Sucrose detection and discrimination estimated by the analysis of psychometric functions with linear and non-linear models

Abstract

Sucrose detection and discrimination thresholds were determined by conducting two alternative forced-choice tests with aqueous solutions. The standard models probit and logit, and non-linear were fitted to the empirical psychometric functions. 0, 7 and 15 g/L sucrose solutions were used as standard stimuli in the detection and discrimination experiments (two levels) respectively. Comparison stimuli consisted of aqueous sucrose solutions with concentrations from 0.5 to 25 g/L. Observed absolute threshold was around 2.96 g/L and the difference limen for 7 and 15 g/L standard stimuli was around 2.98 g/L and 5.29 g/L, respectively. Calculated Weber fractions for 7 and 15 g/L standard stimuli were 0.42 and 0.35, respectively. Judges’ performance in the discrimination experiment was similar to that observed in the detection experiment. Goodness of fit for probit and logit models was similar. The non-linear model showed a lower average error and demonstrated an excellent predictive ability.