Welcome to the Schizophrenia Resource Centre

Welcome, this website is intended for healthcare professionals in EMEA with an interest in the treatment of schizophrenia. By clicking the link below you are declaring and confirming that you are a healthcare professional

You are here

10 year course of IQ in first-episode psychosis: Relationship between duration of psychosis and long-term intellectual trajectories

Psychiatry Research, 3, 225, pages 515 - 521


A substantial proportion of patients suffering from schizophrenia-spectrum disorders (SSDs) exhibit a general intellectual impairment at illness onset, but the subsequent intellectual course remains unclear. Relationships between accumulated time in psychosis and long-term intellectual functioning are largely uninvestigated, but may identify subgroups with different intellectual trajectories. Eighty-nine first-episode psychosis patients were investigated on IQ at baseline and at 10-years follow-up. Total time in psychosis was defined as two separate variables; Duration of psychosis before start of treatment (i.e. duration of untreated psychosis: DUP), and duration of psychosis after start of treatment (DAT). The sample was divided in three equal groups based on DUP and DAT, respectively. To investigate if diagnosis could separate IQ-trajectories beyond that of psychotic duration, two diagnostic categories were defined: core versus non-core SSDs. No significant change in IQ was found for the total sample. Intellectual course was not related to DUP or stringency of diagnostic category. However, a subgroup with long DAT demonstrated a significant intellectual decline, mainly associated with a weaker performance on test of immediate verbal recall/working memory (WAIS-R Digit Span). This indicates a relationship between accumulated duration of psychosis and long-term intellectual course, irrespective of diagnostic category, in a significant subgroup of patients.



  • We examine the 10 year course of IQ in 89 first-episode psychosis patients.
  • We explore the relationship between course of IQ and illness severity.
  • Accumulated duration of psychosisafterstart of treatment is related to IQ-decline.
  • Course of IQ is not related to duration of psychosisbeforestart of treatment.
  • We identify a considerable subgroup of patients in need of targeted treatment.

Keywords: Neurocognition, Psychosis, Longitudinal, Subgroups, Positive symptoms, WAIS-R.

1. Introduction

Deficits in global measures of cognition are commonly reported in patients with chronic schizophrenia, such as lower intelligence quotient (IQ) ( Heinrichs and Zakzanis, 1998 ). This is observed at the first episode ( Mesholam-Gately et al., 2009 ), and even during the prodromal (Fusar-Poli et al, 2012 and Giuliano et al, 2012) and premorbid periods ( Woodberry et al., 2008 ), and indications of an intellectual decline have been reported from childhood into adulthood in schizophrenia patients ( Meier et al., 2013 ). However, the long-term progression of intellectual functioning in adultsafterillness onset remains controversial. Also, the relationship among intellectual course, symptomatic severity, and schizophrenia spectrum disorders (SSDs) remains unclear.

Whether cognitive deficits develop up to, or after, the onset of psychosis was recently investigated in a meta-analysis of 25 studies including ultra-high risk and first episode psychosis (FEP) patients ( Bora and Murray, 2013 ). No evidence of a decline was found, instead both patient groups and healthy controls improved significantly over time. The only exception was in verbal working memory, where improvement was non-significant for FEP but pronounced in healthy controls, indicating a different course for selected cognitive domains in FEP. However, a majority of the studies in this meta-analysis had relatively short follow-up intervals, between 1 and 2 years. Studies of the intellectual course in FEP-patients spanning 5 years or more after start of treatment are relatively rare. One meta-analysis of eight studies ( Hedman et al., 2013 ) with follow-up durations between 1 and 10 years have reported less improvement in IQ in schizophrenia patients compared to healthy controls over time. However, this meta-analysis included studies of both chronic, early onset and FEP, thus complicating interpretation of results. Studies of decline have also been reported, at least for subgroups of older and institutionalized patients ( Kurtz, 2005 ).

Heterogeneity in cognitive expression is a key aspect of SSDs, and schizophrenia has long been characterized by its lack of a uniform signature pattern of neurocognitive impairment ( Kremen et al., 2004 ). Also heterogeneity in clinical and symptom dimensions is frequently reported ( Collin et al., 2012 ). The relationship between cognition and symptomatology has mainly been linked to negative symptoms, although a link to positive symptoms has also been described (Braw et al, 2008 and Ventura et al, 2010).

Earlier research has proposed that active psychosis reflects a period of disease progression insofar as many patients may not return to their previous level of function, and treatment refractoriness may emerge (Lieberman et al, 1996 and Wyatt, 1991). Alternatively, a small cognitive decline may occur during active psychosis, followed by a slow recovery to almost the same level as before the psychosis. If this is the case, then cognitive functioning would decline minimally after each psychotic episode, a process of which the effect can only be measured clearly after several episodes ( Becker et al., 2010 ). Furthermore, different neurocognitive domains may be differently affected by illness progression ( Frommann et al., 2011 ); an issue that also awaits more longitudinal research.

To identify associations between illness progression and intellectual course may be a useful approach to help in understanding the relationship between the two, if any.

Attempts to test the hypothesis of such a relationship are complicated by lack of sufficiently long follow-up intervals, adequate data of symptom progression, as well as the potential confounding effects of medication. Although some antipsychotic medications may give a small improvement in neurocognitive function ( Keefe et al., 2006 ), first generation anti-psychotics are not found to affect cognition to any meaningful degree ( Kahn and Keefe, 2013 ). Thus, the main challenge in the pursuit to investigate a possible connection between intellectual course and illness progression appears to be multi-assessment studies with follow-up intervals that extend beyond the first few years after onset.

In a previous study from the Scandinavian TIPS-project, we found that the course of verbal learning and working memory was negatively related to the number of psychotic episodes accumulated during the first 5 years follow-up ( Barder et al., 2013a ).

No studies have to our knowledge investigated the association between accumulated time in psychosis and intellectual course over a period extending 10 years after the start of treatment.

In the present study we investigate the 10 year course of IQ in a sample of FEP-patients, and explore whether it is related to illness severity as defined by the accumulated time in psychosis.

We aim to investigate whether the longitudinal course of IQ is differently related to duration of psychotic symptomsafterinitiation of treatment (duration of psychosis after start of treatment; DAT) compared to duration of psychotic symptomsbeforeinitiation of treatment (duration of untreated psychosis; DUP).

More specifically, we ask: Does estimated IQ remain stable 10 years after start of treatment? Is the accumulated duration of psychosis before (DUP) or after start of treatment (DAT), related to long-term course of IQ? Are subtests within the IQ-measure differentially related to the long-term course of illness? Finally, does duration of psychosis before or after start of treatment (DUP or DAT) separate long-term intellectual trajectories beyond the categories defined by diagnostic criteria?

