Fig. 5: Heterogenous tuning to eye position.
From: Mouse entorhinal cortex encodes a diverse repertoire of self-motion signals
ai Model-derived tuning curves (mean ± sem of 10 model folds) for cells significantly encoding horizontal eye position (Eh; Supplementary Figs 5–8). The stability (“S”) and the class of the tuning curve (1° = linear, etc.) are indicated at top. aii Bottom: Spiking from an Eh-encoding cell. Gray trace, Eh. Black dots, spikes. Top: The associated tuning curve. aiii Summary of tuning curve classes. Gray bars indicate nasal/temporal preference. aiv Preferred position for cells non-linearly encoding Eh. av Stability of tuning to Eh. Boxplots: linearly tuned cells were significantly more stable than those exhibiting greater curvature (median [1st–3rd quartile]; 1° = 0.93 [0.85–0.98]; 2° = 0.87 [0.71–0.95]; >2° = 0.66 [0.15–0.83]; Χ2 = 69.59, P = 7.7e−16, df = 2; 1° versus 2°, P = 2.6e−5, 1° versus > 2°, P = 1.8e−15, 2° versus > 2°, P = 4.1e−8; Kruskal–Wallis test with post hoc two-sided Wilcoxon rank-sum comparisons with α = 0.0167 after Bonferroni correction for multiple comparisons). bi Tuning curves for cells encoding vertical eye position (Ev). bii Spiking from an Ev-encoding cell. biii Summary of tuning curve classes. Bar color indicates upward/downward preference. biv Preferred position for cells non-linearly tuned to Ev. bv Stability of tuning to Ev. Boxplots: linearly or quadratically tuned cells were significantly more stable than those exhibiting greater curvature (median [1st–3rd quartile]; 1° = 0.86 [0.18–0.97]; 2° = 0.86 [0.53–0.94]; >2° = 0.56 [0.04–0.84]; Χ2 = 17.17, P = 1.9e−4, df = 2; 1° versus 2° P = 0.87, 1° versus > 2° P = 0.012, 2° versus > 2° P = 0.0001; Kruskal–Wallis test with post hoc two-sided Wilcoxon rank-sum comparisons with α = 0.0167 after Bonferroni correction for multiple comparisons). Boxplots: box, interquartile range; solid line, median; whiskers, range; outliers plotted separately.
