Long term implicit and explicit memory for briefly studied words

Long term implicit and explicit memory for briefly studied words Lee Averell (Lee.Averell@newcastle.edu.au) School of Psychology, University of Newcastle. Andrew Heathcote (Andrew.Heathcote@newcastle.edu.au) School of Psychology, University of Newcastle. Abstract Memories fade over time, but do they disappear altogether? The persistence of overlearned material has been regarded as evidence for permanent memories (Bahrick, 1984). However there seems a general consensus in the cognitive literature that briefly studied stimuli disappear from memory altogether (Wixted, 2004ab). We present evidence from implicit (stem completion) and explicit (stem cued recall) memory tasks showing well above chance performance 28 days after only one brief study event. Retention measured by both the implicit and explicit tasks was stable at the same level from seven days to 28 days. Our results question the consensus about the fate of memories for briefly studied stimuli. Keywords: Forgetting; long term memory; recall; context. The Persistence of Long Term Memory Whether or not stored memories are permanent is a subject of debate in both the cognitive and neuroscience communities. Here we investigate the issue of memory permanence and the way in which forgetting progresses over time in both explicit and implicit memory paradigms (Jacoby, Toth & Yonelinas, 1993; Squire, 1994). We first review previous pertinent findings and then report data from a stem-cued recall (explicit) and a stem completion (implicit) task testing memory at delays ranging from one minute to one month. The results are discussed with reference to novel context reinstatement techniques used in our experiment and their effects on mnemonic interference. A number of studies have reported accurate explicit recall of information memorized years and even decades ago. Bahrick’s (1984) seminal study showed that people could remember Spanish language words learned in a classroom for up to 50 years despite not using Spanish since the end of their high school education. He concluded that some memories are relatively permanent, fading only because of the detrimental effects of aging. More recently Squire (1989) found a similar pattern to Bahrick with memory for television programs from over 15 years prior to testing. Schmidt, Peck, Paas and van Breukelen (2000) showed that people could remember street names from suburbs they lived in as a child. Noice and Noice (2002) showed that actors could recall aspects of a Shakespearean script up to 28 years after playing the role. Results such as these support Checlie’s (2006) statement that: “the inability … to account for permanent retention is a serious failing” (p.36). However, these studies of long term retention may be criticized on a number of grounds. They employed a cross sectional design that precludes investigation of the way in which individual memory traces are forgotten. Further, given the meaningful nature of the material remembered in these studies it is possible that participants were reminded of and rehearsed the material in the study-test lag. It would, therefore, be desirable to examine retention in a longitudinal design of material that is less likely to be rehearsed. The stimuli used in these studies were also initially very well learned. Hence, even if these studies demonstrate very long term retention, memory permanence may be restricted to such “overlearned” material. Bahrick (1984) suggested that participation in multiple spaced Spanish language courses in high school was a strong contributing factor to his results. Also, the effect of overlearning was evident in the differences in long term retention between higher achieving (A) students and less accomplished (B) students, with (A) students showing stronger retention in both the short and long term. Evidence for very long term retention of stimuli studied only once has been reported in the implicit memory literature. Slomon, Hayman, Ohta, Law and Tulving (1988, exp 2) found word fragment completion was primed at above chance levels by study of a list of words 16 months previously. However, while the study list was only presented once, study instructions encouraged a very thorough encoding; participants were asked to copy the study list words onto a piece of paper, rate the familiarity of each word and prepare for a memory test on the words. In a cued recall test of explicit memory Runquist (1983, 1986, 1987) showed that participants could recall words up to two weeks after an initial brief exposure to words when associative cues are given. However, in these experiments some study words were presented in multiple testing sessions to assess the effects of retesting on recall of untested items. This may have aided recall of the items tested only once, leading to inflated estimates of long term retention. Some studies of shorter term retention for briefly studied words have provided evidence of stable memory after an initial period of forgetting (i.e., an asymptote in the retention function). McBride and Dosher (1997) tested implicit and explicit memory with a stem completion task using study-test lags from one minute to one hour. They found that both types of memory declined over the first 15 minutes before stabilizing above chance with implicit performance below explicit performance. However, even though their analysis assumed an asymptote, they doubted their results supported memory permanence, saying “further decline would be visible in hours or days” (p.380). This