Model comparison
| AIC | BIC | Prediction (% correct) | |
|---|---|---|---|
| Rescorla–Wagner rule | 1.114±0.03 | 1.119±0.03 | 73.6±0.5 |
| SP model | 1.096±0.04 | 1.101±0.04 | 74.2±0.1 |
| Logistic regression | 1.311±0.01 | 1.314±0.02 | 72.6±0.5 |
Normalized AIC, normalized BIC, and accuracy of the model's prediction for the animal's actual goal choices (% correct) are shown for two RL models (Rescorla–Wagner rule and SP model) and the logistic regression model. Data are mean ± SD (n = 3 animals).