Age-Related Differences in the Impact of Prior Knowledge on Recognition Performance: A Face Recognition Study

ABSTRACT

Background/Study context: The dual-process hypothesis of memory and aging (DPHMA) postulates a decline of recollection with no decline in recognition. While the age-related recollection deficit is well-documented, any age-changes in the familiarity process remain unclear. Some studies have shown that familiar and meaningful material can enhance the recognition performance of older adults. The goal of the present study was to explore the impact of familiar material on age-related recognition-memory decline, using a dual-process approach.

Methods: One hundred participants (50 young adults and 50 older adults) performed two recognition tasks using an unfamiliar and a famous face recognition task. A receiver operating characteristic (ROC) analysis was used to examine the memory processes underlying recognition memory.

Results: The older adults showed lower performance on recognition accuracy and false alarms rate, only in the unfamiliar-face recognition task. A DPHMA-like pattern on unfamiliar-face recognition task was obtained. Prior knowledge in the famous-face recognition task improved recollection for older adults and no age-related deficit was found.

Conclusion: The present study suggests that episodic memory deficits in healthy aging are primarily driven by a recollection deficit while familiarity is relatively well spared. However, this age-related recollection deficit could be alleviated using knowledge-based material.

Notes

1. P(“yes”/target) = P(“yes”/new) + Rt+ (1-Rt)ϕ((d’/2)-c) – (1-Rl)ϕ(-(d’/2)-c) In this equation, Rt is the probability of target recollection (recall to accept); d’ is the difference between the mean of the old-item familiarity distribution and the mean of the new-item familiarity distribution; c is the decision criterion; Rl is the probability of lure (new) recollection (recall to reject); ϕ is the density function of the normal distribution (proportion of targets and lures that exceed the decision criterion c, given d’). We used the DPSD 1 model which used the following parameters: (1) variable parameters = R_old, d’ and c; (2) determined parameter: R_n = 0; (3) constraint: R_old ≥ 0 and v_old = 1.