Longitudinal study of cerebral blood flow in Alzheimer's disease using single photon emission tomography

PSYCHIATRY RESEARCH NEUROIMAGING ELSEVIER

Psychiatry Research: Neuroimaging Section 68 (1997) 133-141

Longitudinal study of cerebral blood flow in Alzheimer's disease using single photon emission tomography P e r m i n d e r S a c h d e v * a,~, R a k e s h G a u r b, H e n r y B r o d a t y b,d, A l e x a n d r a W a l k e r a, S u s a n M e a r e s b, D e b o r a h K o d e r b, W a l t e r H a i n d l c aNeuropsychiatric Institute, The Prince Henry Hospital, P.O. Box 233, Matraville, NSW, Australia hAcademic Department of Psychogeriatrics, The Prince Henry Hospital, P.O. Box 233, Matraville, NSW 2036, Australia CDepartment of Nuclear Medicine, The Prince of Wales Hospital, Sydney, Australia dSchool of Psychiatry, Universityof New South Wales, Sydney, Australia Received 21 February 1995; revised 1 March 1996; accepted 26 May 1996

Abstract

Ten patients with probable Alzheimer's disease were assessed at baseline and a mean 2 years later using a battery of neuropsychological tests, CT scans and Tc99m-HMPAO SPECT scans. The subjects had declined significantly in their functional indices. Cerebral perfusion measures declined in the parietal lobes, left hemisphere and whole brain, but the overall decline did not reach statistical significance. The decline in brain perfusion did not correlate significantly with the decline in various indices of neuropsychological function, either globally or for specific brain regions. The index of cerebral perfusion correlated significantly with global indices of neuropsychological function at baseline but not at follow-up. No single perfusion index was a significant predictor of clinical progression of dementia. © 1997 Elsevier Science Ireland Ltd. All rights reserved Keywords: Alzheimer's disease; SPECT; Cerebral blood flow; Neuropsychological function

I. Introduction Functional neuroimaging is playing an increasingly important role in the evaluation of patients with Alzheimer's disease (AD). Positron emission

*Corresponding author, Neuropsychiatric Institute, The Prince Henry Hospital, P.O. Box 233, Matraville, NSW 2036, Australia. Tel: + 61 2 6945426; fax: +61 2 6945747.

tomography (PET), single photon emission computerised tomography (SPECT), and xenon blood flow studies have all demonstrated the fairly consistent findings of unilateral or bilateral temporo-parietal deficits in cerebral blood flow and glucose metabolism in the early stages, and more diffuse deficits in the later stages of the disease (Duara et al., 1986; Ebmeier et al., 1987; Johnson et al., 1987, 1988). Owing to its lesser expense and ease of availability in comparison

0925-4927/97/$17.00 © 1997 Elsevier Science Ireland Ltd. All rights reserved PII S 0 9 2 5 - 4 9 2 7 ( 9 6 ) 02792-8

134

P. Sachdev et al. /Psychiatry Research: Neuroimaging 68 (1997) 133 141

with PET, and better spatial resolution than the xenon technique, SPECT scanning is preferentially attracting interest as the functional modality of choice in the assessment and follow-up of patients with AD. Even though SPECT scanning can be adapted to the study of neurotransmitter function (e.g. acetylcholine receptor function), the majority of the SPECT studies of AD relate to the measurement of regional cerebral blood flow (rCBF). A number of SPECT studies of rCBF in AD have been published (Beats et al., 1991). These studies have usually been single SPECT examinations of patients, in baseline or activated states, with attempts to either demonstrate the utility of the SPECT findings in the diagnosis of AD, or to correlate the rCBF data with global or focal neuropsychological function. With regard to the utility of SPECT in distinguishing AD from other dementias or normal control subjects, published studies have generally been positive (Bonte et al., 1986; Cohen et al., 1986; Ebmeier et al., 1987; Johnson et al., 1987, 1988), but it is controversial whether SPECT can distinguish early AD from normal aging (Dekosky et al., 1990; Perani et al., 1988; Muller et al., 1986). The study of the correlation between rCBF and neuropsychological function has produced less consistent results. Some studies have reported a significant correlation between the degree of hypoperfusion in the parietal areas and the severity of dementia (Jagust et al., 1987; Johnson et al., 1988; DeKosky et al., 1990). However, not all studies have supported this association (Foster et al., 1984). The relationship between rCBF in specific brain regions and neuropsychological function ostensibly localisable to those regions has been less consistently demonstrated (Cutler, 1988; Goldenberg et al., 1989). Since previous research has relied on cross-sectional examinations, it would be useful to determine whether serial assessments demonstrate a covariation of rCBF and neuropsychological indices. This design also permits the examination of the role of SPECT scanning in the evaluation of progression in AD, an aspect that, to our knowledge, has not been previously studied. In this study, we report the results from 10 patients who had a baseline clinical, neuropsycho-

