Abstract

Normative data from the German adaptation of the Neuropsychological Assessment Battery were used to examine age-related differences in 6 executive function tasks. A multivariate analysis of variance was employed to investigate the differences in performance in 484 participants aged 18–99 years. The coefficient of variation was calculated to compare the heterogeneity of scores between 10 age groups. Analyses showed an increase in the dispersion of scores with age, varying from 7% to 289%, in all subtests. Furthermore, age-dependent heterogeneity appeared to be associated with age-dependent decline because the subtests with the greatest increase in dispersion (i.e., Mazes, Planning, and Categories) also exhibited the greatest decrease in mean scores. In contrast, scores for the subtests Letter Fluency, Word Generation, and Judgment had the lowest increase in dispersion with the lowest decrease in mean scores. Consequently, the results presented here show a pattern of age-related differences in executive functioning that is consistent with the concept of crystallized and fluid intelligence.

Introduction

There is no consensus on the definition of executive functions (EFs) (Reynolds & Horton, 2008; Salthouse, 2005; Strauss, Sherman, & Spreen, 2006); the term is rather a collective name for the higher mental processes responsible for the coordination of cognitive operations (Alvarez & Emory, 2006; Elliott, 2003; Salthouse, Atkinson, & Berish, 2003; Wecker, Kramer, Wisniewski, Delis, & Kaplan, 2000). In general, EFs are believed to represent abilities that are crucial for adapting to changing environments (Anderson, Jacobs, & Anderson, 2008; De Luca et al., 2003) and coping with novel tasks (Duncan, Burgess, & Emslie, 1995). Therefore, EFs are essential for independent living and their impairment affects all aspects of behavior. Consequently, in cases of brain impairment, the precise assessment of EFs is of crucial importance (Lezak, Howieson, Bigler, & Tranel, 2012). In addition, EFs are some of the earliest functions to deteriorate with normal aging (Bryan & Luszcz, 2000; De Luca et al., 2003); therefore, they have been investigated as potential mediators of age-related cognitive decline (Levine, Stuss, & Milberg, 1997; Parkin, 1997; Salthouse et al., 2003; Troyer, Graves, & Cullum, 2007).

In recent years, the factors that contribute to successful and unsuccessful cognitive aging have received considerable attention (see Christensen et al., 1999b; Lindenberger & Baltes, 1997; Salthouse, 2014; Schaie, 2005; Ylikoski et al., 1999). Salthouse (2014) proposed several potential moderators of changes in cognition including demographic characteristics, health, sensory ability, mood, personality, self-efficacy, and lifestyle. The majority of these factors have yet to be fully explored; the exception being demographic characteristics such as age, sex, and education, which significantly influence cognitive performance (Lezak et al., 2012; Pauls, Petermann & Lepach, 2013; Salthouse, 2014; Strauss et al., 2006). In particular, the influence of age on cognition is well studied; moreover, it is widely accepted that cognitive decline occurs with advancing age (Ardila, 2007; Salthouse, 2014). An accurate description of the general pattern of skill deterioration with age has been proposed by Salthouse (2010b): On the one hand, there are skills that are products of processes conducted in the past, such as vocabulary or general information; the ability to perform these functions usually increases until people are in their 60s. On the other hand, there are skills that result from processes that are performed at the time of assessment, which involve manipulations or transformations of abstract or familiar material; the ability to perform these functions shows an almost linear decline from early adulthood.

When investigating age-related changes in cognition, the focus of attention is often directed to the differences in the mean or other measures of central tendency. However, using methods that focus on mean-level changes, individual differences in cognitive performance may be overlooked (Nelson & Dannefer, 1992). For instance, the average performance may decline with age, but some individuals may demonstrate substantial change in cognition, whereas others may change very little (Christensen et al., 1999b). Consequently, studying the differences between individuals may provide insight into the magnitude of diversity in cognitive performance and help explain why some individuals change in cognition, whereas others do not. It should be also noted that there is a distinction in research between interindividual variability, which refers to differences in cognitive performance among individuals, and intraindividual variability, which refers to differences within a single individual, either across tests within a single session or within one task across different sessions (Christensen et al., 1999a; Reckess, Varvaris, Gordon, & Schretlen, 2014). Studies on intraindividual variability with their focus on the distribution of scores across different tasks or changes in test scores within individuals over time are more informative with regards to the individual characteristics and developmental course within one cohort, whereas studies on interindividual variability focus on the differences in performance between individuals, either within one or several cohorts and, thus, may provide information on whether there are similar deterioration patterns for particular abilities common for different individuals.

Some studies have reported that interindividual differences in cognition increase with age and that there exists a meaningful skill-related pattern, which is similar to that identified by Salthouse (2010b). For example, visual-spatial perception, attention, speed, and memory were found to be more heterogeneous in older participants than in younger participants, which was in contrast to verbal or number-related skills (Ardila, 2007; Christensen et al., 1994, 1999b; Daseking & Petermann, 2013; Morse, 1993; Ryan, Sattler, & Lopez, 2000; Wisdom, Mignogna, & Collins, 2012). These two clusters of cognitive functions are also frequently referred to as fluid (Gf) and crystallized intelligence (Gc), originally classified by Horn and Cattell (1966, 1967) and later integrated by Carroll in his Three-Stratum Model (Carroll, 1993). At present, they are included in the Cattell–Horn–Carroll theory of cognitive abilities, which is currently the most popular hierarchical model of intelligence (Kaufman, 2009; Keith & Reynolds, 2010; McGrew, 2009; Newton & McGrew, 2010).

Methods of analysis for exploring interindividual variability in cognition vary greatly.

