Rapid Throughput Operant Taste Discrimination: Signal Detection Theory (SDT) and Concentration‐Response (CR) Analyses for Quantifying Sucrose Taste Sensitivities in Individual Human Subjects

Abstract

Most human taste discrimination experiments have focused on determination of detection thresholds, which are taken as indications of the sensitivity of the taste sensory system. However, detection thresholds reported in the literature can vary substantially across and within studies, and there is little evidence to support repeatability of the values obtained. Two factors that likely contribute to the observed variability are 1) inherent low throughput, which would impact assay robustness, and 2) insufficient accounting for subject bias in the analysis. Low throughput capacity of traditional taste discrimination methods all but precludes quantification by the more broadly applicable measures of discriminability (d′) and subject bias (β) obtained by SDT, as well as the ability to conduct CR analyses. We previously have reported a novel technology for automated rapid throughput operant taste discrimination using human subjects, called the TāStation®. We applied this technology to compare sensitivities to low “threshold” concentrations of sucrose determined by SDT with EC50s obtained by CR analysis in the same subjects. Small volumes (200 μl) of sucrose solutions or water, distributed in replicate wells of a 96‐well plate, were automatically drawn by an electronic pipette from single randomly selected wells on each trial. Subjects then removed the pipette and self‐administered the stimuli to the tongue. Subjects were trained through an interactive game‐like algorithm to associate the taste of sucrose with coordinates on the right sided of a touch‐sensitive display (TSD), and water on corresponding coordinates reflected to the opposite side. Correct responses on control trials were reinforced by the appearance of a virtual poker chip carrying remunerative value (which accumulated across the 96 trials); incorrect responses were penalized by a reduction in the cumulative score. The rapid generation of large robust data sets enabled an assessment of repeatability for SDT values of d′ and β, and EC50s derived from CR functions. We found that d′ and β were relatively stable across tests for some individuals but varied for other subjects. Less variability was observed in the EC50s for all subjects. For example, in three tests each occurring on separate days, 5 mM sucrose d′ values for subject M1200 were 1.18, 1.07, 0.94; those for subject M1194 were 0.56, −0.08, and 1.29. EC50s obtained for subject M1200 were 22, 27, and 24 mM, and for subject M1194 were 16, 20, and 27 mM. The repeatability of the EC50s within subjects revealed two distinct phenotypes for sucrose potency across subjects, with one group exhibiting EC50s that ranged narrowly around 25 mM, and a second group around 60 mM. These differences in sucrose potencies across subjects suggest the possibility of alternative structural features in the sucrose binding site of the T1R2/R3 receptor. Our results further imply that EC50 values, obtained from quantitative characterization of the entire concentrations‐response function for taste, are more reliable and more informative than measures of discriminability at the lowest taste‐active concentrations (i.e., “thresholds”).Support or Funding InformationEntirely funded by Opertech Bio