logical and SPECT examination and were followed up with a repeat assessment 2-3 years later. 2. Methods 2.1. Subjects

Twenty-seven consecutive patients attending the Memory Disorders Clinic at The Prince Henry Hospital in Sydney, Australia who met the DSMIII-R criteria for dementia (American Psychiatric Association, 1987) and NINCDS-ADRDA criteria for probable AD (McKhann et al., 1984) had a detailed baseline clinical assessment and a SPECT scan in 1990-91. An attempt was made to review these subjects in 1994. Of the 27 subjects, 5 had died, 3 had moved to distant places, and 9 subjects or their caretakers refused to participate in the follow-up study. Of the latter, 6 patients were in nursing homes. The main reasons given for the refusal were the patient's poor physical and mental condition, difficulty with transportation required for X-ray computerised tomographic (CT) and SPECT scans, and caretaker's apathy or antipathy towards research. Ten patients participated in the follow-up study and had repeat CT and SPECT scans. 2.2. Baseline assessment

The baseline evaluations were performed using our Memory Disorders Clinic protocol (Brodaty, 1990). This includes a detailed history from informants and subjects, physical examination, psychiatric assessment, screening tests for physical illness, and the administration of the following scales: Mini Mental State Examination (MMSE) (Folstein et al., 1975), Orientation-InformationMemory-Concentration (OIMC) and Dementia Scale (Blessed et al., 1968), Clinical Dementia Rating Scale (Hughes et al., 1982), 21-item Hamilton Rating Scale for Depression (Hamilton, 1967), Hachinski Ischaemic Scale (Hachinski et al., 1975), Activities of Daily Living (ADL) (Katz and Apkon, 1976) and Instrumental Activities of Daily Living (IADL) (Lawton and Brodie, 1969) scales. Additionally, subjects received a compre-

P. Sachdeu et al. / Psychiatry Research: Neuroimaging 68 (1997) 133-141

hensive neuropsychological assessment by a clinical psychologist, and CT and SPECT scans. 2.3. Follow-up assessment

The baseline neuropsychological and functional assessment, including CT and SPECT scans, was repeated. 2.4. Neuropsychological assessment

The test battery comprised the following tests: Logical Memory I and II subtests of the Wechsler Memory Scale-Revised (WMS-R) (Wechsler, 1987), Auditory Verbal Learning Test (AVLT) (Rey, 1964), Visual Reproduction I and II subtests of WMS-R (Wechsler, 1987), Similarities subtest of Wechsler Adult Intelligence Scale-Revised (Wechsler, 1981), and Controlled Oral Word Association Test (Benton et al., 1983). The scores on certain tests that ostensibly examined function localisable to specific neuroanatomical regions were used to compute indices for analysis. The tests used were: for left temporal lobe, verbal memory (logical memory - - immediate and delayed) and Rey Auditory Verbal Learning Test; for right temporal lobe, visual reproduction - - immediate and delayed; for frontal lobe, similarities and verbal fluency. Raw scores for appropriate tests were converted to standardized scores or z scores, and these were added to provide a z index for each ROI. The sum of z scores on all these tests yielded a global neuropsyehological function index. 2.5. C T scans

The scans were performed on a GE-9800 scanner, and 10-mm thick cuts were obtained from the base of the brain to the apex parallel to the orbito-meatal line. The CT scans, which were available only on film, were rated globally on a 0-3 (0 nil, 1 mild, 2 moderate, 3 severe abnormality) scale for both ventricular dilatation and sulcal dilatation by a neuropsychiatrist (PS), with the names and dates of scans being covered.

