Facebook YouTube Twitter Linkedin
Facebook YouTube Twitter Linkedin


7 April 2018

Exploring participant-level trajectories of cognitive performance in patients with schizophrenia

Cambridge Cognition and Boehringer Ingelheim have been investigating cognitive performance among patients with schizophrenia, with a view to improving future pro-cognitive drug trial designs. Our Director of Neuroscience, Dr Kiri Granger, presented the latest findings at ISCTM and SIRS.


It remains unclear whether the lack of clinical trial success and drug approval for cognitive impairment associated with schizophrenia (CIAS) is due to compounds being ineffective, or whether trial methodology itself has been a limiting factor in successfully demonstrating the efficacy of these agents.

Cognitive deficits are a common, but not consistent, feature of schizophrenia. Indeed, whilst most patients exhibit some general cognitive impairment compared to antecedent expectations, up to a quarter display cognitive performance in the ‘normal’ range.

The heterogeneous cognitive profile associated with schizophrenia poses potential problems for pro-cognitive drug trials. Including cognitively ‘normal’ patients within a sample may inflate baseline scores, thus reducing the scope to see improvement between treatment and placebo groups.

In order to examine this potential issue, researchers from Cambridge Cognition and Boehringer Ingelheim investigated participant-level trajectories of cognitive performance among patients with schizophrenia.



A post-hoc analysis was conducted of existing trial data from 463 patients with schizophrenia who participated in a randomized, double-blind, placebo-controlled, phase II, pro-cognitive drug trial.

Patients met established diagnosis for schizophrenia (DSM-5), were clinically stable (non-acute) and had no more than moderate severity ratings on the Positive and Negative Syndrome Scale (PANSS).

Participants completed the Cambridge Neuropsychological Test Automated Battery (CANTAB) and the MATRICS Consensus Cognitive Battery (MCCB), at 4 separate time points:

  1. Screening
  2. Baseline
  3. Week 6
  4. Week 12



Approximately 25% of the overall sample were performing within a clinically normal cognitive range at screening, as determined by the MCCB composite score (t score >35-40).

Patients who performed poorer on CANTAB PAL (>10 errors) at screening demonstrated the potential for an improvement in performance across baseline, week 6 and week 12 time points.

In contrast, the top 25% of performers on PAL (<10 errors) at screening demonstrated the potential for stability in performance across the remainder of the study visits.

How performance changed across the time points for these low and high performers is illustrated in the figures below at both the individual (participant trajectories, figure 1A), and group (figure 1B) level.

Figure 1. (A) Individual and (B) Mean PAL total errors adjusted (PALTEA) scores for participants plotted over 4 testing sessions colour coded by PALTEA scores at screening visit. Lower scores indicate better performance. Orange lines specify individuals who scored less than 10 errors at screening: high scorers. Whilst blue lines specify individuals who scored more than 10 errors at screening: low scorers.



There was substantial variability in cognitive performance among the schizophrenia patients who were sampled. Furthermore, a substantial subsample of patients showed clinically ‘normal’ cognitive performance.

The study suggests that patients who exhibited impaired performance on CANTAB PAL at screening had the potential to improve in cognitive performance across the time-points.

This finding brings into question whether the inclusion of unimpaired patients in pro-cognitive drug trials minimizes the chance to detect drug efficacy.


View poster 

Get in touch 

Tags : clinical trials | cognitive testing | schizophrenia | cognitive impairment | cias | pro-cognitive

Author portrait

Dr Kiri Granger