A natural way of exploring it is to analyze the dispersion in the test scores between different age ranges (Ardila, 2007). The standard deviation (SD) as the root of variance is a measure of dispersion which is frequently used because of its independence from the unit of measurement. The comparison of standard deviations between age groups has been frequently used in the studies on the differences in interindividual variability (Rönnlund & Nilsson, 2006; Salthouse, 2010a; Schaie, 1994). However, the standard deviation should not be analyzed without taking into account the associated mean, particularly when comparing the dispersion between different populations. The coefficient of variation (CV) [(SD/mean) × 100] (Bartlett, 1946; Hendricks & Robey, 1936; Pearson, 1896; Yablokov, 1974), also known as the “percentage of the mean” (Ardila, 2007; Daseking & Petermann, 2013), is a more meaningful ratio for dispersion with regard to the heterogeneity of scores because it does not represent the standard deviation alone but rather the percentage of standard deviation in the mean (Ardila, 2007; Daseking & Petermann, 2013; Morse, 1993; Wisdom et al., 2012). The relation of the standard deviation with the mean is in so far relevant as the mean levels in scores may vary across age ranges and so change the meaning of the associated dispersions. Consequently, the CV, as the percentage of the standard deviation in the mean, allows for comparisons within and between populations (Lande, 1977), which has been already utilized in the analyses of age-related variability in cognition, particularly regarding Wechsler intelligence scales (Ardila, 2007; Daseking & Petermann, 2013; Matarazzo, 1972; Wisdom et al., 2012).

The calculation of the mean and standard deviation raw scores as well as the CV for different age groups is the first step for establishing the extent of variability of scores across age. The analysis of raw scores is a prerequisite for estimating the impact of age on cognition. Age-adjusted standard scores do not allow for meaningful comparisons between age groups in regards to the differences in performance. Only an observed decrease or increase in the mean or standard deviation raw scores may show whether there are age-related differences in cognition. However, the amount of change in the CV that occurs when comparing the age groups with the highest and lowest mean raw scores is more informative in regards to the heterogeneity in cognitive decline (Wisdom et al., 2012). In particular, the comparison between the extent of cognitive decline, as measured by the percentage decrease in the mean, and the change in the heterogeneity of scores, as measured by the percentage increase in the dispersion, may help better understand the changes in cognition over time.

Current research findings regarding the relationship between EFs and aging are inconclusive (Mejia, Pineda, Alvarez, & Ardila, 1998; Wecker et al., 2000). This may be due to the different assessment tools and age ranges that were used in the analyses. Several studies have explored the differences between limited age ranges (e.g., Boone, Miller, Lesser, Hill, & D'Elia, 1990; Brennan, Welsh, & Fisher, 1997; Raz, Gunning-Dixon, Head, Dupuis, & Acker, 1998), but only a few have investigated developmental trajectories of EFs across a life span (De Luca et al., 2003; Reynolds & Horton, 2008; Salthouse et al., 2003). Although some studies have focused on interindividual variability in executive functioning, particularly regarding the factors that contribute to successful and unsuccessful aging (Mejia et al., 1998; Ylikoski et al., 1999), there is a lack of research dealing with the differences in interindividual variability according to age. This issue should be better investigated as the magnitude of variability in performance may provide essential information for neuropsychological assessment. For instance, if there are only little differences between individuals, even small deviations from the normal performance of the standardization sample could be considered as an indication of an impairment. In contrast, if there is high variability in performance in normal population, a relatively great deviation from the mean does not necessarily indicate any pathological impairment. Therefore, the aim of the present study was to explore age-related differences in EFs in adults and to investigate interindividual variability across a large age range. It was assumed that the mean scores of EF subtests would decrease with advancing age, whereas the dispersions would increase with age. Furthermore, it was hypothesized that there is some skill-related pattern of deterioration in EF subtests; that is, the subtests related to fluid intelligence would be associated with substantial decreases in the mean scores and substantial increases in the dispersion from early adulthood. In contrast, the subtests related to crystallized intelligence would be associated with an increase in the mean scores, even in late adulthood, but only a small increase in the dispersion.

Methods

Participants

The sample consisted of 484 normal adults aged 18–99 years. Table 1 presents the demographic composition of the sample. These participants were recruited from four different sites in Germany for norming the German adaptation of the Neuropsychological Assessment Battery (NAB; see Materials for details) (Petermann, Jäncke, & Waldmann, 2016). Data collection occurred between February 2014 and February 2015, under the involvement of the authors of the current study. The data used in the present report originated from subjects who were administered the complete form 1 of NAB on the first occasion. The administration of the NAB took 3 hr on average. All participants were screened and eliminated if they had a history of known cardiovascular, neurological, or psychiatric pathology. Written informed consent was obtained from each participant prior to the administration of the battery.

Table 1.

Demographic composition of the sample

Age group Mean age ± SD N Men Women 
18–29 22.58 ± 3.17 55 26 29 
30–39 33.73 ± 2.88 49 26 23 
40–49 46.11 ± 2.22 47 22 25 
50–59 55.00 ± 2.91 58 27 31 
60–64 61.96 ± 1.54 49 22 27 
65–69 66.82 ± 1.59 50 24 26 
70–74 71.67 ± 1.48 54 27 27 
75–79 76.65 ± 1.40 46 23 23 
80–84 81.40 ± 1.40 42 13 29 
85–99 88.49 ± 3.74 35 16 19 
Age group Mean age ± SD N Men Women 
18–29 22.58 ± 3.17 55 26 29 
30–39 33.73 ± 2.88 49 26 23 
40–49 46.11 ± 2.22 47 22 25 
50–59 55.00 ± 2.91 58 27 31 
60–64 61.96 ± 1.54 49 22 27 
65–69 66.82 ± 1.59 50 24 26 
70–74 71.67 ± 1.48 54 27 27 
75–79 76.65 ± 1.40 46 23 23 
80–84 81.40 ± 1.40 42 13 29 
85–99 88.49 ± 3.74 35 16 19 