135

2.6. S P E C T scans

These scans were performed with 99m-TCHMPAO, using HMPAO reconstituted with freshly milked 99m-TC generator eluate and normal saline according to the manufacturer's instructions (Amersham International). A dose of 600 MBq was injected in an antecubital vein while the patient was supine with eyes closed in a softly lighted and quiet room. Ten min after the injection, imaging was performed on a GE Starcam 4000 x CT camera with a high resolution collimator. Data were acquired at 30 s/image in 128 x 128 matrix, using an elliptical orbit rotating 360 ° CCW for 64 steps and a x l zoom with a 20% symmetrical window. Reconstruction was performed using Butterworth prefilter with a critical frequency factor of 0.4, power factor of 8.0 and a ramp back projection filter. No attenuation correction was used. The resulting brain image comprised 128 x 128 x 45 voxels, with voxel size being 3.2 mm 3. SPECT scan data were transferred to a UNIX workstation and analysed using the software package A N A L Y Z E T M Version 6 (Biomedical Imaging Resource, Mayo Foundation) (Robb et al., 1991). The contiguous transaxial slices were superimposed to construct a 3D brain image, which was then corrected for horizontal and vertical rotations in two phases. The image was resliced interactively to provide 3 transaxial slices, 1.5 cm thick each, that most closely corresponded to the horizontal sections at z coordinates +20 mm, + 12 mm and + 1 mm above the intercommisural line on the brain atlas by Talairach and Tournoux (1988). These brain slices were selected as they most prominently included the brain areas of interest. One image corresponding to - 2 0 mm below the intercommisural line was used for the study of the cerebellum. An automated thresholding technique was used to outline the external margin of the cortex, as well as the boundaries of the ventricles in each slice. A line was drawn to divide each slice into left and right hemispheres. In order to outline the regions of interest (ROD, templates were prepared from the atlas, corresponding to each slice,

136

P. Sachdev et al. ~Psychiatry Research: Neuroimaging 68 (1997) 133-141

and the size of the display was proportionately altered to correspond to the template. The boundaries for the frontal, parietal, temporal and occipital lobes and the basal ganglia were traced manually on each of the 3 slices (Fig. 1). The cerebellum was similarly outlined on the fourth slice. For each ROI and cerebellum, the area, volume and number of counts per pixel were measured. To normalise the measurements, we divided the mean counts per pixel for each ROI by the mean counts per pixel for the cerebellum, thereby producing a mean ROI/cerebellar ratio referred to as the perfusion index for the region. Cerebellar activity was used for normalization because it is only minimally involved neuropathologically in the early stages of AD, and cerebellar perfusion has been reported to show no difference between normal controls and even severely demented AD patients (Johnson et al., 1987).

2. 7. Statistical analysis Since only 10 of the original 27 had a follow-up assessment, those who were followed up were compared with those who were not (n = 17) for their baseline characteristics using multivariate analysis of variance (MANOVA) and post-hoc tests. For the longitudinal data, a repeated mea-

sures analysis of variance design examined the effect of time (baseline and follow-up) and ROI (five perfusion indices averaged for right and left: frontal, parietal, temporal, occipital and basal ganglia) and time by ROI interaction. As it was possible that the rate of change in rCBF differed with age, the analysis was repeated with age as a covariate. Where the effects were significant or approached significance, post hoc tests were performed using paired t-tests for repeated measures. Asymmetry indices, as left/right ratios for each of the ROI, were analysed separately. Spearman's coefficients were used for the examination of correlations between perfusion indices and measures of dementia severity as well as neuropsychological deficits. These were performed for baseline and follow-up data separately, and for measures of change between the two assessments for both perfusion and neuropsychological performance. A multiple regression analysis was performed to determine which rCBF indices best predicted severity of dementia at baseline or its change. 3. Results

The mean interval between baseline (T l) and follow-up (T2) assessments was 24.0 months (S.D.

Fig. 1. SPECT images from a patient with 3 transaxial slices corresponding to the horizontal sections at z coordinates + 20 mm, + 12 mm and + 1 mm above the intercommissural line on the Talairach and Tournoux (1988) atlas. The regions of interest are drawn in using ANALYZE. F, frontal; T, temporal; P, parietal; O, occipital lobes; BG, basal ganglia; V, ventricles.

