White matter signal hyperintensities in the brains of patients with late paraphrenia and the normal, community-living elderly

i

White Matter Signal Hyperintensities in the Brains of Patients with Late Paraphrenia and the Normal, Community-Living Elderly Robert Howard, Timothy Cox, Osvaldo Almeida, Richard Mullen, Phillipa Graves, Adrienne Reveley, and Raymond Levy

We determined the prevalence and anatomical location of areas of white matter hyperintensi~.' visualized by magnetic resonance imaging in the brains af 38 late paraphrenic patients with an onset of psychotic illness after the age of 60 and 31 healthy aged communi~' volunteers. All degrees of white matter signal hyperintensit3' were very common in both groups, and there was no excess of such changes in the brain of patients. Periventricular white matter and subcortical grey matter hyperintensities were significantly associated with both measured diastolic and systolic blood pressure in patients and control subjects. Periventricular and deep white matter, together with subcortical grey matter hyperintensities, were significantly associated with increased age. The excess of such presumed brain-imaging abnormalities previously reported in patients with an onset of psychosis late in life mav be a consequence of earlier authors' failure to include examination of appropriate communi U control populations and to carefully exclude patients with evidence of stroke. Key W o r d s : White matter, hyperintensities, late paraphrenia, late onset schizophrenia

Introduction Among populations of elderly subjects who are examined with magnetic resonance imaging (MRI), areas of signal hyperintensity (SH) within white matter are a common finding. Such areas of SH are reported to occur in from 30% (Bradley et al 1984) to 90% (Awad et al 1987) of the nonpsychiatric elderly population. They appear to be associated with increasing age (Bondareffet al 1990; Kobari et al 1990: Zubenko et al 1990: Deicken et al 1991) and hypertension From the Section of Old Age Psychiatry. Institute of Psychiatr,, (RH. OA, RM, RL ~; Phillip Hams Magnetic Resonance Centre. Department of Radiological Sciences, Guy's Hospital (PG); and Maudsley Hospital (AR ~, London, U K. Address reprint requests to Robert Howard, Section of Old Age Psychiatry. Institute of Psychiatry. DeCrespigny Park, Camberwell, London SE5 8AF. UK. Received May 17, 1993; revised August 16, 1994.

© 1995 Society of Biological P',ychiatr~,

(Bondareff et al 1990; Deicken et al 1991). There are now numerous reports of associations with a late onset of depression (Krishnan et al 1988; Zubenko et al 1990; Coffey et al 1990; Churchill et al 1991; Lesser et al 1991; Rabins et al 1991; Howard et al 1993) and schizophrenic symptoms (Breitner et al 1990; Lesser et al 1991; Miller et al 1986, 1989, ~991, 1992). The precise clinical significance and the underlying neuropathological nature of such areas of SH remain unclear. "Caps" of signal hyperintensity anterior to the frontal horns of the lateral ventricles, together with patches of white matter signal abnormality which are distinct from periventricular changes, are commoner in patients who have sustained cerebral infarction than among elderly controls whose MRI indication was for evaluation of headache or 0006-32231951509.50 SSDI 0006-3223(94)00248-2

