Academic journal article Psychonomic Bulletin & Review

What Form of Memory Underlies Novelty Preferences?

Academic journal article Psychonomic Bulletin & Review

What Form of Memory Underlies Novelty Preferences?

Article excerpt

Novelty preferences (longer fixations on new stimuli than on previously presented stimuli) are widely used to assess memory in nonverbal populations, such as human infants and experimental animals, yet important questions remain about the nature of the processes that underlie them. We used a classical conditioning paradigm to test whether novelty preferences reflect (1) a stimulus-driven bias toward novelty in visual selective attention or (2) explicit memory for old stimuli. Results indicated that conditioning affected adults' looking behavior in the visual paired comparison, but not their recognition memory judgments. Furthermore, the typically observed novelty preference occurred only when a bias toward novelty had no competition from a bias toward salience due to conditioning. These results suggest that novelty preferences may reflect attentional processes and implicit memory to a greater degree than explicit memory, a finding with important implications for understanding memory in nonverbal populations and the development of memory in humans.

A widely used method for testing memory in preverbal infants and nonhuman primates, as well as in adults, is the visual paired comparison (VPC). In this paradigm, two visual stimuli are presented simultaneously-one that was presented previously and one that is novel-and the length of time that observers look at each stimulus is measured. Human infants (e.g., Fantz, 1964), adults (e.g., McKee & Squire, 1993), and nonhuman primates (e.g., Nemanic, Alvarado, & Bachevalier, 2004) typically look longer at novel than at old (familiar) stimuli. This novelty preference reflects the influence of memory on subsequent visual processing and has great utility as a nonverbal measure of memory, but important questions remain about the nature of the processes that underlie it. In this article, we report a new familiarity-to-novelty shift within VPC trials and describe tests of what form of memory underlies novelty preferences. These results have important implications for understanding the ontogeny of memory in humans.

Familiarity-to-Novehy Shifts Within Individual Trials

The typical dependent measure examined in the WC is the novelty score, which is computed as the proportion of time participants look at the novel stimulus within VPC trials, each of which may last 5 sec or more. The main utility of a large novelty score (>50%) averaged over multiple trials is to infer memory for the old stimuli, but some attention has been paid to changes in looking behavior over time. In particular, research with infants, in which one item in a pair remains the same across test trials, while the other item changes, has shown that infants tend to look longer at repeated (familiar) stimuli during early test trials, but longer at novel stimuli during later test trials (e.g., Roder, Bushnell, & Sasseville, 2000; Rose, Gottfried, Melloy-Carminar, & Bridger, 1982).

Recent computational modeling of the VPC suggests that this familiarity-to-novelty shift in looking behavior may also take place within individual test trials (Sirois & Mareschal, 2004). Although no previous study has examined infant looking behavior within test trials, data from one study suggest that adults may show a tendency to look first at the familiar stimulus very early in the trial (during roughly the first second), followed by a tendency to look longer at the novel stimulus during the rest of the trial (Manns, Stark, & Squire, 2000). Since the tendency to look first at the familiar picture was not tested statistically or discussed by those authors, we tested this effect in the present study.

In Manns et al. (2000), the tendency to look first at the familiar picture was relatively small in magnitude, possibly because Manns et al. used preexperimentally familiar stimuli (i.e., color photographs of common objects). It is important to note that Sirois and Mareschal's (2004) models predict a larger effect of first looks to old stimuli with fewer previous presentations. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.