SEARCH

SEARCH BY CITATION

Keywords:

  • Alzheimer’s disease (AD);
  • IQ;
  • Kanji;
  • National Adult Reading Test (NART);
  • Wechsler Adult Intelligence Scale–Revised (WAIS-R)

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFERENCES
  9. Appendix

Abstract  The National Adult Reading Test (NART) is widely used as a measure of premorbid IQ of the English-speaking patients with dementia. The purpose of the present study was to develop a Japanese version of the NART (JART), using 50 Japanese irregular words, all of which are Kanji (ideographic script) compound words. Reading performance based on JART and IQ as measured by the Wechsler Adult Intelligence Scale–Revised (WAIS-R) was examined in a sample of 100 normal elderly (NE) persons and in 70 age-, sex-, and education-matched patients with Alzheimer’s disease (AD). The NE group was randomly divided into the NE calculation group (n = 50) and the NE validation group (n = 50). Using the NE calculation group, a linear regression equation was obtained in which the observed full-scale IQ (FSIQ) was regressed on the reading errors of the JART. When the regressed equation computed from the NE calculation group was applied to the NE validation group, the predicted FSIQ adequately fit the observed FSIQ (R2 = 0.78). Further, independent t-tests showed that the JART-predicted IQs were not significantly different between the NE and AD groups, whereas the AD group performed worse in the observed IQs. The reading ability of Kanji compound words is well-preserved in Japanese patients with AD. The JART is a valid scale for evaluating premorbid IQ in patients with AD.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFERENCES
  9. Appendix

When assessing current cognitive function of patients with dementia, knowledge of premorbid IQ is essential. However, premorbid test data are rarely available. The National Adult Reading Test (NART), a reading test of 50 irregularly spelled words in English (e.g. ache), has promise as an assessment tool for the determination of premorbid intellectual function.1,2 Because the words are irregular, intelligent guesswork will not provide the correct pronunciation. For this reason, it has been argued that performance depends more on previous knowledge than on current cognitive capacity.3,4 The NART is reported to be among the most reliable tests in clinical use,5,6 and the use of NART has been extended to the assessment of acquired IQ prior to the development of illness other than dementia, including schizophrenia.7,8

European languages have a great deal in common because they all belong to the Indo-European family of languages and have evolved from a single source. Therefore, applying NART to European languages other than English may be easier than applying it to languages that belong to other families, such as Japanese. Revised versions of NART for other European languages, including Dutch and Italian, have been developed and validated in clinical settings.9,10

The Japanese language differs significantly from European languages in its reading and writing systems.11 Written Japanese is composed of two distinct orthographic characters: Kanji, the ideographic script used to represent lexical morphemes, and Kana, a syllabic script that includes two syllabaries, Hiragana and Katakana. In Japanese standard academic training, more than 1000 characters of Kanji script are taught in elementary school and junior high school. Kanji is perceptually more complex and demanding than English script in that the Kanji words are constructed from a much larger number of basic perceptual elements than the 26 letters of the alphabet used to construct English words. Kanji script has its own meaning, and it generally contains fewer phonemic factors than English counterparts (i.e. if one does not know the Kanji script, reading of it by guesswork is difficult).

Kanji was introduced into Japan from China. However, unlike Chinese Kanji, a single Japanese Kanji character often has a number of different pronunciations depending on its orthographic context. For example, the single Kanji symbol that appears at the top of Fig. 1 signifies ‘hand’, but it can be read as either te or shu, according to the orthographic context, as seen in the second line in Fig. 1. Kanji characters are often used as components of a Kanji compound word. Each Kanji symbol provides unique information about the pronunciation of its component Kanji characters, and the orthography–phonology relationship is word-specific in Kanji compound words. This means that the successful translation from orthography to phonology of Kanji compound words requires word-specific lexical procedures, making Kanji compound words approximately comparable to irregular words in English.

image

Figure 1. Examples of a single Kanji word (composed of one Kanji character) and Kanji compound words (composed of more than one character).

