Primate ventral striatum maintains neural representations of the value of previously rewarded objects for habitual seeking

The ventral striatum (VS) is considered a key region that flexibly updates recent changes in reward values for habit learning. However, this update process may not serve to maintain learned habitual behaviors, which are insensitive to value changes. Here, using fMRI in humans and single-unit electrophysiology in macaque monkeys we report another role of the primate VS: that the value memory subserving habitual seeking is stably maintained in the VS. Days after object-value associative learning, human and monkey VS continue to show increased responses to previously rewarded objects, even when no immediate reward outcomes are expected. The similarity of neural response patterns to each rewarded object increases after learning among participants who display habitual seeking. Our data show that long-term memory of high-valued objects is retained as a single representation in the VS and may be utilized to evaluate visual stimuli automatically to guide habitual behavior.