Fig. 7: Expanding orientation bandwidth improves visual perception.

a Illustration of the visual discrimination task. The mouse initialized the trial by triggering the lick detection, followed by a 500 ms long ITI with a gray screen, followed by a visual cue (white cross) for 700 ms. Two stimuli were then shown next to each other, and the mouse had to touch one stimulus to report a choice. Mice then received a water reward at a central lick spout after touching the target stimulus. Triggering the lick detection (even if no reward was given) initiated the next trial. b Example stimuli for the spatial frequency discrimination task with different orientation bandwidths (left) and the orientation discrimination task with different spatial frequency bandwidths (right). c Spatial frequency discrimination performance for different orientation bandwidths. Spatial frequency discrimination performance was significantly higher for larger orientation bandwidth (n = 46 sessions from 6 mice). d Same as in panel c) but for the orientation discrimination performance with different spatial frequency bandwidth stimuli. (n = 25,17,19 sessions respectively.) e Psychometric curves for orientation discrimination performance at different target-distractor orientation differences. Colors show psychometric curves for different spatial frequency bandwidths. Expanding the spatial frequency bandwidth did not affect the maximal discrimination performance or discrimination thresholds. The horizontal dashed line shows the 72.7% discrimination threshold. Error bars are centered at the colored circles and show standard deviation (n = 30,18 and 24 sessions, respectively). f Discrimination thresholds for the three spatial frequency bandwidths showed no significant differences (n = 30, 18, and 24 sessions, respectively). Box plots indicate the median (horizontal line), interquartile range (box bounds: 25th–75th percentiles), and whiskers (1.5 × interquartile range). Stars in panels (c, d, and f) show the significance of a two-sided Wilcoxon signed rank test against the narrow condition. The significance threshold was α = 0.025 after Bonferroni correction for performing 2 tests. Panel (a) was created in BioRender. Balla, E. https://BioRender.com/g05l789 (2025).