2. Methods

2.1. The TIPS project

The present report originates from the Early Treatment and Intervention in Psychosis Study (TIPS), a prospective longitudinal study of the relationship between duration of untreated psychosis (DUP) and outcome in FEP. The study was carried out in four Scandinavian health care sectors; Oslo, Stavanger, Haugesund and Roskilde.

A total of 301 patients were included in the TIPS study. The patients were 15–65 years of age, met the DSM-IV criteria for non-organic psychosis, and were actively psychotic without previously receiving adequate antipsychotic treatment.

Research initiative was approved by the Regional Committee for Medical Research Ethics Health Region II and Health Region East in Norway, the Norwegian Data Inspectorate and the regional committee for science ethics region Sjælland, Denmark. Data were obtained in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki). After complete description of the study all participants gave written informed consent.

All patients were included in a defined treatment program ( Melle et al., 2004 ). Symptom ratings were obtained at the start of treatment, at 3 months, and at 1-year follow up. Patients were tested neuropsychologically for the first time after remission of the psychotic symptoms (defined as a score lower than 4 on the PANSS ( Kay et al., 1987 ) psychosis items, defined next paragraph) or following the 3-month follow-up, and re-assessments were performed four times during the follow-up. The IQ-measures were obtained at the baseline and the 10-year follow-up assessment, only.

2.2. Measures

2.2.1. Clinical instruments

The structured clinical interview for the DSM-IV (SCID) ( First et al., 1995 ) was used for diagnostic purposes. Trained clinical research personnel carried out diagnostic evaluations. Symptom levels were assessed with the Positive and Negative Syndrome Scale, PANSS ( Kay et al., 1987 ).

DUP was measured as the time from the first onset of psychotic symptoms (defined as the first week with a PANSS score of 4 or more on one of the Positive scale items P1-delusions, P3-halluciantory behavior, P5-grandiosity, P6-suspiciousness or General scale item 9-unusual thought content) to the start of first adequate treatment of psychosis (defined as start of adequate antipsychotic medication or admission to hospital for treatment of acute psychosis). All available sources, including semi-structured personal interviews with patients and relatives were used to ascertain the length of this period. In the rare case of previous short, self-remitting psychotic episodes, the lengths of these episodes were added to the DUP ( Melle et al., 2004 ).

Longitudinal measures were defined as follows: duration of psychosis: the number of weeks per year with a score of four or higher on any of the PANSS items above. Patients were followed along during the 2 first years. Data for year 3, 4 and 5 were collected retrospectively at the 5 year follow-up, and for the years 6–10 retrospectively at the 10 year follow-up. Data collected retrospectively at 5 and 10 year follow-up were based on patient information and all available clinical information, such as hospital discharge summaries and information form care givers where available.

Premorbid functioning was measured using the Premorbid Adjustment Scale (PAS) ( Cannon-Spoor et al., 1982 ). A previous analysis identified two premorbid dimensions: a social dimension consisting of PAS items social isolation and peer relationships, and an academic dimension which comprise school performance and school adaptation ( Larsen et al., 2004 ).

Alcohol and drug use was measured with the Clinician Alcohol/Drug Use Scale ( Drake et al., 1990 ) which assesses abuse on a scale from 1 (no use) to 5 (dependence with institutionalizing). Abuse was defined as a score of 3 or more.

Satisfactory inter-rater reliability was found with overall agreement for DSM-IV diagnostic categories at baseline, Kappa: 0.76. PANSS: ICC (1, 1): 0.88 for positive symptoms, 0.76 for negative symptoms, and 0.53 for general symptoms.

2.2.2. Neurocognitive measures

All patients were assessed with a comprehensive neurocognitive test battery at baseline ( Friis et al., 2002 ), and samples were re-assessed at 2-year ( Rund et al., 2007 ), 5-year ( Barder et al., 2013a ) and 10-year follow-up ( Barder et al., 2013b ).

The subtests Similarities, Block Design, and Digit Span from WAIS-R ( Wechsler, 1981 ) were used to calculate an IQ-estimate at baseline and at 10 year follow-up.

2.3. Subjects

Two hundred and thirteen patients completed the three subtests from WAIS-R at the baseline assessment, and of these, 78 patients were reassessed with the same measures at the 10 year follow-up, and had data on duration of psychosis from all assessments during the follow-up period. Eleven patients had baseline neuropsychological testing later than the baseline cut-off of 9 months after start of treatment (mean 12.5 months), but were also re-tested at 10 year follow-up. The mean time between start of treatment and the 10 year follow-up assessment was 10.3 years for the patients with baseline testing at 1 year follow-up (mean=3746 days, S.D.=291 days), as was close to the mean time for patients with baseline testing within 9 months (mean=3763 days, S.D.=341 days). Hence, the sample of 11 patients with late baseline test was included in the follow-up sample, giving a total of 89 patients.

The group of patients who completed the baseline assessment but dropped out during the follow-up, or did not complete the 10 year follow-up assessment will be referred to as the remaining sample (n=135).

Baseline demographic and clinical characteristics for the two samples are presented in Table 1 .

Table 1 Demographic and clinical characteristics at baseline.

Variable Follow-up sample (n=89) Remaining sample (n=135) χ²/ t/U
Age (M, S.D.) 28.0 (9.5) 28.5 (10.2) t(222)=0.4, p=0.681
Gender (Male) 47 (53%) 80 (59%) χ² (1, 224)=0.9, p=0.340
Education (M, S.D.) 12.3 (3.0) 11.8 (2.2) t(213)=1.4, p=0.191
IQ estimate (M, S.D.) 100.2 (10.2) 97.9 (10.8) t(219)=1.6, p=0.106
Duration of untreated psychosis (weeks) (Median, range) 8 (0–520) 16 (0–966) U=4973.5, p=0.029lowast
PANSS Positive score 20.2 (5.5) 20.4 (5.7) t(221)=0.3, p=0.795
PANSS Negative score 14.9 (7.3) 15.6 (6.7) t(221)=0.7, p=0.465
PAS social level childhood (M, S.D.) 1.1 (1.2) 1.0 (1.1) t(222)=0.4, p=0.705
PAS change in social level (M, S.D.) 0.9 (1.5) 0.7 (1.4) t(222)=0.8, p=0.401
PAS academic level childhood (M, S.D.) 1.7 (1.2) 1.8 (1.3) t(222)=0.5, p=0.652
PAS change in academic level (M, S.D.) 0.6 (1.4) 0.7 (1.2) t(222)=0.8, p=0.428
Alcohol abuse (N, %) 10 (11%) 18 (13%) χ² (1, 223)=0.3, p=0.628
Drug abuse (N, %) 16 (18%) 32 (24%) χ² (1, 223)=1.1, p=0.294
Diagnosis (DSM-IV)     χ² (6, 224)=3.9, p=0.687
- Schizophrenia 22 (25%) 38 (28%)
- Affective psychosis w/mood incongr. symp. 14 (16%) 16 (12%)
- Schizoaffective 9 (10%) 19 (14%)
- Schizophreniform 22 (25%) 30 (22%)
- Delusional disorder 4 (4%) 11 (8%)
- Brief psychotic episode 8 (9%) 7 (5%)
- Psychotic disorder NOS 10 (11%) 14 (10%)
 Total 89 135

