Summary of behavioral performances for mice in imaging experiments.

Size: px

Start display at page:

Download "Summary of behavioral performances for mice in imaging experiments."

Sabrina Evans
5 years ago
Views:

1 Supplementary Figure 1 Summary of behavioral performances for mice in imaging experiments. (a) Task performance for mice during M2 imaging experiments. Open triangles, individual experiments. Filled triangles, mean±s.e.m. n = 9 sessions from 5 mice. (b) Same as (a) for ALM imaging experiments. n = 8 sessions from 4 mice. (c) Same as (a) for V1 imaging experiments. n = 4 sessions from 2 mice.

3 Supplementary Figure 2 Slower first lick times for trials that were incongruent with the stimulus response contingencies of the prior block. (a) The difference in mean first lick times (see Methods for definition) between left and right responses for perseverative hit trials during sound block. Open triangle, individual experiment. Filled triangle, mean±s.e.m. This plot shows that a mouse could respond faster to either left or right lick port. The differences could be due to a combination of factors including placement of the lick ports and the internal bias of the animal. (b) Schematic illustrating perseverative hit, other hit, and perseverative error trials. (c) Based on the analysis in (a), either left or right is denoted as the fast direction. The mean first lick times for the fast direction are plotted for perseverative hit (phit), other hit (ohit), and perseverative error (perr) trials occurring during pre-switch (20 trials before block switch) or post-switch (20 trials after block switch) condition. The Wilcoxon signed-rank test was used to assess differences between phit trials and ohit or perr trials in the same block type. We also tested sound versus action blocks, and found no significant differences for any comparisons of the same trial types (p>0.05, Wilcoxon signed-rank test). Line, individual experiments. Filled triangles, mean±s.e.m (d) Same as (c) except reporting the average of mean first lick time for the fast and slow directions for each experiment. There were no significant differences for any comparisons of the same trial types between sound and action blocks (p>0.05, Wilcoxon signed-rank test). (e) Same as Fig. 1e, except using only the 20 trials pre-switch. (f) Same as Fig. 1e, except using only the 20 trials post-switch. n = 9 sessions from 5 mice.

5 Supplementary Figure 3 Video tracking of whisker and hindpaw positions during the adaptive decision-making task. (a) Still frame from a video taken using an infrared webcam. One whisker on each side of the face was painted using a phosphorescent paint marker (cat. #222-C-GO, Marvy Uchida). Whiskers were illuminated by two infrared light-emitting diodes (IR-LED). A third IR-LED was programmed to turn on 1 s before auditory cues, so the video could be synchronized to the task events. When the room light was turned off, the reflected infrared light could be visualized by the webcam. (b) Mean position of the left and right whiskers for different trial conditions for subject M12. These traces were averaged across pre-switch, correct trials. For this analysis, each whisker was converted to a binary mask based on pixel intensity. The centroid of the mask was located in each frame, and then a line was fitted between the centroid and a fixed point on the face. The angle between the fitted line and the horizontal axis is reported, with positive deflections being protractions and negative deflections being retractions of the whisker. (c) Same as (b) for another subject, M13. (d) Still frame from a video showing a painted hind paw. The webcam was positioned under the animal, to record footage through the acrylic tube. The synchronization IR-LED is visible. (e) Mean x- and y- positions of the right hindpaw for different trial conditions for subject M12. Hindpaw and whisker data were obtained from two different behavioral sessions.

Supplementary Figure 4 Approximating the spread of muscimol during inactivation experiments by injecting low-molecular-weight fluorescein.

6 Supplementary Figure 4 Approximating the spread of muscimol during inactivation experiments by injecting low-molecular-weight fluorescein. Fluorescein (disodium salt, MW=412; #S25328, Fisher Scientific) was injected bilaterally into M2, at the same concentration and volume used for muscimol (5 mm, 46 nl per hemisphere). Animal was euthanized at about the time when behavioral testing would have occurred (1 hr after injections) via transcardial perfusion, and the brain was immediately sectioned with a vibratome. Using a widefield fluorescence microscope, five images were taken for each brain slice and then stitched together to generate the figures. The image series was fitted to wireframes from the Paxinos and Franklin mouse atlas. These results showed a spread of ~1 mm along the anterior-posterior axis. The fluorescein dye was largely restricted to M2 and Cg1.

