Fig. 2: Comparing visual and spatial coding on test stimuli.

a, Stimuli in the test 1 session (see timeline in Fig. 1b). b, Lick rasters for an example mouse (blue dots represent reward delivery). c, Anticipatory licking behaviour in test 1 (n = 5 mice). d, Example neural responses from the medial region in a mouse from the task cohort, sorted by preferred position in the leaf1 corridor on held-out trials (top), and scatter plots of preferred positions in the leaf1 corridor versus all other corridors (bottom). e, Correlation from panel d (bottom) for each mouse in the task cohort (n = 5 mice). f, Same as panel e but only for the leaf2 and leaf1 corridors, shown across regions (n = 5 task mice; n = 7 unsupervised mice; n = 9 naive mice, 11 sessions). g,h, Example leaf1-selective (g) and circle1-selective (h) neurons from the medial area shown as a population average for each trial. i, Projections onto the coding direction in the medial area, defined as the difference between the leaf1-selective and circle1-selective populations in panels g,h. Time course (left, one mouse), average over trials (right, five mice) and definition of similarity index for circle2 (inset) are shown. a.u., arbitrary units. j, Similarity index for new stimuli of the same mice as in panel f. All data are mean ± s.e.m. **P < 0.01 and ***P < 0.001 from paired, two-sided t-tests.