Materials

NAB is a modular battery of neuropsychological tests developed for the assessment of cognitive functions in adults with known or suspected disorders of the central nervous system (White & Stern, 2003). Both the original NAB and its German adaptation consist of a screening module and five domain-specific modules (i.e., attention, language, memory, spatial, and EFs) (Buczylowska, Bornschlegl, Daseking, Jäncke, & Petermann, 2013). Each participant was administered all six NAB modules. In this study, we were particularly interested in the Executive Functions Module; therefore, an overview of this module is provided in Table 2, and all of the information referring to particular subtests is based on the manual of the original NAB (Stern & White, 2003) and a report on the German NAB adaption (Buczylowska et al., 2013). All four original subtests of the Executive Functions Module have been translated into German and, if necessary, adjusted to the standard conditions in German-speaking countries. It should also be noted that the Executive Functions Module of the German NAB adaptation contains two additional subtests: Planning (“Planen”) is based on the “Bogenhausener Planungstest” (von Cramon, 1988; von Cramon, Matthes-von Cramon, & Mai, 1991), an experimental assessment tool particularly suitable for the diagnostics of complex planning skills in the context of daily living, whereas Letter Fluency (“Wortflüssigkeit”) was newly designed by the authors of the German NAB adaptation and is based on the scheme of a well-known and proven concept of verbal fluency (Lezak et al., 2012; Strauss et al., 2006). In addition, for the German NAB adaptation, the Judgment subtest has been shortened from original 10 to 8 items.

Table 2.

Description of the Executive Functions Module subtests of the German NAB adaptation (Petermann et al., 2016)

Subtest Description Function 
Planninga The examinee is asked to put five typical daily living time-restricted assignments into the correct order during a fixed period of time (max. 15 min) Planning, problem solving, implementing strategies, and mental flexibility 
Mazes The examinee completes seven timed paper–pencil mazes of increasing difficulty Planning, impulse control, and psychomotor speed 
Letter Fluencyb The examinee creates as many words as possible with specified initial letter during 120 s Verbal fluency and generativity 
Judgment The examinee answers eight judgment questions pertaining to daily living issues connected to home safety, health, and medicine Judgment and decisional capacity in daily living situations 
Categories The examinee is asked to generate different two-group categories based on photographs and verbal information about six people Concept formation, cognitive response set, mental flexibility, and generativity 
Word Generation Timed task in which the examinee generates three-letter words based on a visually presented group of eight letters (three vowels and five consonants) Verbal fluency and generativity 
Subtest Description Function 
Planninga The examinee is asked to put five typical daily living time-restricted assignments into the correct order during a fixed period of time (max. 15 min) Planning, problem solving, implementing strategies, and mental flexibility 
Mazes The examinee completes seven timed paper–pencil mazes of increasing difficulty Planning, impulse control, and psychomotor speed 
Letter Fluencyb The examinee creates as many words as possible with specified initial letter during 120 s Verbal fluency and generativity 
Judgment The examinee answers eight judgment questions pertaining to daily living issues connected to home safety, health, and medicine Judgment and decisional capacity in daily living situations 
Categories The examinee is asked to generate different two-group categories based on photographs and verbal information about six people Concept formation, cognitive response set, mental flexibility, and generativity 
Word Generation Timed task in which the examinee generates three-letter words based on a visually presented group of eight letters (three vowels and five consonants) Verbal fluency and generativity 

Notes:aPlanning (German “Planen”).

bLetter Fluency (German “Wortflüssigkeit”) are additional subtests added to the German NAB adaptation.

Procedure and Statistical Analysis

The raw scores of the German NAB standardization sample were used to compare 10 age groups (18–29, 30–39, 40–49, 50–59, 60–64, 65–69, 70–74, 75–79, 80–84, and 85–99) on all six subtests of the Executive Functions Module. For each age group, the mean and standard deviation were calculated. In the next step, the CV [(SD/mean) × 100] was calculated. To assess the extent of decline as well as the variability in scores for each subtest over time, both the percentage decrease in the mean and the percentage increase in the dispersion [(amount of change/highest mean or associated CV) × 100] were calculated. Statistical analysis was performed using SPSS (version 22). A two-way multivariate analysis of variance (MANOVA) was applied to all subtests to calculate the differences in performance according to age group and gender, and Bonferroni-type post hoc comparisons were also conducted. Prior to MANOVA, the index “variance inflation factor” (VIF) was calculated to check for the multicollinearity among the six subtests. The resultant low values of the VIF (<2) excluded the risk of multicollinearity (O'brien, 2007) and so permitted to conduct a MANOVA.

Results

Analysis of variance revealed a highly significant main effect of age for all subtests. The highest effect sizes were for Mazes, F(9, 465) = 65.34, p < .001, η2 = 0.53, and Categories, F(9, 465) = 27.25, p < .001, η2 = 0.33, followed by Planning, F(9, 465) = 12.73, p < .001, η2 = 0.19, Judgment, F(9, 465) = 10.23, p < .001, η2 = 0.16, Letter Fluency, F(9, 465) = 5.16, p < .001, η2 = 0.09, and Word Generation, F(9, 465) = 3.71, p < .001, η2 = 0.07.

Gender significantly influenced the score on Mazes, F(1, 465) = 16.54, p < .001, η2 = 0.03, men outperformed women, and Letter Fluency, F(1, 465) = 17.59, p < .001, η2 = 0.04, women outperformed men, although the effect sizes were small. There was no significant age group-by-gender interaction across all six subtests.

Table 3 presents descriptive statistics for all 6 subtests and all 10 age groups on which further descriptive analyses (Tables 4 and 5), including variations in the mean and in the CV, are based.

Table 3.