P. Sachdev et al. / Psychiatry Research: Neuroirnaging 68 (1997) 133-141

4.8; range 19-33 months). Subjects who were (n = 10) and were not (n = 17) followed up did not differ significantly in age, sex, and MMSE, Blessed OIMC and IADL scores, but differed significantly on Clinical Dementia Rating (P < 0.05), Blessed Dementia Scale ( P < 0.05), and A D L scores (P < 0.001) at baseline, with subjects not followed up having slightly more severe dementia. Of the group not followed up, a sub-group of 11 could be identified who either died (n = 5) or ended up in nursing homes (n = 6). This subgroup, when compared with the follow-up group (n = 10), was slightly older (mean age 75.6 year, F = 1.26, P = 0.27) and had a more severe dementia at baseline as indicated by scores on the MMSE (mean (S.D.) 16.6 (4.8), F = 6.31, P = 0.02), Blessed OIMC (mean (S.D.) 20.3 (4.2), F = 11.39, P = 0.003), Clinical Dementia Rating Scale (mean (S.D.) 1.3 (0.7) F = 6.22, P = 0.022) and Blessed Dementia Scale (mean (S.D.) 8.1 (5.0), F = 7.07, P = 0.015). It was considered useful to compare the SPECT measures for the two groups to examine the effect of severity on rCBF indices. The frontal, parietal, temporal, occipital and basal ganglia perfusion indices (means of right and left) were entered into a MANOVA to determine the difference between the follow-up (n = 10) and the 'severe' (n---11) groups at baseline. The differences were non-significant (P = 0.60). The whole brain perfusion index also did not differ between the two groups ( F = 0.51, P = 0.48). The follow-up group comprised 4 men and 6 women, whose ages ranged from 42 to 83 years (mean 70.7, S.D. 12.5 years). At T1, their mean estimated duration of illness was 3.48 years (S.D. 2.6), mean Hamilton score 5.22 (S.D. 5.6), and mean Hachiniski score was 1.5 (S.D. 1.2). At T2, there had been a significant deterioration on MMSE (P < 0.001) and Blessed Dementia Scale scores (P < 0.001) (Table 1). On neuropsychological assessments, a significant decline was seen on the global index as well as the frontal lobe indices. CT scan scores showed a significant increase in ventricular size but not in sulcal width. For the longitudinal data, the whole brain perfusion ratio declined between Tj and T 2 (P < 0.05). The repeated measures MANOVA using

137

time (2 levels) and region (5 levels) revealed within-subject effects that were significant for region (df = 4,36; F = 26.0, P < 0.001), non-significant for time (dr = 1,9; F = 1.41, P = 0.266), and approaching significance for the time by region interaction (dr = 4,36; F = 2.20, P = 0.089). The inclusion of age as a covariate did not affect this result. Even though the time by region effect only tended toward significance, given the small number of subjects overall, we proceeded to post-hoc tests using paired t-tests, the results of which are presented in Table 1. These revealed a significant decline in perfusion in the parietal, left hemisphere and whole brain indices when not corrected for multiple comparisons, with perfusion ratios of other ROIs not declining significantly. The rCBF demonstrated some left-right asymmetry, with the left hemispheric measures tending to be lower than the right. The perfusion indices were significantly different for the left and right temporal lobes (P < 0.01) and left and right hemispheres (P < 0.05) at both assessments, while basal ganglia perfusion was significantly asymmetrical at T 2 only. The indices of asymmetry did not significantly change from T 1 to T 2. The relationship of neuropsychological and perfusion indices was examined in a number of ways. The decline in the whole brain perfusion index did not correlate significantly with the decline in MMSE, BDS and global neuropsychological indices (Table 2), nor with increase in sulcal atrophy and ventricular size. The change in whole brain perfusion index, on the other hand, correlated significantly with the change in ventricular size on CT scans ( r = 0.6383, P < 0.05). There was no significant correlation between the change in left temporal, right temporal and frontal lobe peffusion indices and the z indices of verbal memory, visual memory and frontal lobe function, respectively, although the correlations between the change in left temporal perfusion and that in verbal memory ( r = 0.53) and the MMSE (r = 0.57) approached significance (Table 2). An unexpected result was a significant correlation between verbal memory change and the perfusion change in the right temporal lobe (r = 0.75). Correlations were also determined separately at T~ and T~ between neuropsychological and

138

P. Sachdev et al. / Psychiatry Research: Neuroimaging 68 (1997) 133-141

Table 1 A comparison of baseline and follow-up measures on 10 Alzheimer's disease subjects Variables

Baseline (T I )

Follow-up (T 2)

Mean (S.D.)

