Skip to main content
. Author manuscript; available in PMC: 2019 Oct 1.
Published in final edited form as: Curr Biol. 2019 Aug 29;29(18):3029–3040.e4. doi: 10.1016/j.cub.2019.07.068

Figure 4. The Distribution of Sensorimotor Features and Their Utility for Predicting Trial Type and Choice.

Figure 4.

(A–F) Histogram of feature values and classifiers for predicting trial type using (A) roll angle, (B) whisk latency, (C) cue latency, (D) touch count, (E) radial distance at touch, and (F) angle at touch.

(G) Trial type prediction performance of logistic classifiers for all mice based on each of the six features. Bars represent SEM. p values; Wilcoxon signed-rank test against shuffled models. Touch trials only are shown. See also Figure S1.

(H) Choice prediction performance of logistic classifiers for all mice trained on pole position, each of the six features, or all six features combined. Bars represent SEM. Touch trials only are shown. values; Wilcoxon signed-rank test against shuffled models. See also Figure S2.