Differential Cortical Atrophy in Subgroups of Mild Cognitive Impairment

ORIGINAL CONTRIBUTION Differential Cortical Atrophy in Subgroups of Mild Cognitive Impairment

Sandra Bell-McGinty, PhD; Oscar L. Lopez, MD; Carolyn Cidis Meltzer, MD; Joelle M. Scanlon, PhD; Ellen M. Whyte, MD; Steven T. DeKosky, MD; James T. Becker, PhD

Objective: To compare gray matter brain volumes in jects. Compared with patients with MCI-MCD, patients patients diagnosed with subtypes of mild cognitive im- with MCI-A had significant volume loss of the left ento- pairment (MCI) (those with a focal amnestic disorder and rhinal cortex and inferior parietal lobe. Compared with those with more diffuse cognitive dysfunction) with those patients with MCI-A, patients with MCI-MCD had sig- of elderly controls. nificantly reduced volume of the right inferior frontal gyrus, right middle temporal gyrus, and bilateral superior Design: Magnetic resonance imaging volumetric study temporal gyrus. Patients with MCI who progressed to Alz- of MCI subgroups (MCI-amnestic [MCI-A], and MCI- heimer disease during follow-up (mean interval 2 years, multiple cognitive domain [MCI-MCD]) using a whole maximum 4.5 years), showed greater atrophy in the left brain voxel-based analysis. entorhinal cortex, bilateral superior temporal gyri, and right inferior frontal gyrus compared with those who did Setting: Referral dementia clinic. not progress.

Patients: Thirty-seven patients with MCI (age range, Conclusions: These data provide evidence of distinct 49-85 years; MCI-A, n=9; MCI-MCD, n=28) and 47 con- brain structural abnormalities in 2 groups of patients with trol subjects (age range, 55-81 years). MCI. While both have mesial temporal and cortical volume loss, those with a focal memory deficit have more Main Outcome Measures: Volumetric anatomical mag- involvement of the mesial temporal structures and less netic resonance imaging differences between MCI sub- involvement of the neocortical heteromodal association groups and normal controls, and between patients with areas than those patients with MCI with diffuse cogni- MCI who progressed to dementia. Magnetic resonance tive dysfunction. Thus, MCI may represent a more het- imaging scans were analyzed using statistical software erogeneous group than currently conceived, possibly re- SPM99. flecting 2 different etiological processes to dementia. These data also suggest that these structural abnormalities pre- Results: Overall, the patients with MCI had signifi- cede the development of Alzheimer disease. cantly decreased volume in the hippocampus and middle temporal gyrus, bilaterally, compared with control sub- Arch Neurol. 2005;62:1393-1397

HE TRANSITIONAL STATE BE- particularly severe among those patients tween normal aging and with MCI who progress to AD compared Author Affiliations: 6,8 Departments of Psychiatry Alzheimer disease (AD), with those who do not. (Drs Bell-McGinty, Lopez, mild cognitive impair- Even though MCI diagnosis relies pri- Meltzer, Scanlon, Whyte, ment (MCI), has become a marily on the presence of memory dysfunc- DeKosky, and Becker), focusT of research owing to the develop- tion, a growing number of studies have con- Neurology (Drs Lopez, ment of effective pharmacotherapy aimed cluded that performance in other cognitive DeKosky, and Becker), at altering the natural history of the dis- domains is often not entirely normal.1,9 Radiology (Dr Meltzer), and ease.1,2 A number of brain structural ab- While some patients exhibit an isolated Psychology (Dr Becker), normalities have been identified among pa- memory problem, others can have altered Neuropsychology Research tients with MCI with abnormal memory, neuropsychological test performance in Program, Functional Imaging including significant reduction in the vol- multiple cognitive areas.1,9 The purpose of Research Program, and the ume of the hippocampus,3,4 medial occipi- this study was to compare regional gray Mental Health Intervention 5 Research Center for Late-Life totemporal lobe, parahippocampal gy- matter brain volumes in 2 subtypes of pa- Mood Disorders, University of rus, entorhinal cortex, superior temporal tients with MCI using a whole brain voxel- Pittsburgh Medical Center, gyrus, and anterior cingulate gyrus.6,7 based analysis.10 This approach is not bi- Pittsburgh, Pa. These morphological abnormalities are ased to a specific brain region and permits

