Supplementary Figure 9: Relationships between bouton activity and visual flow or running speed.

(a) Proportions of dLGN and LP boutons with significant visual flow tuning curves of different shapes obtained in the open-loop condition (significance test excluding stationary periods of running speed < 3 cm s⁻¹, see Methods). Boutons either increased (incr.) or decreased (decr.) their activity with increasing visual flow speed, or they exhibited band-pass tuning curves (BP). Shown are mean proportions and s.e.m over sessions. A higher fraction of dLGN than LP boutons preferred high visual flow speed (P = 0.0054), while more LP boutons preferred low visual flow speed (P = 0.0001). (b) Proportions of dLGN and LP boutons with significant running speed tuning curves of different shapes obtained in the open-loop condition. No differences between dLGN and LP were observed (P-values > 0.05). (c) Left: examples of running speed tuning curves in the dark. Right: proportions of significant running speed tuning curves with different shapes as in b, obtained in the dark. No differences between dLGN and LP were observed (P-values > 0.05). (d) Proportions of boutons significantly informative about running speed in the dark (random forests decoder, PP > 0.16, see Methods) were not different between dLGN and LP (P = 0.10). (e) Examples of RS−VF tuning curves of dLGN and LP boutons in the open loop condition of dLGN and LP boutons (left) and proportions of dLGN and LP boutons with significant RS−VF tuning curves of different shapes (right) obtained in the open-loop condition. Open-loop: dLGN, n = 18 sess., LP, n = 31 sess.; dark: dLGN, n = 18 sess.; LP, n = 29 sess.; dLGN = 8 mice, LP = 10 mice. **, P < 0.01; ***, P < 0.001, Wilcoxon rank-sum test. Error bars indicate s.e.m.