Is it possible to use cognitive testing, which is brief, easily performed, reliable, and noninvasive, to identify individuals at risk for psychosis?
“George” is a 15-year-old male in the 10th grade who is being evaluated by his school psychologist for declining academic performance. Last year, he earned grades of all As and Bs, but this year he has mostly Bs and Cs, with one D in an advanced math class. George has a 4-year history of generalized anxiety disorder, for which he takes a low-dose selective serotonin reuptake inhibitor. He has no active medical problems. He does not currently use tobacco, alcohol, or illicit drugs but admits to trying marijuana once at a party in the past year. There is no history of physical, sexual, or verbal abuse. He has a maternal aunt with chronic schizophrenia. His mother had gestational diabetes mellitus during her pregnancy with him, which was otherwise without complications. She states that he had frequent infections as a child and seemed to “get sick easily.” He has been slightly more irritable recently. George reports having a few close friends.
On interview, his affect is slightly restricted, but it brightens when he talks about his hobby (origami). He reported hearing whispering voices on a few occasions at night, which he attributed to noises outside his bedroom window. He reports this academic year it is harder for him to concentrate at school. Cognitive testing was notable for impairment in processing speed (about half standard deviation) and borderline impairment in visual learning compared with age- and sex-matched peers. His mother is concerned about his risk of schizophrenia.
Cognitive deficits are a well-replicated finding across the course of illness, including the psychosis prodrome, in schizophrenia.1,2 There is evidence for cognitive decline in patients at clinical high risk (CHR) for psychosis compared with controls, with larger effect sizes in patients who transition to a psychotic disorder.3 There is also meta-analytic evidence for broad cognitive impairments in memory, attention, verbal and visuospatial abilities, social cognition, and executive function in CHR patients.3 Furthermore, cognitive deficits in CHR are independent predictors of conversion to psychosis and other outcomes, including role and social functioning.4,5 Although it is not yet possible to definitively predict patients at CHR who will subsequently develop a psychotic disorder, this area of research offers visions of secondary and even primary prevention of psychosis. Is it possible to use cognitive testing, which is brief, easily performed, reliable, and noninvasive, to identify individuals at risk for psychosis? Similarly, among those at CHR, can cognitive testing predict who will convert to psychosis? What role does cognitive testing play in the clinical care of patients like George?
Cui and colleagues6 aimed to relate baseline cognitive performance to a broad range of clinical outcomes at 1-year follow-up in patients at CHR of psychosis and controls from the ShangHai-At-Risk-for-Psychosis (SHARP) study, which launched in China in 2010. They hypothesized that patients with CHR who converted to psychosis (CHR-C) would have greater cognitive impairments than those who did not convert (CHR-NC), and both CHR groups would have greater cognitive impairments than healthy controls. They also analyzed outcomes in patients at CHR based on clinical status (trichotomized as remission, symptomatic, or poor outcome).
A total of 217 participants at CHR and 133 healthy controls from the SHARP study completed the MATRICS Consensus Cognitive Battery (MCCB) at baseline and provided complete clinical data. These patients represented more than 90% of the patients who provided 1-year assessment data. The mean age of participants was aged 18.6 years; 47% were male; and all were lifetime psychotropic naïve at study entry. Subsequently, 80% of patients at CHR received antipsychotic medications. Inclusion criteria for all participants were aged 13 to 45 years, at least 6 years of education, and an available legal guardian for patients 18 years or younger. Participants with an IQ less than 70, severe somatic disease, and lifetime substance dependence were excluded. Psychotic disorder or prodromal symptoms in healthy controls were ruled out by structured clinical interview.
Raw scores on the MCCB were converted to Z-scores based on healthy control scores. The CHR group was divided into CHR-NC (n = 155) and CHR-C (n = 41) based on the 1-year follow-up assessment. In other analyses, the CHR group was trichotomized to 3 clinical outcomes at 1-year: CHR-remission (n = 102), CHR-symptomatic (n = 37), and CHR-poor outcome (n = 57). The CHR-remission group included patients with spontaneous remission or improvement with antipsychotic medication to mildly impaired ranges or better (Structured Interview for Prodromal Syndromes [SIPS] positive symptom scores ≤ 2) and global function at least 60 at follow-up. The CHR-symptomatic group included medicated participants with SIPS positive symptom scores of 3 to 5 and global function of at least 60 at follow-up. The CHR-poor outcome group included those who converted to psychosis (n = 41) and medicated patients with unremitting positive symptoms (SIPS positive symptom scores of 3 to 5) and poor global function (less than 60) at follow-up (n = 16). Data were analyzed using multivariate analysis of variance with Bonferroni-adjusted post hoc pair-wise comparisons between groups for MCCB scores. Effect sizes were calculated using Cohen’s d.
There were no significant differences regarding age, sex, or education between CHR and healthy control groups. Cognitive scores were significantly lower than those for healthy controls on all MCCB subsets for all patients at CHR (effect sizes d = 0.37-0.95), as well as the CHR-C (effect sizes d = 0.43-1.30) and CHR-NC subgroups (effect sizes d = 0.39-0.91) (Table). Participants in the CHR-C group performed significantly worse than those in the CHR-NC group on the Trail Making Test A (small-to-medium effect size, d=0.38) and the Brief Visuospatial Memory Test-Revised (medium-to-large effect size, d = 0.69).
Compared with controls, CHR-remission and CHR-poor outcome scores were lower for all MCCB subsets (medium- to large-effect sizes for poor outcome, d = 0.35-1.91), and CHR-symptomatic scores were lower for 6 of 8 MCCB subsets (effect sizes d = 0.50-93). Furthermore, Brief Visuospatial Memory Test-Revised scores were significantly lower in the CHR-poor outcome versus CHR-remission group (medium effect size, d = 0.53).
