Task parameters of the Continuous Performance Test: Influence of systematic variations on error rates within a non-clinical population

Mark S Glafke, Purdue University

Abstract

The Continuous Performance Test (CPT) paradigms, CPT-AX and CPT-Repeated Letter are two common CPT formats that have been used to provide an assessment of individual’s ability to sustain attention (Parasuraman, 1982). Previous research has shown that these two paradigms are likely to elicit specific types of errors such as commission errors (more likely in CPT-AX condition) and omission errors (more likely in CPT-RL). What remains unclear is how attentional demands change (and subsequently error rates) as a function of varying key CPT task parameters such as proportion of targets (target density) and length of task within each of these target paradigms. In order to examine these potential influences, two CPTs (CPT-AX and CPT-RL) were constructed with two levels of target density (8% and 33%) and task duration (3 and 6 min.). A normative sample of 59 undergraduate students at Purdue University completed a series of eight CPT trial combinations representing the two levels of each of these three task parameters under examination. Results supported expectations that participants would commit more omission errors on the CPT-RL vs. CPT-AX paradigms. An interaction between target paradigm and target density revealed that commission errors were higher on the CPT-AX paradigms, but only when density increased. In the low density condition, commission errors were actually higher (but not significant) in the CPT-RL condition. An effect for the order of presentation of CPT paradigms also influenced omission errors. Being presented with the CPT-AX trials first was related to poorer performance (higher omission errors) on the CPT-RL trials. These results indicate that these CPT paradigms can be structured to elicit different behavioral responses. These results are discussed in light of existing theories of attention and implications of these results for clinical applications.

Degree

Ph.D.

Advisors

Rollock, Purdue University.

Subject Area

Psychology|Experimental psychology

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS