- Email: [email protected]
Influence of education on the benton visual retention test performance as mediated by a strategic search component
Brain and Cognition 53 (2003) 408–411 www.elsevier.com/locate/b&c
Influence of education on the benton visual retention test performance as mediated by a strategic search component Nicolas Le Carret,a Constant Rainville,b Nathalie Lechevallier,a Sylviane Lafont,c Luc Letenneur,a and Colette Fabrigoulea a b
INSERM U.330, Universit e Victor S egalen, Bordeaux, France Institut Universitaire de G eriatrie de Montr eal, Quebec, Canada c INRETS, Universit e Claude Bernard, Lyon, France Accepted 7 May 2003
Abstract Age-related cognitive decline has been reported by several studies. However, little investigations have dealt with the effect of education on this decline. In the present study, we examined the influence of educational level on visual working memory, evaluated by the Benton Visual Retention Test (BVRT in recognition format) in 829 elderly participants of the PAQUID study. A multivariate linear model suggested that the effect of education on BVRT performance was not mediated by visual discrimination abilities suggesting that it was mainly supported by better executive abilities. Moreover, the analysis of success and error location suggested that subjects with higher educational level use a more exhaustive exploration strategy during the recognition phase than subjects with lower educational level, which permit them to better perform. The ability of high educational level subjects to use more efficient strategies may participate to the Ôcognitive reserveÕ capacity. Ó 2003 Published by Elsevier Inc.
1. Introduction Previous investigations have shown that aging alters cognitive performances (Christensen, Henderson, Griffiths, & Levings, 1997; Stankov, 1988). Some results have suggested that age-related changes in cognition may be the result of slowed information processing speed (Salthouse, 1976) associated to a selective disruption of executive functions (OÕSullivan et al., 2001). However, this age-associated decline seems to be lessened in subjects with a high educational level, education being a factor promoting the maintenance of cognitive efficiency (Snowdon et al., 1996). In a previous analysis of the PAQUID study, we have shown that education had a global impact on neuropsychological test performances of elderly people (Le Carret et al., submitted) and that this impact seemed to be mediated by specific enhancement of two cognitive components: the conceptualization abilities and the controlled executive processes. In the PAQUID study, the Benton Visual Retention Test 0278-2626/$ - see front matter Ó 2003 Published by Elsevier Inc. doi:10.1016/S0278-2626(03)00155-6
(BVRT in recognition) was strongly influenced by education. In other studies, the recognition format is rarely used but Shichita, Shuichi, Ohashi, and Matuzaki (1986) using the BVRT copy format, also showed that education was a significant predictor of age-associated change in BVRT performance. To explain the educational influence on the recognition format, it is important to consider the two major cognitive components involved in the test. It implicates visuo-spatial discrimination abilities and executive aspect of working memory because it is necessary to compare the four recognition figures with the figure stored in working memory. The aim of this study is to understand if the effect of educational level is mediated by an increase of executive abilities or by an increase of visual discrimination abilities. In this purpose, we analyzed at the third follow-up of the communitybased PAQUID study, the performances on BVRT and on a visual discrimination test in relation with age and educational level in elderly people, using quantitative and qualitative approaches.
N. Le Carret et al. / Brain and Cognition 53 (2003) 408–411
409
2. Method
2.3. Analysis procedure
2.1. Participants
A multiple linear model was used to analyze the mean BVRT score according to several explanatory variables: age, gender, and educational level. Education was recorded on four levels: no schooling or primary school level (equivalent to 0–5 years of schooling), short secondary school level (equivalent to 6–9 years of schooling), long secondary school level (equivalent to 10–12 years of schooling), and university level (over 12 years of schooling). In the French educational system, each educational level is validated by a diploma which allows the access to the next educational level. In the analysis, we have created 3 variables considering the validation or not of each educational level by a diploma. In this way, level 1 was Ôno schoolingÕ pooled with Ôprimary school education but no diplomaÕ. It was the reference class for comparison with the other educational levels. Level 2 was Ôprimary school validated by a diplomaÕ or Ôshort secondary educationÕ without validation by diploma. Level 3 consisted of Ôshort secondary education validated by a diplomaÕ pooled with all the superior diploma. In a first step, univariate linear models were fitted to individually study the effect of the explanatory variables on the mean BVRT score. In a second step, significant variables identified in the first step were included in a multivariate linear model to study the effect of the explanatory variables, adjusted on each other, on the mean BVRT score. In a second multivariate linear model, we have added the VD score as an extra explanatory variable. Then, an analysis of the location of the figure chosen by the subjects was performed in order to assess potential differences in target search strategy according to educational level. First, the percentage of success was calculated for each BVRT recognition square. The use of percentage was necessary because the number of trials in which a given square contains the target is not the same for the different squares. In the case of a wrong response, we calculated the percentage of errors in each square according to the location of the target. For example, if the target is located in the square B, we calculated the percentage of wrong recognition in the square A, C, and D.
The PAQUID (Personnes Agees QUID) study is a cohort of more than 4000 elderly people living in two administrative areas in Southwestern France, longitudinally followed during 10 years. After a baseline screening (T0), subjects were followed one year (T1), three years (T3), five years (T5), eight years (T8), and 10 years later (T10). At each follow-up, a socioeconomic and demographic questionnaire and a neuropsychological battery were administered at home by trained psychologists (Dartigues et al., 1992). For this analysis, we selected a sample of non-demented people, seen at the third follow-up (T5), according to the following criteria. The inclusion criteria were: (1) to have accepted the visit at the third follow-up (T5) of the PAQUID study, (2) to have accepted the two previous follow-ups of the PAQUID study (T1 and T3), (3) to have completed the Mini Mental State Examination (MMSE), the BVRT and the Visual Discrimination Test at the third followup (T5). The exclusion criteria were: (1) to have been classified as demented at any visit (T0, T1, T3, T5), (2) to have presented sequels of cerebral vascular accident at any visit, and (3) to be blind. Finally, 829 subjects were selected, including 376 men and 453 women. The mean age of the sample is equal to 76.89 (SD ¼ 5.12) and is ranged from 67 to 94. The mean MMSE score is equal to 27.85 (SD ¼ 1.94).
2.2. Neuropsychological tests The Benton Visual Retention Test (BVRT) (Benton, 1965) assessed the immediate visual memory. The Form F was used with the M administration: it consists of 15 ink figures composed of geometric shapes. Each target is presented to the subject for 10 s. Then, it is hidden and the subject has to identify the target among three other figures. The four figures of the recognition phase are presented on four squares paired on two lines. Each figure is on a square and each location of the square is identified by the letters A, B, C, and D according to the American reading direction. Total score is the sum of good recognition. Visual Discrimination (VD) is a sub-test of the Visual Gnosis Examination Protocol of Montreal-Toulouse (Agniel, Joanette, Doyon, & Duchein, 1987). It consists of 10 discrimination sheets. A geometric shape is presented at the top of each sheet, and under this target figure, are disposed six figures including a figure identical to the target and five distractors. The subject has to identify the target among the other figures. Before the testing phase, two examples are given to the subject. Total score is the sum of good recognition.
3. Results 3.1. Influence of educational level on the BVRT performance In multivariate analysis, the coefficients reflect the change of points on BVRT score attributable to the effect of each variable, the change being a gain or a loss (minus sign ¼ loss of point). The adjusted mean score increased by 1.22 points (p < :001) in subjects who reached the
410
N. Le Carret et al. / Brain and Cognition 53 (2003) 408–411
educational level 2 compared to subjects who reached the educational level 1. In the same way, the adjusted mean score increased by 1.84 (p < :001) in subjects who reached level 3 compared to subjects who reached level 1. When the VD score was added as an explanatory variable in the multivariate linear model, the effect of education on BVRT score was slightly lessened, but it remained significant: the mean score increased by 0.95 points (p < :001) for the level 2 versus level 1, and by 1.47 points (p < :001) for level 3 versus level 1. 3.2. Analysis of success and errors location according to educational level Table 1 shows the success percentages at the BVRT recognition according to target locations and the three educational levels. Whatever the educational level, the best recognition performance was obtained when the target was in square A. The performance was much lower when the target was located in square D. However, the performance tend to increase with educational level for each location, especially in square D. The analysis of the rate of wrong recognition on each square according to each location of the target and the three educational levels showed that when the targets were located on squares A and B, wrong recognition percentages on the other squares did not vary systematically according to educational level. However, as shown in Table 2, when the target was located on square Table 1 Percentage of success for each location of the target according to educational level Educational level 1
Educational level 2
Educational level 3
91 68
93 80
96 86
72 38
79 52
85 57
Educational level 1: no schooling and primary schooling without diploma. Educational level 2: primary schooling validated by diploma and secondary schooling without validation. Educational level 3: secondary schooling validated by diploma and more extended schooling.
Table 2 Percentage of wrong recognition on bottom locations according to the place of the target and educational level
Educational level 1 Educational level 2 Educational level 3
Target in C—Percentage of wrong recognition in D
Target in D—Percentage of wrong recognition in C
17.48 18.45 23.68
25.26 28.24 35.33
Educational level 1: no schooling and primary schooling without diploma. Educational level 2: primary schooling validated by diploma and secondary schooling without validation. Educational level 3: secondary schooling validated by diploma and more extended schooling.
C, the percentage of wrong recognition on square D increases with educational level (17.48% for level 1 to 23.68% for level 3), and when the target was located on square D the percentage of wrong recognition on square C increases as well with educational level (25.26% for level 1 to 35.33% for level 3).
4. Discussion Previous studies have shown that a high educational level is related to an increase of the BVRT copy performance. The results of the multivariate linear model extended these findings to BVRT recognition performance and showed that subjects with the highest educational level gained 1.84 points on the mean BVRT score compared to subjects with the lowest educational level, independently of age and gender. When the VD score was introduced as an explanatory variable in the analysis the effect of education on the BVRT performance was slightly decreased. Therefore the effect of education on the BVRT performance was only partially mediated by visuospatial perceptual abilities. It suggests that the main effect of education was mediated by a more efficient executive working memory component. This effect of education on BVRT working memory component may be mediated either by better encoding strategy or by better recognition strategy. The test does not allow to analyze encoding strategy but the analysis of the location of successes and errors permitted to get information on some aspects of the recognition strategy. Indeed, the analysis of the rate of success in the four squares containing the targets showed that the rate of success increases whatever the squares according to educational level. However this increase was stronger in the C and D squares than in the A and B squares. The fact that global performance was better in subjects with a higher educational level could be explained by a more efficient encoding of the item in working memory. However, if we admit that subjects examine the top squares before the bottom squares, according to the reading habit, the greater increase of success in the bottom squares in high-educated subjects could also suggest that they observed the four squares before to make their choice. On the contrary, subjects with a low educational level seemed to make a less exhaustive exploration of the items and to examine less systematically the two bottom squares before making their choice. The analysis of error frequency according to the target location is consistent with this interpretation. Indeed, we noticed that subjects with the lowest educational level made less frequently wrong recognition on square C and D than subjects with the highest educational level. Taken together, these results showed that subjects with high educational level gave more recognition responses (whatever wrong or right) on the bottom square,
N. Le Carret et al. / Brain and Cognition 53 (2003) 408–411
suggesting they may explore more frequently these two bottom square before making their choice. In conclusion, these results showed that educational level is associated with a better performance on BVRT in elderly people. This effect seems to be partly mediated by the ability of highly educated subjects to use a more strategic search of the targets in item recognition, and thus, mediated by executive aspect of working memory. This is consistent with previous results that showed that educational level has a strong influence on tests which involve executive components such as categorical fluency, WAIS similarities and digit symbols. This influence of educational level on executive functions may be related to previous results which have suggested that education is implicated in the formation of a Ôcognitive reserveÕ capacity (Letenneur et al., 1999; Snowdon et al., 1996; Stern et al., 1994) which may delay the clinical expression of AlzheimerÕs disease by offsetting the cognitive expression of the pathology (Schmand, Smit, Geerlings, & Lindeboom, 1997). If educational level improves executive functions, it may constitute a major component of the reserve capacity, giving more capacity to highly educated subjects to use compensatory strategies to offset the repercussions of the very first stages of a neurodegenerative process.
References Agniel, A., Joanette, Y., Doyon, B., Duchein, C., et al. (1987). Protocole dÕevaluation des gnosies visuelles Montreal-Toulouse (PEGV). Laboratoire Th.-Alajouanine, Centre de recherche du Centre Hospitalier C^ ote-des-Neiges, Montreal.
