Stability of performance monitoring with prolonged task performance: A study of error-related negativity and error positivity.

Abstract

The use of forced-choice response tasks to study indices of performance monitoring, such as the error-related negativity (ERN) and error positivity (Pe), is common, and such tasks are often used as a part of larger batteries in experimental research. ERN amplitude typically decreases over the course of a single task, but it is unclear whether amplitude changes persist beyond a single task or whether Pe amplitude changes over time. This preregistered study examined how prolonged task performance affects ERN and Pe amplitude across two study batteries, each with three different tasks. We predicted ERN amplitude would show unique, nonlinear reductions over an individual task and over the task battery, and exploratory analyses were conducted on Pe. Electrophysiological data were recorded during two studies: 156 participants who completed three versions of the flanker task and 161 participants who completed flanker, Go/NoGo, and Stroop tasks. ERN showed unique nonlinear reductions over each flanker task and over the battery of flanker tasks. However, ERN showed a linear reduction in amplitude over the battery of three different tasks, and within-task changes were only observed during the Go/NoGo task, such that ERN increased. Pe generally linearly decreased with prolonged task performance. Variability in ERN and Pe scores generally increased with time, indicating decreases in data quality. Findings suggest that studying ERN and Pe early in a task battery with short tasks is optimal to avoid bias from prolonged performance. Identifying factors affecting ERN and Pe during prolonged performance can help develop optimized paradigms.