Download figure to PowerPoint

Written Japanese is very different from written English. However, if the brain mechanism at work in the lexical–phonological process is essentially the same for all languages, the reading of Kanji compound words is similar to that of the irregular words used in NART, because both can be considered to be phonologically irregular. If the reading of English irregular words is preserved in English-speaking patients with Alzheimer’s disease (AD), the reading of Kanji compound word is to be preserved in Japanese patients with AD. Indeed, some previous studies using a small sample of Japanese patients with AD also suggested that reading of Japanese Kanji was well-preserved in patients with AD.12,13 Moreover, the reading of irregular words is strongly correlated to intelligence level in normal individuals.1,2

The aim of the present study is to develop a Japanese version of NART (JART) by using Kanji compound words. For this purpose, we first develop a regression equation for estimating current IQ from the reading performance of Kanji compound words in a group of normal elderly individuals. Second, we examine the validity of JART-predicted premorbid IQ in individuals with AD.

METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFERENCES
  9. Appendix

Selection of 50 Kanji compound words for JART

A preliminary study was conducted for the purpose of selecting the 50 Kanji compound words that were best suited to predict IQ in a sample of normal elderly persons. Based on the examples described in a report by The National Language Research Institute,14 we initially chose 100 Kanji compound words as the reading stimuli, from which 50 compound words were selected for use in the JART. We then recruited 43 healthy elderly individuals who were registered with the out-placement office for elderly people in an urban area near Tokyo. All were native Japanese-language speakers and living in the community. Informed consent was obtained from each person in the group after the aim of the study was explained. According to the initial interviews, none of the subjects reported having suffered from a major psychiatric or neurological disorder. In that preliminary study, all subjects underwent the reading test with the 100 Kanji compound words, four subtests (i.e. information, picture completion, digit symbol, and similarity) of the Wechsler Adult Intelligence Scale–Revised (WAIS-R),15–17 as well as the Mini-Mental State Examination (MMSE).18,19 The MMSE scale is widely used as an index of cognitive status, and has been validated in Japanese subjects.19 Scores in the MMSE range from 0 to 30, and scores <24 represent cognitive impairment.19 From our results, two of the healthy elderly subjects received scores of <24 on the MMSE and were excluded from further study, leaving a normal elderly group of 41 (11 female, 30 male) with an average age of 69.1 ± 5.3 years. The mean MMSE score was 28.3 ± 1.7 in a range from 24 to 30, and the mean IQ was 111.1 ± 13.4, ranging from 86 to 134. In order to select the items that had the highest correlation with IQ, the correlation (Spearman’s r) between a true or false score (true = 1, false = 0) for reading each of the 100 Kanji compound words and the short-form IQ result was calculated. We then selected the 50 Kanji compound words that had the highest correlation with IQ as the reading stimuli to be included in JART. The number of correct answers for the selected 50 items accounted for 98% of the variance for the 100 items and 71% of the IQ results. The internal consistency (Cronbach’s α)20 of the selected 50 items was 0.96. Thus, we employed these 50 Kanji compound words as reading stimuli for the JART (Appendix I).

Subjects

The subject sample for the present study comprised 106 normal elderly (NE) individuals and 74 age-, education-, and sex-matched AD patients.

In the NE group, 15 subjects participated in the previous investigation to determine the 50 Kanji compound words for the use in JART. All were healthy individuals who had registered with an out-placement office for elderly people in an urban area near Tokyo. All members of the NE group were native Japanese-language speakers and living in the community. Exclusion criteria for the NE group were: (i) past history of receiving any form of psychiatric treatment; (ii) past history of head injury with loss of consciousness; (iii) MMSE score <24;18,19 or (iv) Center for Epidemiologic Studies Depression (CES-D) Scale score >15.21,22 Among the subjects in the NE group, one reported a past history of medical treatment for depression, none had a history of head injury with loss of consciousness, two had an MMSE score <24, and three had a CES-D score >15. These subjects were excluded from our standardized analysis, leaving the NE group consisting of 100 individuals.

The AD patients were recruited from four geriatric hospitals located in an urban area near Tokyo. All patients (n = 74) underwent a clinical evaluation, including computed tomography (CT) imaging, magnetic resonance imaging (MRI), and a neuropsychological assessment including the MMSE.18,19 Following the diagnostic process including a review of the history of each patient with both the patient and caregiver, the attending physician diagnosed AD according to the Diagnostic and Statistical Manual of Mental Disorders, 4th edition.23 If a patient’s MMSE score exceeded the cut-off (i.e. 23 points), we administered the Wechsler Memory Scale–Revised (WMS-R)24,25 to exclude cases of mild cognitive impairment (MCI) from the AD group. We excluded patients whose scaled scores on delayed recall of WMS-R were >70 points each in further analysis. This procedure left an AD group of 70 individuals.

