Author response image 2. Decoding performance during exploration in all three datasets as a function of the training time and corresponding number of laps used to build the encoding model.

(a) Median in-arm decoding error. (b) Fraction of position estimates that were located in the correct arm.