Figure 1. Landmark-dependent navigation task in virtual reality.
(A) Schematic of experimental setup: mice are head-fixed atop a cylindrical treadmill with two computer screens covering most of the animal’s field of view. A reward spout with attached lick-sensor delivers rewards. (B) Task design. Animals learned to locate hidden reward zones at a fixed distance from one of two salient visual cues acting as landmarks. The two landmarks were interleaved within a session, either randomly or in blocks of 5. After each trial animals were placed in a ‘black box’ (screens turn black) for at least 3 s. The randomized starting location ranged from 50 to 150 cm before the landmark. (C) Licking behavior of the same animal at novice and expert stage. Expert animals (bottom) lick close to the reward zones once they have learned the spatial relationship between the visual cue and reward location. (D) The Task Score was calculated as the difference in first lick location (averaged across trials) between short and long trials. (E) Relationship between trial start and first lick locations for one example session. Experimental design ensured that alternative strategies, such as using an internal odometer, could not be used to accurately find rewards. (F) RSC inactivation experiment. VGAT-Cre mice were injected with flexed Channelrhodopsin-2 (left). Stimulation light was delivered through skull-mounted ferrules on a random subset of trials (middle). During inactivation trials, task score was reduced significantly (right).
Figure 1—figure supplement 1. Running and licking behavior in naive animals and during optogenetic silencing.