411
Benton, A. (1965). Manuel pour lÕapplication du test de retention visuelle. Applications cliniques et experimentales (2nd ed.). Paris: Centre de Psychologie Appliquee. Christensen, H., Henderson, A. S., Griffiths, K., & Levings, C. (1997). Does ageing inevitably lead to declines in cognitive performance? A longitudinal study of elite academics. Personality and Individual Differences, 23(1), 67–78. Dartigues, J. F., Gagnon, M., Barberger-Gateau, P., Letenneur, L., Commenges, D., & Sauvel, C., et al. (1992). The PAQUID epidemiological program on brain ageing. Neuroepidemiology, 11(1), 14–18. Letenneur, L., Gilleron, V., Commenges, D., Helmer, C., Orgogozo, J. M., & Dartigues, J. F. (1999). Are sex and educational level independent predictors of dementia and AlzheimerÕs disease? Incidence data from the PAQUID project. Journal of Neurology, Neurosurgery & Psychiatry, 66(2), 177–183. OÕSullivan, M., Jones, D., Summers, P., Morris, R., Williams, S., & Markus, H. (2001). Evidence for cortical ‘‘disconnection’’ as a mechanism of age-related cognitive decline. Neurology, 57(4), 632– 638. Salthouse, T. (1976). Speed and age: multiple rate of age decline. Experimental Aging Research, 2(4), 349–359. Schmand, B., Smit, J. H., Geerlings, M. I., & Lindeboom, J. (1997). The effects of intelligence and education on the development of dementia. A test of the brain reserve hypothesis. Psychological Medicine, 27, 1337–1344. Shichita, K., Shuichi, H., Ohashi, Y., & Matuzaki, T. (1986). Memory changes in the benton visual retention test between ages 70 and 75. Journal of Gerontology, 41, 385–386. Snowdon, D. A., Kemper, S. J., Mortimer, J. A., Greiner, L. H., Wekstein, D. R., & Markesbery, W. R. (1996). Linguistic ability in early life and cognitive function and AlzheimerÕs disease in later life. Journal of the American Medical Association, 275(7), 528–532. Stankov, L. (1988). Aging, attention, and intelligence. Psychology and Aging, 3(1), 59–74. Stern, Y., Gurland, B., Tatemichi, T. K., Tang, M. X., Wilder, D., & Mayeux, R. (1994). Influence of education and occupation on the incidence of AlzheimerÕs disease. Journal of the American Medical Association, 271(13), 1004–1010.
Influence of education on the benton visual retention test performance as mediated by a strategic search component Nicolas Le Carret,a Constant Rainville,b Nathalie Lechevallier,a Sylviane Lafont,c Luc Letenneur,a and Colette Fabrigoulea a b
INSERM U.330, Universit e Victor S egalen, Bordeaux, France Institut Universitaire de G eriatrie de Montr eal, Quebec, Canada c INRETS, Universit e Claude Bernard, Lyon, France Accepted 7 May 2003
Abstract Age-related cognitive decline has been reported by several studies. However, little investigations have dealt with the effect of education on this decline. In the present study, we examined the influence of educational level on visual working memory, evaluated by the Benton Visual Retention Test (BVRT in recognition format) in 829 elderly participants of the PAQUID study. A multivariate linear model suggested that the effect of education on BVRT performance was not mediated by visual discrimination abilities suggesting that it was mainly supported by better executive abilities. Moreover, the analysis of success and error location suggested that subjects with higher educational level use a more exhaustive exploration strategy during the recognition phase than subjects with lower educational level, which permit them to better perform. The ability of high educational level subjects to use more efficient strategies may participate to the Ôcognitive reserveÕ capacity. Ó 2003 Published by Elsevier Inc.
