Skip to main content
. 2022 Jan 31;6(4):555–564. doi: 10.1038/s41562-021-01263-w

Extended Data Fig. 6. Model recovery results for Exp. 1 with full feedback (cf. Figure 3e).

Extended Data Fig. 6

Simple Q-learning (models Q1/Q2) could not be confidently distinguished from models Q1*/Q2*. However, symmetric (Q1/Q1*) and asymmetric learning (Q2/Q2*) were distinguished relatively well. See Methods: Model- and parameter recovery for details.