Article Figures & Data
Figures
- Figure 1.
Scn2a+/− mice learn the foraging task as well as WT littermates. A, Schematic of the behavioral box. Animals self-initiate trials in the center nosepoke and subsequently choose between the left and right port for reward. B, Timeline of behavioral training. On days 1 and 2, animals learn to receive reward from only the left or right port, 100 trials per day. Day 3 onward, animals learn to receive reward from both the left and right ports. Reward probabilities across ports are 0% and 100%, and the reward contingencies switch after 100 trials or after 9/10 correct choices. Animals train for 10 d or until they complete two sessions >80% correct, whichever comes first. After training, animals are put on the full behavioral task (details in Materials and Methods). Pink denotes the section used for analysis (in C–E, day 1 corresponds to day 3 of the training timeline). C, Probability of choosing the 100% baited port across training sessions across all animals (left), males (center), or females (right). Scn2a+/− mice in cyan and WT littermates in black. Bars are means ± SEM. All animals: p = 0.382 for genotype and day; males: p = 0.379 for genotype and day; females: p = 0.451 for genotype and day, repeated measures two-way ANOVA. D, Probability of choosing the new 100% baited port after the reward contingences were reversed on day 1 of the training block across all animals (left), males (center), and females (right). Scn2a+/− mice in cyan and WT littermates in black. Bars are means ± SEM. All animals: p = 0.950 for genotype and trial; males: p = 0.950 for genotype and trial; females: p = 0.630 for genotype and trial, repeated measures two-way ANOVA. E, Probability of choosing the new 100% baited port after the reward contingences were reversed on the final training day for each mouse across all animals (left), males (center), and females (right). Note that only 10 trials are plotted here as animals are well-trained and experience uncued block changes after fewer trials than on day 1. Scn2a+/− mice in cyan and WT littermates in black. Bars are means ± SEM. All animals: p = 0.002 for genotype and trial; males: p = 0.014 for genotype and trial; females: p = 0.140 for genotype and trial, repeated measures two-way ANOVA.
- Figure 2.
Scn2a+/− mice perform the foraging task as well as WT littermates. A, Timeline of the behavioral training. Pink denotes section used for analysis. B, Schematic of the behavioral box. In this task, ports are set to 15% or 60% reward probabilities. C, Choice behavior from an example animal (green, 10-trial moving average) in response to changes in reward contingencies [black, ratio of left to right side reward probabilities; e.g., 60/(15 + 60) or 15/(15 + 60)]. D, Probability of choosing the new 60% baited port after the reward contingencies were reversed across all animals (left), males (center), and females (right). Scn2a+/− mice in cyan and WT littermates in black. Bars are means ± SEM. All animals: p = 0.543 for genotype and trial; males: p = 0.046 for genotype and p = 0.682 for genotype and trial; females: p = 0.680 for genotype and trial, repeated measures two-way ANOVA. E, Intertrial interval (ITI), measured as the time from when the animal left a side port to when it next entered the center poke, broken down by genotype and previous trial outcome for all animals (left), males (center), and females (right). Circles are individual animals. All WT, previously rewarded: median 1.8 s, interquartile range (IQR) 1.4–2.4 s; all Scn2a+/−, previously rewarded: median 1.5 s, IQR 1.2–2.1 s; all WT, previously unrewarded: median 1.1 s, IQR 0.7–1.4 s; all Scn2a+/−, previously unrewarded: median 1.0 s, IQR 0.7–1.3 s; male WT, previously rewarded: median 1.7 s, IQR 1.3–2.0 s; male Scn2a+/−, previously rewarded: median 1.5 s, IQR 1.2–1.8 s; male WT, previously unrewarded: median 1.0 s, IQR 0.7–1.2 s; male Scn2a+/−, previously unrewarded: median 0.9 s, IQR 0.6–1.1 s; female WT, previously rewarded: median 2.1 s, IQR 1.5–2.7 s; female Scn2a+/−, previously rewarded: median 1.5 s, IQR 1.2–3.2 s; female WT, previously unrewarded: median 1.3 s, IQR 0.8–1.7 s; female Scn2a+/−, previously unrewarded: median 1.4 s, IQR 0.8–1.4 s. Rewarded versus unrewarded, control: p < 0.001; rewarded versus unrewarded, Scn2a+/−: p = 0.002. Rewarded versus unrewarded, male controls: p = 0.005; rewarded versus unrewarded, Scn2a+/− males: p = 0.001. Two-way ANOVAs followed by pairwise Mann–Whitney tests and Bonferroni correction. F, The fraction of win-stay and lose-switch trials for all mice (left), males (center), and females (right). Circles are individual animals. All WT, win-stay: median 0.90, IQR 0.87–0.92; all Scn2a+/−, win-stay: median 0.92, IQR 0.89–0.94; all WT, lose-switch: median 0.44, IQR 0.37–0.50; all Scn2a lose-switch: median 0.38, IQR 0.30–0.45; male WT, win-stay: median 0.91, IQR 0.88–0.94; male Scn2a+/−, win-stay: median 0.91, IQR 0.89–0.94; male WT, lose-switch: median 0.38, IQR 0.31–0.48; male Scn2a lose-switch: median 0.41, IQR 0.31–0.45; female WT, win-stay: median 0.89, IQR 0.86–0.91; female Scn2a+/−, win-stay: median 0.92, IQR 0.87–0.95; female WT, lose-switch: median 0.45, IQR 0.38–0.41; female Scn2a lose-switch: median 0.34, IQR 0.25–0.48. No difference across genotypes, two-way ANOVAs followed by pairwise Mann–Whitney tests and Bonferroni correction. See Extended Data Figure 2-1 for behavioral data on a similar, head-fixed version of the task.
