Schizophrenia Research: Cognition 28 (2022) 100222 8 Melbourne Neuropsychiatry Centre, University of Melbourne & Bolt, L.K., Amminger, G.P., Farhall, J., McGorry, P.D., Nelson, B., Markulev, C., Yuen, H. P., Schäfer, M.R., Mossaheb, N., Schlögelhofer, M., 2019. Neurocognition as a Melbourne Health, Carlton South, Vic, Australia. predictor of transition to psychotic disorder and functional outcomes in ultra-high 9 University of Basel, Faculty of Medicine, Basel, Switzerland. risk participants: findings from the NEURAPRO randomized clinical trial. Schizophr. 10 LiNC - Lab Interdisciplinar Neurociências Clínicas, Depto Psi­ Res. 206, 67–74. quiatria, Escola Paulista de Medicina, Universidade Federal de Bora, E., Murray, R.M., 2014. Meta-analysis of cognitive deficits in ultra-high risk to psychosis and first-episode psychosis: do the cognitive deficits progress over, or São Paulo – UNIFESP. after, the onset of psychosis? Schizophr. Bull. 40, 744–755. 11 Depto Psiquiatria, Escola Paulista de Medicina, Universidade Bora, E., Lin, A., Wood, S., Yung, A., McGorry, P., Pantelis, C., 2014. Cognitive deficits in Federal de São Paulo – UNIFESP. youth with familial and clinical high risk to psychosis: a systematic review and meta- analysis. Acta Psychiatr. Scand. 130, 1–15. 12 Departament de Psicologia Clínica i de la Salut (Universitat Bowie, C.R., Harvey, P.D., 2005. Cognition in schizophrenia: impairments, determinants, Autònoma de Barcelona), Fundació Sanitària Sant Pere Claver and functional importance. Psychiatr. Clin. 28, 613–633. (Spain), Spanish Mental Health Research Network (CIBERSAM). Carrión, R.E., McLaughlin, D., Goldberg, T.E., Auther, A.M., Olsen, R.H., Olvet, D.M., Correll, C.U., Cornblatt, B.A., 2013. Prediction of functional outcome in individuals 13 CONACYT-Dirección de Investigaciones Epidemiológicas y Psi­ at clinical high risk for psychosis. JAMA Psychiatry 70, 1133–1142. cosociales, Instituto Nacional de Psiquiatría Ramón de la Fuente Catalan, A., Salazar de Pablo, G., Aymerich, C., Damiani, S., Sordi, V., Radua, J., Muñiz (México). Oliver, D., McGuire, P., Giuliano, A.J., Stone, W.S., Fusar-Poli, P., 2021. Neurocognitive functioning in individuals at clinical high risk for psychosis: a 14 Departament de Psicologia Clínica i de la Salut (Universitat systematic review and meta-analysis. JAMA Psychiat. 78 (8), 859–867. Autònoma de Barcelona). de Herdt, A., Wampers, M., Vancampfort, D., de Hert, M., Vanhees, L., Demunter, H., van 15 Department of Psychology, University of Illinois at Urbana- Bouwel, L., Brunner, E., Probst, M., 2013. Neurocognition in clinical high risk young adults who did or did not convert to a first schizophrenic psychosis: a meta-analysis. Champaign (USA). Schizophr. Res. 149, 48–55. 16 University Paris Descartes, Hôpital Sainte-Anne, C’JAAD, Service Delaney, R.C., Prevey, M.L., Cramer, J., Mattson, R.H., Group, V.E.C.S.R., 1992. Test- Hospitalo- Universitaire, Inserm U894, Institut de Psychiatrie retest comparability and control subject data for the rey-auditory verbal learning test (CNRS 3557) Paris, France. and rey-osterrieth/Taylor complex figures, 7, 523–528. Dickson, H., Laurens, K.R., Cullen, A.E., Hodgins, S., 2012. Meta-analyses of cognitive 17 University of Paris, GHU Psychiatrie et Neurosciences of Paris, and motor function in youth aged 16 years and younger who subsequently develop Sainte-Anne, C’JAAD, pôle PEPIT, Inserm 1266, Institut de Psy­ schizophrenia. Psychol. Med. 42, 743–755. chiatrie (CNRS 3557) Paris, France First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B., 2002. Structured Clinical Interview for DSM-IV-TR Axis I Disorders, Research Version, Patient Edition (SCID-1/P). 