OBJECTIVES The ability to calculate, which is an important aspect of social daily living, is commonly impaired in patients with Alzheimer’s disease even early in the course of the disease. Dyscalculia is often accompanied by focal brain damage, and has been argued to be an independent sign localised around the left temporoparietal region. However, the region most responsible for dyscalculia in Alzheimer’s disease has not been determined. The relation between calculation ability and regional cerebral glucose metabolism in Alzheimer’s disease was therefore examined.
METHODS The calculation ability, In 91 patients with probable Alzheimer’s disease of minimal to moderate severity, was assessed using the arithmetic subtest of the Wechsler adult intelligence scale-revised and the performance correlated with regional cerebral glucose metabolism determined by18F-fluorodeoxyglucose and PET.
RESULTS Regional glucose metabolism in the left inferior parietal lobule and in the left inferior temporal gyrus was significantly correlated with the calculation performance irrespective of age, sex, education, and severity of disease.
CONCLUSIONS The results suggest that dysfunction of the left inferior parietal lobule and the left inferior temporal gyrus plays an important part in producing dyscalculia in patients with Alzheimer’s disease.
- Alzheimer’s disease
- left inferior parietal lobule
Statistics from Altmetric.com
The ability to calculate is important for many aspects of social daily living, especially in shopping, financing, and business affairs. Dyscalculia is often accompanied by focal brain damage, and has been argued to be an independent sign localised around the left temporoparietal region.1-3 Impairment of calculation ability is also a common symptom in patients with Alzheimer’s disease, which can appear early in the course of the disease.3-5However, the pathological basis that is most relevant to dyscalculia in Alzheimer’s disease is still unclear. The purpose of this study was to determine whether there are any anatomical correlates of dyscalculia by using PET of regional cerebral glucose metabolism (CMRglc).
According to the following criteria, 91 Japanese patients with Alzheimer’s disease were selected from those who were given a short term admission to our hospital for examination between February 1994 and October 1997. After a complete description of the study was given to the subjects and their relatives, written informed consent was obtained. The inclusion criteria were (1) the NINCDS/ADRDA criteria for probable Alzheimer’s disease,6 and (2) minimal to moderate functional severity. The exclusion criteria were (1) complications of other neurological diseases or illness, (2) any evidence of focal brain lesions on MRI and of cerebral arterial occlusive lesions on MR angiography, (3) the presence of severe aphasia, severe cognitive, attentional, or behavioural disorders that would make assessment of calculation and PET examination difficult (⩾6 on the digit span subtest of the Wechsler adult intelligence scale-revised (WAIS-R),7 ⩾8 on the repetition subscale, and ⩾6 on the comprehension subscale of the western aphasia battery (WAB)8), and (4) left handed or ambidextrous.
The mean (SD) age of the patients was 69.3 (SD 8.2) years and the mean educational level was 9.2 (SD 2.3) years. There were 65 women and 26 men. Functional severity as rated by the clinical dementia rating scale (CDR)9 was 0.5 (minimal) in 16 patients, 1 (mild) in 59 patients, and 2 (moderate) in 16 patients. The mean mini mental state examination (MMSE)10 score was 19.6 (SD 4.1).
A semiquantitative assessment of calculation was done by using the arithmetic subtest of the WAIS-R7 and the calculation subtest of the WAB.8 The WAIS-R arithmetic subtest consists of 18 arithmetical story problems of increasing difficulty. The subjects were asked to answer within time limits ranging from 15 seconds on the first four problems to 60 seconds on the last nine problems. The subjects were not allowed to use paper or pencil. Raw score bonus points were given for particularly rapid responses on the last nine items. The maximum score of the test was 27. The raw score was subjected to the analysis. The normative performance of elderly people determined in our institute was 13.9 (SD 4.1). The WAB calculation subtest consists of 12 mental arithmetic problems that are presented in written forms. The subjects were requested to calculate mentally and either answer orally or choose the correct answer from four alternatives. The subjects were not allowed to use paper or pencil. Two points were given for each correct answer. The normative performance of elderly people determined in our institute was the maximum score (= 24).
The CMRglc was measured by PET and18F-fluorodeoxyglucose within 1 month after the neuropsychological testing as described previously,11 with the patient under resting conditions with eyes closed and ears unplugged. Two to four circular regions of interest of 10 mm diameter were placed on 32 cerebral regions (table). A senior neuroradiologist blinded to the patients’ status was employed in region of interest placement and measurements. To increase the reliability, the values of CMRglc in each region were averaged. To remove the between subject difference in baseline metabolism, analyses were based on normalised values of regional CMRglc (nCMRglc)—that is, on the ratio of the regional CMRglc to the mean value of metabolic rates for glucose in the bilateral primary sensorimotor cortices, where the histological changes are largely unaffected in Alzheimer’s disease.
A Pearson correlation coefficient was initially used to analyse the relation between nCMRglc of each brain region and each calculation test score. Subsequently, a multiple linear regression partial correlation was calculated with each test score as a dependent variable and nCMRglc of each brain region as an independent variable, to eliminate the possible effect of age, sex, education, and CDR, as these variables were incorporated in the model. Because 32 repeated comparisons were involved for each calculation test, we adopted a conservative α level of 0.05/32=0.00156. All statistical analyses were carried out on SAS release 6.10 (SAS Institute Inc, Cary, NC, USA).
