82 Cognitive impairment, especially memory dysfunction, occurs in the course of human aging and precedes the development of Alzheimer’s disease (AD), a degenerative neurologic illness. 1,2 Older individuals with subjective memory complaints frequently test within age-adjusted means on standardized neuropsychological tests; how- ever, recent studies have shown a correlation between mild memory impairments and increased risk for devel- oping AD. 1–3 The hippocampus and associated structures within the medial temporal lobe are critical substrates of mem- ory formation, 4 and progressive hippocampal dysfunction has been proposed as a possible neuroanatomic basis of AD. 5–7 Neuroradiographic evidence for hippocampal for- mation tissue loss has been observed in patients with mild AD 8–12 and in elderly individuals exhibiting cognitive impairment suggestive of possible AD 13–15 ; further, the hippocampal formation was the only temporal lobe sub- volume sensitive to these small cognitive differences. 16 In fact, for older adults with mild cognitive impairment, hippocampal atrophy (HA) can predict whether an indi- vidual with mild cognitive impairment is at short-term risk of subsequently developing dementia. 8 Such early prediction is critical, given that all existing AD therapies work to arrest the progress of the disease rather than reverse its effects. Several previous studies with cognitively normal older adults have shown that HA correlates significantly with performance on memory tests, 17–20 although other studies have found that hippocampal volume does not cor- relate strongly with memory performance. 21,22 Recall of verbal information may be especially sensitive to hip- pocampal volume. 21,23 In one study considering nonde- mented elderly individuals, HA was shown to predict longitudinal decline on tests of delayed paragraph recall. 23 Paragraph delayed recall (PDR) tasks are also particu- larly sensitive to hippocampal region damage resulting from other etiologies. 4 Therefore, it seems plausible that other tasks that are also sensitive to hippocampal region damage might similarly be disrupted in individuals with mild HA, and such hippocampal-sensitive tasks might be disrupted in early AD before more general cognitive deficits are evident. Hippocampal Atrophy Disrupts Transfer Generalization in Nondemented Elderly Catherine E. Myers, PhD, Alan Kluger, PhD, James Golomb, MD, Steven Ferris, PhD, Mony J. de Leon, EdD, Geoffrey Schnirman, PhD, and Mark A. Gluck, PhD ABSTRACT Specific reductions in hippocampal volume in nondemented elderly individuals with mild cognitive impairment have been shown to correlate with future development of Alzheimer’s disease (AD). Hippocampal atrophy (HA) is also correlated with cognitive impairments, leading to the promise of behavioral markers for early AD. Prior theo- retical work has suggested that hippocampal dysfunction may selectively impair generalization involving novel recombinations of familiar stimuli. In this study, nondemented elderly individuals were trained on a series of con- current visual discriminations and were then tested for transfer when stimulus features were recombined in new ways. Presence or absence of HA, revealed by neuroimaging, was not correlated with concurrent discrimination per- formance; however, individuals with mild HA showed significant decreases in transfer performance relative to nonatrophied participants. These preliminary results suggest that even very mild degrees of hippocampal atrophy may be associated with subtle behavioral impairments. (J Geriatr Psychiatry Neurol 2002; 15:82–90). Received December 7, 2000. Received revised July 17, 2001. Accepted for publication August 1, 2001. From the Department of Psychology (Dr. Myers), Rutgers University, Newark, New Jersey; the Department of Psychology (Dr. Kluger), Lehman College/CUNY, New York; the Departments of Psychiatry (Drs. Kluger, Ferris, and de Leon) and Neurology (Dr. Golomb), New York University Medical Center, New York, New York; Nathan Kline Institute (Dr. de Leon), New York, New York; Massachusetts Mental Health Center (Dr. Schnirman), Boston, Massachusetts; and the Center for Molecular and Behavioral Neuroscience (Dr. Gluck), Rutgers University, Newark, New Jersey. This work was partially funded as a pilot subproject under National Institute on Aging (NIA) grant #P30 AG08051 to the NYU Aging and Dementia Research Center (to Dr. Ferris), by NIA grants RO1 AG12101 and RO1 AG13616 (to Dr. de Leon), and by a grant from the Alzheimer’s Association (to Dr. Gluck). Reprint requests: Dr. Catherine E. Myers, Department of Psychology, Rutgers University, Newark, NJ 07102.