UCLA

This dissertation examines how manipulation of value during encoding leads to better memory for high-value items. We use a variant of the value-directed remembering paradigm (Castel, Benjamin, Craik, & Watkins, 2002), a particular context for manipulating value at encoding that provides learners with the opportunity to metacognitively assess and strategically control encoding processes. Prior studies to investigate effects of value on memory have largely focused on how signaling from dopamine-producing reward regions directly facilitates formation of memories in the hippocampus. Such mechanisms seem to have little effect on memory here. Instead, our focus is on the cortically-mediated processes that learners can strategically engage to improve memory. In Chapter 2, we show that the degree of value-related modulation of brain activity in regions typically associated with semantic processing is critical, as selectivity with this brain mechanism correlates with how strongly value affects free recall in young adults. In Chapter 3, we show that a similar mechanism underlies selectivity in older adults. We also find evidence that young adults engage additional brain mechanisms that older adults do not, such as an increased proactive engagement of prefrontal cortex during the cue that precedes the to-be-remembered word when that cue is high-value. These additional mechanisms turn out to be largely ineffective. The fact that age-dependent mechanisms are not effective may relate to why older adults are successful at being selective in this paradigm. In chapter 4, we find that value can also modulate memory via increased deactivations in medial posterior brain regions, in both young and older adults. In addition, a set of behavioral studies applying a dual-process analysis to value-directed remembering shows that high-value items in this paradigm typically show increases in both recollection and familiarity relative to low-value items, consistent with selective application of deep encoding strategies. When the opportunity for free recall with feedback is removed, and, separately, in a subset of participants who report being insensitive to value, only recollection is enhanced, consistent with effects driven by more automatic, possibly dopamine-driven mechanisms, in response to reward.