(REPRINTED) ARCH NEUROL / VOL 62, SEP 2005 WWW.ARCHNEUROL.COM 1393

MCI Subgroups

Controls All MCI MCI-A MCI-MCD No. of patients 47 37 9 28 Mean age ± SD (range), y 66.9 ± 7.3 (55-81) 71.9 ± 7.6 (49-85)* 73.4 ± 4.3 (66-78)† 71.5 ± 8.4 (49-85) Men (%) 27 (57) 17 (46) 4 (44) 13 (46) Caucasian (%) 42 (89) 31 (84) 9 (100) 22 (46) Mean educational level ± SD, y 15.7 ± 2.7 14.1 ± 3.9 13.7 ± 2.1 14.6 ± 4.4 Mean baseline MMSE ± SD 29.4 ± .4 25.5 ± 3.2* 23.1 ± 3.8† 26.3 ± 2.6† Mean baseline MDRS ± SD ‡ 133.7 ± 7.9 130.7 ± 7.5† 134.7 ± 7.9

Abbreviations: MCI, mild cognitive impairment; MCI-A, mild cognitive impairment–amnestic; MCI-MCD, mild cognitive impairment–multiple cognitive domain; MDRS, Mattis Dementia Rating Scale; MMSE, Mini-Mental State Examination. *Patients with MCI were older (t = - 3.04, P = .003) and had lower MMSE scores (t = - 5.86, P = .001) than controls. †Controls vs MCI subtypes; age was different among the 3 groups (F = 4.84, P = .01); patients with MCI-A and MCI-MCD were older than controls; the MMSE scores were different among the 3 groups (F = 3.8, P = .02); patients with MCI-A had lower scores than those with MCI-MCD (t = - 2.85, P = .01); patients with MCI-A had lower MDRS scores than those with MCI-MCD (t = 2.32, P = .02). ‡The MDRS was not administered to the comparison subjects from outside the Alzheimer’s Disease Research Center.

identification of potential unsuspected brain structure ab- main (not including memory), without sufficiently severe cog- normalities,10 allowing for a more comprehensive descrip- nitive impairment or loss of daily living skills to constitute tion of the differences between MCI subtypes. dementia, or 2 abnormal tests in 2 different domains.16

METHODS MRI ACQUISITION AND ANALYSIS

Magnetic resonance imaging scans were conducted using a Signa PATIENTS 1.5 Tesla scanner (GE Medical Systems, Milwaukee, Wis). The MRI of the brain was done within 6 months after the initial evalu- Thirty-seven patients from a group of 200 who met criteria for ation. The spoiled gradient-recalled sequence was designed to MCI, as described below, underwent a volumetric spoiled gra- maximize contrast between gray and white matter (echo time=5 dient-recalled magnetic resonance imaging (MRI) scan. Each 11 milliseconds, repetition time=25 milliseconds, 1.5-mm sec- patient received an extensive evaluation, and was reevalu- tion, 0-mm intersection interval, 40° flip angle). ated on an annual basis with regard to neuropsychiatric status All MRI data were processed using Statistical Parametric Map- to determine whether there was a change in diagnosis. ping (SPM99; Wellcome Department of Cognitive Neurology, Volumetric MRI scans were obtained on 47 older compari- London, England) running in MATLAB (Mathworks, Sher- son subjects from 3 ongoing studies, including the Alzhei- 11 born, Mass). The spoiled gradient recalled images were spa- mer’s Disease Research Center (n=28), the University of tially normalized (Montreal Neurological Institute coordinate Pittsburgh’s Mental Health Intervention Research Center for system; McGill University, Montreal, Quebec), and the tissue Late-Life Mood Disorders,12 and a study of cognitive and cere- 13 segmented using a modified mixture model cluster analysis tech- brovascular consequences of hypertension (n=19). None of nique.17 The segmented gray matter images were then smoothed the controls converted to dementia or MCI within 5 years of using an 8-mm isotropic gaussian kernel. A more complete de- the scan. scription of voxel-based morphometry method can be found in Good et al18 and Ashburner et al.17 NEUROPSYCHOLOGICAL EVALUATION RESULTS The neuropsychological evaluation included the Mini-Mental State Examination (MMSE),14 the Mattis Dementia Rating Scale,15 and measures of 4 cognitive domains: memory, language, vi- The demographic characteristics of all subjects and MCI suospatial/visuoconstructional, and attention/executive func- subgroups are shown in Table 1. Fourteen patients with tions. Details of the neuropsychological battery have been de- MCI (38%) converted to AD during follow-up (mean±SD scribed elsewhere.16 The results of the cognitive tests were follow-up: 45.7±26.5 months). The proportion of pa- classified normal or abnormal (Ͼ1.5 below that of subjects of tients with MCI-A (44%) and MCI-MCD (36%) who con- comparable age and education) based on normative data ob- verted to AD was similar between groups (␹2=0.65, P=.41). tained from the Alzheimer’s Disease Research Center normal The baseline demographic characteristics of those who con- control sample. verted to AD are shown in Table 2. Overall, those who converted to AD did have lower MMSE (t=3.2, P=0.01) MCI CRITERIA and Mattis Dementia Rating Scale (t=4.15, P=.004) scores at the time of study entry compared with nonconverters. Patients with MCI-amnestic (MCI-A) (n=9) required memory deficits, with otherwise normal cognitive function. These patients must have impairments in delayed recall verbal memory, VOXEL-BASED MORPHOMETRY nonverbal memory, or both.16 Patients with MCI-multiple cognitive domain (MCI- Patients with MCI, as a group, had significantly de- MCD) (n=28) required deterioration in at least 1 cognitive do- creased volume in the hippocampus and middle tempo-

