Extended Data Fig. 4: Mice attend to the optogenetic cue.

All panels show reach rate as in Fig. 1. a, Omit the food pellet, but present the cue. Black: Pellet presented. Orange: Pellet omitted on random trials. n = 11805 black trials, 1637 orange trials from 18 mice. We excluded trials when the mouse dislodged the pellet before the cue. b, Omit the cue, but present the food pellet. Black: Cue turns on. Orange: Cue omitted on random trials. n = 3268 black trials, 246 orange trials from 18 mice. c, Compare reaching in response to the cue with reaching in response to the distractor LED. Black: Aligned to cue. Orange: Aligned to distractor LED. n = 3268 black trials, 3268 orange trials from 18 mice. Video frames at right show that distractor LED is brighter than real cue. d, Response to a red light cue in mice expressing the red-activatable opsin ChrimsonR in visual cortex. Poor visual detection of red light in mice, yet the mice still learn to respond to the optogenetic cue. n = 862 orange trials from 3 mice. e, Response to the blue light cue when the visual cortex does not express the activating opsin Channelrhodopsin2, ChR. Orange: Aligned to the cue, from mice that lack ChR in visual cortex. n = 3225 orange trials from 4 mice. f, In mice trained to respond to the blue light optogenetic cue, block the thinned skull to prevent blue light from accessing the brain. Video frames at right show that the blue light turns on but does not penetrate the blocked skull. Black: Control day before blocking the skull. Orange: The next day when we blocked the skull. n = 2357 black trials, 1733 orange trials from 18 mice.