Alternative forms of the parametric Go/No-Go task: development and comparison

The Parametric Go/No-Go (PGNG) task assesses cognitive domains, including executive functioning and processing speed, within three difficulty levels. The PGNG task was originally developed for clinical and research assessments, in only one version (using target consonants, XYZ). The study aims to develop six alternative tri-consonant forms of the preexistent form and to test their equality. We hypothesized performance differences between each difficulty level of a given PGNG version, and that, in contrast, all the tri-consonant alternate versions would be equivalent in performance.

A total of 349 participants were recruited online to complete one of seven task versions - each a unique tri-consonant variation with three difficulty levels. The dependent variables measured were Percentage of Correct Target (PCT), Percentage of Correct Inhibition (PCI), Efficiency, Response Time (RT), and Response Time Coefficient of Variation (RTCoV).

For each version, we found level 1 to be the most challenging compared to level 2, and level 3 to be the most challenging, with the lowest accuracy and the longest response time. No differences were found in RTCoV measure across any of the tri-consonant versions. However, differences between tri-consonant versions were observed depending on the specific performance measures and levels of the PGNG.

As expected, there were differences between difficulty levels on all measures and versions, with performances being globally different in levels 1, 2, and 3. The overall results showed that all the tri-consonant versions were similar on temporal measures (RT and RTCoV). Four tri-consonant versions (XYZ, BCM, NPV, and RST) were similar in terms of performance on all five performance measures. The observed differences are discussed through Baddeley's model of working memory and the implication of multisensory information.