Informed consent was obtained from all subjects after the aim of the study was explained. The protocol was approved by National Center of Neurology and Psychiatry Ethics Committee.

Neuropsychological assessment

All subjects were given the JART, MMSE, and full version of WAIS-R.15,16 As described in the preceding section, JART is a reading test consisting of 50 Kanji compound words. The reading stimuli were randomly printed out for reading. The subjects were asked to read aloud each Kanji compound word. If they did not know the correct reading, they were encouraged to make their best guess.

Validation of regression equation

We randomly divided the NE group into the NE calculation group (n = 50) and the NE validation group (n = 50). Following the method that Nelson reported for the NART,1,2 we regressed the estimated full-scale IQ (FSIQ), verbal IQ (VIQ), and performance IQ (PIQ), respectively, based on the number of errors on the JART in the NE calculation group.

Using these equations, we determined predicted IQ for each individual in the NE validation group. In the NE validation group, the standardized residuals (SR) and standardized errors (SE) were calculated to examine the fitness of the linear regression equation yielded from the NE calculation group. We also examined whether the SR of the predicted IQ results had a specific pattern for the obtained IQs.

Next, we determined predicted IQs for all subjects in the NE and AD groups, and then examined group differences of JART-predicted IQs and observed IQs using independent t-tests.

We used a two-tailed alpha level of 0.05 for significance. The data were analyzed using spss 10.0 for Windows (SPSS, Tokyo, Japan).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFERENCES
  9. Appendix

The characteristics of the three subject groups are shown in Table 1. The groups did not differ significantly in age (F2,167 = 0.926, P = 0.38), sex (χ2 = 0.70, P = 0.70), or years of education (F2,167 = 0.68, P = 0.51). The mean MMSE score in the AD group was 19.6 (range: 11–29). As for the MMSE score, by definition, a significant difference was seen between the NE calculation, NE validation, and AD groups (F2,167 = 205.6, P < 0.001).

Table 1.  Subject characteristics (mean ± SD)
 NE calculation (n = 50)NE validation (n = 50)AD (n = 70)Statistics
  1. AD, Alzheimer’s disease; MMSE, Mini-Mental State Examination18,19; NE, normal elderly.

Age (years)69.6 ± 5.369.7 ± 5.370.9 ± 6.5F2,167 = 0.926, P = 0.38
Female (%)666067χ2 = 0.70, P = 0.70
Years of education11.5 ± 2.611.5 ± 2.312.0 ± 2.6F2,167 = 0.68, P = 0.51
MMSE score28.0 ± 1.728.0 ± 1.619.6 ± 3.6F2,167 = 205.6, P < 0.001

Linear regression equations extracted from the data obtained from the NE calculation group were as follows:

  • JART-predicted WAIS-R FSIQ = 124.1 − 0.964 ×  (no. errors on the JART)
  • JART-predicted WAIS-R VIQ = 127.8 − 1.093 ×  (no. errors on the JART)
  • JART-predicted WAIS-R PIQ = 117.0 − 0.708 ×  (no. errors on the JART)

As shown in Table 2, the correlations (Pearson’s r) between predicted IQ and observed IQs (i.e. FSIQ, VIQ, PIQ) in the NE validation group were 0.88 (P < 0.001), 0.91 (P < 0.001), and 0.68 (P < 0.001), respectively. The JART-predicted FSIQ was significantly correlated with the age-adjusted scaled scores of all 11 subscales of the WAIS-R.

Table 2.  Correlations between JART-predicted IQ and observed IQs in the NE validation group (n = 50)
  • FSIQ, full -scale IQ; JART, Japanese version of National Adult Reading Test; NE, normal elderly; PIQ, Performance IQ; VIQ, Verbal IQ; WAIS-R, Wechsler Adult Intelligence Scale–Revised.

  • Age-adjusted scaled scores.