Descriptive statistics of the NAB Executive Functions Module subtests scores

 Age group
 
18–29 30–39 40–49 50–59 60–64 65–69 70–74 75–79 80–84 85–99 
Planning 
 Mean 9.1 8.4 8.1 7.5 6.9 6.1 6.2 4.9 4.6 3.8 
SD 2.0 2.8 3.1 3.4 3.7 3.6 3.6 3.3 3.3 3.3 
 CV 22.1 33.3 38.9 45.5 53.1 58.9 58.6 66.2 71.9 86.0 
Mazes 
 Mean 21.0 19.3 16.6 14.3 12.8 9.5 8.9 7.2 4.7 4.1 
SD 4.7 5.1 5.1 5.6 5.4 5.6 5.3 5.1 3.4 3.3 
 CV 22.3 26.4 30.5 39.2 41.9 58.9 59.1 70.9 72.2 79.7 
Letter Fluency 
 Mean 16.1 15.8 18.4 18.0 16.9 15.7 15.7 14.1 12.6 12.8 
SD 5.8 5.3 6.4 6.1 6.2 5.7 5.4 5.5 4.7 5.2 
 CV 36.3 33.2 35.0 33.9 36.9 36.5 34.2 39.1 37.5 40.9 
Judgment 
 Mean 12.9 12.8 12.6 12.8 12.1 11.9 11.9 10.8 10.9 10.3 
SD 1.9 1.6 1.7 1.9 2.1 1.9 2.0 2.1 1.9 2.2 
 CV 14.8 12.7 13.5 14.4 17.1 16.3 16.9 19.8 17.8 21.2 
Categories 
 Mean 27.4 25.2 24.0 22.5 18.4 16.1 15.4 11.5 9.5 9.4 
SD 10.8 8.1 8.8 8.8 8.2 8.5 8.2 7.5 6.2 6.2 
 CV 39.6 32.2 36.7 39.0 44.3 52.8 53.5 65.4 64.5 65.5 
Word Generation 
 Mean 7.6 7.4 8.1 7.6 7.9 6.8 6.7 5.9 5.6 6.3 
SD 3.0 2.9 3.4 2.2 2.9 2.9 2.6 2.8 2.7 3.2 
 CV 39.8 39.0 42.1 29.2 37.3 42.3 38.9 47.7 47.4 49.7 
 Age group
 
18–29 30–39 40–49 50–59 60–64 65–69 70–74 75–79 80–84 85–99 
Planning 
 Mean 9.1 8.4 8.1 7.5 6.9 6.1 6.2 4.9 4.6 3.8 
SD 2.0 2.8 3.1 3.4 3.7 3.6 3.6 3.3 3.3 3.3 
 CV 22.1 33.3 38.9 45.5 53.1 58.9 58.6 66.2 71.9 86.0 
Mazes 
 Mean 21.0 19.3 16.6 14.3 12.8 9.5 8.9 7.2 4.7 4.1 
SD 4.7 5.1 5.1 5.6 5.4 5.6 5.3 5.1 3.4 3.3 
 CV 22.3 26.4 30.5 39.2 41.9 58.9 59.1 70.9 72.2 79.7 
Letter Fluency 
 Mean 16.1 15.8 18.4 18.0 16.9 15.7 15.7 14.1 12.6 12.8 
SD 5.8 5.3 6.4 6.1 6.2 5.7 5.4 5.5 4.7 5.2 
 CV 36.3 33.2 35.0 33.9 36.9 36.5 34.2 39.1 37.5 40.9 
Judgment 
 Mean 12.9 12.8 12.6 12.8 12.1 11.9 11.9 10.8 10.9 10.3 
SD 1.9 1.6 1.7 1.9 2.1 1.9 2.0 2.1 1.9 2.2 
 CV 14.8 12.7 13.5 14.4 17.1 16.3 16.9 19.8 17.8 21.2 
Categories 
 Mean 27.4 25.2 24.0 22.5 18.4 16.1 15.4 11.5 9.5 9.4 
SD 10.8 8.1 8.8 8.8 8.2 8.5 8.2 7.5 6.2 6.2 
 CV 39.6 32.2 36.7 39.0 44.3 52.8 53.5 65.4 64.5 65.5 
Word Generation 
 Mean 7.6 7.4 8.1 7.6 7.9 6.8 6.7 5.9 5.6 6.3 
SD 3.0 2.9 3.4 2.2 2.9 2.9 2.6 2.8 2.7 3.2 
 CV 39.8 39.0 42.1 29.2 37.3 42.3 38.9 47.7 47.4 49.7 

Note: CV = (SD/mean) × 100; numbers in bold represent age groups with highest and lowest scores, respectively.

Table 4.

Variation in the mean and in the coefficient of variation (CV)

Subtest Age group with the highest score (G1)
 
Age group with the lowest score (G2)
 
Variation between G1 and G2
 
Age Mean CV Age Mean CV Mean (%) CV (%) 
Planning 18–29 9.1 22.1 85–99 3.8 86.0 58 289 
Mazes 18–29 21.0 22.3 85–99 4.1 79.7 81 258 
Letter Fluency 40–49 18.4 35.0 80–84 12.6 37.5 32 
Judgment 18–29 12.9 14.8 85–99 10.3 21.2 21 43 
Categories 18–29 27.4 39.6 85–99 9.4 65.5 66 65 
Word Generation 40–49 8.1 42.1 80–84 5.6 47.4 31 12 
Subtest Age group with the highest score (G1)
 
Age group with the lowest score (G2)
 
Variation between G1 and G2
 
Age Mean CV Age Mean CV Mean (%) CV (%) 
Planning 18–29 9.1 22.1 85–99 3.8 86.0 58 289 
Mazes 18–29 21.0 22.3 85–99 4.1 79.7 81 258 
Letter Fluency 40–49 18.4 35.0 80–84 12.6 37.5 32 
Judgment 18–29 12.9 14.8 85–99 10.3 21.2 21 43 
Categories 18–29 27.4 39.6 85–99 9.4 65.5 66 65 
Word Generation 40–49 8.1 42.1 80–84 5.6 47.4 31 12 
Table 5.

Percentage changes in mean and CV scores

Subtest Mean (%) Subtest CV (%) 
Mazes 81 Planning 289 
Categories 66 Mazes 258 
Planning 58 Categories 65 
Letter Fluency 32 Judgment 43 
Word Generation 31 Word Generation 12 
Judgment 21 Letter Fluency 
Subtest Mean (%) Subtest CV (%) 
Mazes 81 Planning 289 
Categories 66 Mazes 258 
Planning 58 Categories 65 
Letter Fluency 32 Judgment 43 
Word Generation 31 Word Generation 12 
Judgment 21 Letter Fluency 

Planning

The highest scores for Planning were observed in the 18–29 age group and scores decreased 58% gradually across the entire age range until the lowest scores were observed in the 85–99 age group. Post hoc comparisons revealed significant differences in performance between the 18–29 and 60–99 age groups. There were no significant differences in the performance of the age groups consisting of participants in the 18–59 age range. The CV increased gradually with age from the youngest age group, until it was almost fourfold higher in the oldest age group than in the youngest.

