An instructive reinforcement learning model with a free constant learning rate best explains participants' contingency learning¶

Claim · All claims
Evidence · 1 study · causal r2 · q3
Currency · newest study 2023 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i? Model comparison by BIC established that the fLRI model fitted participants' behaviour better than the dLRI, dfLRI and fLRE models in both experiments. → Francesco Pupillo 2023

Evidence¶

Francesco Pupillo 2023¶

Francesco Pupillo, Javier Ortiz-Tudela, Rasmus Bruckner and Yee Lee Shing. (2023). The effect of prediction error on episodic memory encoding is modulated by the outcome of the predictions. npj Science of Learning 8:18. https://doi.org/10.1038/s41539-023-00166-x

q3 · i? · causal · r2

Computational model comparison across four Rescorla-Wagner variants fitted to learning and encoding data from both experiments, using subject-wise BIC with parameter penalisation. fLRI had the smallest BIC (266.4 in Experiment 1; 783.1 in Experiment 2, per Table 1).

"Model comparison established that the instructive model with the free, constant learning rate (fLRI) explained participants' behaviour better than the other models. The overall BIC of fLRI was the smallest (indicating better fit), and the number of participants for which it was the best model was 20 over a total of 32 participants for Experiment 1 and 23 over a total of 40 participants for Experiment 2."

Discussion¶

Related Claims¶

• A multidimensional IRT model with simple structure fits testlet-based reading data better than both the general polytomous testlet model and the standard 2PL/GPCM — related
• Preschoolers' Chicken game choices are best explained by reward-based learning with a higher learning rate for the competitive opponent — related