The follow-up sample consisted of approximately the same number of men and women; they were in their late twenties and had an estimated average IQ. Symptom scores, diagnostic distribution, premorbid functioning and change from childhood level (PAS), did not differ between the groups at baseline, but the follow-up sample had a significantly shorter DUP. PANSS positive and negative scores at 10 year follow-up were low; 11.2 (4, 6) and 11.6 (4,9), respectively.

2.4. Subgroups defined by duration of untreated psychosis (DUP) and duration of psychosis after start of treatment (DAT)

In order to identify whether it is the timing of psychosis or the length spent in psychosis, total time in psychosis was defined as two separate variables, i.e. as duration of untreated psychosis (DUP) before inclusion in the study, and as duration of psychosis after start of treatment (DAT). The distributions of both the DUP and the DAT-variables were grossly skewed. As we had no clear hypothesis about any specific threshold, we chose to split each variable into three equally sized groups. There was a significantly different distribution of gender across the three DAT-groups; fewer males were found in the group with shortest DAT.

Baseline characteristics of the two groupings of duration are presented in Table 2 .

Table 2 Baseline characteristics for the follow-up sample (N=89) split by duration of untreated psychosis (DUP) and duration of psychosis after start of treatment (DAT).

Duration: Range (M, S.D.) DUP (N=89) DAT (N=89)
  (N=30) (N=30) (N=29) F/χ2 (N=30) (N=29) (N=30) F/χ2
  0–3 weeks (11.0, 22.1) 4–19 weeks (23.6, 40.0) 20–520 weeks (58.7, 111.7)   1–22 weeks (10.3, 5.7) 23–144 weeks (56.7, 30.8) 157–545 weeks (360.0, 128.4)  
Age 27.7 (10.1) 28.7 (10.0) 27.6 (8.5) F (2, 88)=0.13 29.5 (8.7) 29.5 (11.9) 25.0 (6.8) F (2, 88)=2.3
p=0.882 p=0.110
Gender (%male) 13 (43%) 16 (53%) 18 (62%) χ² (2, 89)=2.1 10 (33%) 19 (65%) 18 (60%) χ² (2, 89)=7.1
p=0.353 p=0.029*
Edu. 12.3 (2.3) 12.2 (3.5) 12.3 (3.2) F (2,82)=0.008 12.9 (2.9) 11.5 (3.1) 12.3 (3.0) F (2, 82)=1.8
P=0.992 p=0.175

2.5. Diagnostic subgroups

In order to investigate the discriminating effect of diagnostic category on long-term intellectual course, the sample was split based on the core group of SSDs; schizophrenia, schizophreniform and schizoaffective (hereafter referred to as the Core-SSDs) (n=53) and the non-Core SSDs (affective psychosis with mood-incongruent delusions, delusional disorder, brief psychotic episode, or psychotic disorder not otherwise specified) (n=36) at baseline. The same grouping was made at 10 year follow-up; Core-SSDs (n=67) and non-Core SSDs (n=22). This demonstrates an increase in the number of patients fulfilling the criteria for more severe diagnoses over the course of the follow-up interval.

A Core-SSD diagnosis was significantly more frequent in the group with longest DAT (75%), than the group with second longest DAT (67%) and shortest DAT (37%), as anticipated based on diagnostic criteria.

2.6. Medication

At the 10 year follow-up 62 patients (70%) were using antipsychotic medication whereas 27 were medication free. Fifteen patients (17% of the sample) were using medication through the entire follow-up period. The rate of patients on medication at 10 years did not differ significantly from the remaining sample; 70% versus 71% respectively (χ² (1, 147) =0.02,P=0.894). Dosage of medication was measured as Defined Daily Dosage (DDD): One DDD represents an ordinary maintenance dose of a medication for its standard use. (Guidelines for ATC classification and DDD assignment. WHO Collaborating Centre for Drug Statistics Methodology, Oslo 1998. ISSN 0802-8753, ISBN 82-90312-28-8.)

The DDD of antipsychotic medication at 10 year was not significantly different between the follow-up sample; 0.9 (S.D.: 1.0) and the remaining sample; 1.1 (S.D.: 1.3) (t(144)=0.9,P=0.367). There was no difference in duration of periods with medical treatment between the follow-up and the remaining sample for either assessment point.

2.7. Attrition and missing data

At 10 year follow-up, two persons had missing data on Similarities, one person had missing data on Digit Span, and five persons had missing data on Block Design. Missing data were replaced by the group mean. This applied for less than three percent of the total number of subtests performed. The IQ-estimate was calculated after missing scores were replaced.

At 10 year follow-up, two patients were missing data on duration of psychosis at two of the follow-up assessments, and one patient had missing data on three assessments. We replaced the missing data with the mean duration of psychosis of the remaining years from start of year 2 until follow-up.

2.8. Statistical methods

Analyses were conducted using the statistical package SPSS (PAWS) for Windows (version 18).

Group differences were evaluated witht-tests and Mann–WhitneyU-tests for continuous variables, and chi-square tests for categorical variables.

A one-way repeated measure analysis of variance (ANOVA) was conducted to analyze the course of estimated IQ over the 10 year follow-up interval.

In order to investigate the hypothesis of an association between intellectual course and duration of psychosis, two separate repeated measure ANOVAs were carried out, one for each of the duration-variables. Time was used as within-subjects factor, and DUP and DAT as the between-subjects factor, respectively.

To explore further if the three WAIS-subtests within the IQ-measure were related to the duration-variables, two repeated measures multiple analysis of variance (MANOVA) were performed, and further followed up by separate repeated measure ANOVAs

To explore if a baseline diagnosis of a Core versus non-core SSDs could identify separate intellectual trajectories over time; an ANOVA was performed with Time as within-subjects factor, and the diagnostic groups as between-subject factor. The same analysis was carried out with diagnostic groups at 10 years as the between-subjects factor.

