Differential aging of the medial
temporal lobe
A study of a five-year change
N. Raz, PhD; K.M. Rodrigue, MS; D. Head, PhD; K.M. Kennedy, MS; and J.D. Acker, MD
Abstract—Objective: To test the hypothesis that entorhinal cortex (EC) volume decreases at a slower rate than the
hippocampal (HC) volume in healthy adults, and to examine whether the rate of shrinkage increases with age. Methods:
Volumes of the HC and EC were measured twice on MRI scans of 54 healthy adults (aged 26 to 82 years), with an average
interval of 5 years. Results: Markedly different age trends were noted in the examined regions. The EC showed no
age-related differences on both occasions and only minimal age-related change (0.33%/y). By contrast, the HC exhibited
significant age-related differences at baseline and at follow-up evaluation and decreased at a faster pace of 0.86%/y. Older
participants (aged 50 years) showed increased annual shrinkage of the HC (1.18%) and EC shrinkage (0.53%/y). The rate
of HC volume loss significantly exceeded that of the EC. No EC shrinkage and modest HC volume reduction were observed
in people aged 50 years. Conclusions: Age-related shrinkage occurs in the medial temporal lobes of healthy adults, with
significant hippocampal decline and minimal entorhinal changes. In both regions, the rate of decline accelerates with age,
although the role of pathologic factors in age-related increase of volume loss merits further investigation.
NEUROLOGY 2004;62:433–438
The medial temporal lobe structures—the hippocam-
pus (HC) and the entorhinal cortex (EC)—play an
important role in episodic memory,
1
which declines
with age
2
and is profoundly impaired in patients
with Alzheimer disease (AD).
3
Although neuropatho-
logic changes are found in both regions at the earli-
est stages of AD, the effects of normal aging on EC
and HC are unclear.
4
The estimates of age-related
differences in the HC volume derived from cross-
sectional MRI studies range from large effects
5
(r =
-0.65) to essentially zero
6
(r = 0.05), with median
age-volume correlation of r =-0.31.
7
It is possible
that age-related differences in HC volume are stron-
ger among older adults than among their younger
counterparts
5,8,9
and are more likely in men than in
women, even at younger ages.
10
Although the HC is affected early in the course of
AD, the EC is believed to be the first cerebral struc-
ture to show AD-type pathology.
4,11
Moreover, deteri-
oration of EC may be a preclinical sign of incipient
dementia. For example, attrition of EC neurons in
lamina II is more likely to be discovered in nonde-
mented but cognitively impaired adults than in their
counterparts who died without exhibiting signs of
cognitive decline,
12
and deficient metabolic activity in
the EC of normal elderly persons predicts onset of
memory decline 3 years later.
13
The contrast between
normal and pathologic aging seems stronger in EC
than in HC. Unlike HC, EC volume is not associated
with age in humans,
14
and in nonhuman primates
the number of EC neurons does not change with
age.
15,16
Although reduction in the HC volume is
strongly associated with concurrent AD
17-19
and se-
verity of antemortem AD pathology,
20
the volume of
the EC may be a better predictor of progression from
minimal preclinical impairment to AD.
21-23
The estimates of age-related changes in HC and
EC are based mainly on cross-sectional studies, and
only a handful of longitudinal investigations have
been conducted to date. Those studies revealed longi-
tudinal shrinkage of the HC
24-28
and, in a single sam-
ple of very old adults, reduction in the volumes of the
parahippocampal gyrus (that includes EC) and the
temporal lobes.
26
In one study, longitudinal changes
in EC volume were observed in healthy nonde-
mented elderly persons.
29
Notably, most longitudinal
studies of volumetric brain changes have covered
only short periods
30
and were based on relatively
small samples with age ranges restricted to the later
2 or 3 decades of life.
In this study, we examine longitudinal change in
the medial temporal structures (HC and EC) in a
sample of healthy adults who span a wide age range.
Based on the reviewed literature, we expected that
comparing brain measures taken 5 years apart
would reveal significant reduction in the HC but not
in the EC volume. We also hypothesized that be-
cause of the nonlinearity noted in cross-sectional
From the Institute of Gerontology and Department of Psychology (Dr. Raz, K.M. Rodrigue and K.M. Kennedy), Wayne State University, Detroit, MI;
Department of Psychology (Dr. Head), Washington University, St. Louis, MO; and Baptist Memorial Hospital-East (Dr. Acker), Diagnostic Imaging Center,
Memphis, TN.
This study was supported in part by National Institutes of Health grant AG-11230 to N.R.
Received March 3, 2003. Accepted in final form August 13, 2003.
Address correspondence and reprint requests to Dr. N. Raz, Institute of Gerontology, 87 East Ferry Street, Detroit, MI 48202; e-mail: nraz@wayne.edu
Copyright © 2004 by AAN Enterprises, Inc. 433