In an attempt to determine whether temporal references identified in dreams follow the same temporal distributions as those documented for autobiographical memories, 28 younger women (18-35 years of age) and 30 older women (60-77 years of age) kept a home dream diary for 1 week and then slept 1 night in the laboratory for rapid eye movement sleep dream collection. The following morning, they identified temporal references in their dreams and produced a sample of autobiographical memories using the semantic cuing method. For both groups, there was a linear decrease in temporal references identified in dreams and autobiographical memories with increased remoteness for the last 30 years. As predicted, for the older group, there were similar cubic trends reflecting a disproportionately higher number of both temporal references identified in dreams and autobiographical memories from adolescence/ early adulthood compared with adulthood and childhood. The results support the notion of continuity between waking and dreaming memory processes.
Several authors have discussed how dream processes are related in some way to cognitive and memory processes (Cavallero & Foulkes, 1993; Hobson, 1988). Most notably, Foulkes (1985) stated that dream production mechanisms draw bits and pieces of elements within the vast mnemonic repertoire that contains all the information, knowledge, and personal memories accumulated over an entire lifetime. He further postulated that the cognitive organization of dreams should reflect that of the waking state. It would thus follow that any pattern by which temporal references would be included in dreams would parallel that of autobiographical memory, which has been studied and modeled quite systematically.
Indeed, Rubin, Wetzler, and Nebes (1986) integrated data from four studies on autobiographical memory (Fitzgerald & Lawrence, 1984; Franklin & Holding, 1977; Rubin et al., 1986; Zola-Morgan, Cohen, & Squire, 1983) and elaborated a three-component model to describe the distribution of autobiographical memories across the lifespan. The first component is a retention function referring to the general finding that it is increasingly difficult to recall memories as they become more and more remote. This retention function describes the entire distribution of autobiographical memories recalled by young adults (under 35 years of age), and for the most recent 30 years in the case of middle-aged and older adults. The second component of the model refers to a reminiscence process; that is, adults over the age of 40 have a tendency to recall life experiences dating back to adolescence and/or early adulthood. This component of the model is meant to explain the disproportionate amount of memories of events and activities that originally took place when a person was between 10 and 30 years old-a phenomenon often referred to as a differential resurgence or reminiscence bump of memories. Childhood amnesia constitutes the third component of the model: People of all ages rarely remember events that took place before they were 6 years old.
It was of great interest to determine whether references to past events in dreams share the same characteristics. The existing literature provides very limited information. First, several studies have concluded that a significant proportion of dreams contain elements that refer to experiences of the preceding day, a phenomenon known as day residues (Botman & Crovitz, 1989; Davidson & Kelsey, 1987; Epstein, 1985; Hartmann, 1968; Jouvet, 1979; Marquardt, Bonato, & Hoffman, 1996; Nielsen & Powell, 1992; Verdone, 1965), whereas other studies have focused on dream elements that stem from experiences that took place 1 week prior to the dream, a phenomenon called the dream-lag effect (Nielsen & Powell, 1992). Verdone's laboratory study was the first to systematically identify temporal references in dreams that were related to experiences from a remote past-that is, over 5 years prior to the dream. …