Follow-up analyses of co-variance (ANCOVAs, MANCOVA) were performed in order to control for the effect of gender over time.

3. Results

A small but non-significant improvement in performance of IQ was found in the follow-up sample; 100.2 (S.D.: 10.1) at baseline, and 101.2 (S.D.: 11.1) at 10 year follow-up (F(1, 88)=1.5,p=0.217,η2=0.02).

No interaction was found between time and DUP groups (F(2, 86)=2.2,p=0.116,η2=0.05). However, a significant interaction was found between Time and DAT groups (F(2, 86)=3.7,p=0.028,η2=0.08), as well as a significant main effect of DAT-groups (F(2, 86)=3.6,p=0.032,η2=0.08). No significant differences in IQ were found between the DAT-groups at baseline(F(2, 88)=1.26, p=0.288).

Post-hoc analyses showed that on general IQ, the group with longest DAT performed significantly weaker over time compared to the group with shortest DAT (p=0.034), and the middle group performed intermediate between the two. At 10 year follow-up, the mean IQ was 96.0 (S.D.: 12.3) in the group with longest DAT, and 105.0 (S.D.: 9.1) in the group with shortest DAT.

In order to explore associations amongst the three WAIS-R subtests and duration of psychosis, a MANOVA was performed for each of the DUP- and DAT-variables, respectively. No significant main effect of group, or group interactions were found for DUP-groups. However, significant main effects were found for DAT-groups and Subtests, and significant interactions were found for Time and Subtests, and Time and DAT.

Results from the MANOVA, scaled scores (M/SDs) and the ANOVAs for the three WAIS-R subtests across DAT-groups are presented in Table 3 .

Table 3 WAIS-R subtests from baseline to 10 year follow-up, split by DAT-groups, with Means (S.D.) and ANOVAs reported.

Duration/weeks DAT (N=89)
  1–22 (N=30) 23–144 (N=29) 157–545 (N=30) Group (d.f.=2, 86) Time (d.f.=1, 86) Time×Group (d.f.=2, 86)
  Baseline 10 year Baseline 10 year Baseline 10 year F F F η2
Similarities a 11.1 (2.8) 11.2 (2.5) 10.8 (2.4) 11.0 (2.5) 9.9 (2.7) 9.5 (3.0) 3.6 lowast 0.02 0.4 0.01
Block design a 11.4 (2.5) 11.0 (3.2) 11.2 (3.0) 11.0 (2.7) 10.8 (3.7) 9.8 (3.4) 0.9 4.5 lowast 0.7 0.02
Digit span a 8.7 (2.4) 10.8 (3.0) 8.5 (1.9) 9.6 (3.5) 8.1 (2.1) 8.4 (2.9) 3.0 20.1 lowastlowast 4.1 lowast 0.09

a Scaled scores.

lowast p<0.05.

lowastlowast p<0.01.

MANOVA for the three subtests and the three duration-groups over time: Subtests (F(2, 85)=24.0,p<0.000,η2=0.36. Time (F(1, 86)=1.7,p=0.200,η2=0.02. DAT-groups (F(2, 86)=3.6,p=0.032,η2=0.08. Subtests×Time (F(2, 85)=8.7,p<0.000,η2=0.17. DAT×Time (F(2, 86)=3.7,p=0.028,η2=0.08. DAT×Subtest (F(4, 170)=0.4,p=0.819,η2=0.01. DAT×Subtests×Time (F(4, 170)=0.6,p=0.628,η2=0.2.

Follow-up analyses of the WAIS-R subtests identified only one significant interaction effect; between the subtest Digit Span and DAT (F(2, 86)=4.1,p=0.021,η2=0.09). No interaction effects were found for the other two subtests. The group with longest DAT demonstrated a significantly weaker performance on the subtest Digit Span compared to the group with shortest DAT, and the middle group did not differ significantly from the other two.

Analyzing the course of IQ for diagnostic groups at baseline gave no significant interactions (F(1, 87)=1.8,p=0.181,η2=0.02), neither did the diagnostic grouping at 10 year follow-up (F(1, 87)=2.8,p=0.099,η2=0.03).

All key interactions remained significant after controlling for gender.

With regard to medication, no significant correlations were found in the follow-up sample between IQ and DDD (antipsychotic (r=0.15), antidepressant (r=0.14), or mood stabilizer (r=0.05), nor between DDDs and any of the three WAIS-R subtests at 10 year follow-up.

4. Discussion

We found longitudinal stability of IQ in the total group of 89 FEP patients 10 years after start of treatment. This finding is in agreement with other studies of intellectual course in SSDs (Gold et al, 1999 and Hedman et al, 2013).

No association was found between DUP and course of IQ, which corresponds to an extensive amount of studies reporting no association between DUP and neurocognition at the onset of first episode (Ho et al, 2003, Rund, 2014, and Rund et al, 2004). The present study expands this notion, as the DUP did not differentiate intellectual trajectories over the 10 following years.

However, dividing the FEP-sample into groups based on duration of psychosis after start of treatment (DAT) revealed intellectual trajectories that differed significantly from the stable course found in the total group. The group with longest DAT demonstrated a negative intellectual course, in contrast to the improving trend found in the other two groups.

Two factors may be emphasized regarding the capacity of DAT to distinguish intellectual trajectories over that of DUP; one key factor may be the considerably longer time span among the DAT-groups. More distinct subgroups may emerge over time, and may be more easily identified in still longer follow-up intervals. Second, the core difference between DUP and DAT is the initiation of treatment. The DAT-groups represent patients with different degrees of treatment response and/or treatment adherence.

Of particular clinical interest is the group with the most severe illness progression, characterized by a longer DAT as well as a long-term decline in performance compared with the other two groups.

The finding of an association between an increasing DAT and a weaker performance on a subtest of IQ may help identify clinical subgroups in need of targeted treatment. The group with longest DAT presented with a lower IQ at baseline, although not significantly different from the other DAT-groups until later in the illness progression. Hence, in retrospect from 10-year follow-up, these findings imply that a lower baseline IQ may indicate vulnerability for a more severe course of illness as well as a weaker performance over time on a subtest of IQ tapping working memory. Also of note is that this relationship does not support the idea that psychopathology and cognition are independent domains in psychotic disorders.