(REPRINTED) ARCH NEUROL / VOL 62, SEP 2005 WWW.ARCHNEUROL.COM 1394

©2005 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/25/2021 Table 2. Demographic and Cognitive Characteristics of Patients With MCI Who Converted to AD and Those Who Did Not

MCI-A MCI-MCD

Converters Nonconverters Converters Nonconverters No. of patients 4 5 10 18 Mean age ± SD (range), y 75.7 ± 2.2 (73-78) 71.6 ± 4.9 (66-78) 70.8 ± 7.7 (57-81) 71.8 ± 9.0 (49-85) Men (%) 2 (50) 2 (40) 2 (20) 11 (61) Caucasian (%) 4 (100) 5 (100) 8 (80) 14 (78) Mean educational level ± SD, y 13.0 ± 2.0 14.4 ± 2.1 14.5 ± 4.2 14.7 ± 4.6 Time to conversion* or time to last 24.9 ± 16.3* 19.2 ± 6.2† 31.4 ± 13.3* 36.6 ± 19.7† contact in nonconverters†, mo Mean baseline MMSE ± SD 20 ± 2.0‡ 25.6 ± 2.9 25.6 ± 2.3 26.6 ± 2.6 Mean baseline MDRS ± SD 124 ± 4.2‡ 136 ± 4.3 133.6 ± 6.6 135.4 ± 8.5

Abbreviations: AD, Alzheimer disease; MCI, mild cognitive impairment; MCI-A, mild cognitive impairment–amnestic; MCI-MCD, mild cognitive impairment–multiple cognitive domain; MDRS, Mattis Dementia Rating Scale; MMSE, Mini-Mental State Examination. *Time to conversion (t = .92, P = .37). †Follow-up time (t = 1.48, P = .15). ‡MCI-A converters had lower MMSE (t = 3.2, P = .01) and MDRS (t = 4.15, P = .004) scores than nonconverters. No statistical differences were noted between MCI-MCD converters and nonconverters–, MMSE (t = 1.1, P = .30), or MDRS (t = .59, P = .56) scores.