White Matter Signal Hyperintensities in the Elderly

BIOLPSYCHIATRY

87

1995 ;38:86-91

dizziness (Kertesz et al 1988). "Rims" of hyperintensity around the lateral ventricles, however, are seen in 74% of such elderly controls (Kertesz et al 1988) and have been dubbed "pseudolesions," regarded as normal phenomena in the elderly (Sze et al 1986; Bondareff et ai 1990; Baldwin 1993). Demonstration of the presence of extensive white matter SH for up to 7 years in the brains of subjects without neurological or cognitive deficit (Fein et al 1990) further suggests that even large patches of SH do not necessarily indicate clinically significant central nervous system disease. In the absence of studies of the prevalence of areas of SH in normal community-living elderly people, no true normative data for these phenomena have been available (Baldwin 1993). Exactly what such areas of signal hyperintensity represent neuropathologically has still to be established. SH may indicate the presence of several types of pathology, since it is found in association with a variety of neuropsychiatric conditions, including Alzheimer's disease (Fazekas et al 1987; Mirsen et al 1991 ) and vascular dementia (Kinkel et al 1985; Hershey et al 1987). Attempts to correlate both anteand postmortem MR images with neuropathological examination of brains have been succesful in demonstrating that each of the recognized types of hyperintense signal lesion (rims, caps, punctate lesions, and patchy lesions) have their own distinct neuropathologica[ correlates. Rims are characterised by subependymal gliosis and a loss of the ependymal lining, caps and punctate lesions by dilated perivascular spaces and gliosis, and large patches by myelin pallor and dilated perivascular spaces (Chimowitz et al ! 992). The excess of areas of SH reported from populations of patients with late-onset depression thus most probably reflects subtle cerebrovascular disease and has been implicated as a marker of either a specific risk factor for depression or, alternatively, vascular disturbance consequent upon malnutrition and dehydration during depressed episodes and the hypotensive effects of drug treatment (Coffey et al 1989). The results of MRI studies of white matter hyperintensities in patients with a late-life onset of schizophrenic symptoms have been interpreted by their authors as evidence for structural brain pathology. Evidence of focal structural pathology emerged following the computed tomography (CT) studies of patients with late paraphrenia and late-onset schizophrenia which indicated abnormalities including dilated lateral ventricles and cortical atrophy (Naguib and Levy 1987; Rabins et al 1987; Pearlson et al 1987; Howard et al 1992). Miller et al (1986) reported brain imaging findings on five such patients in whom organic factors had supposedly been excluded by clinical assessment. Four out of the five patients had evidence of old infarct or some other structural abnormality such as normal pressure hydrocephalus. In a further series of eight patients without focal neuro-

logical signs (Breitner et al 1990), all showed significant white matter abnormality or vascular pathology on MRI. Temporoparietal and occipital lesions were particularly prevalent, and little such pathology was evident on the scans of normal controls (Breitner et al 1990). Miller and colleagues (1989, 1991, 1992) have reported the results of structural MRI and regional cerebral blood flow (RCBF) investigations using single-photon emission tomography (SPET) in patients with what they have termed "late-life psychosis". Forty-two percent of nondemented patients with an onset of psychosis after the age of 45 had areas of white matter SH on MRI, compared to only 8% of a healthy age-matched control group. The appearance of large patches of SH was 6 times more likely in the temporal lobes of patients than controls, and 4 times more in the frontal lobes, according to Miller et al (1991). These authors hypothesized that, although insufficient to give rise to focal neurological signs, areas of SH might produce dysfunction in the overlying frontal and temporal cortex and that this contributed to psychotic symptomatology. They acknowledged that, since areas of SH in the occipital lobes might also be implicated, it might not be possible to pinpoint an isolated anatomical white matter lesion that predisposed to psychosis. In a combined MRI and SPET comparison of 18 patients satisfying DSM-III-R (APA 1987) criteria for lateonset schizophrenia with 30 elderly controls, 55% of patients and 7% of controls had structural MRI evidence of cerebrovascular disease (Miller et al 1992). Eighty-three percent of late-onset schizophrenics and 27% of controls had at least one area of RCBF hypoperfusion in temporal or frontal regions. Although areas of hypoperfusion were seen in occipital, frontal, and temporal regions, temporal lobe changes were the most severe and most common, occurring in 72% of the patients (Miller et al 1992). The aims of the present study were to investigate the prevalence and anatomical location of SHs in the brains of a series of patients with late paraphrenia and a communityliving elderly control group. We wanted to determine whether white matter abnormality is indeed more common in individuals with late-life psychoses than among their nonpsychotic peers.