18 Mental Health Center Copenhagen and Center for Clinical Inter­ Biometrics Research, New York State Psychiatric Institute, New York, NY, USA. vention and Neuropsychiatric Schizophrenia Research, CINS, Fiszdon, J., Choi, K., Bell, M., Choi, J., Silverstein, S., 2016. Cognitive remediation for Mental Health Center Glostrup, Mental Health Services in the individuals with psychosis: efficacy and mechanisms of treatment effects. Psychol. Med. 46, 3275–3289. Capital Region of Copenhagen, University of Copenhagen. Fusar-Poli, P., Deste, G., Smieskova, R., Barlati, S., Yung, A.R., Howes, O., Stieglitz, R.-D., 19 Centre for Neuropsychiatric Schizophrenia Research (CNSR) & Vita, A., McGuire, P., Borgwardt, S., 2012. Cognitive functioning in prodromal Centre for Clinical Intervention and Neuropsychiatric Schizo­ psychosis: a meta-analysis. Arch. Gen. Psychiatry 69, 562–571. Giuliano, A.J., Li, H., Mesholam-Gately, R.I., Sorenson, S.M., Woodberry, K.A., phrenia Research (CINS), Mental Health Centre Glostrup, Uni­ Seidman, L.J., 2012. Neurocognition in the psychosis risk syndrome: a quantitative versity of Copenhagen, Glostrup, Denmark. and qualitative review. Curr. Pharm. Des. 18, 399–415. 20 Department of Psychiatry and Psychotherapy, Faculty of Medi­ Glenthøj, L.B., Hjorthøj, C., Kristensen, T.D., Davidson, C.A., Nordentoft, M., 2017. The effect of cognitive remediation in individuals at ultra-high risk for psychosis: a cine and University Hospital, University of Cologne, Cologne, systematic review. NPJ Schizophr. 3, 1–8. Germany. Glenthøj, L.B., Kristensen, T.D., Wenneberg, C., Hjorthøj, C., Nordentoft, M., 2021. 21 Psyberlin, Berlin, Germany. Predictors of remission from the ultra-high risk state for psychosis. Early Interv. 22 Medical University of Vienna, Department of Psychiatry and Psychiatry 15, 104–112. Goldman, H.H., Skodol, A.E., Lave, T.R., 1992. Revising axis V for DSM-IV: a review of Psychotherapy. measures of social functioning. Am. J. Psychiatry 149, 9. 23 Department of Psychiatry and Neuropsychology, School for Green, M.F., Braff, D.L., Kern, R.S., Mintz, J., 2000. Neurocognitive deficits and Mental Health and Neuroscience, Maastricht University Medical functional outcome in schizophrenia: are we measuring the 'right stuff'? Schizophr. Bull. 26, 119–136. Centre, P.O. Box 616, 6200 MD 464 Maastricht, The Netherlands. Green, M.F., Kern, R.S., Heaton, R.K., 2004. Longitudinal studies of cognition and 24 Mondriaan Mental Health Trust, P.O. Box 4436 CX Heerlen, The functional outcome in schizophrenia: implications for MATRICS. Schizophr. Res. 72, Netherlands 41–51. Grimes, K.M., Zanjani, A., Zakzanis, K.K., 2017. Memory impairment and the mediating role of task difficulty in patients with schizophrenia. Psychiatry Clin. Neurosci. 71, 600–611. Hall, R.C., 1995. Global assessment of functioning: a modified scale. Psychosomatics 36, Appendix A. Supplementary data 267–275. Harrison, E., Rose, D., 2006. The european socio-economic classification (ESeC) user Supplementary data to this article can be found online at https://doi. guide [Online]. Available: https://www.iser.essex.ac.uk/files/esec/guide/docs/User org/10.1016/j.scog.2021.100222. Guide.pdf. (Accessed November 2020). Hauser, M., Zhang, J.