17 Norming for the reverse-translated 5-choice continuous performance test (5C-CPT) of attention and cognitive control

Abstract

Objective:

Translatability of preclinical results remains a major obstacle in neuropsychiatric research. Even when cognitive tests in preclinical models show translational validity for human testing, with sensitivity to clinical deficits, there remains the issue of heterogeneity among human participants. Norming of performance on cognitive tasks enable corrections for any differences in performance that may arise from the influence of socioeconomic factors, and thus a more direct comparison with preclinical testing results. The 5-choice continuous performance task (5C-CPT) is a test sensitive to changes in sustained attention and cognitive control in rodent manipulations and clinical populations, including schizophrenia and bipolar disorder. Herein, we present normed results of 5C-CPT data from a cohort of human participants, enabling greater comparison to future clinical and rodent testing.

Participants and Methods:

5C-CPT data were generated from a range of participants from the Translational Methamphetamine AIDS Research Center (n=82) and a study of bipolar disorder (n=45). Participant demographics were as follows: Age M=38.5, SD=16.7, Education: M=14.5, SD=1.9, 45% female, 10% Asian, 17% African American, 27% Hispanic, and 46% non-Hispanic White. We used the test2norm R-package to create norms for each of the major outcomes from the 5C-CPT. Non-normally distributed raw scores were transformed to generate more normally distributed data needed for the norming process. Raw scores were first converted into uniform scaled scores that range from 0-20 where a higher score indicated better performance. We then generated T-score formulas, which are standardized residuals and scaled to have a mean of 50 and standard deviation of 10. The residuals are obtained from regressions, modeled using multiple fractional polynomial method (MFP), which regresses scaled scores on demographic variables, which a user wishes to control for (gender, age, education, ethnicity, etc.). MFP models allow to fit non-linear effects for numeric demographic factors (e.g., age), if such effects exist.

Results:

New, demographically corrected T-score formulas were calculated for each major outcome of the 5C-CPT: reaction time (MCL), reaction time variability (VarRT), dprime, hit rate (HR) and false-alarm rate (FAR). MFP models showed that age had a significant effect on MCL, VarRT, dprime, and HR (all p<0.01), while gender only showed a significant effect for MCL and VarRT (all p<0.05). Interestingly, education and ethnicity did not show a significant effect for any MFP model and none of the demographic factors (age, education, gender, ethnicity) were significant in the model for FAR. As defined in the test2norm package, all scaled scores had a mean of 10 and SD of 3 and all T-scores had a mean of 50 and SD of 10.

Conclusions:

The 5C-CPT is a test of attention and cognitive control available for human testing, reverse-translated from rodent studies. The normative data generated here will enable future comparisons of data without the need for additional control studies. Furthermore, comparing these normative data to manipulations will enable further comparisons to rodent testing, with manipulations relative to baseline becoming more meaningful. Thus, the 5C-CPT is a viable tool for conducting cross-species translational research toward developing novel therapeutics that treat dysfunctional attentional and cognitive control.