Methods

Subjects All of a series of 101 patients with late paraphrenia, recruited from eight psychiatric hospitals over an 18-month period, were asked if they would undergo magnetic resonance imaging (for details of recruitment, see Howard et al 1994). Once 50 had consented, recruitment to the study ceased. All subjects satisfied diagnostic criteria for late paraphrenia (Naguib and Levy 1987); briefly these were:

g8

R. Howard et al

BIOL PSYCHIATRY 1995:38:86 01

I. Onset of persecutory, fantastic, referential, or grandiose delusions with or without hallucinations after the age of 60. 2. Intellectual capacity consistent with normal ageing [Mini-mental state (MMSE; Folstein et al 1975) score of --> 24]. 3. Absence of a primary affective disorder. 4. Psychotic phenomena always occur in setting of clear consciousness. 5. Absence of history or physical signs on examination of stroke, neurological illness, alcohol, or drug abuse. Controls were 35 healthy, community-living individuals recruited from Age Concern lunch clubs, a retirement home, and a church congregation in the south London innercity borough of Camberwell. Exclusion criteria were as follows: I. History of depression or other psychiatric illness necessitating referral to a physician. 2. History or physical signs of stroke, neurological illness, drug or alcohol abuse. 3. Evidence of dementia (MMSE < 25).

Scanning

All subjects and controls were scanned on a Phillips 1.5T Gyroscanner in the Department of Radiological Sciences at Guy's Hospital. Fifteen T2-weighted (TR 2000, TE 20/80) transverse slices were taken through the brain parallel to the floors of the frontal and occipital lobes. Slice thickness was 8 mm, and the interslice gap was 0.8 ram. Scans were rated using a standardized grading system described by Coffey et al (1990) and derived from Fazekas et al (I 987). Separate ratings were made of each of the following: • Periventricular Ioperintensity (absent = 0, "caps" o1

pencil-thin lining = 1. smooth "halo'" = 2, irregular periventricular hyperintensity extending into deep white matter = 3), • Deep white matter h37~erintensi O' (absent = 0, punctate foci = I, beginning confluence of loci = 2, large confluent areas = 3). • Changes in subeortical grey matter nuclei, for example, basal ganglia and thalamus (absent = 0, puncture = 1, multipunctate = 2, diffuse = 3). In addition, the position of areas of white matter highsignal intensity were recorded (frontal, temporal, occipital, pons/medulla/midbrain, and cerebellum) on right and left sides. All scans were reviewed by an experienced consultant neuroradiologist (TC) who was blind to subject status. The films on each subject were viewed on a conventional light box, and all 15 slices of the T2 images were examined in

conjunction with coronal TI and axial proton density images of the subjects. Data were analyzed using SPSS/PC+ (Norusis 1990) with the t test and chi-squared statistics with corrections for multiple comparisons. Pooled variance estimations were used in comparisons of systolic and diastolic blood pressures between patient and control groups. Analysis of variance was used to investigate associations between SH, age, systolic and diastolic blood pressure measurements, and MMSE score.

Results The mean age of patients was 79.84 years (range, 68-94, standard deviation 6.94), and of the controls 79.49 years (range, 68-97, SD 6.22). The ages of the members of these groups were not significantly different (two-tailed t test, t = .22, p = .827). The patient group contained 42 women and 8 men. the control group 30 women and 5 men. Patients and controls were well matched for verbal IQ estimation as measured by the National Adult Reading Test (NART) ( Nelson 1982) (patient mean, 101.55: control mean, 103.39, t = .91, p = .365) and for years of completed full-time education (patient mean, 9.66; control mean, 9.58, t = .31, p---.76). Although all patients underwent MR scanning, images that were considered adequate for the visual ratings were obtained from 38 late paraphrenics and 3 ! controls. As part of a standardized scanning protocol, the T2-weighted images were acquired at the end of a 45-minute scanning sequence. By this time, subjects were frequently becoming restless, and a number of images were spoiled by movement. The experimenters were also more inclined to persuade controls than patients to remain in the scanner despite mild degrees of reported discomfort: these factors account for the lower proportion of rateable images acquired from patients. The prevalence of each grade of periventricular, deep white matter, and subcortical grey matter hyperintensity in patients and controls is shown in Table 1. There was no significant excess of such findings in either group (for periventricular chi squared = 1.96, 3 d t~ two-tailed p = .581; for deep white matter chi squared = 3.685, 3 df, two-tailed p = .298: and for subcortical grey matter chi squared = 2.17, 3 d/i two-tailed p = .5378). There was no significant excess of deep white matter SHs in any of the specific brain regions examined, nor was there any significant right-left asymmetry in the appearance of hyperintensities in patients or control subjects. Some degree of both deep white matter and periventricular hyperintensity was almost universal among patients and controls. At least Grade 1 periventricular hyperintensity