A number of studies have failed to find a link between illness severity and neurocognitive course (Brissos et al, 2011 and Tamminga et al, 1998) but very few (if any) have been able to investigate this relationship over as long as 10 years after the start of treatment, and the finding of a long-term relationship between positive symptoms and a decline in IQ has to the best of our knowledge not previously been described.

Analyses of the subtests within the IQ-measure identified the WAIS-R Digit Span as the subtest largely accountable for discriminating among the DAT-subgroups. The Digit Span reflects immediate verbal recall and working memory in healthy individuals ( Lezak et al., 2004 ) and in schizophrenia patients, and involves temporary on-line storage and mental manipulation of information ( Nuechterlein et al., 2004 ). Moreover, the domain of Working Memory is found to be one of the fundamental neurocognitive factors that predict return to work or school after outpatient clinical stabilization, thereby representing essential deficits that need to be targeted in attempts to prevent chronic disability ( Nuechterlein et al., 2011 ).

Our results on the subtest Digit Span do not necessarily imply a stable decrease in IQ. Performance on the Digit Span may be more vulnerable to the degree of psychosis at the time of testing than the other subtests. However, symptom level was not higher at the 10 year follow-up assessment, so it seems unlikely that the decrease in test scores could be explained (only) by active psychosis at the time of testing.

An alternative discriminating factor for both IQ and illness severity is diagnostic category. The subgroup with longest DAT included a majority of patients with a Core-SSD diagnosis at baseline. Yet, grouping these diagnoses together gave no significant discriminatory effect for long-term intellectual trajectories. Neither did the same grouping based on 10 year diagnosis, which reduces the possibility that diagnostic conversion may explain the lack of a significant interaction effect. Consequently, even if patients with a Core SSD diagnosis constitute a majority of the longest DAT group, these diagnoses are also represented in the groups with shorter DAT, illustrating the clinical heterogeneity within the FEP-sample in general, and within the Core SSD-diagnoses in particular. This finding is in agreement with a meta-analysis reporting similar neurocognitive trajectories across FEP diagnoses in the premorbid and early stages of the disease ( Bora and Murray, 2013 ). Further, a recent study of bipolar and schizophrenia patients found that neurocognitive dysfunction was more determined by a history of psychosis than by DSM-IV diagnostic category or subtype ( Simonsen et al., 2011 ).

In the analyses of DAT we did not control for use of medication, because medication per definition is strongly correlated with illness duration and severity. Also, recent findings have indicated improved cognitive functioning associated with antipsychotic medication (typical and atypical) (Heinrichs, 2007, CATIE InvestigatorsNeurocognitive Working Group, 2007, and Woodward et al, 2005), which implies that the use of antipsychotic medication does not explain the present findings. The lack of significant correlations at 10 years between dosage and IQ, and dosage and WAIS-R subtests, supports this notion.

There may be several potential causal mechanisms behind the relationship of severe illness duration and an intellectual decline. The present study cannot explain these processes, as they are likely to be multifaceted in both timing and patterns. The psychotic process per se may act as a potential negative effect on cognition, although this may not be quantifiable in terms of intellectual assessments until a significant duration of psychosis is accumulated (e.g. (Becker et al, 2010 and Braw et al, 2008). Alternatively, since the group with longest DAT had a non-significant, but slightly lower IQ at baseline, a relative intellectual impairment could have been present before the start of treatment. This may correspond to the concept of a cognitive reserve ( Stern, 2002 ) as applied to neuropsychiatry ( Barnett et al., 2006 ), proposing that the higher the premorbid cognitive ability, the more resilient individuals are to the impact of cerebral dysfunction. It has been proposed that in schizophrenia, better cognitive reserve may result in fewer psychotic symptoms either because of superior reasoning skills or because of the ability to inhibit the abnormal neural processing that mediates psychotic symptoms ( Barnett et al., 2006 ).

Finally, there may be a separate factor influencing both intellectual decline and a longer duration of psychosis. Recent research on susceptibility genes has suggested that there may be a distinct subtype of schizophrenia that is characterized by pervasive cognitive deficit ( Hallmayer et al., 2005 ), and similarly, a potential schizophrenia subgroup associated with less impaired cognitive performance ( Walters et al., 2010 ).

Taken together, this illustrates the heterogeneity of psychotic disorders, which again may demonstrate why research often provides contradictory conclusions of the intellectual course in FEP. Thus, rather than attempting to establish the course of a non-existent “average” patient with a psychotic disorder, it appears valuable to consider the cognitive heterogeneity in patterns and course of impairments, and identify the factors that do or do not co-aggregate and form subgroups over time. Furthermore, as cognitive dysfunction is an important determinant of global functional outcome ( Kahn and Keefe, 2013 ), the findings in the present study emphasize the significance of cognitive remediation strategies, not limited to the early stages of illness.

The present study has several limitations, and a generalization of the findings to a larger population must be made with caution. First, data on duration of psychosis were collected retrospectively, and will automatically imply an element of uncertainty. Although this information is obtained from multiple sources, the data are still based on human recollection, in which a certain degree of uncertainty is unavoidable. Further, the calculation of IQ is based on three subtests only. This is a clear limitation, and in retrospect, several others of the WAIS-R subtests should have been included in the test-battery when the TIPS-study was initiated. Consequently, the main finding of the present study is not the IQ-scores per se, but how they change over time, and how this change is related to illness progression. Our sample is heterogeneous in terms of baseline diagnosis. This heterogeneity enabled an investigation, and subsequently a rejection, of the hypothesis that diagnostic category may discriminate between performance on IQ measures over time. Finally, the present study has no normal controls, and therefore cannot account for possible changes in controls. Still, the question of whether there are subgroups with divergent neurocognitive paths in the sample is not necessarily dependent upon a healthy control group; the patients׳ first assessment constitutes the reference point, and the performances in each subgroup constitute the basis for comparison. A major strength is that the 10 year assessment interval reduces potential practice effects considerably.


The authors thank the participants for their essential contribution to the study. This work was supported by the Norwegian National Research Council (# 133897/320 and #154642/320), the Norwegian Department of Health and Social Affairs , the National Council for Mental Health/Health and Rehabilitation (#1997/41 and #2002/306), Rogaland County and Oslo County (Drs Vaglum, Johannessen, Friis, Larsen, Melle, and Opjordsmoen). Also funded by the Theodore and Vada Stanley Foundation, the Regional Health Research Foundation for Eastern Region, Denmark; Roskilde County, Helsefonden, Lundbeck Pharma, Eli Lilly and Janssen-Cilag Pharmaceuticals, Denmark(Drs Simonsen and Haahr). Also supported by a National Alliance for Research on Schizophrenia and Depression (NARSAD) Distinguished Investigator Award and NIMH Grant MH-01654 (Dr. McGlashan) and a NARSAD Young Investigator Award (Dr. Larsen). Health South East (#2008001), Health West (#200202797-65 (Inge Joa) and #911313 (Regional Centre Clinical Psychosis Research).