Mazes

The 18–29 age group produced the highest scores in the Mazes subtest, whereas the 85–99 group produced the lowest scores, which were only 19% of those observed in the youngest group. Post hoc comparisons revealed significant differences in the performance of the 18–29 age group and all other groups, with the exception of 30- to 39-year olds who themselves showed superior performance to the entire 50–99 age range. The mean score values decreased with age from the youngest age group, whereas the dispersions gradually increased. The CV for the 85–99 age group was ∼3.5-fold higher than that of the 18- to 29-year olds.

Letter Fluency

Scores for Letter Fluency increased up to the 40–49 age group and then decreased; hence, the trend for the entire age range was an inverted U-shaped curve. The scores produced by the 80- to 84-year-old participants decreased by 32% when compared with those produced by the 40- to 49-year olds. Post hoc comparisons revealed that participants in the 40–49 and 50–59 age groups outperformed those within the 75–99 age range. There were no significant differences in performance between the age groups in the 18–74 age range or between the 18–39 age group and any other age group. The dispersions remained stable with only a 7% increase in the CV in the 80–84 group.

Judgment

For Judgment, the highest scores were observed within the 18–59 age range and then decreased 21% across the remaining age range until the lowest scores were observed in the 85–99 age group. Furthermore, post hoc comparisons revealed significant differences between the performance of participants in age groups from the 18–59 range and those in age groups from the 75–99 range. However, there were no significant differences in performance between the age groups within the entire 18–74 age range. There was neither a regular change nor stability in the dispersions across the age groups, but the CV for 85- to 99-year olds was 43% higher than that for 18- to 29-year olds.

Categories

Scores for the Categories subtest were at their highest in the 18–29 age group and they continually decreased from this age group across the entire age range. Consequently, the mean performance of the 85–99 age group decreased 66% when compared with that of the 18–29 age group. Post hoc comparisons revealed that the age groups within the 18–49 range outperformed those within the 60–99 range. There were no significant differences in performance between the age groups within the entire 18–59 range. Dispersions continually increased with age so that the CV for 85- to 99-year olds was 65% higher than that observed for 18- to 29-year olds.

Word Generation

Scores for Word Generation remained stable up to the age of 60–64 years, with the highest scores produced by 40- to 49-year-old participants. After 64 years, a decrease in the performance was observed. For example, scores from the 80–84 age group decreased 31% when compared with those from the 40–49 age group. Post hoc comparisons confirmed the differences in the performance between participants in age groups within the 18–64 range and those in the 80–84 age group, with the exception of those aged 30–39 years. There were no significant differences in performance between the age groups in the entire 18–74 range. Dispersions were relatively stable; a 12% increase in the CV was observed between the groups with the highest and the lowest mean scores.

Discussion

As expected, the results of this study showed a strong impact of age on the performance of participants in EF tasks. The data demonstrate that performance in different subtests of the Executive Functions NAB module peak and decline at different ages. The highest and lowest test scores were generally observed in the youngest and oldest age groups, respectively, with the exception of scores for Letter Fluency and Word Generation. Performance in these latter two tests reached high and low points in the 40–49 and 80–84 age groups, respectively. Thus, 85- to 99-year-old participants achieved higher scores in Letter Fluency and Word Generation than 80- to 84-year olds. This requires an explanation, even if there were no significant differences in the performance between these two age groups. This result could be explained by the selection of older people who participate in scientific studies. Participants used in the current study were administered a test battery that takes 3 hr on average, in participants >70 even almost 3.5 hr. Subjects aged ≥85 who are capable of participating in such an extensive cognitive study usually possess superior states of health and they may also possess superior states of cognition because there is a research-based assumption of the impact of health on cognitive performance (Salthouse, 2004, 2014; Zelinski & Gilewski, 2003). Furthermore, there might be individual features of the participants, which cannot be easily controlled and yet have an impact on test performance.

Although there was in each subtest a particular age group that produced the highest score, statistical analysis showed no significant differences in the mean performance between different age groups with superior performance. Furthermore, the age range of homogenous performance differed according to subtest. For Mazes, there was no difference in mean performance only between the 18–29 and 30–39 age groups. The homogenous age range for the Planning and Categories subtests was 18–59, whereas for Letter Fluency, Judgment, and Word Generation, it was 18–74 years. Thus, the first significant difference in performance between the age group with the highest score and any following age group with a lower score may be used to establish an age-related line of decline. There is a link between the extent of homogenous performance across the age groups and the extent of change in the mean, as measured by the percentage of the amount of change in the highest score. Consequently, the Planning, Categories, and Mazes subtests showed the largest age-related decline, with a decrease in the mean ranging from 58% to 81%, followed by Judgment, Word Generation, and Letter Fluency, which showed a moderate decline ranging from 21% to 32%.

