Extended Data Fig. 3: Additional analysis of the fitted nonlinear selectivity models. | Nature Neuroscience

Extended Data Fig. 3: Additional analysis of the fitted nonlinear selectivity models.

From: Hue selectivity from recurrent circuitry in Drosophila

Extended Data Fig. 3

(a-b) Different sensitivities to the hue angle relative to the preferred direction of tuning for fixed saturation in the nonlinear selectivity model. The gray dashed line shows the linear case and the colored lines show increasing sensitivity to hue, as indicated by an increase in κ. (c) Different sensitivities to saturation for a fixed hue angle. The gray dashed line shows the linear case and the color lines show decreasing sensitivity to saturation, as indicated by a decrease of the α parameter in the nonlinear selectivity model. (d-g) Relative responses of the non-linear selectivity model (colored) and the data (gray) across \(\cos (\theta )\), the cosine of the angle between the preferred tuning direction and the stimulus direction. The dashed line is the fitted curve using the obtained κ. (h-k) Relative responses of the non-linear selectivity model (colored) and the data (gray) across saturation values. The dashed line is the fitted curve using the obtained α. Only responses with a \(\cos (\theta )\) greater than 0.5 are shown. (l) Mean R2 values using a linear model with the additional Rh1 feature (that is the calculated excitation of R1-6 photoreceptors) for all neurons. The error bars indicate the 95% confidence interval of fitting the models to the bootstrapped iterations of the data.

Source data

Back to article page