Mean (S.D.)

MMSE a BDS b Neuropsychological indices: c 1 - - Verbal memory 2 - - Visual memory 3 - - Frontal indices 4 - - Global indices

22.75 (4.31) 3.5 (2.38) -

16.05 (6.85) 6.15 (2.43)

<

0.001 0.001

2.21 (1.08) 1.52 (1.06) 0.92 (0.83) 1.77 (0.57)

- 1.7 (0.53) - 2.03 (0.68)

0.1 0.85 0.01" 0.05"

1 - - Sulcal atrophy 2 - - Ventricular enlargement

1.3 (0.48) 1.3 (0.82)

1.5 (0.52) 1.7 (0.65)

0.16 0.03

Perfusion ratios: e Left frontal Right frontal Left parietal Right parietal Left temporal Right temporal Left occipital Right occipital Left basal ganglia Right basal ganglia Left hemisphere Right hemisphere Whole cerebrum

0.812 (0.06) 0.830 (0.05) 0.813 (0.05) 0.849 (0.06) 0.807 (0.05) 0.850 (0.07) 0.893 (0.07) 0.907 (0.07) 0.925 (0.10) 0.939 (0.10) 0.813 (0.03) 0.843 (0.05) 0.828 (0.05)

0.805 0.813 0.774 0.803 0.781 0.884 0.886 0.884 0.902 0.954 0.778 0.807 0.792

(0.04) (0.04) (0.04) (o.04) (0.05)

Asymmetry indices (left/right ratios) of rCBF Left/right frontal Left/right parietal Left/right temporal f Left/right occipital Left/right basal ganglia Left/right hemisphere f

0.979 0.960 0.952 0.985 0.988 0.966

0.992 0.966 0.956 1.003

(0.05) (0.06) (0.06) (0.03)

- 2.6 (0.95) -

1.48

(1.54)

CT: a

(0.04) (0.07) (0.05) (0.04) (0.09) (0.04)

(0.05)

(0.05) (0.05) (0.09) (0.08) (0.05)

(0.03) (0.02)

0.945 (0.04) g

0.964 (0.05)

0.68 0.37 0.03 * 0.05 * 0.29 0.38 0.75 0.28 0.03 * 0.62 0.02* 0.06 0.03* 0.54 0.54 0.77 0.21 0.13 0.70

a Mini-Mental State Examination (Folstein et al., 1975) - - higher score = better cognition. bBlessed Dementia Scale (Blessed et al., 1968) - - higher score = worse function. Clndex represents sum of z scores of selected tests (see text) - - higher score = better function. X-ray computed tomography. ~Mean region/cerebellum rCBF ratios. fThe asymmetry was significant at both T 1 and T 2 ( P < 0.05). gThe asymmetry was significant at T 2 only ( P < 0.05). * P < 0.05.

cerebral perfusion indices as follows: MMSE, global neuropsychological index and Blessed Dementia Scale score with whole brain perfusion index, frontal neuropsychological with frontal perfusion; verbal memory with left temporal perfusion, and visual memory with right temporal

perfusion indices. Significant correlations were obtained between whole brain perfusion and the MMSE (r = 0.78, P < 0.05) and global neuropsychological index (r = 0.80, P < 0.05) at T v Of the specific brain regions examined, only the correlation between left temporal lobe perfusion and

P. Sachdev et aL / Psychiatry Research: Neuroimaging 68 (1997) 133-141

139

Table 2 Spearman's correlation coefficients between changes in neuropsychological indices and rCBF indices (differences between baseline and follow-up scores) Perfusion ratios"

M MSE b

BDS c

Global n/psych d

Verbal memory ~

Visual memory ~

FrontaP

Whole cerebrum Right hemisphere Left hemisphere Left temporal Right temporal Frontal

