Fig. 1.

Choice task and reward design for studying nutrient influences on monkeys’ choices. (A) Nutrient–choice task. Monkeys chose from two sequentially presented options. Conditioned stimuli predicted different liquid rewards; magnitude bars predicted randomly varying reward amounts. (B) Nutrient–reward design. Liquid rewards differed in sugar and fat concentration. LFLS: low-fat, low-sugar; HFLS: high-fat, low-sugar; LFHS: low-fat, high-sugar; and HFHS: high-fat, high-sugar. Rewards were matched in flavor (peach of blackcurrant) and other ingredients (protein, salt, etc.); HFLS and LFHS were matched in energy content (isocaloric); HFHS had a higher energy content; and LFLS was lowest in energy content. (C) Completed choice trials per testing session in each animal (N: number of sessions). (D) Choice frequencies for each nutrient reward (± SEM), across sessions and animals (N = 55,205 trials). (E) Choice biases for fat and sugar in single sessions. Trial-by-trial choice records of two representative sessions from monkey Ya choosing between a low-nutrient option (yellow) and rewards with added fat (HFLS, green, Top) or sugar (LFHS, blue, Bottom). Upward/downward bars represent choices for high-/low-nutrient rewards; bar height indicates repeated choice counts. Gray curve shows choice frequency for high-nutrient rewards (seven-trial running average). (F) Nutrient–value functions. Choice frequencies for the low-nutrient reference as a function of offered magnitude ratio (LFLS/high-nutrient rewards ± SEM). Indifference points, estimated by inflection points of fitted sigmoid curves, identify relative values of the high-nutrient rewards, measured on the common scale of the low-nutrient reference. (Inset) Relative values of high-fat and high-sugar rewards and their 95% CIs.