Heterogeneity in EFs across age was also explored in the present research. The CV was used as a ratio of dispersion, and the percentage of the amount of change in the highest score was further analyzed to establish the extent of variability across age groups. An increase in dispersion was observed in all subtests, which varied from 7% to 12% for Letter Fluency and Word Generation, 43% to 65% for Judgment and Categories, and 258% to 289% for Mazes and Planning. Similar to the previous research on developmental patterns in intellectual skills (Ardila, 2007; Daseking & Petermann, 2013; Wisdom et al., 2012), age-dependent variability appears to be associated with age-dependent decline because the subtests with the greatest increase in dispersion also exhibited the greatest decrease in mean scores, whereas those with the lowest increase in dispersion exhibited the lowest decrease in mean scores. Furthermore, the language-related Letter Fluency and Word Generation subtests, which had the longest maturation paths and the latest onset of decline, were those that exhibited the smallest decrease in mean scores and the lowest increase in dispersion. Judgment, a subtest that is also partly affected by verbal abilities, followed a similar developmental trajectory. The most heterogeneous pattern of development was observed in Mazes, a visual and highly speed-related task with a short maturation path and rapid decline. These findings are not surprising because the pattern of development in language-related measures versus those underpinned by perceptual-motor abilities is consistent with prior research on intellectual skills (Ardila, 2007; Daseking & Petermann, 2013; Morse, 1993; Ryan et al., 2000; Wisdom et al., 2012) and EFs (Reynolds & Horton, 2008). Hence, the observed results are consistent with the concept of crystallized and fluid intelligence. However, Categories and Planning cannot be clearly assigned to one of these clusters, and this may be due to their predominantly executive component. These tasks require similar skills such as mental flexibility, strategy implementation, and problem solving (Stern & White, 2003; von Cramon et al., 1991). Both subtests also appear to share a similar developmental trajectory with peak performances at 18–29 years and onset of decline, which occurs at >60 years of age with a similar extent of deterioration; however, these subtests differ substantially in the magnitude of variability, with the dispersion in Planning being twofold higher than that of Categories. One plausible explanation for this finding could be the multifactorial nature of these subtests. Although they tap similar abilities, they appear to differ in their demand for attentional capacity. Planning, being a paper–pencil test, tends to reflect its substantial attentional component with the greater variability of scores observed as age advances.

In conclusion, the presented findings highlight potential implications for the interpretation of neuropsychological assessment outcomes and must be taken into account in relation to the development of new EF measures. First, the multifactorial nature of most of the existing tests on EFs (Stuss & Alexander, 2000) should be considered. Expectations with regard to the performance on a particular measure should be adjusted to its underlying skills because the extent of age dependence differs according to the measured function. Moreover, when norming EFs measures, the developmental trajectories of particular abilities must be taken into account; that is, separate norms should be provided for age groups that differ greatly in performance on particular tasks. For example, if changes occur rapidly, tasks should be normed in separate small age ranges, according to the pattern of deterioration. Second, questions remain regarding whether certain types of executive task are more appropriate for detecting age-related cognitive decline than others. It is apparent that functions with a pronounced crystallized component are less sensitive to age-dependent deterioration than those linked to fluid intelligence. This finding might be useful in the assessment of mild, subclinical executive dysfunction in normal adults, which is difficult to detect compared with dysfunctions related to frontal lobe lesions (Bryan & Luszcz, 2000). Third, information about the approximate age at which cognitive decline begins for different cognitive functions could be used to help prevent or reverse age-related changes, for example, by determining the optimum time for implementing interventions (Salthouse, 2009). Fourth, high variability in performance in healthy adults may also change the interpretation guidelines for assessment outcomes in brain-damaged patients; for example, an outcome rated beyond the limits of normal performance may not necessarily mean a lesion-related impairment; it might rather reflect an age-related deterioration pattern. Consequently, base rates for impaired performance on EFs measures in normal population should be established to help enhance clinical diagnosis. Future research should also focus on exploring interindividual variability at different age ranges than those used in the current study. In particular, small age ranges with a sufficient amount of participants across the life span should be investigated. Furthermore, tasks from different EFs measures must be used to better explore age-related differences and the magnitude of variability within executive functioning; especially, the contribution of non-executive skills such as language, psychomotor speed, attention, and memory should be examined.

Limitations

Some caution is warranted with regards to the findings in the present report. Comparing scores at different age ranges within cross-sectional analyses tends to be confounded by cohort influences. As an example, differences exist in the educational level between younger and older participants that may affect cognitive performance. Longitudinal studies are considered to be more informative in terms of cognitive change because they allow within-person comparisons in performance between different occasions (Schaie, 2005). There are discrepancies between cross-sectional and longitudinal age trends. Between-person comparisons usually show gradual declines from early adulthood, whereas within-person comparisons reveal stability or an increase in the performance (Rönnlund & Nilsson, 2006; Salthouse, 2009). One plausible explanation is prior test experience in longitudinal comparisons, which brings into question the purity of this type of study as a measure of cognitive change. Nevertheless, longitudinal data with sensitive measures of change should be used to confirm whether interindividual variability exists in cognitive change and to explore whether only some individuals show decline, whereas others remain stable (Salthouse, 2010b). For a better understanding of age-related trends in cross-sectional and longitudinal comparisons, further investigation of practice and cohort effects will be required.

Conflict of Interest

None declared.

