Binding Costs in Processing Efficiency as Determinants of Cognitive Ability

Abstract

Performance in elementary cognitive tasks is moderately correlated with fluid intelligence and working memory capacity. These correlations are higher for more complex tasks, presumably due to increased demands on working memory capacity. In accordance with the binding hypothesis, which states that working memory capacity reflects the limit of a person’s ability to establish and maintain temporary bindings (e.g., relations between items or relations between items and their context), we manipulated binding requirements (i.e., 2, 4, and 6 relations) in three choice reaction time paradigms (i.e., two comparison tasks, two change detection tasks, and two substitution tasks) measuring mental speed. Response time distributions of 115 participants were analyzed with the diffusion model. Higher binding requirements resulted in generally reduced efficiency of information processing, as indicated by lower drift rates. Additionally, we fitted bi-factor confirmatory factor analysis to the elementary cognitive tasks to separate basal speed and binding requirements of the employed tasks to quantify their specific contributions to working memory capacity, as measured by Recall−1-Back tasks. A latent factor capturing individual differences in binding was incrementally predictive of working memory capacity, over and above a general factor capturing speed. These results indicate that the relation between reaction time tasks and working memory capacity hinges on the complexity of the reaction time tasks. We conclude that binding requirements and, therefore, demands on working memory capacity offer a satisfactory account of task complexity that accounts for a large portion of individual differences in ability.

Intelligence IS Cognitive Flexibility: Why Multilevel Models of Within-Individual Processes Are Needed to Realise This

“…Of course, such tasks do not exist. However, with careful, theory-driven task analyses, the parameters of complexity can be formalised and investigated ( Beckmann 2010 ; Birney and Bowman 2009 ; Ecker et al 2010 ; Goecke et al 2021 ; Halford et al 1998 ).…”

“…Investigations of processes in individual differences research has had a strong focus on understanding mechanisms underlying working memory (WM) in and of itself (e.g., Ecker et al 2010 ; Goecke et al 2021 ; Oberauer and Lewandowsky 2016 ), or as a set of processes common to both WM and Gf (e.g., Ackerman et al 2005 ; Engle et al 1999 ; Oberauer et al 2007 ; Shipstead et al 2016 ). What is common in many of the studies and approaches described in the rest of this section is the combined experimental-correlational methodology—basic processes are proposed, operationalised as individual differences variables and “measured”, and then “validated” as incremental predictors of the latent attribute (e.g., WMC or Gf ).…”

“…In a recent study investigating the role of the working-memory binding hypothesis, Goecke et al ( 2021 ) combined an experimental manipulation of complexity of elementary cognitive tasks (ECTs), also using a bi-factor SEM approach to identify the mechanisms underlying binding demands (e.g., more stimulus-response mappings = greater binding demand) on working memory capacity. This was achieved in a three-step process.…”

“…As we have suggested above, with careful, theory-driven task analyses, the parameters of complexity can be formalised and investigated ( Birney and Bowman 2009 ; Ecker et al 2010 ; Goecke et al 2021 ; Halford et al 1998 ). Multilevel models (MLM) are well-suited for this in that they provide a means to explicate a definition, operationalisation, and assessment of cognitive flexibility that is formally aligned and testable within a common model.…”

“…That is, while these cognitive psychology constructs tend to have articulated process accounts, they were not the panacea to the conceptualisation-operationalisation misalignment of intelligence hoped for. Translating these process-focused constructs into assessments, sometimes referred to as elementary cognitive tasks (ECTs), has psychometric challenges which the traditional latent variable (psychometric) approach to intellectual abilities cannot resolve alone ( Goecke et al 2021 ).…”