Cognitive outcomes are frequently implemented as endpoints in nutrition research. To reduce the number of statistical comparisons it is commonplace for nutrition researchers to combine cognitive test results into a smaller number of broad cognitive abilities. However, there is a clear lack of understanding and consensus as to how best execute this practice. The present paper reviews contemporary models of human cognition and proposes a standardised, evidence-based method for grouping cognitive test data into broader cognitive abilities. Both Carroll's model of human cognitive ability and the Cattell-Horn-Carroll (CHC) model of intelligence provide empirically based taxonomies of human cognition. These models provide a cognitive 'map' that can be used to guide the handling and analysis of cognitive outcomes in nutrition research. Making use of a valid cognitive nomenclature can provide the field of clinical nutrition with a common cognitive language enabling efficient comparisons of cognitive outcomes across studies. This will make it easier for researchers, policymakers and readers to interpret and compare cognitive outcomes for different interventions. Using an empirically derived cognitive nomenclature to guide the creation of cognitive composite scores will ensure that cognitive endpoints are theoretically valid and meaningful. This will increase the generalisability of trial results to the general population. The present review also discusses how the CHC model of cognition can also guide the synthesis of cognitive outcomes in systematic reviews and meta-analysis.
Keywords: Cattell–Horn–Carroll (CHC) model; Clinical trials; Cognition; Nutrition.