Memory in the aging brain: doubly dissociating the contribution of the hippocampus and entorhinal cortex

Abstract

Since the time of Aristotle it has been thought that memories can be divided into two basic types; conscious recollections and familiarity-based judgments. Neuropsychological studies have provided indirect support for this distinction by suggesting that different regions within the human medial temporal lobe (MTL) are involved in these two forms of memory, but none of these studies have demonstrated that these brain regions can be fully dissociated. In a group of nondemented elderly subjects, we found that performance on recall and recognition tests was predicted preferentially by hippocampal and entorhinal volumes, respectively. Structural equation modeling revealed a double dissociation, whereby age-related reductions in hippocampal volume resulted in decreases in recollection, but not familiarity, whereas entorhinal volume was preferentially related to familiarity. The results demonstrate that the forms of episodic memory supported by the human hippocampus and entorhinal cortex can be fully dissociated, and indicate that recollection and familiarity reflect neuroanatomically distinct memory processes.

FIGURE 1

A: Hippocampal (H) and entorhinal (E) volumes were measured using a volumetric T1-wieghted image on a 1.5 T Siemans scanner. Brain volumes for each subject were normalized for brain size by dividing by total intracranial volume. B: Pearson correlations between medial temporal lobe volumes (i.e., hippocampal and entorhinal cortex) and mean recognition and recall scores. Recall was most strongly predicted by hippocampal volume, whereas recognition was most strongly predicted by entorhinal volume.

FIGURE 2

Modeling the effects of aging on recollection and familiarity. The ellipses and rectangles represent latent and manifest variables, respectively. Solid arrows reflect significant standardized regression coefficients and dotted arrows reflect nonsignificant coefficients. The model assumes that recognition relies on recollection and familiarity, whereas recall relies solely on recollection. Recall scores represent proportion of correctly recalled items in short-delay free recall and long-delay free recall. Recognition scores represent the proportion of recognized studied items (hits), correct rejections for related lures, semirelated lures, and unrelated lures. Hippocampal volume (HIP) was allowed to influence Recollection, entorhinal volume (ENT) was allowed to influence familiarity, and aging was allowed to influence HIP.

FIGURE 3

The effects of increasing hippocampal and entorhinal cortex by one standard deviation on recollection and familiarity. Changes in hippocampal volume had larger effects on recollection than familiarity, whereas changes in entorhinal volume had larger effects on familiarity than recollection.