References

Alvarez
J. A.
,
Emory
E.
(
2006
).
Executive function and the frontal lobes: A meta-analytic review
.
Neuropsychological Review
 ,
16
,
17
42
.
Anderson
V.
,
Jacobs
R.
,
Anderson
P.
(Eds.). (
2008
).
Executive functions and the frontal lobes: A lifespan perspective
 .
New York
:
Taylor & Francis
.
Ardila
A.
(
2007
).
Normal aging increases cognitive heterogeneity: Analysis of dispersion in WAIS-III scores across age
.
Archives of Clinical Neuropsychology
 ,
22
,
1003
1011
.
Bartlett
M. S.
(
1946
).
The coefficient of variation
.
Nature
 ,
158
,
520
521
.
Boone
K. B.
,
Miller
B. L.
,
Lesser
I. M.
,
Hill
E.
,
D'Elia
L.
(
1990
).
Performance on frontal lobe tests in healthy, older individuals
.
Developmental Neuropsychology
 ,
6
,
215
223
.
Brennan
M.
,
Welsh
M. C.
,
Fisher
C. B.
(
1997
).
Aging and executive function skills: An examination of a community-dwelling older adult population
.
Perceptual and Motor Skills
 ,
84
,
1187
1197
.
Bryan
J.
,
Luszcz
M. A.
(
2000
).
Measurement of executive function: Considerations for detecting adult age differences
.
Journal of Clinical and Experimental Neuropsychology
 ,
22
,
40
55
.
Buczylowska
D.
,
Bornschlegl
M.
,
Daseking
M.
,
Jäncke
L.
,
Petermann
F.
(
2013
).
Zur deutschen Adaptation der Neuropsychological Assessment Battery (NAB) [German adaption of the Neuropsychological Assessment Battery (NAB)]
.
Zeitschrift für Neuropsychologie
 ,
24
,
217
227
.
Carroll
J. B.
(
1993
).
Human cognitive abilities: A survey of factor-analytic studies
 .
Cambridge
:
Cambridge University Press
.
Christensen
H.
,
Mackinnon
A.
,
Jorm
A. F.
,
Henderson
A. S.
,
Scott
L. R.
,
Korten
A. E.
(
1994
).
Age-differences and iterindividual variation in cognition in community-dwelling elderly
.
Psychology and Aging
 ,
9
,
381
390
.
Christensen
H.
,
Mackinnon
A. J.
,
Korten
A. E.
,
Jorm
A. F.
,
Henderson
A. S.
,
Jacomb
P.
(
1999a
).
Dispersion in cognitive ability as a function of age: A longitudinal study of an elderly community sample
.
Aging, Neuropsychology, and Cognition (Neuropsychology, Development and Cognition: Section B)
 ,
6
,
214
228
.
Christensen
H.
,
Mackinnon
A. J.
,
Korten
A. E.
,
Jorm
A. F.
,
Henderson
A. S.
,
Jacomb
P.
et al
. (
1999b
).
An analysis of diversity in the cognitive performance of elderly community dwellers: Individual differences in change scores as a function of age
.
Psychology and Aging
 ,
14
,
365
379
.
Daseking
M.
,
Petermann
F.
(
2013
).
Analyse von Querschnittsdaten zur Intelligenzentwicklung im Erwachsenenalter: Eine Studie zur deutschsprachigen Version der WAIS-IV [Cross-sectional investigation of the German Wechsler Adult Intelligence Scale—Fourth Edition (WAIS-IV).]
.
Zeitschrift für Neuropsychologie
 ,
24
,
149
160
.
De Luca
C. R.
,
Wood
S. J.
,
Anderson
V.
,
Buchanan
J. A.
,
Proffitt
T. M.
,
Mahony
K.
et al
. (
2003
).
Normative data from the CANTAB. I: Development of executive function over the lifespan
.
Journal of Clinical and Experimental Neuropsychology
 ,
25
,
242
254
.
Duncan
J.
,
Burgess
P. W.
,
Emslie
H.
(
1995
).
Fluid intelligence after frontal lobe lesions
.
Neuropsychologia
 ,
33
,
261
268
.
Elliott
R.
(
2003
).
Executive functions and their disorders
.
British Medical Bulletin
 ,
65
,
49
59
.
Hendricks
W. A.
,
Robey
K. W.
(
1936
).
The sampling distribution of the coefficient of variation
.
Annals of Mathematical Statistics
 ,
7
,
129
144
.
Horn
J. L.
,
Cattell
R. B.
(
1966
).
Refinement test of theory of fluid and crystallized general intelligences
.
Journal of Educational Psychology
 ,
57
,
253
270
.
Horn
J. L.
,
Cattell
R. B.
(
1967
).
Age differences in fluid and crystallized intelligence
.
Acta Psychologica
 ,
26
,
107
129
.
Kaufman
A. S.
(
2009
).
IQ testing 101
 .
New York
:
Springer Publishing
.
Keith
T. Z.
,
Reynolds
M. R.
(
2010
).
Cattell-Horn-Carroll abilities and cognitive tests: What we've learned from 20 years of research
.
Psychology in the Schools
 ,
47
,
635
650
.
Lande
R.
(
1977
).
On comparing coefficients of variation
.
Systematic Zoology
 ,
26
,
214
217
.
Levine
B.
,
Stuss
D. T.
,
Milberg
W. P.
(
1997
).
Effects of aging on conditional associative learning: Process analyses and comparison with focal frontal lesions
.
Neuropsychology
 ,
11
,
367
381
.
Lezak
M. D.
,
Howieson
D. B.
,
Bigler
E. D.
,
Tranel
D.
(
2012
).
Neuropsychological assessment
  (
5th ed.
).
Oxford
:
Oxford University Press
.
Lindenberger
U.
,
Baltes
P. B.
(
1997
).
Intellectual functioning in old and very old age: Cross-sectional results from the Berlin Aging Study
.
Psychology and Aging
 ,
12
,
410
432
.
Matarazzo
J. D.
(
1972
).
Wechsler's measurement and appraisal of adult intelligence: Fifth and enlarged edition
 .
New York
:
Oxford University Press
.
McGrew
K. S.
(
2009
).
CHC theory and the human cognitive abilities project: Standing on the shoulders of the giants of psychometric intelligence research
.
Intelligence
 ,
37
,
1
10
.
Mejia
S.
,
Pineda
D.
,
Alvarez
L. M.
,
Ardila
A.
(
1998
).
Individual differences in memory and executive function abilities during normal aging
.
International Journal of Neuroscience
 ,
95
,
271
284
.
Morse
C. K.
(
1993
).
Does variability increase with age—An archival study of cognitive measures
.
Psychology and Aging
 ,
8
,
156
164
.
Nelson
E. A.
,
Dannefer
D.
(
1992
).
Aged heterogeneity—Fact or fiction—The fate of diversity in gerontological research
.
Gerontologist
 ,
32
,
17
23
.
Newton
J. H.
,
McGrew
K. S.
(
2010
).
Introduction to the special issue: Current research in Cattell-Horn-Carroll-based assessment
.
Psychology in the Schools
 ,
47
,
621
634
.
O'brien
R. M.
(
2007
).
A caution regarding rules of thumb for variance inflation factors
.
Quality & Quantity
 ,
41
,
673
690
.
Parkin
A. J.
(
1997
).
Normal age-related memory loss and its relation to frontal lobe dysfunction
. In
Rabbit
P.
(Ed.),
Methodology of frontal and executive function
  (pp.
177
190
).
Hove, UK
:
Psychology Press
.
Pauls
F.
,
Petermann
F.
,
Lepach
A. C.
(
2013
).
Gender differences in episodic memory and visual working memory including the effects of age
.
Memory
 ,
21
,
857
874
.
Pearson
K.
(
1896
).
Mathematical contributions to the theory of evolution. III. Regression, heredity, and panmixia
 .
London
:
Royal Society of London
.
Petermann
F.
,
Jäncke
L.
,
Waldmann
H. C.
(Eds.). (
2016
).
The Neuropsychological Assessment Battery (NAB). German adaptation
 .
Bern
:
Hogrefe
.
Raz
N.
,
Gunning-Dixon
F. M.
,
Head
D.
,
Dupuis
J. H.
,
Acker
J. D.
(
1998
).
Neuroanatomical correlates of cognitive aging: Evidence from structural magnetic resonance imaging
.
Neuropsychology
 ,
12
,
95
114
.
Reckess
G. Z.
,
Varvaris
M.
,
Gordon
B.
,
Schretlen
D. J.
(
2014
).
Within-person distributions of neuropsychological test scores as a function of dementia severity
.
Neuropsychology
 ,
28
,
254
260
.
Reynolds
C. R.
,
Horton
A. M.
Jr.
(
2008
).
Assessing executive functions: A life-span perspective
.
Psychology in the Schools
 ,
45
,
875
892
.
Rönnlund
M.
,
Nilsson
L.-G.
(
2006
).
Adult life-span patterns in WAIS-R Block Design performance: Cross-sectional versus longitudinal age gradients and relations to demographic factors
.
Intelligence
 ,
34
,
63
78
.
Ryan
J. J.
,
Sattler
J. M.
,
Lopez
S. J.
(
2000
).
Age effects on Wechsler Adult Intelligence Scale-III subtests
.
Archives of Clinical Neuropsychology
 ,
15
,
311
317
.
Salthouse
T. A.
(
2004
).
Localizing age-related individual differences in a hierarchical structure
.
Intelligence
 ,
32
.
Salthouse
T. A.
(
2005
).
Relations between cognitive abilities and measures of executive functioning
.
Neuropsychology
 ,
19
,
532
545
.
Salthouse
T. A.
(
2009
).
When does age-related cognitive decline begin?
Neurobiology of Aging
 ,
30
,
507
514
.
Salthouse
T. A.
(
2010a
).
Major issues in cognitive aging
 .
New York
:
Oxford University Press
.
Salthouse
T. A.
(
2010b
).
Selective review of cognitive aging
.
Journal of International Neuropsychological Society
 ,
16
,
754
760
.
Salthouse
T. A.
(
2014
).
Correlates of cognitive change
.
Journal of Experimental Psychology: General
 ,
143
,
1026
1048
.
Salthouse
T. A.
,
Atkinson
T. M.
,
Berish
D. E.
(
2003
).
Executive functioning as a potential mediator of age-related cognitive decline in normal adults
.
Journal of Experimental Psychology: General
 ,
132
,
566
594
.
Schaie
K. W.
(
1994
).
The course of adult intellectual-development
.
American Psychologist
 ,
49
,
304
313
.
Schaie
K. W.
(
2005
).
What can we learn from longitudinal studies of adult development?
Research in Human Development
 ,
2
,
133
158
.
Stern
R. A.
,
White
T.
(
2003
).
The Neuropsychological Assessment Battery (NAB). Administration, scoring, and interpretation manual
 .
Lutz, FL
:
Psychological Assessment Resources
.
Strauss
E.
,
Sherman
E. M. S.
,
Spreen
O.
(
2006
).
A compendium of neuropsychological tests: Administration, norms, and commentary
  (
3rd ed.
).
Oxford
:
Oxford University Press
.
Stuss
D. T.
,
Alexander
M. P.
(
2000
).
Executive functions and the frontal lobes: A conceptual view
.
Psychological Research
 ,
63
,
289
298
.
Troyer
A. K.
,
Graves
R. E.
,
Cullum
C. M.
(
2007
).
Executive functioning as a mediator of the relationship between age and episodic memory in healthy aging
.
Aging, Neuropsychology, and Cognition
 ,
1
,
45
53
.
von Cramon
D. Y.
(
1988
).
Planen und Handeln
. In
von Cramon
D. Y.
,
Zihl
J.
(Eds.),
Neuropsychologische Rehabilitation. Grundlagen-Diagnostik-Behandlungsverfahren
 
