Fig. 7: Neurons that direct reorientation behaviors are broadly dysregulated in tdc-1 mutant animals.

a, Comparisons of activity features across neurons in WT and tdc-1 animals recorded through whole-brain imaging. Each row is a neuron; neurons with fewer than four recordings in either genotype were excluded from analyses (rows shaded black). From left to right, column variables are as follows: (1) neuron n. (2) Overall dynamic range for that neuron (calculated as the s.d. of F/F_mean activity). Higher values show that this neuron’s activity changes more during a recording. From top to bottom, P = 0.0004, <0.0001, <0.0001. (3) Forward encoding strength (the slope of this neuron’s tuning to velocity, as defined in ref. ¹⁵). Positive values show a neuron is forward encoding, negative values indicate reverse encoding. All significant P = <0.0001. (4) Activity at the reversal end and during the postreversal turn. This is calculated as the average activity in the ~5 s around the ends of reversals followed by ventral turns. All significant P = <0.0001. (5) Median half-decay time or encoding timescale (as in ref. ¹⁵). Color bar is on a log scale. From top to bottom, P = <0.0001, 0.0004. (6) Expression levels of the five tyramine receptors (data from ref. ⁴⁸). Black indicates that receptor expression is not detected. Differences in each category between WT and tdc-1 were determined through a two-sided Wilcoxon rank-sum test with Bonferroni correction. b, Quantification of behavior in WT and tdc-1(n3419) animals during whole-brain imaging. Two-sided Wilcoxon rank-sum test with Bonferroni correction (from left to right, P = 0.025, 0.014, 0.0034). n = 17 tdc-1 and 32 WT recordings. Data show mean ± s.e.m. c, Activity of forward-associated neurons in WT and tdc-1 animals aligned to forward run starts. Dashed black line shows at run start; red shading shows the reversal. n = 275–762 runs (n values on the plot show the number of recordings with data for that neuron). Wilcoxon rank-sum test with Bonferroni correction comparing activity between genotypes during the run (black asterisks) and reversal (red asterisks; P = <0.0001, <0.0001, 0.0003, <0.0001, 0.0021). Data show mean ± 95% CI. d, Aligned activity of head-steering circuit neurons at reversal ends, as in Fig. 2b. Only reversals followed by ventral turns are shown. n = 328–523 reorientations (n values on the plot show the number of recordings per genotype with data for that neuron). Two-sided Wilcoxon rank-sum test with Bonferroni correction, comparing ~5 s (one head swing) before or after the reversal end (P = <0.0001, <0.0001, <0.0001, <0.0001). Data are mean ± 95% CI. e, Z-scored neuron activity aligned to head curvature, as in Fig. 2a. As head-curvature frequencies vary across time and animals, activity is aligned to the crossing from dorsal (positive) to ventral (negative) and vice versa, as in Fig. 2a. Head curvature for both genotypes is shown on the right. n = 329–468 time windows of forward movement, n = 82–130 time windows of reverse movement (n values on the plot show the number of recordings per genotype with data for that neuron). Data are mean ± 95% CI. f, Connectivity of RIM and the head-steering network, using data from ref. ¹². g, Mock traces of RIM and the neurons of the head-steering network (reorientations in red), showing each neuron’s stereotyped changes across the behavior. These traces are drawn based on actual traces shown in Extended Data Fig. 8b. The first column lists behavioral features that we have shown affect each neuron’s activity. The second column shows the behavioral features we have shown are affected when these neurons are manipulated, either through optogenetics or cell-silencing/ablation experiments. For all panels, significance is noted as: NS (not significant), *P < 0.05, **P < 0.01, ***P < 0.001 and **P < 0.0001. For panels with multiple comparisons, symbols denote Bonferroni-adjusted P values.