0.19 0.29 0.21 0.57* 0.55 0.69"*

0.19 0.41 0.11 0.21 0.09 0.21

0.28 0.11 0.33 0.30 0.55 0.16

0.47 0.43 0.43 0.53* 0.75 * * 0.40

0.19 0.01 0.21 0.15 0.34 0.09

0.10 0.03 0.00 0.08 0.00 0.30

Region/cerebellum perfusion ratios. bMini-Mental State Examination (Folstein et al., 1975). Blessed Dementia Scale (Blessed et al., 1968). °GIobal neuropsychological index. Indices calculated from z scores (see text). * P < 0.1. **P < 0.05.

verbal memory at T I was significant ( r = 0.72, P < 0.05). We used multiple regression analysis to determine whether perfusion indices at T~, or any absolute change or rates (per month) of change in perfusion indices predicted outcome at T 2 as measured by MMSE, BDS or neuropsychological indices. The results were negative, with no perfusion index at T~ or T l-T z change emerging to be a significant predictor of clinical progression of dementia. 4. Discussion

We assessed 10 patients with AD at baseline and a mean of 2 years later. In this period, they demonstrated progression of their illness on most parameters of function. The MMSE scores had declined by a mean 3.5 points per year, which is consistent with previous reports (Burns et al., 1990; Teri et al., 1990). The mean annual increase in the BDS score was 1.38 ( + 0.95), which is less than the 3.5 + 3.7 reported by Lucca et al. (1993). It is interesting that the changes in specific neuropsychological functions were smaller. Indices of frontal function deteriorated significantly, but no change was observed in visual memory at all. There are a number of possible explanations

for the lack of observable change in visual memory: the low sensitivity of the tests used, a slow progression of impairment in this function, or a floor effect. The differential changes in the different neuropsychological functions studied make it important that we examine the relationship between rCBF indices and a range of neuropsychological functions. Some rCBF measures showed significant change over this 2-year period. The cerebellar blood flow did not show an overall change, supporting our decision to use the cerebellar values to normalize the data. The occipital/cerebellar ratios also did not demonstrate any change. While some decline was seen in frontal and temporal perfusion, the changes were small and non-significant for our sample. The possibility of a type II error owing to a small sample size cannot be ruled out. The most significant change was seen in the parietal perfusion, with a greater change seen in the left hemisphere. The finding of an overall reduction in CBF over time is according to expectation from cross-sectional studies, and is consistent with the longitudinal studies of cerebral metabolism in AD using PET scanning (Duara et al., 1986; Jagust et al., 1988; Haxby et al., 1990). It has been suggested that late-onset AD may be associated with greater deficits on the left side, as was the case in

140

P. Sachdev et al. /Psychiatry Research: Neuroimaging 68 (1997) 133-141

our subjects (Koss et al., 1985; Burns et al., 1989). The basis for this asymmetrical decline in rCBF or metabolic activity is not understood. The finding that the parietal rCBF showed the most significant change over time is in agreement with the report of Jagust et al. (1988) who reported a similar finding for cerebral metabolism. It is interesting that changes in whole brain perfusion indices did not correlate significantly with functional decline, but did correlate with the change in ventricular size on CT scan. The CT scan measures did not show significant correlation with changes in any of the functional parameters. Changes in specific neuropsychological functions did not correlate significantly with changes in the corresponding brain regions. While this may again be an effect of the small sample size, there are other possible explanations as well. The correspondence between certain cognitive functions and certain brain lobes is, at best, an approximation, as recent evidence points to multiple brain regions being involved with even 'simple' functions (Posner et al., 1988). Further, the changes in performance on the neuropsychological tests are not linear, and therefore correspondence with rCBF change becomes unlikely. There is some evidence that metabolic changes may precede neuropsychological dysfunction (Haxby et al., 1990), arguing for a functional reserve in the brain. This would suggest that neuropsychological dysfunction may follow rather than co-occur with rCBF changes, resulting in low correlations. Our finding that some of the significant correlations between rCBF and neuropsychological function at baseline were no longer significant at follow-up may have a similar explanation. Some major limitations of our study should be pointed out. The significant attrition of our original sample led to a small sample size. Consequently, our findings can best be considered preliminary and should be followed up with a larger study. Our SPECT scans were performed on a single-headed camera, leading to a relatively low resolution. Recent improvements in technology, which include the availability of triple-headed, brain-dedicated cameras, have considerably improved the resolution of SPECT images, thus making the study of rCBF and neuropsychological function more powerful.