The authors report no competing interests.


  • Barder et al., 2013a H.E. Barder, K. Sundet, B.R. Rund, J. Evensen, U. Haahr, W. Ten Velden Hegelstad, I. Joa, J.O. Johannessen, H. Langeveld, T.K. Larsen, I. Melle, S. Opjordsmoen, J.I. Rossberg, E. Simonsen, P. Vaglum, T. McGlashan, S. Friis. Neurocognitive development in first episode psychosis 5 years follow-up: associations between illness severity and cognitive course. Schizophrenia Research. 2013;149(1–3):63-69 10.1016/j.schres.2013.06.016 Crossref
  • Barder et al., 2013b H.E. Barder, K. Sundet, B.R. Rund, J. Evensen, U. Haahr, W. Ten Velden Hegelstad, I. Joa, J.O. Johannessen, J. Langeveld, T.K. Larsen, I. Melle, S. Opjordsmoen, J.I. Rossberg, E. Simonsen, P. Vaglum, T. McGlashan, S. Friis. Ten year neurocognitive trajectories in first-episode psychosis. Frontiers in Human Neuroscience. 2013;7:643 10.3389/fnhum.2013.00643
  • Barnett et al., 2006 J.H. Barnett, C.H. Salmond, P.B. Jones, B.J. Sahakian. Cognitive reserve in neuropsychiatry. Psychological Medicine. 2006;36(8):1053-1064 10.1017/S0033291706007501 Crossref
  • Becker et al., 2010 H.E. Becker, D.H. Nieman, S. Wiltink, P.M. Dingemans, J.R. van de Fliert, E. Velthorst, L. de Haan, T.A. van Amelsvoort, D.H. Linszen. Neurocognitive functioning before and after the first psychotic episode: does psychosis result in cognitive deterioration?. Psychological Medicine. 2010;40(10):1599-1606 10.1017/S0033291710000048 Crossref
  • Bora and Murray, 2013 E. Bora, R.M. Murray. Meta-analysis of cognitive deficits in ultra-high risk to psychosis and first-episode psychosis: do the cognitive deficits progress over, or after, the onset of psychosis?. Schizophrenia Bulletin. 2013;40(4):744-755 10.1093/schbul/sbt085
  • Braw et al., 2008 Y. Braw, Y. Bloch, S. Mendelovich, G. Ratzoni, G. Gal, H. Harari, A. Tripto, Y. Levkovitz. Cognition in young schizophrenia outpatients: comparison of first-episode with multiepisode patients. Schizophrenia Bulletin. 2008;34(3):544-554 10.1093/schbul/sbm115
  • Brissos et al., 2011 S. Brissos, V.V. Dias, V. Balanza-Martinez, A.I. Carita, M.L. Figueira. Symptomatic remission in schizophrenia patients: relationship with social functioning, quality of life, and neurocognitive performance. Schizophrenia Research. 2011;129(2–3):133-136 10.1016/j.schres.2011.04.001 Crossref
  • Cannon-Spoor et al., 1982 H.E. Cannon-Spoor, S.G. Potkin, R.J. Wyatt. Measurement of premorbid adjustment in chronic schizophrenia. Schizophrenia Bulletin. 1982;8(3):470-484 10.1093/schbul/8.3.470 Crossref
  • Collin et al., 2012 G. Collin, E.M. Derks, N.E. van Haren, H.G. Schnack, H.E. Hulshoff Pol, R.S. Kahn, W. Cahn. Symptom dimensions are associated with progressive brain volume changes in schizophrenia. Schizophrenia Research. 2012;138(2–3):171-176 10.1016/j.schres.2012.03.036 Crossref
  • Drake et al., 1990 R.E. Drake, F.C. Osher, D.L. Noordsy, S.C. Hurlbut, G.B. Teague, M.S. Beaudett. Diagnosis of alcohol use disorders in schizophrenia. Schizophrenia Bulletin. 1990;16(1):57-67 10.1093/schbul/16.1.57 Crossref
  • First et al., 1995 M.B. First, R.L. Spitzer, M. Gibbon, J.B.W. Williams. Structured Clinical Interview for DSM-IV Axis I Disorders, Patient Edition (SCID-P) (New York State Psychiatric Institute, Biometrics Research, New York, 1995) (version 2)
  • Friis et al., 2002 S. Friis, K. Sundet, B.R. Rund, P. Vaglum, T.H. McGlashan. Neurocognitive dimensions characterising patients with first-episode psychosis. British Journal of Psychiatry Supplement. 2002;43:S85-S90 10.1192/bjp.181.43.s85
  • Frommann et al., 2011 I. Frommann, R. Pukrop, J. Brinkmeyer, A. Bechdolf, S. Ruhrmann, J. Berning, P. Decker, M. Riedel, H.J. Moller, W. Wolwer, W. Gaebel, J. Klosterkotter, W. Maier, M. Wagner. Neuropsychological profiles in different at-risk states of psychosis: executive control impairment in the early—and additional memory dysfunction in the late—prodromal state. Schizophrenia Bulletin. 2011;37(4):861-873 10.1093/schbul/sbp155 Crossref
  • Fusar-Poli et al., 2012 P. Fusar-Poli, G. Deste, R. Smieskova, S. Barlati, A.R. Yung, O. Howes, R.D. Stieglitz, A. Vita, P. McGuire, S. Borgwardt. Cognitive functioning in prodromal psychosis: a meta-analysis. Archives of General Psychiatry. 2012;69(6):562-571 10.1001/archgenpsychiatry.2011.1592
  • Giuliano et al., 2012 A.J. Giuliano, H. Li, R.I. Mesholam-Gately, S.M. Sorenson, K.A. Woodberry, L.J. Seidman. Neurocognition in the psychosis risk syndrome: a quantitative and qualitative review. Current Pharmaceutical Design. 2012;18(4):399-415 10.2174/138161212799316019
  • Gold et al., 1999 S. Gold, S. Arndt, P. Nopoulos, D.S. O’Leary, N.C. Andreasen. Longitudinal study of cognitive function in first-episode and recent-onset schizophrenia. American Journal of Psychiatry. 1999;156(9):1342-1348 (doi: http://www.ncbi.nlm.nih.gov/pubmed/10484943)
  • Hallmayer et al., 2005 J.F. Hallmayer, L. Kalaydjieva, J. Badcock, M. Dragovic, S. Howell, P.T. Michie, D. Rock, D. Vile, R. Williams, E.H. Corder, K. Hollingsworth, A. Jablensky. Genetic evidence for a distinct subtype of schizophrenia characterized by pervasive cognitive deficit. American Journal of Human Genetics. 2005;77(3):468-476 10.1086/432816 Crossref
  • Hedman et al., 2013 A.M. Hedman, N.E. van Haren, C.G. van Baal, R.S. Kahn, H.E. Hulshoff Pol. IQ change over time in schizophrenia and healthy individuals: a meta-analysis. Schizophrenia Research. 2013;146(1–3):201-208 10.1016/j.schres.2013.01.027 Crossref
  • Heinrichs, 2007 R.W. Heinrichs. Cognitive improvement in response to antipsychotic drugs: neurocognitive effects of antipsychotic medications in patients with chronic schizophrenia in the CATIE Trial. Archives of General Psychiatry. 2007;64(6):631-632 10.1001/archpsyc.64.6.631 Crossref
  • Heinrichs and Zakzanis, 1998 R.W. Heinrichs, K.K. Zakzanis. Neurocognitive deficit in schizophrenia: a quantitative review of the evidence. Neuropsychology. 1998;12(3):426-445 10.1037//0894-4105.12.3.426 Crossref
  • Ho et al., 2003 B.C. Ho, D. Alicata, J. Ward, D.J. Moser, D.S. O’Leary, S. Arndt, N.C. Andreasen. Untreated initial psychosis: relation to cognitive deficits and brain morphology in first-episode schizophrenia. American Journal of Psychiatry. 2003;160(1):142-148 10.1176/appi.ajp.160.1.142 Crossref
  • Kahn and Keefe, 2013 R.S. Kahn, R.S. Keefe. Schizophrenia is a cognitive illness: time for a change in focus. JAMA Psychiatry. 2013;70(10):1107-1112 10.1001/jamapsychiatry.2013.155 Crossref
  • Kay et al., 1987 S.R. Kay, A. Fiszbein, L.A. Opler. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophrenia Bulletin. 1987;13(2):261-276 10.1093/schbul/13.2.261 Crossref