ral gyrus, bilaterally. In addition, the left inferior parietal, left middle frontal, and right superior frontal volumes Table 3. Gray Matter Atrophy in MCI and MCI Subgroups were also reduced in patients with MCI compared with Compared With Normal Controls control subjects (Table 3). x, y, z Patients with MCI-A had significantly reduced vol- Region Coordinates* z Score ume of the mesial temporal lobe on the right, including the hippocampus, entorhinal cortex, and amygdala MCI vs control Right hippocampus 28, − 25, − 9 4.82 (Figure, A) compared with control subjects. In addi- Left hippocampus − 21, − 13, − 20 4.22 tion, reduced volume was observed in the left inferior pa- Right middle temporal 47, − 6, − 11 4.47 rietal, inferior and middle frontal, and superior tempo- Left middle temporal − 46, − 7, − 9 4.58 ral gyri. Patients diagnosed with MCI-MCD had significant Left inferior parietal − 36, − 60, 47 4.68 bilateral volume loss of the hippocampus, middle and su- Left middle frontal − 36, 62, 0 4.62 perior temporal, and inferior frontal gyri compared with Right superior frontal 9, 16, 52 4.73 MCI-A vs control controls (Figure, B). In addition, the left inferior pari- Right hippocampus 26, − 27, − 1 5.04 etal gyrus and the right superior frontal gyrus were sig- Right entorhinal 17, − 6, − 26 4.00 nificantly decreased in patients with MCI-MCD com- Right amygdala 19, − 3, − 17 3.54 pared with controls (Table 3). Left superior temporal − 52, − 28, 16 3.74 Compared with patients with MCI-MCD, patients with Left inferior parietal − 49, − 60, 49 4.21 MCI-A had significantly greater volume loss in the left Left inferior frontal − 54, 16, 7 3.92 inferior parietal lobe (PϽ.001) and the left entorhinal/ Left middle frontal − 47, 11, 46 4.30 Ͻ MCI-MCD vs control perirhinal cortex (P .01). By contrast, compared with Right hippocampus 26, − 20, − 11 4.38 patients with MCI-A, patients with MCI-MCD had sig- Left hippocampus − 22, − 13, − 20 4.34 nificantly greater volume loss in the right inferior fron- Right middle temporal 47, − 6, − 11 4.29 tal gyrus, right middle temporal gyrus, and superior tem- Left middle temporal − 49, − 7, − 8 4.36 poral gyrus, bilaterally (Table 4). Right superior temporal 50, − 29, 9 4.76 Among all patients with MCI who progressed to AD, Left superior temporal − 53, − 36, 13 3.35 Left inferior parietal − 60, − 28, 34 4.63 there was greater baseline atrophy in the left entorhinal Right inferior frontal 54, 0, 11 4.79 cortex, bilateral superior temporal gyri, and right infe- Left inferior frontal − 41, 16, 4 4.46 rior frontal gyrus (Table 5). Partial correlation analy- Right superior frontal 8, 21, 52 4.69 ses, controlling for age, showed significant positive cor- relations between the MMSE score and the volumes of Abbreviations: MCI, mild cognitive impairment; MCI-A, mild cognitive the left entorhinal cortex (r=0.35; df=34; P=.04) and right impairment-amnestic; MCI-MCD, mild cognitive impairment-multiple cognitive domain. inferior frontal gyrus (r=0.36; df=34; P=.03). *Coordinates of peak difference within cluster, obtained from statistical software SPM99 using the Montreal Neurological Institute template. COMMENT as in the amygdala, and in the neocortex. Patients with The present study demonstrates distinct brain struc- MCI-MCD showed more diffuse and extensive volume tural abnormalities in 2 subgroups of patients with MCI. loss in the neocortical heteromodal association, with less Specifically, patients with MCI-A have atrophy in the hip- involvement of the medial temporal lobe structures com- pocampus and entorhinal cortex, as expected,3-7 as well pared with those diagnosed with MCI-A. The MCI sub-

(REPRINTED) ARCH NEUROL / VOL 62, SEP 2005 WWW.ARCHNEUROL.COM 1395

x, y, z Region Coordinates* z Score Regions of Less Volume in Patients With MCI-A Compared With Patients With MCI-MCD Left inferior parietal − 46, − 62, 49 3.97 Left entorhinal/perirhinal − 26, 6, − 34 2.78 Regions of Less Volume in Patients With MCI-MCD Compared With Patients With MCI-A Right middle temporal 55, − 12, − 18 3.87 Right superior temporal 62, − 21, 7 4.29 Left superior temporal − 52, − 22, 2 3.66 Right inferior frontal 29, 24, 16 3.75

Abbreviations: MCI, mild cognitive impairment; MCI-A, mild cognitive impairment–amnestic; MCI-MCD, mild cognitive impairment–multiple cognitive domain.

Table 5. Locations of Significant Differences in Gray Matter Volume Between MCI Converters and Nonconverters

x, y, z Region Coordinates* z Score Left entorhinal − 24, − 4, − 40 3.76 Right superior temporal 44, 13, − 18 4.15 Left superior temporal − 45, 12, − 11 3.95 Right inferior frontal 54, 20, 22 3.99

Abbreviation: MCI, mild cognitive impairment. *Coordinates of peak difference within cluster, obtained from statistical software SPM99 using the Montreal Neurological Institute template.