  • *

    P < 0.05;

  • **

    P < 0.01;

  • ***

    P < 0.001.

Observed IQs
 FSIQ0.88***
 VIQ0.91***
 PIQ0.68***
Observed WAIS-R scores
 Information0.78***
 Digit span0.64***
 Vocabulary0.87***
 Arithmetic0.58***
 Comprehension0.73***
 Similarity0.80***
 Picture completion0.60***
 Picture arrangement0.66***
 Block design0.47***
 Object assembly0.42**
 Digit symbol0.33*

In the NE validation group, JART-predicted FSIQ was significantly associated with years of education (Pearson’s r = 0.34, P = 0.015). Predicted FSIQ, VIQ, and PIQ accounted for 78%, 84%, and 46% of the variance in observed FSIQ, VIQ, and PIQ, respectively. Using the FSIQ equation, we found that 96% (48/50) of the NE validation group showed their SR within ±2, and the SR of the predicted IQ showed no specific patterns for the number of errors on the JART. The SE of the predicted FSIQ, VIQ, and PIQ were 6.27, 6.02, and 9.06, respectively.

When the equation was applied to the AD group, 84% (59/70) of the subjects had a higher predicted FSIQ than the observed FSIQ (Fig. 2). In subjects with AD, the average discrepancy between estimated IQ and observed FSIQ was 11.2, whereas that of the NE validation group was 0.3.

image

Figure 2. Distribution of Japanese version of the National Adult Reading Test (JART)-predicted full-scale IQ (FSIQ) and observed FSIQ.

Download figure to PowerPoint

Table 3 presents data from the AD group and the entire NE group. Independent t-tests show that members of the AD group had lower scores than those in the NE group for observed FSIQ (t = 4.75, P < 0.001), VIQ (t = 3.52, P = 0.001), and PIQ (t = 5.23, P < 0.001). However, the JART-predicted IQs were not significantly different between the two groups (t = 0.79, P = 0.93).

Table 3.  JART-predicted IQs and observed IQs (mean ± SD)
 NE (n = 100)AD (n = 70)tP
  1. AD, Alzheimer’s disease; FSIQ, full-scale IQ; JART, Japanese version of Adult Reading Test; NE, normal elderly; PIQ, Performance IQ; VIQ, Verbal IQ.

JART-predicted FSIQ102.1 ± 12.1102.0 ± 12.00.79  0.93
Observed FSIQ102.0 ± 13.4 90.8 ± 14.74.75<0.001
JART-predicted VIQ102.9 ± 13.7102.7 ± 13.50.79  0.93
Observed VIQ102.2 ± 14.9 93.4 ± 14.03.52<0.001
JART-predicted PIQ100.9 ± 8.9100.8 ± 8.70.79  0.93
Observed PIQ101.4 ± 11.9 88.5 ± 16.35.23<0.001

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFERENCES
  9. Appendix

The first aim of the present study was to develop a valid regression equation for estimating current IQ from the JART results, which assumed the reading ability of Japanese Kanji (ideographic script) compound words, in normal elderly. We accomplished this aim by dividing the NE subjects into two groups: the calculation group and the validation group. The linear regression equation developed with the calculation group was well-fitted to the validation group. In the validation group, JART-predicted IQ had a higher correlation with observed IQ and was also significantly related to years of education. These results suggest that the equation is valid for estimating current intelligence level in normal elderly individuals.

The second aim was to examine the validity of JART-predicted premorbid IQs in patients with AD. The present AD group was compared with an age-, sex-, and education-matched NE group for both observed IQs and JART-predicted IQs. We found that the observed IQ score was significantly different between the two groups, in that the mean IQ of the AD group was significantly lower than that of the NE group. In contrast, JART-predicted IQs were strikingly similar. In subjects with AD, the average discrepancy between estimated IQ and observed FSIQ was 11.2, whereas that of the NE validation group was 0.3. Individuals in both groups were assumed to have similar levels of intelligence because they did not differ in terms of age, sex, and education. This suggests that the mean FSIQ of the AD group was reduced by 11.2 points from the premorbid level. These results support the validity of JART-predicted IQ as an index of premorbid IQ in individuals with AD. The discrepancy is similar to those found in previous studies using the NART.4,26,27 For example, Nelson and O’Connell reported that their patients with brain atrophy had a 12.15-point discrepancy between NART-predicted IQ and observed IQ.4 Paque and Warrington reported a 10.2-point discrepancy between NART IQ and verbal IQ in initial assessments in their study using patients with dementia.27