The authors noted that an educated, mostly urban Chinese sample of patients at CHR had cognitive impairments of a similar magnitude to a western prodromal consortium (the North American Prodrome Longitudinal Study [NAPLS]), with evidence for greater impairments in those who converted to psychosis at 1-year follow-up (versus non-converters). Findings advance our understanding of cognitive heterogeneity in patients at CHR of psychosis. The relatively small sample size of some CHR subgroups and 1-year (versus a longer follow-up) duration were noted as potential study limitations.
Matters to consider
An important question follows: what is the evidence for the feasibility and efficacy of strategies or treatment for cognitive impairments in patients at CHR of psychosis? Do they forestall or decrease conversion to psychosis, or are they associated with improvements in other outcomes? To date, there are small studies that have investigated cognitive interventions in this population. Several studies have investigated computer-based cognitive training in participants with CHR.7-10 Rauchensteiner and colleagues7 found that cognitive training was associated with significant improvements in verbal learning and memory. Another study by Hooker et al8 found significant improvement in processing speed and trend-level improvements in visual learning, visual memory, and global cognition. By contrast, a small double-blind randomized trial by Piskulic and colleagues9 did not find significant effects on cognition between cognitive training and an active control (computer games). A larger double-blind randomized controlled trial of the same interventions used in the negative trial by Piskulic et al9 found significant improvements in verbal memory in the cognitive training (versus computer games) group. Recently, a randomized trial of an integrated social- and neuro-cognitive remediation intervention that included a combination of individual, group, in-person, and computer-based training—versus enriched acceptance and commitment therapy—found improvements in social cognition and a trend for improved reaction time.11,12 Overall, these studies demonstrate the feasibility of and point to potential efficacy of cognitive interventions in CHR. Future studies are needed to investigate the optimal type and “dose” of cognitive training in patients at CHR, as well as the impact on symptoms, functioning, and conversion to psychosis.
This study demonstrates the feasibility of harmonizing measures across different geographical locations for the reliable and valid assessment of cognition in patients at CHR of psychosis. Furthermore, findings replicate, and therefore reinforce, earlier studies, as the pattern and magnitude of cognitive impairments in patients at CHR were similar between the SHARP study and the NAPLS consortium.
Based on these data, cognitive impairments represent an important, viable target for intervention in this population.
An earlier version of this article titled "Cognition and Outcomes in Clinical High-Risk for Psychosis" was posted ahead of print. -Ed
Dr Miller is professor, Department of Psychiatry and Health Behavior, Augusta University, Augusta, GA. He is the schizophrenia section chief for Psychiatric TimesTM. The author reports that he receives research support from Augusta University, the National Institute of Mental Health, the Brain and Behavior Research Foundation, and the Stanley Medical Research Institute.
1. Agnew-Blais J, Seidman LJ. Neurocognition in youth and young adults under age 30 at familial risk for schizophrenia: a quantitative and qualitative review. Cogn Neuropsychiatry. 2013;18(1-2):44-82.
2. Gur RC, Calkins ME, Satterthwaite TD, et al. Neurocognitive growth charting in psychosis spectrum youths. JAMA Psychiatry. 2014;71(4):366-374.
3. Fusar-Poli P, Deste G, Smieskova R, et al. Cognitive functioning in prodromal psychosis: a meta-analysis. Arch Gen Psychiatry. 2012;69:562-571.
4. Bolt LK, Amminger GP, Farhall J, et al. Neurocognition as a predictor of transition to psychotic disorder and functional outcomes in ultrahigh risk participants: findings from the NEURAPRO randomized clinical trial. Schizophr Res. 2019;206:67-74.
5. Seidman LJ, Shapiro DI, Stone WS, et al. Association of neurocognition with transition to psychosis: Basline functioning in the second phase of the North American Prodrome Longitudinal Study. JAMA Psychiatry. 2016;73:1239-1248.
6. Cui H, Giuliano AJ, Zhang T, et al. Cognitive dysfunction in a psychotropic medication-naïve, clinical high-risk sample from the ShangHai-At-Risk-for_psychosis (SHARP) study: associations with clinical outcomes. Schizophr Res. 2020.
7. Rauchensteiner S, Kawohl W, Ozgurdal S, et al. Testperformance after cognitive training in persons at risk mental state of schizophrenia and patients with schizophrenia. Psychiatry Res. 2011;185:334-339.
8. Hooker CI, Carol EE, Eisenstein TJ, et al. A pilot study of cognitive training in clinical high risk for psychosis: initial evidence of cognitive benefit. Schizophr Res. 2014;157:314-316.
9. Piskulic D, Barbato M, Liu L, Addington J. Pilot study of cognitive remediation therapy on cognition in young people at clinical high risk of psychosis. Psychiatry Res. 2015;225:93-98.
10. Loewy R, Fisher M, Schlosser DA, et al. Intensive Auditory Cognitive Training Improves Verbal Memory in Adolescents and Young Adults at Clinical High Risk for Psychosis. Schizophr Bull. 2016;42:S118-26.
11. Friedman-Yakoobian MS, Parrish EM, Eack SM, et al. Neurocognitive and social cognitive training for youth at clinical high risk (CHR) for psychosis: A randomized controlled feasibility trial. Schizophr Res. 2020.
12. Friedman-Yakoobian MS, Parrish EM, Thomas A, et al. An integrated neurocognitive and social-cognitive treatment for youth at clinical high risk for psychosis: Cognition for Learning and for Understanding Everyday Social Situations (CLUES). Schizophr Res. 2019;208:55-59.❒