sociated with decreased volume in both the frontal and Figure. Statistical software SPM99 results, controlling for current entorhinal areas. chronological age (PϽ.001), showing regions of reduced gray matter The structural changes associated with the MCI syn- volume in patients with MCI-A (A) and patients with MCI-MCD compared drome are more diffuse than previously thought. Al- with the comparison subjects (B), controlling for current age. The top rows though the hippocampus has been the most studied area show results projected onto an averaged template brain. The bottom rows show results projected in the coronal plane onto a mean image of 37 patients in MCI cases, more recent studies have shown that struc- with MCI (A, Bottom left image=right hippocampus, y=-23; bottom right tural lesions (eg, neurofibrillary tangles, neuritic plaques) image=right entorhinal/amygdala, y=-4.B, Bottom left image=bilateral are more widely distributed in MCI, and include the neo- hippocampus, y=-18; bottom right image=bilateral middle temporal and 20,21 right inferior frontal, y=-5). Images are displayed in neurological convention. cortex and limbic areas. Furthermore, amyloid de- posits were identified in vivo in the frontal and parietal lobes in patients with mild AD, and practically no amy- groups share a region of atrophy in the inferior frontal loid was seen in the medial temporal lobe structures.22 cortex, specifically BA 44/45, suggesting that this re- Therefore, the structural abnormalities in MCI cases, es- gion may be important in the clinical presentation of MCI, pecially in those who progress to AD, are not limited to and perhaps, reflecting the impending defect in re- the mesial temporal areas. trieval from semantic memory. This study is limited by the small sample of MCI Our data also suggest that atrophy in specific cortical subtypes, particularly those who progressed to AD. regions precede the development of dementia in pa- Nevertheless, our study shows that there are at least 2 tients with MCI. Specifically, patients who developed AD subtypes of MCI, those with the more traditionally during follow-up had significantly decreased volume of defined memory deficit and those with more diffuse the left entorhinal cortex, bilateral superior temporal gyri, cognitive impairment, each presenting with distinct and right inferior frontal gyrus at study entry. Volume brain structural abnormalities. The therapeutic impli- reduction of the entorhinal cortex and superior tempo- cations of our findings need to be explored, as differ- ral gyrus has been noted in recent studies;6,19 however, ences in brain abnormalities may be associated with our findings suggest additional structural abnormalities variations in disease course and treatment response. within the inferior frontal cortex of those patients with Better understanding of these subtypes may enhance MCI who progress to AD. The degree of cognitive im- our knowledge of the relationship between normal pairment, as measured by the MMSE, appears to be as- aging and dementia.

