Separating sensory and cognitive components of decision making in a study of salt taste

Abstract

Separating sensory from non-sensory components of decision making allows us to develop more accurate chemical models of perception. Identification experiments with feedback provide probabilities that a stimulus will be correctly identified or confused with another stimulus. From these probabilities, stimulus distributions (sensory component) and decision boundaries (non-sensory component) can be estimated using Decision Boundary (DB) theory, a multidimensional generalization of signal detection theory. Stimuli consisted of nine deionized water-gum systems including one non-ionic gum level and two ionic gum levels, to which three levels of NaCl or NaCl + KCl were added. Twelve subjects identified ten replicates of randomly presented samples, judged the degrees of similarity between all pairs, and rated taste intensities. The DB model accounted for 99.2% of the variance of the data, a significantly better fit than deterministic multidimensional scaling models. Two sensory dimensions (due to salt and gum levels/type) were not perceptually correlated within stimuli. DB theory is applicable to basic and applied research in both academic and food industry settings. The DB identification study was easy to manage and subjects were motivated by the task, perhaps contributing to the success of the method.