Figure 6.

Examining a possibility of cross-credit assignment on retrospective inference trials

(A) We reexamined trial transitions of the type used to study MBCA (Figure 2D). Here, however, we considered only retrospective inference format trial n, wherein the common outcome (e.g., monkey; designated as common on trial n + 1) appeared second. If credit from the first reward (related to garlic) is assigned to the second outcome, a reward for the first outcome promotes choice generalization.

(B) Choice-generalization probability as a function of the first outcome’s feedback (reward versus non-reward). Following reward, the generalization probability was numerically lower. Error bars correspond to SEM across participants calculated separately in each condition (n = 38).

See also Figure S6 for further analyses.