-P., Sheridan, E.M., Burdick, K.E., Mogil, R., Kane, J.M., Auther, A., Carrión, R.E., Cornblatt, B.A., Correll, C.U., 2017. Neuropsychological test References performance to enhance identification of subjects at clinical high risk for psychosis and to be most promising for predictive algorithms for conversion to psychosis: a meta-analysis. J. Clin. Psychiatry 78. Aas, M., Dazzan, P., Mondelli, V., Melle, I., Murray, R.M., Pariante, C.M., 2014. Henry, J., Crawford, J., 2005. A meta-analytic review of verbal fluency deficits in A systematic review of cognitive function in first-episode psychosis, including a schizophrenia relative to other neurocognitive deficits. Cogn. Neuropsychiatry 10, discussion on childhood trauma, stress, and inflammation. Front. Psychiatry 4, 182. 1–33. Addington, J., Barbato, M., 2012. The role of cognitive functioning in the outcome of Kahn, R.S., Keefe, R.S., 2013. Schizophrenia is a cognitive illness: time for a change in those at clinical high risk for developing psychosis. Epidemiol. Psychiatric Sci. 21, focus. JAMA Psychiatry 70, 1107–1112. 335–342. Lee, T.Y., Shin, Y.S., Shin, N.Y., Kim, S.N., Jang, J.H., Kang, D.-H., Kwon, J.S., 2014. Addington, J., Farris, M., Stowkowy, J., Santesteban-Echarri, O., Metzak, P., Kalathil, M. Neurocognitive function as a possible marker for remission from clinical high risk for S., 2019a. Predictors of transition to psychosis in individuals at clinical high risk. psychosis. Schizophr. Res. 153, 48–53. Curr. Psychiatry Rep. 21, 1–10. Lejeune, J.A., Northrop, A., Kurtz, M.M., 2021. A meta-analysis of cognitive remediation Addington, J., Stowkowy, J., Liu, L., Cadenhead, K.S., Cannon, T.D., Cornblatt, B.A., for schizophrenia: efficacy and the role of participant and treatment factors. McGlashan, T.H., Perkins, D.O., Seidman, L.J., Tsuang, M.T., 2019b. Clinical and Lin, A., Wood, S., Nelson, B., Brewer, W., Spiliotacopoulos, D., Bruxner, A., Broussard, C., functional characteristics of youth at clinical high-risk for psychosis who do not Pantelis, C., Yung, A., 2011. Neurocognitive predictors of functional outcome two to transition to psychosis. Psychol. Med. 49, 1670–1677. 13 years after identification as ultra-high risk for psychosis. Schizophr. Res. 132, Becker, H.E., Nieman, D.H., Dingemans, P.M., van de Fliert, J.R., de Haan, L., Linszen, D. 1–7. H., 2010. Verbal fluency as a possible predictor for psychosis. Eur. Psychiatry 25, Maccabe, J.H., Brébion, G., Reichenberg, A., Ganguly, T., McKenna, P.J., Murray, R.M., 105–110. David, A.S., 2012. Superior intellectual ability in schizophrenia: neuropsychological Blyler, C.R., Gold, J.M., Iannone, V.N., Buchanan, R.W., 2000. Short form of the WAIS-III characteristics. Neuropsychology 26, 181. for use with patients with schizophrenia. Schizophr. Res. 46, 209–215. 