Quantitative prediction of the bitterness suppression of elemental diets by various flavors using a taste sensor

The purpose of the study was to develop a method for the quantitative prediction of the bitterness suppression of elemental diets by various flavors and to predict the optimum composition of such elemental diets for oral administration using a multichannel taste sensor.

We examined the effects of varying the volume of water used for dilution and of adding varying quantities of five flavors (pineapple, apple, milky coffee, powdered green tea, and banana) on the bitterness of the elemental diet, Aminoreban EN. Gustatory sensation tests with human volunteers (n = 9) and measurements using the artificial taste sensor were performed on 50 g Aminoreban EN dissolved in various volumes (140), 180, 220, 260, 300, 420, 660, 1140, and 2100 ml) of water, and on 50 g Aminoreban EN dissolved in 180 ml of water with the addition of 3-9 g of various flavors for taste masking.

In gustatory sensation tests, the relationship between the logarithmic values of the volumes of water used for dilution and the bitterness intensity scores awarded by the volunteers proved to be linear. The addition of flavors also reduced the bitterness of elemental diets in gustatory sensation tests; the magnitude of this effect was, in decreasing order, apple, pineapple, milky coffee, powdered green tea, and banana. With the artificial taste sensor, large changes of membrane potential in channel 1, caused by adsorption (CPA values, corresponding to a bitter aftertaste), were observed for Aminoreban EN but not for any of the flavors. There was a good correlation between the CPA values in channel 1 and the results of the human gustatory tests, indicating that the taste sensor is capable of evaluating not only the bitterness of Aminoreban EN itself but also the bitterness-suppressing effect of the five flavors, which contained many elements such as organic acids and flavor components, and the effect of dilution (by water) on this bitterness. Using regression analysis of data derived from the taste sensor and from human gustatory data for four representative points, we were able to predict the bitterness of 50 g Aminoreban EN solutions diluted with various volumes of water (14-300 ml), with or without the addition of a selected flavor.

Even though this prediction method does not offer perfect simulation of human taste sensations, the artificial taste sensor may be useful for predicting the bitterness intensity of elemental diets containing various flavors in the absence of results from full gustatory sensation tests.