  • CATIE InvestigatorsNeurocognitive Working Group, 2007 R.S. Keefe, R.M. Bilder, S.M. Davis, P.D. Harvey, B.W. Palmer, J.M. Gold, H.Y. Meltzer, M.F. Green, G. Capuano, T.S. Stroup, J.P. McEvoy, M.S. Swartz, R.A. Rosenheck, D.O. Perkins, C.E. Davis, J.K. Hsiao, J.A. Lieberman, CATIE Investigators, Neurocognitive Working Group. Neurocognitive effects of antipsychotic medications in patients with chronic schizophrenia in the CATIE Trial. Archives of General Psychiatry. 2007;64(6):633-647 10.1001/archpsyc.64.6.633 Crossref
  • Keefe et al., 2006 R.S. Keefe, L.J. Seidman, B.K. Christensen, R.M. Hamer, T. Sharma, M.M. Sitskoorn, S.L. Rock, S. Woolson, M. Tohen, G.D. Tollefson, T.M. Sanger, J.A. Lieberman, H.R. Group. Long-term neurocognitive effects of olanzapine or low-dose haloperidol in first-episode psychosis. Biological Psychiatry. 2006;59(2):97-105 10.1016/j.biopsych.2005.06.022 Crossref
  • Kremen et al., 2004 W.S. Kremen, L.J. Seidman, S.V. Faraone, R. Toomey, M.T. Tsuang. Heterogeneity of schizophrenia: a study of individual neuropsychological profiles. Schizophrenia Research. 2004;71(2–3):307-321 10.1016/j.schres.2004.02.022 Crossref
  • Kurtz, 2005 M.M. Kurtz. Neurocognitive impairment across the lifespan in schizophrenia: an update. Schizophrenia Research. 2005;74(1):15-26 10.1016/j.schres.2004.07.005 Crossref
  • Larsen et al., 2004 T.K. Larsen, S. Friis, U. Haahr, J.O. Johannessen, I. Melle, S. Opjordsmoen, B.R. Rund, E. Simonsen, P.V. Vaglum, T.H. McGlashan. Premorbid adjustment in first-episode non-affective psychosis: distinct patterns of pre-onset course. British Journal of Psychiatry. 2004;185:108-115 10.1192/bjp.185.2.108 Crossref
  • Lezak et al., 2004 M.D. Lezak, D.B. Howieson, D.W. Looring. Neuropsychological Assessment (Oxford University Press, New York, 2004)
  • Lieberman et al., 1996 J.A. Lieberman, A.R. Koreen, M. Chakos, B. Sheitman, M. Woerner, J.M. Alvir, R. Bilder. Factors influencing treatment response and outcome of first-episode schizophrenia: implications for understanding the pathophysiology of schizophrenia. Journal of Clinical Psychiatry. 1996;57(Suppl. 9):S5-S9 (doi: http://www.ncbi.nlm.nih.gov/pubmed/8823344)
  • Meier et al., 2013 M.H. Meier, A. Caspi, A. Reichenberg, R.S. Keefe, H.L. Fisher, H. Harrington, R. Houts, R. Poulton, T.E. Moffitt. Neuropsychological decline in schizophrenia from the premorbid to the postonset period: evidence from a population-representative longitudinal study. American Journal of Psychiatry. 2013;171(1):91-101 10.1176/appi.ajp.2013.12111438
  • Melle et al., 2004 I. Melle, T.K. Larsen, U. Haahr, S. Friis, J.O. Johannessen, S. Opjordsmoen, E. Simonsen, B.R. Rund, P. Vaglum, T. McGlashan. Reducing the duration of untreated first-episode psychosis: effects on clinical presentation. Archives of General Psychiatry. 2004;61(2):143-150 10.1001/archpsyc.61.2.143 Crossref
  • Mesholam-Gately et al., 2009 R.I. Mesholam-Gately, A.J. Giuliano, K.P. Goff, S.V. Faraone, L.J. Seidman. Neurocognition in first-episode schizophrenia: a meta-analytic review. Neuropsychology. 2009;23(3):315-336 10.1037/a0014708 Crossref
  • Nuechterlein et al., 2004 K.H. Nuechterlein, D.M. Barch, J.M. Gold, T.E. Goldberg, M.F. Green, R.K. Heaton. Identification of separable cognitive factors in schizophrenia. Schizophrenia Research. 2004;72(1):29-39 10.1016/j.schres.2004.09.007 Crossref
  • Nuechterlein et al., 2011 K.H. Nuechterlein, K.L. Subotnik, M.F. Green, J. Ventura, R.F. Asarnow, M.J. Gitlin, C.M. Yee, D. Gretchen-Doorly, J. Mintz. Neurocognitive predictors of work outcome in recent-onset schizophrenia. Schizophrenia Bulletin. 2011;37(Suppl. 2):S33-S40 10.1093/schbul/sbr084 Crossref
  • Rund, 2014 B.R. Rund. Does active psychosis cause neurobiological pathology? A critical review of the neurotoxicity hypothesis. Psychological Medicine. 2014;40(8):1577-1590 10.1017/S0033291713002341 Crossref
  • Rund et al., 2007 B.R. Rund, I. Melle, S. Friis, J.O. Johannessen, T.K. Larsen, L.J. Midboe, S. Opjordsmoen, E. Simonsen, P. Vaglum, T. McGlashan. The course of neurocognitive functioning in first-episode psychosis and its relation to premorbid adjustment, duration of untreated psychosis, and relapse. Schizophrenia Research. 2007;91(1–3):132-140 10.1016/j.schres.2006.11.030 Crossref
  • Rund et al., 2004 B.R. Rund, I. Melle, S. Friis, T.K. Larsen, L.J. Midboe, S. Opjordsmoen, E. Simonsen, P. Vaglum, T. McGlashan. Neurocognitive dysfunction in first-episode psychosis: correlates with symptoms, premorbid adjustment, and duration of untreated psychosis. American Journal of Psychiatry. 2004;161(3):466-472 10.1176/appi.ajp.161.3.466 Crossref
  • Simonsen et al., 2011 C. Simonsen, K. Sundet, A. Vaskinn, A.B. Birkenaes, J.A. Engh, A. Faerden, H. Jonsdottir, P.A. Ringen, S. Opjordsmoen, I. Melle, S. Friis, O.A. Andreassen. Neurocognitive dysfunction in bipolar and schizophrenia spectrum disorders depends on history of psychosis rather than diagnostic group. Schizophrenia Bulletin. 2011;37(1):73-83 10.1093/schbul/sbp034 Crossref
  • Stern, 2002 Y. Stern. What is cognitive reserve? Theory and research application of the reserve concept. Journal of the International Neuropsychological Society. 2002;8(3):448-460 (doi: http://www.ncbi.nlm.nih.gov/pubmed/11939702) Crossref
  • Tamminga et al., 1998 C.A. Tamminga, R.W. Buchanan, J.M. Gold. The role of negative symptoms and cognitive dysfunction in schizophrenia outcome. International Clinical Psychopharmacology. 1998;13(Suppl. 3):S21-S26 10.1097/00004850-199803003-00004 Crossref
  • Ventura et al., 2010 J. Ventura, A.D. Thames, R.C. Wood, L.H. Guzik, G.S. Hellemann. Disorganization and reality distortion in schizophrenia: a meta-analysis of the relationship between positive symptoms and neurocognitive deficits. Schizophrenia Research. 2010;121(1–3):1-14 10.1016/j.schres.2010.05.033
  • Walters et al., 2010 J.T. Walters, A. Corvin, M.J. Owen, H. Williams, M. Dragovic, E.M. Quinn, R. Judge, D.J. Smith, N. Norton, I. Giegling, A.M. Hartmann, H.J. Moller, P. Muglia, V. Moskvina, S. Dwyer, T. O’Donoghue, B. Morar, M. Cooper, D. Chandler, A. Jablensky, M. Gill, L. Kaladjieva, D.W. Morris, M.C. O’Donovan, D. Rujescu, G. Donohoe. Psychosis susceptibility gene ZNF804A and cognitive performance in schizophrenia. Archives of General Psychiatry. 2010;67(7):692-700 10.1001/archgenpsychiatry.2010.81 Crossref
  • Wechsler, 1981 D. Wechsler. Wechsler Adult Intelligence Scale-Revised (New York Psychological Corp, New York, 1981)
  • Woodberry et al., 2008 K.A. Woodberry, A.J. Giuliano, L.J. Seidman. Premorbid IQ in schizophrenia: a meta-analytic review. American Journal of Psychiatry. 2008;165(5):579-587 10.1176/appi.ajp.2008.07081242 Crossref
  • Woodward et al., 2005 N.D. Woodward, S.E. Purdon, H.Y. Meltzer, D.H. Zald. A meta-analysis of neuropsychological change to clozapine, olanzapine, quetiapine, and risperidone in schizophrenia. International Journal of Neuropsychopharmacology. 2005;8(3):457-472 10.1017/S146114570500516X Crossref
  • Wyatt, 1991 R.J. Wyatt. Neuroleptics and the natural course of schizophrenia. Schizophrenia Bulletin. 1991;17(2):325-351 10.1093/schbul/17.2.325 Crossref