Patients with AD have a common pattern of cognitive decline progression. Memory deficit is typically the first cognitive deficit to appear in AD, followed by problems with overall level of intellectual functioning (e.g. problem solving) and visuospatial deficits.28 In general, memory function is divided into various components: the major dichotomy is between episodic memory and semantic memory. Kanji reading ability involves semantic memory function, which relies on the retrieval of word information once acquired. Some studies showed that, in the process of Alzheimer’s-type dementia, semantic memory tasks are relatively preserved in contrast with the marked decline in performance on episodic memory tasks.6,12,29 For example, Sasanuma et al. found that their subjects with dementia had near-normal ability to read Kanji words aloud until the very advanced stage of the disease, and hypothesized that Kanji reading is an involuntary, automatic processing of semantic information that has been highly overlearned and less vulnerable to any semantic degradation.12 McGurn et al. suggested that the aspects of semantic memory involved in pronunciation remain relatively intact in persons with dementia.6 Graham et al. showed that their subjects with AD did not differ from controls in semantic memory tasks.29 The present results in Kanji reading performance could thus be explained in terms of the relative invulnerability of semantic memory in mild AD.

When interpreting our results in terms of reading mechanism, direct comparisons between Japanese Kanji and English words are difficult because the two languages are very different in both writing and reading. Kanji words are perceptually more complex and demanding than English words, in that they are formed from a much larger number of basic perceptual elements than the 26 letters of the English alphabet. This complexity makes Kanji script (and Kanji compound words) more semantic than English words, which seem to be more phonemic.30,31 Indeed, a recent functional MRI study demonstrated that cognitive strategies for reading Chinese Kanji differ from those for words based on the English alphabet.31 That study also showed that Chinese reading disability is characterized by a dysfunction of the neural circuits responsible for the mapping of orthography to phonology and orthography to semantics, and suggested differences in the methods utilized to read Chinese Kanji characters and alphabet-based words. Although the present results showed that the reading of phonologically irregular words in Japanese was well-preserved, as with native speakers of English, the difference in the reading mechanisms between the two languages was not clear and additional study is required.

When utilizing the JART, some caution must be used regarding the limitations of the test. First, our data show that JART was relatively poor in predicting performance IQ. This finding is similar to those seen in previous studies with NART.2 Therefore, one must be cautious when using JART to estimate premorbid performance IQ. However, the present study shows that JART is a reliable measure for estimation of FSIQ and VIQ. Second, the use of regression procedures is limited in terms of range of predicted scores. Our regression equation yielded a possible JART-predicted FSIQ range from 75.9 to 124.1, which is more limited than that of the original NART, which has a possible range of 68.6–130.6. Therefore, the JART-predicted FSIQ scores for subjects with very high intelligence are underestimated, while those of subjects with very low intelligence are overestimated. The third limitation is a selection bias. In the NE group we did not perform any neuropsychological tests other than the MMSE to detect individuals with MCI. It is possible that individuals with MCI were included in the NE group. As for the AD group, their mean MMSE score was 19.6, suggesting that the present AD sample was in the relatively mild stage of dementia. As some studies using the NART have suggested, JART-based prediction may have limited applicability for patients with advanced dementia.32–34

Evaluating premorbid IQ in those with neurological or psychiatric disorders has obvious potential in clinical settings. The deviation of observed IQ from estimated premorbid IQ can indicate the extent to which the disease has progressed or the severity of the condition. Kanji compound words are shown to be useful in estimating premorbid IQ in individuals with AD in a Japanese population, and the use of JART can be extended, in the same way as the original NART, to investigate premorbid IQ in various medical conditions such as schizophrenia.

ACKNOWLEDGMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFERENCES
  9. Appendix

We gratefully acknowledge Dr Hazel E. Nelson for her useful comments. We thank Nori Takei, Department of Psychiatry and Neurology, Hamamatsu University School of Medicine; Masako Fujii, Head of Traumatic Brain Injury (TBI) Rehabilitation Center; Kuniaki Tanaka, Saitama Prefectural Psychiatric Hospital; Takashi Asada, Department of Psychiatry, Institute of Clinical Medicine, University of Tsukuba; Noriko Yamamoto, Chiba University School of Nursing; and Yuki Miyamoto, Kumiko Fujita, Masahiro Sakata, Riichi Someya, and Miki Uetsuki, Graduate School of Medicine, University of Tokyo, for their helpful support and assistance with this study.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFERENCES
  9. Appendix
  • 1
    Nelson HE. National Adult Reading Test (NART). NFER-Nelson, Windsor, UK, 1982.
  • 2
    Nelson HE, Willison JR. National Adult Reading Test (NART), 2nd edn. NFER-Nelson, Windsor, UK, 1991.
  • 3
    Nelson HE, McKenna P. The use of current reading ability in the assessment of dementia. Br. J. Soc. Clin. Psychol. 1975; 14: 259267.
  • 4
    Nelson HE, O’Connell A. Dementia: the estimation of premorbid intelligence levels using the New Adult Reading Test. Cortex 1978; 14: 234244.
  • 5
    Spreen O, Straus E. General intellectual ability and assessment of premorbid intelligence. In: SpreenO, StrausE (eds). A Compendium of Neuropsychological Tests; Administration, Norms, and Commentary, 2nd edn. Oxford University Press, Oxford, 1998; 7582.
  • 6
    McGurn B, Starr JM, Topfer JA et al. Pronunciation of irregular words is preserved in dementia, validating premorbid IQ estimation. Neurology 2004; 62: 11841186.
  • 7
    Crawford JR, Besson JAO, Bremner M, Ebmeier KP, Cochrane HB, Kirkwood K. Estimation of premorbid intelligence in schizophrenia. Br. J. Psychiatry 1992; 161: 6974.
  • 8
    Tracy JI, McGrory AC, Josiassen RC, Monaco CA. A comparison of reading and demographic-based estimates of premorbid intelligence in schizophrenia. Schizophr. Res. 1996; 22: 103109.
  • 9
    Schmand B, Bakker D, Saan R, Louman J. The Dutch Reading Test for adults: a measure of premorbid intelligence level. Tijdschr. Gerontol. Geriatr. 1991; 22: 1519 (in Dutch).
  • 10
    Colombo L, Brivio C, Benaglio I, Siri S, Cappa SF. Alzheimer patient’s ability to read words with irregular stress. Cortex 2000; 36: 703714.
  • 11
    Patterson KE. Basic process of reading: do they differ in Japanese and English? Jpn. J. Neuropsychol. 1990; 6: 414.
  • 12
    Sasanuma S, Sakuma N, Kitano K. Reading Kanji without semantics: evidence from a longitudinal study of dementia. Cogn. Neuropsychol. 1992; 9: 465486.
  • 13
    Nakamura K, Meguro K, Yamazaki H et al. Kanji-predominant alexia in advanced Alzheimer’s disease. Acta Neurol. Scand. 1998; 97: 237243.
  • 14
    National Language Research Institute. Vocabulary and Chinese Characters in Ninety Magazines of Today. National Language Research Institute, Tokyo, 1963 (in Japanese).
  • 15
    Wechsler D. Manual for the Wechsler Adult Intelligence Scale-Revised. Psychological Corporation, New York, 1981.
  • 16
    Shinagawa F, Kobayashi S, Fujita K, Maekawa H. Japanese Wechsler Adult Intelligence Scale-Revised. Nihon Bunka Kagakusha, Tokyo, 1990 (in Japanese).
  • 17
    Misawa G, Kobayashi S, Fujita K, Maekawa H, Dairoku H. Japanese Wechsler Adult Intelligence Scale-Revised Short Forms. Nihon Bunka Kagakusha, Tokyo, 1993 (in Japanese).
  • 18
    Folstein MF, Folstein SE, McHugh PR. ‘Mini-Mental State’: a practical method for grading the cognitive state for the clinician. J. Psychiatr. Res. 1975; 12: 189198.
  • 19
    Mori E, Mitani Y, Yamadori A. The validity of Japanese version of Mini-Mental State Examination in patients with nervous disorder. Shinkei Shinri 1985; 1: 210 (in Japanese).
  • 20
    Cronbach L. Coefficient alpha and internal consistency of tests. Psychometrica 1951; 16: 297334.
  • 21
    Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Appl. Psychol. Measure. 1977; 1: 385401.
  • 22
    Shima S, Shikano T, Kitamura T, Asai M. New self-rating scales for depression. Clin. Psychiatry 1985; 27: 717723 (in Japanese).
  • 23
    American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th edn. American Psychiatric Association, Washington DC, 1994.
  • 24
    Wechsler D. Manual for the Wechsler Memory Scale–Revised. Psychological Corporation, NewYork, 1987.
  • 25
    Sugishita M. Japanese version of Wechsler Memory Scale–Revised. Nihon Bunka Kagakusha, Tokyo, 2001 (in Japanese).
  • 26
    Maddrey AM, Cullum CM, Weiner MF, Filley CM. Premorbid intelligence estimation and level of dementia in Alzheimer’s disease. J. Int. Neuropsychol. Soc. 1996; 2: 551555.
  • 27
    Paque L, Warrington EK. A longitudinal study of reading ability in patients suffering from dementia. J. Int. Neuropsychol. Soc. 1995; 1: 517524.
  • 28
    Zec RF. Neuropsychological function in Alzheimer’s disease. In: ParksRW, ZecRF, WilsonRS (eds). Neuropsychology of Alzheimer’s Disease and Other Dementias. Oxford University Press, New York, 1993; 1013.
  • 29
    Graham NL, Emery T, Hodges JR. Distinctive cognitive profiles in Alzheimer’s disease and subcortical vascular dementia. J. Neurol. Neurosurg. Psychiatry 2004; 75: 6171.
  • 30
    Hirose H. An investigation of the recognition process for jukugo by use of priming paradigms. Jpn. J. Psychol. 1992; 63: 303309 (in Japanese with English abstract).
  • 31
    Siok WT, Perfetti CA, Jin A, Tan LH. Biological abnormality of impaired reading is constrained by culture. Nature 2004; 431: 7176.
  • 32
    Fromm D, Holland AL, Nebes RD, Oakley MA. A longitudinal study of word-reading ability in Alzheimer’s disease: evidence from the National Adult Reading Test. Cortex 1991; 27: 367376.
  • 33
    Patterson K, Graham N, Hodges J. Reading ability of the Alzheimer’s type: a preserved ability? Neuropsychology 1994; 8: 395407.
  • 34
    O’Carroll RE, Prentice N, Murray C, Van Beck M, Ebmeier KP, Goodwin GM. Further evidence that reading ability is not preserved in Alzheimer’s disease. Br. J. Psychiatry 1995; 167: 759662.

