Extended Data Fig. 2: Error at the periphery and heading error of the mouse in different light and search conditions.
a, Illustration of two measures of homing error: error at the periphery (left) and heading error. Error at the periphery is the angle between the animal’s position when it reaches the periphery of the arena after pressing the lever, the center of the arena, and the center of the bridge. Heading error is calculated along the homing path (at 50 Hz) and is the angle between two vectors with origin at the animal’s position, one pointing in the animal’s direction of movement and the other pointing towards the bridge. The median heading error is calculated for each trial. b, Distribution of the error at the periphery (left) and the heading error (right) for homing paths in light and dark trials. c, Median error at the periphery (left) or median heading error (right) for each mouse in light and dark trials (N = 17 mice, two-sided Wilcoxon signed-rank test, error at the periphery: P = 1.526 ×10−5, heading error: P = 1.526 ×10−5). Box plots show median (center line), first and third quartiles (box bounds), and 1.5 × interquartile range (whiskers). d, Distribution of the error at the periphery (left) and heading error (right) for dark trials with short and long search paths. Trials within a session were divided into short and long search paths depending on whether they were below or above the median search path length of the session. e, Median error at the periphery (left) or heading error (right) for each mouse during light and dark trials with short and long search paths (N = 17 mice, two-sided Wilcoxon signed-rank test, error at the periphery, light: P = 1.526 ×10−5, dark: P = 2.136 ×10−4; heading error, light: P = 6.561 ×10−4, dark: P = 1.526 ×10−5). Box plots show median (center line), first and third quartiles (box bounds), and 1.5 × interquartile range (whiskers). f, Effect of search path length during trials in which the lever was located near the center of the arena. We considered only trials in which the lever was located within a 10 ×30 cm rectangle centered on the arena center (left). Error at the periphery for each mouse for long and short search paths in light and dark trials (right, N = 17 mice, two-sided Wilcoxon signed-rank test, error at the periphery, light: P = 5.569 ×10−3, dark: P = 6.653 ×10−3). Box plots show median (center line), first and third quartiles (box bounds), and 1.5 × interquartile range (whiskers). ****P < 0.0001, ***P < 0.001, ** 0.001 < P < 0.01.