1. Introduction Previous investigations have shown that aging alters cognitive performances (Christensen, Henderson, Griffiths, & Levings, 1997; Stankov, 1988). Some results have suggested that age-related changes in cognition may be the result of slowed information processing speed (Salthouse, 1976) associated to a selective disruption of executive functions (OÕSullivan et al., 2001). However, this age-associated decline seems to be lessened in subjects with a high educational level, education being a factor promoting the maintenance of cognitive efficiency (Snowdon et al., 1996). In a previous analysis of the PAQUID study, we have shown that education had a global impact on neuropsychological test performances of elderly people (Le Carret et al., submitted) and that this impact seemed to be mediated by specific enhancement of two cognitive components: the conceptualization abilities and the controlled executive processes. In the PAQUID study, the Benton Visual Retention Test 0278-2626/$ - see front matter Ó 2003 Published by Elsevier Inc. doi:10.1016/S0278-2626(03)00155-6
(BVRT in recognition) was strongly influenced by education. In other studies, the recognition format is rarely used but Shichita, Shuichi, Ohashi, and Matuzaki (1986) using the BVRT copy format, also showed that education was a significant predictor of age-associated change in BVRT performance. To explain the educational influence on the recognition format, it is important to consider the two major cognitive components involved in the test. It implicates visuo-spatial discrimination abilities and executive aspect of working memory because it is necessary to compare the four recognition figures with the figure stored in working memory. The aim of this study is to understand if the effect of educational level is mediated by an increase of executive abilities or by an increase of visual discrimination abilities. In this purpose, we analyzed at the third follow-up of the communitybased PAQUID study, the performances on BVRT and on a visual discrimination test in relation with age and educational level in elderly people, using quantitative and qualitative approaches.
N. Le Carret et al. / Brain and Cognition 53 (2003) 408–411
409
2. Method
2.3. Analysis procedure
2.1. Participants
A multiple linear model was used to analyze the mean BVRT score according to several explanatory variables: age, gender, and educational level. Education was recorded on four levels: no schooling or primary school level (equivalent to 0–5 years of schooling), short secondary school level (equivalent to 6–9 years of schooling), long secondary school level (equivalent to 10–12 years of schooling), and university level (over 12 years of schooling). In the French educational system, each educational level is validated by a diploma which allows the access to the next educational level. In the analysis, we have created 3 variables considering the validation or not of each educational level by a diploma. In this way, level 1 was Ôno schoolingÕ pooled with Ôprimary school education but no diplomaÕ. It was the reference class for comparison with the other educational levels. Level 2 was Ôprimary school validated by a diplomaÕ or Ôshort secondary educationÕ without validation by diploma. Level 3 consisted of Ôshort secondary education validated by a diplomaÕ pooled with all the superior diploma. In a first step, univariate linear models were fitted to individually study the effect of the explanatory variables on the mean BVRT score. In a second step, significant variables identified in the first step were included in a multivariate linear model to study the effect of the explanatory variables, adjusted on each other, on the mean BVRT score. In a second multivariate linear model, we have added the VD score as an extra explanatory variable. Then, an analysis of the location of the figure chosen by the subjects was performed in order to assess potential differences in target search strategy according to educational level. First, the percentage of success was calculated for each BVRT recognition square. The use of percentage was necessary because the number of trials in which a given square contains the target is not the same for the different squares. In the case of a wrong response, we calculated the percentage of errors in each square according to the location of the target. For example, if the target is located in the square B, we calculated the percentage of wrong recognition in the square A, C, and D.
The PAQUID (Personnes Agees QUID) study is a cohort of more than 4000 elderly people living in two administrative areas in Southwestern France, longitudinally followed during 10 years. After a baseline screening (T0), subjects were followed one year (T1), three years (T3), five years (T5), eight years (T8), and 10 years later (T10). At each follow-up, a socioeconomic and demographic questionnaire and a neuropsychological battery were administered at home by trained psychologists (Dartigues et al., 1992). For this analysis, we selected a sample of non-demented people, seen at the third follow-up (T5), according to the following criteria. The inclusion criteria were: (1) to have accepted the visit at the third follow-up (T5) of the PAQUID study, (2) to have accepted the two previous follow-ups of the PAQUID study (T1 and T3), (3) to have completed the Mini Mental State Examination (MMSE), the BVRT and the Visual Discrimination Test at the third followup (T5). The exclusion criteria were: (1) to have been classified as demented at any visit (T0, T1, T3, T5), (2) to have presented sequels of cerebral vascular accident at any visit, and (3) to be blind. Finally, 829 subjects were selected, including 376 men and 453 women. The mean age of the sample is equal to 76.89 (SD ¼ 5.12) and is ranged from 67 to 94. The mean MMSE score is equal to 27.85 (SD ¼ 1.94).