The mean score of the WAIS- R arithmetic test was 5.7 (SD 2.5). The Pearson correlation analyses showed that the test scores correlated significantly with nCMRglc in the left inferior parietal lobule (r=0.405, p<0.0001) and in the left inferior temporal gyrus (r=0.381, p=0.0002). Even after controlling for the effects of age, sex, educational attainment, and CDR in the multiple regression analyses, these correlations remained significant (table 1). No other significant region was noted. Scatter plots illustrate the relations between the arithmetic test score and nCMRglc in both regions (figure). The mean score of the WAB calculation test was 22.7 (SD 2.5). Because more than two thirds of the patients (n=65) achieved the full score on this test, it was not further analysed.
In the present study, dyscalculia in patients with minimal to moderate Alzheimer’s disease was evident in the performance of the WAIS-R arithmetic story problem test but less perceptible in the performance on the WAB calculation problem test. The WAIS-R story problems assess the knowledge of and the ability to apply arithmetical operations without testing symbol recognition or spatial dyscalculia. On the other hand, the performance on the WAB calculation subtest is strongly and adversely affected by an inability to recognise symbols and spatial dyscalculia as well as by poor arithmetical skills.12 Our results indicated that in patients with minimal to moderate Alzheimer’s disease, knowledge and skill of arithmetic operations are the main features of the dyscalculia, whereas symbol recognition disability and spatial dyscalculia are largely insignificant attributes. Attentional, language, and memory impairments may affect the test performance. However, in the present study, as patients with attentional and language impairments severe enough to interfere with the performance of the tests were not included, involvement of these factors is unlikely to be significant. Although verbal short term memory is a prerequisite for maintenance of arithmetical stories, sentence repetition was well preserved in all patients.
The main finding of the present study is that dyscalculia in Alzheimer’s disease was significantly correlated with glucose hypometabolism in the left inferior parietal lobule and in the left inferior temporal gyrus. Only a few studies have explored the relation between dyscalculia and regional brain dysfunction in patients with Alzheimer’s disease and they have varying methodological limitations. These include spurious correlations caused by severity of dementia that should be factored out, the lack of statistical power with small sample size, and inappropriate application of statistical analysis including multiple comparisons. Previous studies using single photon emission computed tomography (SPECT) or PET have shown conflicting results; they have shown a correlation with hypoperfusion in the left posterior temporal and parietal regions,13 with hypoperfusion in the left parieto-occipital region,14 or with hypoperfusion in the right anterior inferior frontal and anterior superior parietal regions,15 or have failed to show correlations with any specific regions.16 However, based on many studies of patients with focal brain damage, dyscalculia has been localised to lesions in the left temporoparietal region including the left inferior parietal lobule,3 4 although there is no specific region uniquely underlying calculation impairment.17 The role of the posterior parietal cortex in arithmetical calculation was also shown by activation studies in normal subjects.18 19Roland et al,18 measuring regional cerebral blood flow with the intracarotid 133Xe injection technique, found that the blood flow increased in the bilateral angular cortices during a silent serial subtraction task. Rueckert et al,19 by using functional MRI, showed a significant cortical activation in bilateral posterior parietal cortices as well as premotor and prefrontal cortices during a silent serial subtraction task. Our findings that dyscalculia in patients with Alzheimer’s disease is attributable to dysfunction of the inferior parietal lobule is consistent with those findings of dyscalculia in patients with focal brain damage and those in brain activation studies in normal subjects.
A peculiar finding in this study is that the glucose metabolism in the left inferior temporal gyrus was associated with the performance on the WAIS-R story problems. Little direct evidence is available from lesion studies and functional neuroimaging studies indicating a role of this region in calculation. The left inferior temporal gyrus is involved in semantic memory processing.20 Therefore, the association between the left inferior temporal function and the calculation performance is likely attributable to this region’s function, which subserves the knowledge of arithmetical operations that are essential for solving arithmetical story problems. Kennedy et al 21 reported a pedigree with familial Alzheimer’s disease, in which neuropsychological profiles were characterised by an initial memory deficit and early dyscalculia. In those patients, severe dyscalculia was evident on the WAIS-R arithmetic subtest, and glucose hypometabolism in the left inferior temporal gyrus was prominent. Their findings are consistent with the results of our correlational analysis.
In conclusion, the results of the present study suggest that damage to the left inferior parietal lobule, where dyscalculia has commonly been localised by studies of patients with focal brain damage and by functional MRI studies, also plays an important part in impairing calculation ability in patients with Alzheimer’s disease. The present regional metabolic study successfully demonstrated the functional neuroanatomy of calculation impairment in Alzheimer’s disease. This result shows, in agreement with a previous study,22 how studies of brain-behaviour relations in patients with Alzheimer’s disease may be useful for determining the cortical localisation of subtle cognitive functions.
We thank Dr Hajime Kitagaki (Division of Neuroimaging Research) for his help in various parts of the study, and Yoko Takatsuki and Akitsugu Tokimasa (Neurorehabilitation Service) for their technical assistance. This study was supported by the Medical Research Fund of Hyogo Medical Association (1997).
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.