a Division of Mental Health and Addiction, Oslo University Hospital, N-0407 Oslo, Norway

b Department of Psychology, University of Oslo, P.O. 1094 Blindern, 0317 Oslo, Norway

c Vestre Viken Hospital Trust, 3004, Drammen, Norway

d Institute of Clinical Medicine, University of Oslo, P.O. 1171 Blindern, 0318 Oslo, Norway

e Psychiatric Research Unit, Zealand Region, Psychiatry Roskilde, Roskilde University and University of Copenhagen, Smedegade 10-16, 4000 Roskilde, Denmark

f Early Psychosis Intervention Center, Psychiatry Roskilde, Region Zealand, Smedegade 1-16, 4000 Roskilde, Denmark

g Regional Centre for Clinical Research in Psychosis, Psychiatric Division, Stavanger University Hospital, P.O. 8100, 4068 Stavanger, Norway

h Faculty of social sciences, University of Stavanger, Norway

i Institute of Psychiatry, University of Bergen, Jonas Lies vei 65, 5021 Bergen, Norway

j Department of Behavioral Sciences in Medicine, University of Oslo, P.O. 1130 Blindern, 0318 Oslo, Norway

k Department of Psychiatry, Yale University School of Medicine, Yale Psychiatric Research at Congress Place, 301 Cedar Street, New Haven, CT 06519, USA

lowast Corresponding author at: KG Jebsen Centre for Psychosis Research, Division of Mental Health and Addiction, Oslo University Hospital, Ullevaal, PO Box 4956 Nydalen, 0424 Oslo, Norway. Tel.: +47 98465289, +47 23027350; fax: +47 23027333.