Repeatability of taste recognition threshold measurements with QUEST and quick Yes-No
Taste perception, though vital for nutrient sensing, has long been overlooked in psychological and clinical sensory assessments. This can, at least in part, be attributed to challenges associated with the handling of liquid, perishable stimuli, but also with scarce efforts to optimize testing procedures to be more time-efficient. We have previously introduced an adaptive, QUEST-based procedure to measure taste sensitivity thresholds that can be quicker than other existing approaches, yet similarly reliable. Despite its advantages, this procedure lacks experimental control of false alarms (i.e., response bias) and psychometric function slope. Variations of these parameters, however, may influence the threshold estimate likewise, raising the question as to whether a procedure assessing threshold, false-alarm rate, and slope simultaneously might be able to produce threshold estimates with higher repeatability, i.e., smaller variation between repeated measurements of the same participant.
Here, we compared the performance of QUEST with a method that allows measurement of false-alarm rates and slopes, quick Yes-No (qYN), in a test-retest design for citric acid, sodium chloride, quinine hydrochloride, and sucrose recognition thresholds using complementary measures of repeatability, namely test-retest correlations and coefficients of repeatability. Both threshold procedures yielded largely overlapping thresholds with good repeatability between measurements.