Can the post-error effect mask age-related differences in congruency conditions when education and overall accuracy are controlled for?

Objective: Age-related differences in stimulus-response congruency tasks have been attributed to older adults' greater difficulties in handling the irrelevant spatial-dimensional overlap between stimulus and response. However, performance on congruency tasks may also be influenced by the previous trial accuracy (i.e. post-error effect), which may affect young and older adults differently. The main objective of this study was to analyse age-related differences in the post-error effect as a function of congruency. In addition, we examined the meditational role of the Gratton effect on the age-related differences in the post-error slowing (PES) and post-error increased accuracy (PIA) as a function of congruency.

Method: The sample comprised 165 healthy adult participants with diverse educational attainment, divided into five age groups. Participants performed a spatial stimulus-response congruency task. Age-related differences in the post-error effect were analysed for each congruency condition taking into account educational attainment and overall accuracy. Statistical procedures were used to neutralize age-related processing speed effects on the PES.

Results: PES was observed across all age groups, except the Very old group (aged 85-98 years), and it was not related to congruency condition. PIA was observed across age groups in all congruency conditions and was slightly higher in incongruent trials. Evidence of simultaneous PES and PIA was found for young participants and older participants under 85 years. The Very old group did not need to significantly slow down their responses after errors to improve accuracy. No age- related difference was found in the influence of the Gratton effect on PES or PIA as a function of congruency.

Conclusions: PES and PIA were observed in young adults and older adults under 85 years old. Evidence of simultaneous PES and PIA in the young and older age group (except for the Very old) indicates that the post-error effect can be interpreted in terms of recruitment of additional resources to prevent subsequent errors. Slightly higher accuracy was observed in the incongruent condition in post-error trials relative to pre-error correct trials across age groups.