Appendix

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFERENCES
  9. Appendix

APPENDIX I: LIST OF 50 KANJI COMPOUND WORDS USED IN JART

Kanji compound wordCorrect readingKanji compound wordCorrect reading
inline imagetou-jiinline imagei-ka
inline imagesan-mainline imageu-chi-wa
inline imageta-ba-koinline imagee-shaku
inline imagemi-zu-sa-kiinline imagema-ne
inline imagehoto-to-gisuinline imagebatteki
inline imagei-saiinline imageka-ka-shi
inline imagechou-hatsuinline imageza-ko
inline imageou-nininline imagemi-ko
inline imageki-beninline imageo-kan
inline imageo-ya-jiinline imagehan-sa
inline imagei-bu-kiinline imageka-sha-ku
inline imagetan-kainline imagesa-su-ga
inline imagea-hi-ruinline imageyu-en
inline imagee-seinline imageji-gi
inline imageakkeinline imagetan-de-ki
inline imagehi-souinline imagezei-ja-ku
inline imagena-ni-to-zoinline imagehi-ma-wa-ri
inline imagekan-juinline imageou-you
inline imageu-nu-bo-reinline imagea-ji-sa-i
inline imageta-so-ga-reinline imageno-ren
inline imagefu-guinline imageshi-ke
inline imageri-kouinline imagene-tsu-zou
inline imagebou-keiinline imageku-ru-mi
inline imagecha-bu-daiinline imagera-tsu-wan
inline imageka-tsu-aiinline imagehi-i-ki