2.2. Neuropsychological tests The Benton Visual Retention Test (BVRT) (Benton, 1965) assessed the immediate visual memory. The Form F was used with the M administration: it consists of 15 ink figures composed of geometric shapes. Each target is presented to the subject for 10 s. Then, it is hidden and the subject has to identify the target among three other figures. The four figures of the recognition phase are presented on four squares paired on two lines. Each figure is on a square and each location of the square is identified by the letters A, B, C, and D according to the American reading direction. Total score is the sum of good recognition. Visual Discrimination (VD) is a sub-test of the Visual Gnosis Examination Protocol of Montreal-Toulouse (Agniel, Joanette, Doyon, & Duchein, 1987). It consists of 10 discrimination sheets. A geometric shape is presented at the top of each sheet, and under this target figure, are disposed six figures including a figure identical to the target and five distractors. The subject has to identify the target among the other figures. Before the testing phase, two examples are given to the subject. Total score is the sum of good recognition.
3. Results 3.1. Influence of educational level on the BVRT performance In multivariate analysis, the coefficients reflect the change of points on BVRT score attributable to the effect of each variable, the change being a gain or a loss (minus sign ¼ loss of point). The adjusted mean score increased by 1.22 points (p < :001) in subjects who reached the
410
N. Le Carret et al. / Brain and Cognition 53 (2003) 408–411
educational level 2 compared to subjects who reached the educational level 1. In the same way, the adjusted mean score increased by 1.84 (p < :001) in subjects who reached level 3 compared to subjects who reached level 1. When the VD score was added as an explanatory variable in the multivariate linear model, the effect of education on BVRT score was slightly lessened, but it remained significant: the mean score increased by 0.95 points (p < :001) for the level 2 versus level 1, and by 1.47 points (p < :001) for level 3 versus level 1. 3.2. Analysis of success and errors location according to educational level Table 1 shows the success percentages at the BVRT recognition according to target locations and the three educational levels. Whatever the educational level, the best recognition performance was obtained when the target was in square A. The performance was much lower when the target was located in square D. However, the performance tend to increase with educational level for each location, especially in square D. The analysis of the rate of wrong recognition on each square according to each location of the target and the three educational levels showed that when the targets were located on squares A and B, wrong recognition percentages on the other squares did not vary systematically according to educational level. However, as shown in Table 2, when the target was located on square Table 1 Percentage of success for each location of the target according to educational level Educational level 1
Educational level 2
Educational level 3
91 68
93 80
96 86
72 38
79 52
85 57
Educational level 1: no schooling and primary schooling without diploma. Educational level 2: primary schooling validated by diploma and secondary schooling without validation. Educational level 3: secondary schooling validated by diploma and more extended schooling.
Table 2 Percentage of wrong recognition on bottom locations according to the place of the target and educational level
Educational level 1 Educational level 2 Educational level 3
Target in C—Percentage of wrong recognition in D
Target in D—Percentage of wrong recognition in C
17.48 18.45 23.68
25.26 28.24 35.33
Educational level 1: no schooling and primary schooling without diploma. Educational level 2: primary schooling validated by diploma and secondary schooling without validation. Educational level 3: secondary schooling validated by diploma and more extended schooling.
C, the percentage of wrong recognition on square D increases with educational level (17.48% for level 1 to 23.68% for level 3), and when the target was located on square D the percentage of wrong recognition on square C increases as well with educational level (25.26% for level 1 to 35.33% for level 3).