White Matter Signal Hyperintensities in the Elderly

BIOLPSYCHIATRY 1995:38:86-91

89

Table 1. Periventricular and Deep White Matter and Subcortical Grey Matter Abnormalities in Late Paraphrenic Patients and Controls

Peri ventricular

Deep white matter

Subcortical g re.', matter

Controls In=31) c/c

Late paraphrenics (n=38) %

Grade 0 I 2 3 Grade 0 I 2 3

3.3 29.0 32.3 35.4 16.1 29.0 35.5 19.4

2.6 31.6 18.4 47.4 7.9 23.7 28.9 39.5

Grade 0 I 2 3

71.0 16.1 9.7 3.2

65.8 26.3 7.9 0

was demonstrated in 96.7% of subjects, and 88.4% had some degree of deep white matter hyperintensity. Mean systolic blood pressures on the day of scanning were 149.29 mmHg in controls and 144.18 mmHg in patients. Mean diastolic blood pressures were 86.57 mmHg in controls and 87.65 mmHg in patients. Differences between patients and controls on these parameters were not significant (pooled variance estimate t = -.62, two-tailed p = .538). Analysis of variance ( A N O V A ) demonstrated that the severity of periventricular hyperintensities in all subjects was significantly associated with increasing diastolic blood pressure (DBP) ( F = 2.88, two-tailed p = .043) and systolic blood pressure (SBP) ( F = 2.75, two-tailed p = .050). The presence of hyperintensities in subcortical grey matter was also associated with increasing DBP ( F = 3.256, two-tailed p = .027) and SBP ( F = 3.753, two-tailed p = .015); However, the severity of deep white matter changes was not significantly associated with increased DBP ( F = 1.258, two-tailed p = .296) or SBP ( F = 1.382, two-tailedp = .256). Increasing age in both patients and controls was strongly associated with the degree of periventricular SH ( F = 5.599, two-tailed p = .002), deep white matter SH ( F = 3.779, two-tailed p = .015), and subcortical grey matter SH I F = 4.472, two-tailed p = .006). Mini-mental state [MMSE (Folstein et al (1975)] score in both patient and control subjects was not significantly associated with periventricular SH (F = .926, two-tailed p = .433), deep white matter SH [F = .377. two-tailed p = .770t, or subcortical grey matter SH (F = .963, two-tailed p = .416).

Discussion The most striking finding of this study is the apparent ubiquitous nature of areas of white matter SH visualized in the

Chi square/p

1.96.3 d]] p = .581

3.69, 3 (,(fip = .298

2.17,3d/]p=.538

brains of both patient and control groups. Since the control group represents one of the first true elderly community samples to undergo MRI, previous studies in which controls have either been younger than patients (Breitner et al 1990), recruited by newspaper advertisement (Rabins et al 1991; Miller et al 1991 )+ or for whom no details are given regarding their recruitment (Deicken et al 1991; Krull et al 1991) may have led to previous underestimates of the prevalence of these changes. Periventricular "rims," or the presence of a pencil-thin line, of white matter hyperintensity around the lateral ventricles have previously been regarded as normal findings in the healthy elderly (Sze et al 1986; Bondareffet al 1990), but our results suggest that more severe grade 2 or 3 (Fazekas et al 1987: Coffey et al 1989) periventricular changes are present in as many as 70% of healthy elderly individuals and do not have specificity for patients with psychosis. This would further appear to be true for white matter hyperintensities which are distinct from periventrieular areas, since such grade 2 or 3 (Fazekas et al 1986; Coffey et al 1987) changes were found in 55% of controls and 68.4% of patients. Earlier studies have indicated impressive differences in deep white matter appearance between patients with "late-life psychosis" and controls. White matter lesions in the temporal lobes of patients have been reported to be six times more prevalent than in controls (Miller et al 1991 )+and excesses of hyperintensity in frontal, occipital and parietal areas are also described (Breitner et al 1990: Miller et al 1989 and 1991 ). How might the disparity between the findings of earlier authors and the current study be explained? We have already alluded to the fact that careful choice of control subjects is the most important determinant of the apparent prevalence of white matter hyperintensities among normals, and to our belief that recruitment of controls for the current study from among the pa-

