Decision and navigation in mouse parietal cortex
- University College London, United Kingdom
Figures
Figure 1 with 1 supplement
Imaging PPC activity during a navigation task driven by visual decisions.
(a) Schematic view of the experimental setup. The monitors are positioned at 90 deg relative to each other, spanning 270 deg of the horizontal field of view. (b) Schematic of the virtual …
Figure 1—figure supplement 1
Location of imaged neurons relative to somatosensory and primary visual cortices.
Outlines of visual, primary visual, somatosensory and motor cortices were derived from Allen Institute atlas, and aligned to coordinates relative to bregma – [0, 0]. Magenta circle represents the …
Figure 2 with 1 supplement
Predicting the responses of PPC neurons based on position and heading.
(a,b) Activity () of an example neuron, plotted in pseudocolor, for trajectories that ended in leftward (a) or rightward (b) choices. Columns correspond to different stimulus contrasts and sides …
Figure 2—figure supplement 1
Single-trial responses and model predictions for three example cells.
(a) Responses (top) and model predictions (bottom) for the cell in Figure 2h. Format as in Figure 2b,d. (b,c) same as a, for the example cells in Figure 2i–j.
Figure 3
Position and heading are sufficient to explain the dependence of responses on choice.
(a,b) The position-heading model correctly predicts the sequential choice-selective activations seen in the data (compare to Figure 1k–n, where cells are arranged in the same order). (c,d) The …
Figure 4 with 3 supplements
Effects of adding decision as an explicit variable.
(a) Average responses of the example cell shown previously (Figure 2a–g) in the range of positions (z) of 60–80 cm, where the cell responds maximally, for trials ending in leftward choices (red) or …
Figure 4—figure supplement 1
Quality of fits by position-heading model across all neurons in individual mice, measured by the correlation between the trial-averaged raw data and the model predictions.
(a–g) The seven individual mice, each with the symbol used to denote it in Figure 4. The genetic backgrounds of the mice were C57bl/6 (a,b), Camk2a-TTA;Ai93;Emx1 (c–e), and Ai95;Slc17a7-Cre (f,g). …
Figure 4—figure supplement 2
Assessing the role of visual factors.
(a) In the playback condition, the estimated position-heading fields of PPC neurons were either much weaker (example cell 1) than during the task behavior, or completely absent (example cell 2) (b) …
Figure 4—figure supplement 3
Assessing the role of motor factors.
(a) The position-heading model performs better than the alternative motor model, as measured by cross-validated explained variance. (b) The explained variance of the position-heading model is …
Figure 5 with 1 supplement
Decoding animal position and choice from neural activity.
(a) The posterior estimate of the position closely follows the actual trajectory of the animal. Different rows represent different trials; different columns represent different moments in the trial. …
Figure 5—figure supplement 1
Full trajectory decoding from a sequence of posterior distribution estimates.
(a) Posterior distribution estimated from PPC population activity at a specific time t and z= z(t). (b) Probability distribution of the heading angle is estimated by calculating an integral of …
Videos
Video 1
A mouse performing some trials of the task.
https://doi.org/10.7554/eLife.42583.005
Video 2
Frame-by-frame decoding of mouse position from PPC population activity.
The main rectangle represents the position of the mouse in the coordinates of position (z) vs. heading angle (θ). For each trial, the contrast and position of the stimulus shown in that trial is …
Additional files
-
Transparent reporting form
- https://doi.org/10.7554/eLife.42583.016