In spite of these limitations, our study supports the utility of SPECT scanning in the longitudinal study of decline in AD. It demonstrates that the cerebral CBF indices, in particular parietal lobe indices, do significantly decline with the progression of the illness. Furthermore, our study supports the published literature in pointing to great heterogeneity in the changes in different parameters of brain function as dementia progresses. Overall assessment should therefore incorporate measures from multiple modalities.

Acknowledgements The authors are grateful to the following colleagues for assistance in data collection and analysis: Ahmad Aniss PhD, Kit-Yun Chee FRANZCP and Clayton Frater ANMT. Doreen Hanlon and Debbie-Maree Bargallie assisted with manuscript preparation, and the Medical Illustration Unit of the University of New South Wales with the figure.

References American Psychiatric Association (1987) DSM-III-R: Diagnostic and Statistical Manual of Mental Disorders. 3rd edn. revised. American Psychiatric Association, Washington DC. Beats, B., Burns, A. and Levy, R. (1991) Single photon emission tomography in dementia, lnt J Geriatr Psychiatry 6, 57-62. Benton, A.L., Hamsher, K., Varney, N.R. and Spreen, O. (1983) Contributions to Neuropsychological Assessment. Oxford University Press, New York. Blessed, G., Tomlinson, B.E. and Roth, M. (1968)The association between quantitative measures of dementia and of senile changes in cerebral grey matter of elderly subjects. Br J Psychiatry 114, 797-811. Bonte, F.J., Ross, E.D., Chehabi, H.H. and Devous, M.D. (1986) SPECT study of regional cerebral blood flow in Alzheimer's disease. J Comput Assist Tomogr 5, 549-564. Brodaty, H. (1990) Low diagnostic yield in a memory disorder clinic. Int Psychogeriatr 2, 149-159. Burns, A., Philpot, M.P., Costa, D.C., Ell, P.J. and Levy, R. (1989) The investigation of Alzheimer's disease with single photon emission tomography. J Neurol Neurosurg Psychiatry 52, 248-253. Burns, A., Jacoby, R. and Levy, R. (1990) Progression of cognitive impairment in Alzheimer's disease. J Am Geriatr Soc 39, 39-45. Cohen, M.B., Graham, L.S., Lake, R., Metter, E.J., Fitten, J., Kulkarni, M.K., Sevrin, R., Yamada, Z., Chang, C.C., Woodruff, N. and Kling, A.S. (1986) Diagnosis of Alzheimer's

P. Sachde v et al. /Psychiatry Research: Neuroimaging 68 (1997) 133-141

disease and multi infarct dementia by tomographic imaging of iodine-123 IMP. J Nllcl Med 27, 769-774. Cutler, N.R. (1988) Cognitive and brain imaging measures of Alzheimer's disease. Neurobiol Aging 9, 90-90. Dekosky, S.T., Shih, W., Schmitt, F.A,, Coupal, J. and Kirkpatrick, C. (1990) Assessing utility of single photon emission tomography (SPECT) scan in Alzheimer's disease: correlation with cognitive severity. Alzheimer Dis Assoc Disord 4, 14-23. Duara, R., Grady, C., Haxby, J., Sundaram, M., Cutler, N.R., Heston, L., Moore, A., Schalgeter, N., Larson, S., Rapoport, S.I. (1986) Positron emission tomography in Alzheimer disease. Arch Neurol 36, 879-887. Ebmeier, K.P., Besson, J.A.O., Crawford, J.R., Palin, A.N., Gemmel, H.G., Sharp, P.F., Chercyman, G.R. and Smith, F.W. (1987) Nuclear magnetic imaging and single photon emission tomography with radio-iodine labelled compounds in the diagnosis of dementia. Acta Psychiatr Scand 75, 549-556. Folstein, M.F., Folstein, S.E. and McHugh, P.R. (1975) 'Mini Mental State': a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189-198. Foster, N.L., Chase, T.N., Mansi, L., Brooks, R., Fedio, P., Patronas, N.J. and Dichiro, G. (1984) Cortical abnormalities in Alzheimer's disease. Ann Neurol 16, 649-654. Goldenberg, G., Podreka, I., Suess, E. and Deecke, L. (1989) The cerebral localization of neuropsychological impairment in Alzheimer's disease: a SPECT study. J Neurol 236, 131-138. Hachiniski, V.C., lliff, L.D. and Zilkha, E. (1975) Cerebral blood flow in dementia. Arch Neurol 32, 632-637. Hamilton, M. (1967) Development of a rating scale for primary depressive illness. BrJ Soc Clin Psychol 6, 278-296. Haxby, J.V., Grady, C.L., Koss, E., Horwitz, B., Heston, L., Schapiro, M., Friedland, R.P. and Rapoport, S.I. (1990) Longitudinal study of cerebral metabolic asymmetries and associated neuropsychological patterns in early dementia of the Alzheimer type. Arch Neurol 47, 753-760. Hughes, C.P., Berg, L., Danziger, W.L., Cohen, L.A. and Martin, R.L. (1982) A new clinical rating scale for the staging of dementia. BrJ Psychiatry 140, 1225-1230. Jagust, J., Friedland, P., Budinger, F., Koss, E. and Ober, B. (1988) Longitudinal studies of regional cerebral metabolism in Alzheimer's disease. Arch Neurol 38, 909-912. Jagust, W.J., Budinger, T.F. and Reed, B.R. (1987) Diagnosis of dementia with single photon emission computer tomography. Arch Neurol 44, 258-262. Johnson, K.A., Holman, B.L., Muller, S.P., Rosen, T.J., English, R., Nagel, J.S. and Growdon, J.H. (1988) Single photon emission computed tomography in Alzheimer's disease. Arch Neurol 45, 392-396.

