Research suggests that age-related changes are apparent in autobiographical memories when the qualitative content of the memories is examined. For example, older adults retrieve overly general memories (i.e. not restricted to a single event) relative to young adults' specific memories (the overgenerality effect) and rate their memories as more positive than young adults (the positivity effect). The majority of studies reporting these effects instruct participants to retrieve specific memories, thereby requiring participants to maintain task goals and inhibit inappropriate responses. Because these processes are impaired in healthy aging, these requirements may contribute to age-related differences in memory retrieval. To isolate underlying differences in memory representations, the current project utilized an autobiographical memory paradigm in which instructions were manipulated to separate age-related differences in underlying representations from the ability to follow task instructions. Music was selected as a retrieval cue due to its unique capability to elicit specific emotional memories without explicit retrieval instructions. Experiment 1 compared young and older adults' autobiographical memories under restricted and unrestricted retrieval conditions. The age-related overgenerality effect was reduced when participants were provided with unrestricted instructions, suggesting that this effect partially reflects a difficulty maintaining task instructions. The positivity effect was not observed in any condition, suggesting that memories cued by music may not exhibit the effect. Experiment 2 utilized event-related functional neuroimaging to examine age-related changes to the neural networks recruited during emotional and specific autobiographical memory retrieval. Young and older adults engaged many of the same regions during retrieval of specific events, including inferior and superior parietal lobes, medial temporal lobe, and prefrontal cortex, suggesting that retrieval in both groups may rely on similar underlying cognitive processes. During emotional memory retrieval, older adults recruited the medial prefrontal cortex to a greater extent than young adults. Importantly, although a positivity effect was not observed in the behavioral data, such age-related neural differences suggest that young and older adults process emotional memories differently. The current study provides valuable insight into how memory representations change with time and experience, highlighting circumstances that exaggerate and diminish age-related changes, and neural differences that exist when behavior is equivalent.