(REPRINTED) ARCH NEUROL / VOL 62, SEP 2005 WWW.ARCHNEUROL.COM 1396

©2005 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/25/2021 Accepted for Publication: January 4, 2005. the medial occipitotemporal, inferior, and middle temporal gyri in non- Correspondence: Oscar L. Lopez, MD, Neuropsychol- demented elderly predict decline to Alzheimer’s disease. Neurobiol Aging. 2000; 21:19-26. ogy Research Program, Suite 830, Oxford Bldg, 3501 6. Killiany RJ, Gomez-Isla T, Moss M, et al. Use of structural magnetic resonance Forbes Ave, Pittsburgh, PA 15213 (lopezol@upmc imaging to predict who will get Alzheimer’s disease. Ann Neurol. 2000;47:430- .edu). 439. Author Contributions: Study concept and design: Bell- 7. Visser PJ, Scheltens P, Verhey FR, et al. Medial temporal lobe atrophy and memory McGinty, Lopez, Meltzer, Scanlon, DeKosky, and Becker. dysfunction as predictors for dementia in subjects with mild cognitive impairment. J Neurol. 1999;246:477-485. Acquisition of data: Bell-McGinty, Lopez, Meltzer, and 8. Visser PJ, Verhey FR, Hofman PA, Scheltens P, Jolles J. Medial temporal lobe Becker. Analysis and interpretation of data: Bell- atrophy predicts Alzheimer’s disease in patients with minor cognitive impairment. McGinty, Lopez, Meltzer, Scanlon, Whyte, DeKosky, and J Neurol Neurosurg Psychiatry. 2002;72:491-497. Becker. Drafting of the manuscript: Bell-McGinty, Lopez, 9. Hanninen T, Hallikainen M, Koivisto K, et al. Decline of frontal lobe functions in DeKosky, and Becker. Critical revision of the manuscript subjects with age-associated memory impairment. Neurology. 1997;48:148- 153. for important intellectual content: Bell-McGinty, Meltzer, 10. Ashburner J, Friston KJ. Voxel-based morphometry: the methods. Neuroimage. Scanlon, Whyte, DeKosky, and Becker. Statistical analy- 2000;11:805-821. sis: Lopez and Becker. Obtained funding: Bell-McGinty and 11. Lopez OL, Becker JT, Klunk W, et al. Research evaluation and diagnosis of prob- DeKosky. Administrative, technical, and material sup- able Alzheimer’s disease over the last two decades: I. Neurology. 2000;55: port: Bell-McGinty, Meltzer, and DeKosky. Study super- 1854-1862. vision: Lopez, Meltzer, DeKosky, and Becker. 12. Mulsant BH, Pollock BG, Nebes RD, et al. A double-blind randomized comparison of nortriptyline and paroxetine in the treatment of late-life depression: 6-week Funding/Support: This research was supported in part outcome. J Clin Psychiatry. 1999;60(suppl 20):16-20. by grants AG05133 and AG20098 from the National In- 13. Jennings JR, Muldoon MF, Ryan CM, et al. Cerebral blood flow in hyperten- stitute on Aging, Bethesda, Md; and HL57529 from the sives: an initial report of reduced and compensatory blood flow responses dur- National Heart, Lung, and Blood Institute, Bethesda. Drs ing performance of two cognitive tasks. Hypertension. 1998;31:1216-1222. Bell-McGinty, Scanlon, and Whyte are postdoctoral re- 14. Folstein MF, Folstein SE, McHugh PR. Mini-mental state: a practical method grad- ing the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12: search fellows in the Department of Psychiatry, Univer- 189-198. sity of Pittsburgh Medical Center, Pittsburgh, Pa (T32 15. Mattis S. Mental status examination for organic mental syndrome in the elderly MH19986). Dr Becker is the recipient of a Research Sci- patient. In: Bellak L, Karuso TB, eds. Geriatric Psychiatry. New York, NY: Grune entist Development Award, Level-II (K02 MH01077). & Stratton; 1976. 16. Lopez OL, Becker JT, Sweet RA. Non-cognitive symptoms in mild cognitive impairment subjects. Neurocase. 2005;11:65-71. REFERENCES 17. Ashburner J, Friston K. Multimodal image coregistration and partitioning: a uni- fied framework. Neuroimage. 1997;6:209-217. 1. Petersen RC, Smith GE, Waring SC, Ivnik RJ, Tangalos EG, Kokmen E. Mild cog- 18. Good CD, Johnsrude IS, Ashburner J, Henson RN, Friston KJ, Frackowiak RS. nitive impairment: clinical characterization and outcome. Arch Neurol. 1999; A voxel-based morphometric study of aging in 465 normal adult human beings. 56:303-308. Neuroimage. 2001;14:21-36. 2. Morris JC, Storandt M, Miller JP, et al. Mild cognitive impairment represents early- 19. Killiany RJ, Hyman BT, Gomez-Isla T, et al. MRI measures of entorhinal cortex stage Alzheimer disease. Arch Neurol. 2001;58:397-405. vs hippocampus in preclinical AD. Neurology. 2002;58:1188-1196. 3. Wolf H, Grunwald M, Kruggel F, et al. Hippocampal volume discriminates be- 20. Price JL, Morris JC. Tangles and plaques in nondemented aging and “preclini- tween normal cognition: questionable and mild dementia in the elderly. Neuro- cal” Alzheimer’s disease. Ann Neurol. 1999;45:358-368. biol Aging. 2001;22:177-186. 21. Kordower JH, Chu Y, Stebbins GT, et al. Loss and atrophy of layer II entorhinal 4. Soininen HS, Partanen K, Pitkanen A, et al. Volumetric MRI analysis of the amyg- cortex neurons in elderly people with mild cognitive impairment. Ann Neurol. dala and the hippocampus in subjects with age-associated memory impair- 2001;49:202-213. ment: correlation to visual and verbal memory. Neurology. 1994;44:1660-1668. 22. Klunk WE, Engler H, Nordberg A, et al. Imaging brain amyloid in Alzheimer’s dis- 5. Convit A, de Asis J, de Leon MJ, Tarshish CY, De Santi S, Rusinek H. Atrophy of ease with Pittsburgh Compond-B. Ann Neurol. 2004;55:306-319.

(REPRINTED) ARCH NEUROL / VOL 62, SEP 2005 WWW.ARCHNEUROL.COM 1397