7 Supplementary Figure 5 Bilateral inactivation of M2 has no effect on mean lick rate or first lick response time. (a) Same analysis as Fig. 1e using data from saline- (black, red, blue) and muscimol-injected (overlaid in green) mice. Lick rates in each rule-sensory-motor combination were compared between saline and muscimol conditions for each 0.1 s bin, and the bars at the top of the panels denote significant differences (p<0.01, t-test; only 4 bins had significant difference by this measure, and none were consecutive bins). Line, mean. Shading, ±s.e.m. (b) Same analysis as Supplementary Fig. 2c, using data from the saline- (black, red) and muscimol-injected (green) mice. The Wilcoxon signed-rank test was used to assess differences between saline and muscimol conditions. n = 11 mice.

8 Supplementary Figure 6 Relationship between task parameters and neural/behavioral transitions. (a) Ensemble transition trials plotted against the behavioral transition trial from the same block (see Methods for definition of transition trials). Each point represents one block switch. Bold circle, median value. Left panel: n = 33 switches to a sound block from 9 sessions from 5 mice. Right panel: n = 35 switches to a spatial block from 9 sessions from 5 mice. Symbols denote the different experimental sessions. (b) Same as (a), except symbols denote the subjects tested. (c) Same as (a), except symbol sizes denote the number of other errors in the block. (d) Same as (a), except symbol sizes denote the number of perseverative errors in the block.

9 Supplementary Figure 7 Strategy-specific differences in decoding accuracy and neural trajectory separation are not due to differences in trial conditions. (a, c) Same analysis as performed for Fig. 5d except using only the subset of trials with matched trial conditions. (b, d) Same analysis as performed for Fig. 5e except using only the subset of trials with matched trial conditions. (e, f) Same analysis as performed for Fig. 6b except using only the subset of trials with matched trial conditions. The Wilcoxon signed-rank test was used to test for significance, and the p-values are noted in the figure. n = 9 sessions from 5 mice.

10 Supplementary Figure 8 Choice selectivity for neurons in M2, ALM and V1. (a) Selectivity for current choice was quantified by calculating the normalized ΔF/F difference during pre-switch, sound-guided trials. Each row represents an M2 neuron. Only M2 neurons that were significant for current choice, C(n), at t = 3 s from response are plotted. Cells are sorted by their choice selectivity. n = 9 sessions from 5 mice. (b) Same as (a) for ALM. n = 8 sessions from 4 mice. (c) Same as (a) for V1. Neurons that were significant for current choice, C(n), at t = 1 s from response are plotted. n = 4 sessions from 2 mice.

11 Supplementary Figure 9 Summary of behavioral and neural results. A summary of the task, behavioral performance, and neural dynamics. Recovery of behavioral performance following a switch was approximated as an exponential function for this schematic. The plots were generated based on the median values of behavioral transition trial (sound: 19.0, action: 17.5; dotted lines), neural transition trial (sound: 5.1, action: 14.6), midpoint trial (sound: 4.0, action: 10.4), steepness (sound: 1.02, action: 0.35), and range (sound: 0.36, action: 0.38).

12 Experiment Subject ID Trials Block Cells switches M2 imaging L M2 imaging M M2 imaging M M2 imaging L M2 imaging M M2 imaging M M2 imaging M M2 imaging M M2 imaging L Experiment Subject ID Trials Block switches Treatment Order of treatment Inactivation FA MUS 1 Inactivation FA VEH 2 Inactivation FA VEH 1 Inactivation FA MUS 2 Inactivation MUS 1 Inactivation VEH 2 Inactivation VEH 1 Inactivation MUS 2 Inactivation L VEH 1 Inactivation L MUS 2 Inactivation L MUS 1 Inactivation L VEH 2 Inactivation M MUS 1 Inactivation M VEH 2 Inactivation M VEH 1 Inactivation M MUS 2 Inactivation M MUS 1 Inactivation M VEH 2 Inactivation M VEH 1 Inactivation M MUS 2 Inactivation M MUS 1 Inactivation M VEH 2 Experiment Subject ID Trials Block Cells switches ALM imaging M ALM imaging M ALM imaging M ALM imaging M ALM imaging M ALM imaging M ALM imaging M ALM imaging M

13 Experiment Subject ID Trials Block Cells switches V1 imaging M V1 imaging M V1 imaging M V1 imaging M Experiment Subject ID Trials Block switches Whisker M tracking Whisker M tracking Hindpaw tracking M Supplementary Table 1: Mice used in this study

SUPPLEMENTARY INFORMATION

SUPPLEMENTARY INFORMATION doi:1.138/nature1139 a d Whisker angle (deg) Whisking repeatability Control Muscimol.4.3.2.1 -.1 8 4-4 1 2 3 4 Performance (d') Pole 8 4-4 1 2 3 4 5 Time (s) b Mean protraction angle (deg) e Hit rate (p