4. Discussion Previous studies have shown that a high educational level is related to an increase of the BVRT copy performance. The results of the multivariate linear model extended these findings to BVRT recognition performance and showed that subjects with the highest educational level gained 1.84 points on the mean BVRT score compared to subjects with the lowest educational level, independently of age and gender. When the VD score was introduced as an explanatory variable in the analysis the effect of education on the BVRT performance was slightly decreased. Therefore the effect of education on the BVRT performance was only partially mediated by visuospatial perceptual abilities. It suggests that the main effect of education was mediated by a more efficient executive working memory component. This effect of education on BVRT working memory component may be mediated either by better encoding strategy or by better recognition strategy. The test does not allow to analyze encoding strategy but the analysis of the location of successes and errors permitted to get information on some aspects of the recognition strategy. Indeed, the analysis of the rate of success in the four squares containing the targets showed that the rate of success increases whatever the squares according to educational level. However this increase was stronger in the C and D squares than in the A and B squares. The fact that global performance was better in subjects with a higher educational level could be explained by a more efficient encoding of the item in working memory. However, if we admit that subjects examine the top squares before the bottom squares, according to the reading habit, the greater increase of success in the bottom squares in high-educated subjects could also suggest that they observed the four squares before to make their choice. On the contrary, subjects with a low educational level seemed to make a less exhaustive exploration of the items and to examine less systematically the two bottom squares before making their choice. The analysis of error frequency according to the target location is consistent with this interpretation. Indeed, we noticed that subjects with the lowest educational level made less frequently wrong recognition on square C and D than subjects with the highest educational level. Taken together, these results showed that subjects with high educational level gave more recognition responses (whatever wrong or right) on the bottom square,
N. Le Carret et al. / Brain and Cognition 53 (2003) 408–411
suggesting they may explore more frequently these two bottom square before making their choice. In conclusion, these results showed that educational level is associated with a better performance on BVRT in elderly people. This effect seems to be partly mediated by the ability of highly educated subjects to use a more strategic search of the targets in item recognition, and thus, mediated by executive aspect of working memory. This is consistent with previous results that showed that educational level has a strong influence on tests which involve executive components such as categorical fluency, WAIS similarities and digit symbols. This influence of educational level on executive functions may be related to previous results which have suggested that education is implicated in the formation of a Ôcognitive reserveÕ capacity (Letenneur et al., 1999; Snowdon et al., 1996; Stern et al., 1994) which may delay the clinical expression of AlzheimerÕs disease by offsetting the cognitive expression of the pathology (Schmand, Smit, Geerlings, & Lindeboom, 1997). If educational level improves executive functions, it may constitute a major component of the reserve capacity, giving more capacity to highly educated subjects to use compensatory strategies to offset the repercussions of the very first stages of a neurodegenerative process.
References Agniel, A., Joanette, Y., Doyon, B., Duchein, C., et al. (1987). Protocole dÕevaluation des gnosies visuelles Montreal-Toulouse (PEGV). Laboratoire Th.-Alajouanine, Centre de recherche du Centre Hospitalier C^ ote-des-Neiges, Montreal.
411
Benton, A. (1965). Manuel pour lÕapplication du test de retention visuelle. Applications cliniques et experimentales (2nd ed.). Paris: Centre de Psychologie Appliquee. Christensen, H., Henderson, A. S., Griffiths, K., & Levings, C. (1997). Does ageing inevitably lead to declines in cognitive performance? A longitudinal study of elite academics. Personality and Individual Differences, 23(1), 67–78. Dartigues, J. F., Gagnon, M., Barberger-Gateau, P., Letenneur, L., Commenges, D., & Sauvel, C., et al. (1992). The PAQUID epidemiological program on brain ageing. Neuroepidemiology, 11(1), 14–18. Letenneur, L., Gilleron, V., Commenges, D., Helmer, C., Orgogozo, J. M., & Dartigues, J. F. (1999). Are sex and educational level independent predictors of dementia and AlzheimerÕs disease? Incidence data from the PAQUID project. Journal of Neurology, Neurosurgery & Psychiatry, 66(2), 177–183. OÕSullivan, M., Jones, D., Summers, P., Morris, R., Williams, S., & Markus, H. (2001). Evidence for cortical ‘‘disconnection’’ as a mechanism of age-related cognitive decline. Neurology, 57(4), 632– 638. Salthouse, T. (1976). Speed and age: multiple rate of age decline. Experimental Aging Research, 2(4), 349–359. Schmand, B., Smit, J. H., Geerlings, M. I., & Lindeboom, J. (1997). The effects of intelligence and education on the development of dementia. A test of the brain reserve hypothesis. Psychological Medicine, 27, 1337–1344. Shichita, K., Shuichi, H., Ohashi, Y., & Matuzaki, T. (1986). Memory changes in the benton visual retention test between ages 70 and 75. Journal of Gerontology, 41, 385–386. Snowdon, D. A., Kemper, S. J., Mortimer, J. A., Greiner, L. H., Wekstein, D. R., & Markesbery, W. R. (1996). Linguistic ability in early life and cognitive function and AlzheimerÕs disease in later life. Journal of the American Medical Association, 275(7), 528–532. Stankov, L. (1988). Aging, attention, and intelligence. Psychology and Aging, 3(1), 59–74. Stern, Y., Gurland, B., Tatemichi, T. K., Tang, M. X., Wilder, D., & Mayeux, R. (1994). Influence of education and occupation on the incidence of AlzheimerÕs disease. Journal of the American Medical Association, 271(13), 1004–1010.