Go/No-Go Task

Problems with inference come in many forms. Flexibility is often discussed in terms of p-hacking, deliberate practices to increase the probability of observing statistical significance. Here is a neglected area that also suffers from lacking standardization: measurement validity.

What is the Go/No-Go Task?

The Go/No-Go task is a computerized test used to assess inhibitory control, a cognitive process that enables humans to rapidly cancel motor activity even after its initiation. Each trial consists of one stimulus either indicating either "Go" (respond) or "No-Go". Thus, prepotent motor activity is elicited, which then sometimes has to be inhibited. The response accuracy of each No-Go trial is used as a measure for inhibitory control.

How the Go/No-Go Task is Flexible

The most critical component in designing an inhibitory control task is to ensure that motor activity is elicited on each trial, so that No-Go trials truly test inhibitory control. In order to elicit prepotent motor activity ("Go"), which then sometimes has to be inhibited ("No-Go"), two parameters of the experimental design are important: First, the relative proportion of Go and No-Go trials should be set so that No-Go trials are less frequent, making it strategically beneficial to initiate a Go response on every trial. Second, trials should be presented at a rapid pace, so that responses will have to be made quickly (Wessel, 2017). However, despite these well-known considerations, about 40% of published Go/No-Go experiments have equiprobable Go and No-Go trials. Additionally, while most studies present stimuli at relatively rapid pace, almost 20% of studies use maximum trial durations greater than 4000 ms, which allows a conscious response to the stimulus instead of a prepotent response or its immediate inhibition.

Why the Flexibility is Problematic

It is important to emphasize that the No-Go trial proportion and stimulus duration are always set a priori, not after data collection. Thus, the problems with flexibility and lack of standardization across the Go/No-Go literature are not related to p-hacking; rather, they concern the validity of the Go/No-Go task. If motor activity is not elicited on each trial, the task will not test the ability to withhold a response tendency, but will instead contrast conditions in which a response is made (Go trials) with conditions in which no response is ever initiated (No-Go trials). Thus, equiprobable presentation of Go and No-Go trials, and excessively slow presentation, both pose a threat to the Go/No-Go task's validity.

How the Go/No-Go Task Could be Used Better

Studies with the Go/No-Go Task need to be extremely carefully designed to ensure that inhibitory control is actually necessary to perform the No-Go trials within the respective tasks. If that is not the case, any differential effects that No-Go trials have in the context of such studies cannot be interpreted as effects of inhibitory control activity. It is recommended that trial durations should be shorter than 1,500 ms with a No-Go trial proportion of less than 20%.

Source: Wessel, J. R. (2017). Prepotent motor activity and inhibitory control demands in different variants of the Go/No-go paradigm. Psychophysiology. doi: 10.1111/psyp.12871