141

Johnson, K.A., Muller, S.P., Walshe, T.M., English, R.J. and Holman, B.L. (1987) Cerebral perfusion imaging in Alzheimer's disease, use of single photon emission computed tomography and Iofetmine hydrochloride 1-123. Arch Neurol 44, 165-168. Katz, S. and Apkon, C.A. (1976) A measure of primary sociobiological functions, lnt J Health Serv 6, 493-507. Koss, E., Friedland, R., Ober, B. and Jagust, W. (1985) Differences in lateral hemispheric asymmetries of glucose utilization between early and late onset Alzheimer's type dementia. Am J Psychiatry 142, 638-640. Lawton, M.P. and Brodie, E.M. (1969) Assessment of older people: self maintaining and instrumental activity of daily living. J Gerontol 9, 179-186. Lucca, U., Comelli, M., Tettamanti, M., Tiraboschi, P.S. and Pagnoli, A. (1993) Rate of progression and prognostic factors in Alzheimer's disease: a prospective study. J Am Geriatr Soc 41, 45-49. McKhann, G., Drachman, D., Folstein, M., Katzman, R., Price, D. and Stadlan, E. (1984) Clinical diagnosis of Alzheimer's disease: report of the NINCDS-ADRDA work group under the auspices of the Department of Health and Human Services Task Force in Alzheimer's disease. Arch Neurol 34, 939-944. Muller, S.P., Johnson, K.A., Hamil, D., English, R.J., Nagel, S.J., Ichise, M. et al. (1986) Assessment of 1-123 IMP SPECT in mild/moderate and severe AIzheimer's disease. J Nucl Med 27, 889-894. Perani, D., Dipiero, V., Vallar, G., Cappa, S., Messa, C., Boffini, G. et al, (1988) Technetium-99 HMPAO-SPECT study of regional cerebral perfusion in early Alzheimer's disease. J Nucl Med 29, 1507-1514. Posner, M.I., Petersen, S.E., Fox, P.T. and Raichle, M.E. (1988) Localization of cognitive operations in the human brain. Science 240, 1627-1631. Rey, A. (1964) L'examen clinique en psychologie. Paris: Presses Universitaires de France. Robb, R.A. and Hanson, D.P. (1991) A software system for interactive and quantitative visualization of multidimensional biomedical images. Australasian Phys Eng Sere Med 14, 9-30. Talairach, J. and Tournoux, P. (1988) Co-Planar Stereotaxic Atlas of Human Brain. Thieme Medical Publisher Inc., New York. Teri, L., Hughes, J. and Lorson, E.B. (1990) Cognitive deterioration in Alzheimer's disease: behaviour and health factors. J Gerontol 45, 58-63. Wechsler, D. (1981) Wechsler Adult Intelligence Scale-Revised Manual. The Psychological Corporation, New York. Wechsler, D. (1987) Wechsler Memory Scale-Revised Manual. The Psychological Corporation, New York.