Strategy-dependent coding of reward and delay during temporal discounting in macaque dorsolateral prefrontal cortex

Some rewards are worth the wait, while others are not. This behavior is governed by temporal discounting, whereby delays reduce the value of future rewards, such that only sufficiently large rewards justify long waits. Individuals differ widely in their strategies to weigh temporally delayed rewards. Prefrontal neurons are known to encode reward and delay; however, it remains unclear whether these neural representations reflect a common temporal discounting computation process, or whether they depend on the individual behavioral strategies for trading off reward and time. We trained two monkeys to choose whether to accept or reject delayed reward offers by releasing a lever at different times contingent on a go signal. We analyzed how the unforeseen go signals, which introduced different additional delays before an offer can be accepted, changed animals' temporal discounting decisions. Behavioral modeling revealed that the two monkeys adopted distinct strategies in re-evaluating the original reward offer in response to the go signal. Monkey S re-calculated the discounted value of the original offer by accounting for the additional delay associated with the go signal, whereas Monkey T treated the go signal as having additive values without re-evaluating the original offer. Although neurons in the dorsolateral prefrontal cortex (dlPFC) of both monkeys encoded reward and delay, we found important differences in how neurons represented these factors, which reflected monkeys' respective behavioral strategies. Our results revealed a strategy-dependent coding of reward and delay in dlPFC.