A confirmatory factor analysis of the MATRICS consensus cognitive battery in severe mental illness

Abstract

Background

A recent factor analysis of Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) Consensus Cognitive Battery (MCCB) data in a sample of schizophrenia-spectrum patients described a three factor model representing processing speed, attention/working memory, and learning; the social cognition measure was excluded from the analysis. The current analyses sought to replicate a three factor structure of the MCCB in a larger, more diagnostically diverse sample of participants.

Methods

Confirmatory factor analyses were performed to evaluate the factor structure of the MCCB in 300 outpatients with severe mental illness (54.3% schizophrenia-spectrum diagnoses) who were participants in three studies of cognitive remediation. Exploratory analyses were conducted to evaluate the concurrent prediction of symptoms from MCCB factor scores.

Results

The three factor structure of the MCCB demonstrated stronger model fit (χ² = 14.53, p = 0.75, SRMR = 0.02, RMSEA = 0.0, CFI = 1.00, NNFI = 1.01) than the unifactoral structure. Poorer cognitive performance across all three MCCB factors was significantly correlated with more severe overall and disorganization symptoms from the PANSS and BPRS, and less severe affective symptoms (e.g., depression, anxiety) in the overall sample.

Conclusions

A three factor structure of the MCCB—composed of processing speed, attention/working memory, and learning—was replicated in a heterogeneous sample of persons with severe mental illness. Cognitive performance on the MCCB factors is associated with clinical symptoms.

Introduction

Impaired cognitive functioning is common among individuals with schizophrenia and other severe mental illnesses (SMI) (Martínez-Arán et al., 2004, Porter et al., 2003, Reichenberg and Harvey, 2007) and has a significant impact on everyday functioning (Austin et al., 2001, Bowie et al., 2006, Martinez-Aran et al., 2007). The Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) Consensus Cognitive Battery (MCCB) (Nuechterlein et al., 2008) is a reliable and valid comprehensive assessment of cognition containing a total of ten cognitive tests assessing the domains of processing speed, verbal learning, visual learning, verbal and nonverbal working memory, reasoning/problem solving, attention/vigilance, and social cognition, and has become a standard assessment in studies evaluating the impact of cognitive treatments (Burdick et al., 2011, Kessler et al., 2013, McGurk et al., 2015), and those designed to determine relationships between cognitive performance and community functioning in schizophrenia and other SMI (Green et al., 2008, Green et al., 2011, Keefe et al., 2011, Kern et al., 2011). Because the MCCB is being used in diagnostically heterogeneous samples (Burdick et al., 2011, Kessler et al., 2013), its factor structure in the broader SMI population is of interest.

Two studies have investigated the factor structure of the MCCB (Burton et al., 2013, Harvey et al., 2013). The most recent study (Burton et al., 2013) identified a three factor structure in people with schizophrenia-spectrum disorders, including processing speed, attention/working memory, and learning, in contrast to a unifactorial model reflecting global cognitive ability (Nuechterlein et al., 2008). The factor structure analysis by Harvey et al. (2013) examined whether cognitive ability measured by traditional neuropsychological tests (i.e., MCCB) and measures of functional capacity reflect a single ability construct, which they, in fact, demonstrated.

The current study sought to extend the factor analytic findings of Burton et al. and Harvey et al. by employing a confirmatory factor analysis to evaluate the factor structure of the MCCB in a heterogeneous sample of persons with schizophrenia-spectrum disorders and mood disorders. In addition, we explored the relationships between MCCB factor scores and clinical symptoms.