7 E.P. Hedges et al. Schizophrenia Research: Cognition 28 (2022) 100222 Mallett, R., 1997. Sociodemographic schedule. Section of Social Psychiatry, Institute of Seidman, L.J., Shapiro, D.I., Stone, W.S., Woodberry, K.A., Ronzio, A., Cornblatt, B.A., Psychiatry, London. Addington, J., Bearden, C.E., Cadenhead, K.S., Cannon, T.D., 2016. Association of Menghini-Müller, S., Studerus, E., Ittig, S., Valmaggia, L.R., Kempton, M.J., Van der neurocognition with transition to psychosis: baseline functioning in the second phase Gaag, M., De Haan, L., Nelson, B., Bressan, R.A., Barrantes-Vidal, N., Jantac, C., of the North American prodrome longitudinal study. JAMA Psychiatry 73, Nordentoft, M., Ruhrmann, S., Sachs, G., Rutten, B.P., Van Os, J., Riecher- 1239–1248. Rössler, A., Group, E.-G.H.R.S., 2020. Sex differences in cognitive functioning of Sheffield, J.M., Karcher, N.R., Barch, D.M., 2018. Cognitive deficits in psychotic patients at-risk for psychosis and healthy controls: Results from the European Gene- disorders: a lifespan perspective. Neuropsychol. Rev. 28, 509–533. Environment Interactions study, 63, e25. Simon, A.E., Grädel, M., Cattapan-Ludewig, K., Gruber, K., Ballinari, P., Roth, B., Mesholam-Gately, R.I., Giuliano, A.J., Goff, K.P., Faraone, S.V., Seidman, L.J., 2009. Umbricht, D., 2012. Cognitive functioning in at-risk mental states for psychosis and Neurocognition in first-episode schizophrenia: a meta-analytic review. 2-year clinical outcome. Schizophr. Res. 142, 108–115. Neuropsychology 23, 315. van der Gaag, M., van den Berg, D., Ising, H., 2019. CBT in the prevention of psychosis Niendam, T.A., Bearden, C.E., Zinberg, J., Johnson, J.K., O'Brien, M., Cannon, T.D., and other severe mental disorders in patients with an at risk mental state: a review 2007. The course of neurocognition and social functioning in individuals at ultra and proposed next steps. Schizophr. Res. 203, 88–93. high risk for psychosis. Schizophr. Bull. 33, 772–781. van Os, S., Rutten, B., Myin-Germeys, I., 2014. European network of national networks Pukrop, R., Schultze-Lutter, F., Ruhrmann, S., Brockhaus-Dumke, A., Tendolkar, I., studying gene-environment InteractionsIdentifying gene–environment interactions Bechdolf, A., Matuschek, E., Klosterkötter, J., 2006. Neurocognitive functioning in in schizophrenia: contemporary challenges for integrated, large-scale investigations. subjects at risk for a first episode of psychosis compared with first-and multiple- Schizophr. Bull. 40, 729–736. episode schizophrenia. J. Clin. Exp. Neuropsychol. 28, 1388–1407. Velthorst, E., Levine, S.Z., Henquet, C., de Haan, L., van Os, J., Myin-Germeys, I., Salazar de Pablo, G., Radua, J., Pereira, J., Bonoldi, I., Arienti, V., Besana, F., Soardo, L., Reichenberg, A., 2013. To cut a short test even shorter: reliability and validity of a Cabras, A., Fortea, L., Catalan, A., Vaquerizo-Serrano, J., Coronelli, F., Kaur, S., da brief assessment of intellectual ability in schizophrenia—a control-case family study. Silva, J., Shin, J.I., Solmi, M., Brondino, N., Politi, P., McGuire, P., Fusar-Poli, P., Cogn. Neuropsychiatry 18, 574–593. 2021. Probability of transition to psychosis in individuals at clinical high risk: an Yung, A.R., Yung, A.R., Pan Yuen, H., McGorry, P.D., Phillips, L.J., Kelly, D., updated meta-analysis. JAMA Psychiat. 78 (9), 970–978. Dell'olio, M., Francey, S.M., Cosgrave, E.M., Killackey, E., 2005. Mapping the onset Schaefer, J., Giangrande, E., Weinberger, D.R., Dickinson, D., 2013. The global cognitive of psychosis: the comprehensive assessment of at-risk mental states. Aust. N. Z. J. impairment in schizophrenia: consistent over decades and around the world. Psychiatry 39, 964–971. Schizophr. Res. 150, 42–50. 8