Changes in familiarity and recollection across the lifespan: an ERP perspective

Abstract

The ability to recognize previous experience depends on two neurocognitive processes, familiarity, fast-acting and relatively automatic, and recollection, slower-acting and more effortful. Familiarity appears to mature relatively early in development and is maintained with aging, whereas recollection shows protracted development and deteriorates with aging. To assess this model, ERP and behavioral data were recorded in children (9-10 years), adolescents (13-14), young (20-30) and older (65-85) adults during a recognition memory task in which the same items were studied and tested over four cycles. Participants decided whether each item was old or new and then whether the decision was associated with (Remember, R) or without (Know, K) contextual detail. Memory sensitivity was greatest in young adults, although all groups showed increases in memory sensitivity and R judgments with repetition. Familiarity-based processes (mid-frontal episodic memory, EM, effect) appeared to be used by adolescents, young and older adults, but apparently not to the same extent by children. Recollection-based processes (parietal EM effect) were recruited by children, adolescents and young adults, but to a much lesser extent by older adults. Repetition enhanced the parietal effect in all but older adults. However, post-hoc analyses indicated that reduced recollective processing was confined to poor-performing older adults. By contrast, children appeared to rely mainly on recollection concordant with their conservative decision criteria across tests. We conclude that episodic-memory development reflects the increasingly flexible and interchangeable use of familiarity and recollection with a breakdown in the latter at older ages, perhaps limited to poor-performing older adults.

Figure 1

Grand mean waveforms associated with hits and CRs in Tests 1&2 and Tests 3&4 for children and adolescents (1A) and young and older adults (1B). The data are depicted at the midline Fz site where the mid-frontal EM effect was largest in children, adolescents and young adults. The data at F3 and F4 are also illustrated, because older adults showed greater negative-going activity to Hits than CRs at the F3 scalp site and the largest magnitude mid-frontal EM effect at F4. ERP waveforms are also depicted at left (P3), midline (Pz) and right (P4) parietal scalp locations (all age groups) to illustrate the early-onset and subsequent parietal EM effects. Light gray shading denotes the mid-frontal EM effect, black shading the early-onset parietal effect, dark gray shading the parietal EM effect, and cross hatching the left-frontal negativity observed only in older adults. Note that, in order to demonstrate the negative-going ERP effects for older adults, the scale has been increased for this group. Arrows mark stimulus onset, with timelines every 300 ms.

Figure 1

Grand mean waveforms associated with hits and CRs in Tests 1&2 and Tests 3&4 for children and adolescents (1A) and young and older adults (1B). The data are depicted at the midline Fz site where the mid-frontal EM effect was largest in children, adolescents and young adults. The data at F3 and F4 are also illustrated, because older adults showed greater negative-going activity to Hits than CRs at the F3 scalp site and the largest magnitude mid-frontal EM effect at F4. ERP waveforms are also depicted at left (P3), midline (Pz) and right (P4) parietal scalp locations (all age groups) to illustrate the early-onset and subsequent parietal EM effects. Light gray shading denotes the mid-frontal EM effect, black shading the early-onset parietal effect, dark gray shading the parietal EM effect, and cross hatching the left-frontal negativity observed only in older adults. Note that, in order to demonstrate the negative-going ERP effects for older adults, the scale has been increased for this group. Arrows mark stimulus onset, with timelines every 300 ms.

Figure 2

Scalp topographies of the 300-500 ms and 500-700 ms regions (based on the Hit-CR difference means) putatively reflecting, respectively, familiarity- and recollection-based processes. The two left-hand columns represent Tests 1&2 and the two right-hand columns, Tests 3&4. For the children, note that the mid-frontal scalp map is not depicted for the 300-500 ms region (top left), as the difference between old and new items was not reliable (see text). Note also the different calibrations (D mV) to the right of the scalp maps for each of the four age groups. Unshaded regions reflect positivity; shaded regions reflect negative activity.

Figure 3

Grand mean waveforms associated with hits and CRs (across all 4 test blocks) for older-adult participants in the High- and Low-Pr subgroups. The early-onset and subsequent parietal EM effects are shaded, respectively in black and gray fill. Cross hatching indicates the left-frontal negativity which was reliable in the Low-Pr but not the High-Pr subgroup. Arrows mark stimulus onset, with timelines every 300 ms.

Figure 4

Schematic of the experimental paradigm showing both study and test trials.