• aging
• memory loss
• hippocampus

Memory complaints are very common in elderly people. In 1958 Kral¹ had already introduced the terms “benign and malignant senescent forgetfulness,” attempting to differentiate between “normal” and pathological decline in memory performance during aging. Since that time the discussion has been continuing and many definitions for a suspected “transitional” stage between normal aging and dementia have been proposed. In the currently widely used criteria of “mild cognitive impairment” (MCI) by Petersen et al,² one of the five key elements is objective memory impairment on testing, corrected for age and education, suggesting that age has an influence on memory performance. Yet little is known about the cognitive and structural changes that occur in the normal brain with aging. The study by Lyle et al in this issue tries to define magnetic resonance imaging (MRI) correlates of normal brain aging.³

The investigators carried out cranial MRI in 102 very old community dwelling individuals (aged 82 to 94 years). Using a multivariate model with hippocampal size, estimated visually and volumetrically, sociodemographic factors, and age related disease as predictors of memory performance, they found that left hippocampal measurements were predictive, especially on delayed retention of verbal material. Even in a subgroup of cognitively healthy elderly people (clinical dementia rating (CDR) = 0), left hippocampal measures remained predictive of delayed retention of verbal information. For those fond of visual rating of hippocampal atrophy, the results support the view that visual and volumetric estimates of hippocampal size perform equally well.⁴

These results strengthen the already widely known relation between the hippocampus and memory, but what do they tell us about (normal) aging? When comparing the cognitively healthy group (n = 57) with the group as a whole, the difference in influence of age on memory is striking. Age was a significant predictive factor in memory performance in the group as a whole, but lost its predictive value in the cognitively healthy group. In the light of this, is it possible that (a part of) the subgroup of CDR ⩾0.5 was mildly demented, and although these subjects were selected from the community, can they still be regarded as representatives of “normal” aging? If not, this might be considered an indication that age predicts memory deficits through the age related presence of hippocampal pathology. In this respect it would have been interesting to compare hippocampal sizes between the two groups. It has been shown repeatedly that hippocampal atrophy is a sensitive marker of early Alzheimer’s disease pathology. Jack et al have observed a decline in volume of all medial temporal lobe structures, including the hippocampus, with advancing age, and were able to discriminate normal aging from early dementia among subjects with a CDR of 0.5.⁵

Studies correlating MRI and pathology in the very old are much needed. Hippocampal atrophy may be sensitive to Alzheimer’s disease⁶ but lacks specificity, and the pathological basis of radiologically observed hippocampal atrophy in normal aging is still largely unknown. It seems reasonable to assume that besides Alzheimer’s disease pathology there may be other processes in aging that target the hippocampus, such as hippocampal sclerosis or vascular damage.

A better understanding of the basic mechanisms of cognitive function and its anatomical correlates in normal brain aging will help clinicians to diagnose and treat memory disorders in the elderly and possibly to prevent dementia as the ultimate consequence of longevity.⁷