MIT Press

We considered four ways in which the reward probabilities

p^{i}

are set, illustrated schematically in Figure 1c. First, we considered stable environments in which reward probabilities were constant. We also considered conditions of 1 reversal and 3 reversals where the payout probabilities were reversed to

1 - p^{i}

once in the middle of the task (second display in Figure 1c) or three times at equal intervals (third display in Figure 1c). In stable, 1 reversal and 3 reversals conditions, the initial probabilities

p^{i}

at the start of the task were sampled at intervals of 0.1 in the range [0.05, 0.95] such that

p^{1} \neq p^{2}

⁠, and we tested all possible combinations of these probabilities (45 probability pairs). Unless otherwise noted, results are averaged across these initial probabilities.

Table 1:

Learning Rates in the Confirmation and Alternative Models.

	Chosen Option $i$		Unchosen Option $j \neq i$
Model	$δ_{t}^{i} > 0$	$δ_{t}^{i} < 0$	$δ_{t}^{j} > 0$	$δ_{t}^{j} < 0$
Confirmation model	$α^{C}$	$α^{D}$	$α^{D}$	$α^{C}$
Valence model	$α^{+}$	$α^{-}$	$α^{+}$	$α^{-}$
Hybrid model	$α^{+}$	$α^{-}$	$α^{=}$	$α^{=}$
Partial feedback	$α^{+}$	$α^{-}$	—	—

	Chosen Option $i$		Unchosen Option $j \neq i$
Model	$δ_{t}^{i} > 0$	$δ_{t}^{i} < 0$	$δ_{t}^{j} > 0$	$δ_{t}^{j} < 0$
Confirmation model	$α^{C}$	$α^{D}$	$α^{D}$	$α^{C}$
Valence model	$α^{+}$	$α^{-}$	$α^{+}$	$α^{-}$
Hybrid model	$α^{+}$	$α^{-}$	$α^{=}$	$α^{=}$
Partial feedback	$α^{+}$	$α^{-}$	—	—

Note: To make the table easier to read, $α^{C}$ and $α^{+}$ are highlighted in bold.

View Large

This Feature Is Available To Subscribers Only