Modelling the correlation between two putative inhibition tasks: a simulation approach

Behavioural studies of individual differences have shown mild but significant correlations in performance on tasks that require the withholding of a response to a prepotent stimulus, i.e., on so-called response inhibition tasks. Several computational models of response inhibition tasks have been developed, but the dominant models of such tasks have been produced in isolation of each other. Consequently they fail to present a coherent unitary picture of response inhibition. In this paper we consider two established interactive activation models of distinct response inhibition tasks – the stop signal task and the Stroop task – and explore potential mechanisms within those models that might underlie the observed behavioural correlation. Only one plausible account of the correlation emerges: that it results from shared mechanisms of attentional bias. This account does not map onto the classical concept of response inhibition. It is concluded that either the accepted models are flawed or that the concept of response inhibition as applied to these tasks is misleading (and hence counterproductive). More generally the work may be taken to support an architectural approach to modelling, albeit at the level of interactive activation models, rather than the more traditional production system models