90

BIOLPSYCHIATRY

R, Howard et al

1995:38:86-91

tients' community peers was unique. A second reason lot less dramatic brain-imaging differences between our patients and controls, compared with some earlier reports may have been the selection of patients. Some previously reported studies have emanated from departments of neurology (Miller et al 1989, 1991, 1992); these workers may have been referred a very different population of late-life psychotics to those seen by our own Section. Certainly the cases we see have a much higher age at onset of psychosis [mean, 74.1; range, 60-93; SD 8.2 years (Howard et al 1994)], compared to onsets typically around the age of 60 in some previously published studies (e.g,, Miller et al 1991 ). The diagnosis of late paraphrenia has no direct counterpart in DSM-III-R (1987): most cases will be subsumed under the diagnosis of paranoid schizophrenia by the criteria of ICD-10 ( W H O 1992), with about 30% satisfying diagnostic criteria for delusional disorder (Quintal et al 1991). In a CT comparison of patients with late-onset psychoses, divided into "paraphrenics'" and "paranoids" by the presence of hallucinations, clinically unsuspected cerebral infarction was significantly more common in the paranoid group (Flint et al 1991 ); this led the authors to conclude that cerebral organicity is an important risk factor for late-onset paranoia. Since none of the late paraphrenics in the current study had MRI evidence of previous cerebral infarction, once again we believe that our different results are explained largely by our patterns of patient referral and careful exclusion of individuals with a history or physical signs of stroke. Although neuropathological examinations of areas of brain identified as containing areas of SH have indicated that arteriosclerotic changes are not invariably associated

with such lesions (Chimowitz et al 1992), the links with both risk factors for (Bondareff et al 1990; Deicken et al 1991), and established cerebrovascular disease (Kertesz et al 1988), are strong. The positive associations between measured DBP and SBP and the presence of periventricular and subcortical grey matter hyperintensities found in the current study are entirely consistent with earlier reports and indicate the apparent importance of hypertensive cerebrovascular changes in the genesis of signal hyperintensities. Areas of white matter SHs may not therefore be more prevalent in patients with an onset of psychosis in late life than in the normal elderly. The differences (if any) between the brain imaging appearances of these populations would appear to be more subtle than previously has been suggested. In patients whose late paraphrenic or late-onset schizophrenic psychoses have an onset late in life, brainimaging abnormalities (such as frontal or temporal lobe W M L s ) are more likely to be neurodegenerative than neurodevelopmental in origin; however, the further role of genetic and social factors in the determination of possible vulnerability to psychosis after mild brain injury in this group is unclear (Castle and Howard 1992). W e are currently collecting data on genetic risk, detailed neuropsychological testing, and quantified volumetric brain-MRI examination of late paraphrenic patients. W e hope to establish more clearly the nature of the relationship between the schizophrenic psychoses of early and late life, in terms of predisposing factors and associated biological features.

The project was funded by a grant from the Mental Health Foundation.RH is supportedby the Eleanor Peel Trust and OA by CNPq (Brazil).