- Figure 3.
Varying block length did not affect Scn2a+/− mouse performance. A, Left, Plot of reward contingencies for a session with a block length of 40 and 15%/60% reward probabilities. Right, Probability of choosing the new 60% baited port after the reward contingencies were reversed across all animals, males, and females. Bars are means ± SEM. All animals: p = 0.644 for genotype and trial; males: p = 0.997 for genotype and trial; females: p = 0.638 for genotype and trial, repeated measures two-way ANOVA. B, Left, Plot of reward contingencies for a session with a block length of 100 and 15%/60% reward probabilities. Right, Probability of choosing the new 60% baited port after the reward contingencies were reversed across all animals, males, and females. Scn2a+/− mice in cyan and WT littermates in black. Bars are means ± SEM. All animals: p = 0.157 for genotype and trial; males: p = 0.573 for genotype and trial; females: p = 0.213 for genotype and trial, repeated measures two-way ANOVA.
- Figure 4.
Varying reward contingencies did not affect Scn2a+/− mouse performance. A, Left, Plot of reward contingencies for a session with a block length of 80 and 30% versus 30%, 15% versus 45%, and 10% versus 50% reward probabilities. Right, Probability of choosing the high reward port (when applicable) for the 10% versus 50%, 15% versus 45%, and 30% versus 30% contingencies across all animals. Scn2a+/− mice in cyan and WT littermates in black. Bars are means ± SEM 10% versus 50%: p = 0.351 for genotype and trial; 15% versus 45%: p = 0.689 for genotype and trial; 30% versus 30%: p = 0.150 for genotype and trial, repeated measures two-way ANOVA. B, Left, Plot of reward contingencies for a session with a block length of 80 and 15% versus 85%, 25 versus 75%, and 50% versus 50% reward probabilities. Right, Probability of choosing the high reward port (when applicable) for the 15% versus 85%, 25% versus 75%, and 50% versus 50% contingencies across all animals. Bars are means ± SEM 15% versus 85%: p = 0.843 for genotype and trial; 25% versus 75%: p = 0.669 for genotype and trial; 50% versus 50%: p = 0.318 for genotype and trial, repeated measures two-way ANOVA.
- Figure 5.
Scn2a+/− mice with seizures did not perform differently on behavioral tasks. A, Same as Figure 1C. Probability of choosing the 100% baited port across training sessions across all animals. N = 3 Scn2a+/− mice (2 males, 1 female) with observed home cage seizures overlaid in purple (animals are included in N = 24 mean data). The days where individual animals experienced seizures are denoted with symbols. B, Same as Figure 1D. Probability of choosing the new 100% baited port after the reward contingences were reversed on day 1 of the training block across all animals. N = 3 Scn2a+/− mice with observed home cage seizures overlaid in purple (animals are included in N = 24 mean data). C, Same as Figure 1E. Probability of choosing the new 100% baited port after the reward contingences were reversed on the final training day for each mouse across all animals. N = 3 Scn2a+/− mice with observed home cage seizures overlaid in purple (animals are included in N = 24 mean data). D, Same as Figure 3A. Probability of choosing the new 60% baited port after the reward contingencies were reversed (block length of 40 trials) across all animals. N = 3 Scn2a+/− mice with observed home cage seizures overlaid in purple (animals are included in N = 12 mean data).
Extended Data
Extended Data Figure 2-1
Performance of Scn2a+/- mice and wild-type littermates in a head-fixed version of the probabilistic reward task. A, Schematic of the behavioral task. On each trial, a head-fixed mouse makes a choice via a left or right tongue lick following an auditory go cue. Depending on the reward probabilities, the choice may lead to a water reward. Trials are organized in blocks with each block having a different set of reward probabilities, including “70:10” (70% chance of reward on the left side; 10% on the right) and “10:70” (10% on the left; 70% on the right). Uncued block switches occur when the animal satisfies the switching condition, which is to choose the side with higher reward probability for 10 trials and then perform an additional random number of trials drawn from a truncated exponential distribution. B, Performance in an example session for a Scn2a+/- mouse. Top, Reward probabilities. Bottom, Choice behavior and the outcome for each trial. Red bar, Left choice. Blue bar, Right choice. Black bar, Reward. C, Similar to B for a wild-type littermate mouse. D, Probability of choosing the side with high reward probability after a block transition for Scn2a+/- mice (blue) and wild-type littermates (black). Mean ± SEM. N = 3 Scn2a+/+ mice (2 male, 1 female) and N = 3 Scn2a+/- mice (1 male, 2 female). E, Fraction of trials in which mice employ win-stay or lose-switch strategies. Circle, Individual animals. Mean ± SEM. Download Figure 2-1, EPS file.
In this issue
