The cartoon illustrates the behavioral arena and the protocol used to examine short-term taste memories. Head-restrained mice, placed on a running wheel, were given random presentations of bitter (1 mM quinine) or sweet stimulus (5 mM AceK) from a sampling spout (middle bottle, labeled T). After a variable delay period of 1-9 s, an odor cue was presented to signal the end of delay and mark the response window (2 s). The animals were trained to report the identity of the taste stimulus by going to the left or right spout for a water reward (labeled W). Correct responses required the mice to be capable of holding the memory of the taste stimuli during the variable delay and to properly match them to the odor cue. Each session had 90-130 trials. Odor A, n-amyl acetate; odor B, benzaldehyde (see text for details). The two boxes detail the taste-odor combinations used in the four- and three-arm behavioral paradigms (seeSTAR Methods); also shown are the corresponding differences in training time and performance (Figure S1).