References American Psychiatric Association ( 1987): DiagnosticamlStatistical Manual t~]Mental Disorders, 3rd ed rev., Washington DC: American Psychiatric Association. Awad IA, Spetzler RF, Hodak JA. Awad CA, Williams F, Carey R ( 1987): Incidental lesions noted on magnetic resonance imaging of the brain: Prevalence and clinical significance in various age groups. Neurosurgery 20:222-226. Baldwin RC (1993): Late life depression and structural brain changes: A review of recent magnetic resonance imaging research, lnt J Geriatr Psychiatry 8:115-123. Bondareff W, Raval J, Woo B, Hauser DL, Colleni PM (1990): Magnetic resonance imaging and the severity of dementia in older adults. Arch Gen Psychiatry 47:47-51. Bradley WG. Waluch V, Brant-Zawadzki M, Yadley RA. Wyckoff RR (1984): Patchy periventricular white matter lesions in the elderly: A common observation during NMR imaging. Noninvasive Imaging 1:35-41. Breitner JCS, Husain MM, Figiel GS, Krishman KRR, Boyko OB (1990): Cerebral white matter disease in late-onset paranoid psychosis. Biol Psychiat O"28:266-274.

Castle D, Howard R ( 1992): What do we know about the aetiology of late-onset schizophrenia? Eur Psychiatry 7:99-108. Chimowitz MI, Estes ML, Furlan AJ, Awad IA (1992): Further observations on the pathology of subcortical lesions identified on magnetic resonance imaging. Arch Neuro149:747-752. Churchill CM, Priolo CV, Nemeroff CB, Krishnan KRR, Breitner JCS ( 1991 ): Occult subcortical magnetic resonance findings in elderly depressives, lnt J Geriatr Psychiato' 6:213-216. Coffey CE, Figiel GS, Djang WT, Saunders WB, Weiner RD (1989): White matter hyperintensity on magnetic resonance imaging: Clinical and neuroanatomic correlates in the depressed elderly. J Neuropsychiatr3' 2:135-144. Coffey CE, Figiel GS, Djang WT, Weiner RD ( 1990): Subcortical hyperintensity on magnetic resonance imaging: A comparison of normal and depressed elderly subjects. Am J Psychiato' 147:187-189. Deicken RF, Reus VI, Manfredi L, Wolkowitz OM (1991): MRI deep white matter intensity in a psychiatric population. Biol P~sTchiatry 29:918-922.

White Matter Signal Hyperintensities in the Elderly

BIOLPSYCHIATRY

91

1995:38:86-91

Fazekas F, Chawluck JB, Alavi A, Hurtig HI, Zimmerman R (1987): MR signal abnormalities at I.ST in Alzheimer's dementia and normal aging. Am J Neuroradiol 8:421-426. Fein G, van Dyke C, Davenport L, et al (1990): Preservation of normal cognitive functioningin elderly subjects with extensive white-matter lesions of long duration. Arch Gen Psychiatry 47:220-223. Flint AJ, Rifat SL, Eastwood MR ( 1991 ): Late-onset paranoia: Distinct from paraphrenia? Int J Geriatr Psychiatry"6:103-109. Folstein MF, Folstein SE, McHugh PR (1975): Mini-mental state: A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12:189-197. Hershey LA, Modic MT, Greenough PG, Jaffe DF ( 1987): Magnetic resonance imaging in vascular dementia. Neurology 37:29-36. Howard R, Almeida O, Levy R (1994): Phenomenology, demography and diagnosis in late paraphrenia. Psychol Med 24:397-410. Howard RJ, Beats B, Forstl H, Graves P, Bingham J, Levy R (1993) White matter changes in late onset depression: A magnetic resonance imaging study, lnt J Geriatr Psychiatry 8:183185. Howard RJ, Forstl H, Naguib M, Burns A, Levy R (1992): Firstrank symptoms of Schneider in [ate paraphrenia: Cortical structural correlates. Br J Psychiatry 160:108-109. Hunt AL, Orrison WW, Yeo RA, et al (1989): Clinical significance of MRI white matter lesions in the elderly. Neurology 39:1470-1474. Kertesz A, Black SE. Tokar G, Benke T, Carr T, Nicholson L (1988): Periventricular and subcortical hyperintensities on magnetic resonance imaging: "Rims, caps and unidentified bright objects." Arch Neuro14 5 :404-408. Kinkel WR, Jacobs L. Polachini I, Bates V, Heffner RR (1985): Subcortical arteriosclerotic encephalopathy (Binswanger's disease): Computed tomographic, nuclear magnetic resonance and clinical correlations. Arch Neuro142:951-959. Kobari M, Meyers JS, Ichijo M ( 1990): Leuko-araiosis, cerebral atrophy and cerebral perfusion in normal aging. Arch Neurol 47:161-165. Krishnan KKR, Goli V, Ellinwood EH, France RD, Blazer DG, Nemeroff CB (1988): Leukoencephalopathy in patients diagnosed as major depressive. Biol Psychiatry 23:519-522. Krull AJ, Press G, Dupont R, Harris MJ, Jeste D (1991): Brain imaging in late-onset schizophrenia and related psychoses, lnt J Geriatr Psychiatry 6:651-658. Lesser IM, Miller BL, Boone KB, et al (1991): Brain injury and cognitive function in late-onset psychotic depression. J Neuropsychiat O' Clin Neurosci 3:33-40.