[Neuropsychological rehabilitation. Basics-diagnostics-treatment methods]
(pp.
248
263
).
Berlin
:
Springer
.
von Cramon
D. Y.
,
Matthes-von Cramon
G.
,
Mai
N.
(
1991
).
Problem-solving deficits in brain-injured patients: A therapeutic approach
.
Neuropsychological Rehabilitation
 ,
1
,
45
64
.
Wecker
N. S.
,
Kramer
J. H.
,
Wisniewski
A.
,
Delis
D. C.
,
Kaplan
E.
(
2000
).
Age effects on executive ability
.
Neuropsychology
 ,
14
,
409
414
.
White
T.
,
Stern
R. A.
(
2003
).
The Neuropsychological Assessment Battery (NAB). Psychometric and technical manual
 .
Lutz, FL
:
Psychological Assessment Resources
.
Wisdom
N. M.
,
Mignogna
J.
,
Collins
R. L.
(
2012
).
Variability in Wechsler Adult Intelligence Scale-IV subtest performance across age
.
Archives of Clinical Neuropsychology
 ,
27
,
389
397
.
Yablokov
A. V.
(
1974
).
Variability of mammals
 .
New Delhi
:
Amerind Publishing
.
Ylikoski
R.
,
Ylikoski
A.
,
Keskivaara
P.
,
Tilvis
R.
,
Sulkava
R.
,
Erkinjuntti
T.
(
1999
).
Heterogeneity of cognitive profiles in aging: Successful aging, normal aging, and individuals at risk for cognitive decline
.
European Journal of Neurology
 ,
6
,
645
652
.
Zelinski
E. M.
,
Gilewski
M. J.
(
2003
).
Effects of demographic and health variables on Rasch scaled cognitive scores
.
Journal of Aging and Health
 ,
15
,
435
464
.