Miller BL, Benson F, Cummings JL, Neshkes R (1986) Late-life paraphrenia: An organic delusional syndrome. J Clin Psychiatry 47:204-207. Miller BL, Lesser IM, Boone K, et al (1989): Brain white matter lesions and psychosis. BrJ Psychiatry 155:73-78. Miller BL, Lesser IM, Boone K, Hill E, Mehringer CM, Wong K (1991): Brain lesions and cognitive function in late-life psychosis. Br J Psychiatry 158:76-82. Miller BL, Lesser IM, Mena I, et al (1992): Regional cerebral blood flow in late-life-onset psychosis. Neuropsyehiato' Neuropsychol Behav Neurol 5:132-137. Mirsen TR, Lee DH, Wong CJ, et al ( 1991 ): Clinical correlates of white matter changes on magnetic resonance imaging scans of the brain. Arch Neuro148:1015-1021. Naguib M, Levy R ( 1987): Late paraphrenia: Neuropsychological impairment and structural brain abnormalities on computed tomography. Int J Geriatr Psychiatr)' 2:83-90. Nelson HE (1982): The National Adult Reading Test: Windsor NFER-Nelson. Norusis M (1990): SPSSIPC+ statistics 4.0 for the IBM PC/XT/AT and PSI2. Gorinchem, Netherlands: SPSS International. Pearlson GD, Garbacz D, Tompkins RH, Ahn HO, Rabins PV (1987): Lateral cerebral ventricular size in late onset schizophrenia. In Schizophrenia and Aging Miller NE, Cohen GD (eds), New York: Guildford Press. Quintal M, Day-Cody D, Levy R (1991): Late paraphrenia and ICD- 10. lnt J Geriatr Psychiat~" 6:111-116. Rabins P, Pearlson G, Aylward E, Kumar AJ, Dowell K (1991): Cortical magnetic resonance imaging changes in elderly inpatients with major depression. Am J Psychiat~' 148:617-620. Rabins P, Pearlson G, Jayaram G, Steele C, Tune L (1987): Increased ventricle-to-brainratio in late-onset schizophrenia. Am J Psychiatr 3" 144:1216-1218. Sze G, De Armond SJ, Brant-Zawadzki M, Davis RL, Norman D, Newton T ( 1986): Foci of MRI signal (pseudo lesions) anterior to the frontal horns: Histologic correlations of a normal finding. Am J Radio1147:331-337. World Health Organization (1992): The ICD-IO Classification of Mental and Behavioural Disorders. Geneva, Switzerland: World Health Organization. Zubenko GS, Sullivan P, Nelson JP, Belle SH, Huff J, Wolf GL ( 1990): Brain imaging abnormalities in mental disorders of late life.Arch Neuro147:1107-1111.