Abstract
In this article, we examined memory for pictures and words in middle-age (45-59 years), young-old (60-74 years), old-old (75-89 years) and the oldest-old adults (90-97 years) in the Louisiana Healthy Aging Study. Stimulus items were presented and retention was tested in a blocked order where half of the participants studied 16 simple line drawings and the other half studied matching words during acquisition. Free recall and recognition followed. In the next acquisition/test block, a new set of items was used where the stimulus format was changed relative to the first block. Results yielded pictorial superiority effects in both retention measures for all age groups. Follow-up analyses of clustering in free recall revealed a greater number of categories were accessed (which reflects participants' retrieval plan) and more items were recalled per category (which reflects participants' encoding strategy) when pictures served as stimuli compared to words. Cognitive status and working memory span were correlated with picture and word recall. Regression analyses confirmed that these individual difference variables accounted for significant age-related variance in recall. These data strongly suggest that the oldest-old can utilize nonverbal memory codes to support long-term retention as effectively as do younger adults.
Keywords: Episodic memory, healthy aging, oldest-old, picture superiority effect, verbal recall
Age-related declines in episodic memory are well documented in the cognitive aging literature (for reviews, Craik, 2000; Zacks, Hasher, & Li, 2000). Much of the prior work on memory aging consists of studies where younger (20s to 30s) and older adults (60s to 70s) were compared on various episodic memory tasks. In recent years, measures of episodic memory have been included in multidisciplinary studies of the oldest-old that assess cognitive, psychosocial, and physiologic variables (see Bäckman, Small, Wahlin, & Larsson, 2000, for review). Prior work confirms that the oldest-old benefit from task-relevant memory support in the form of prior knowledge (e.g., Wahlin et al., 1993) and retrieval cues (e.g., Bäckman & Wahlin, 1995). Other evidence has shown that demographic and life style variables influence the extent to which older adults benefit from cognitive support to enhance episodic memory (Hill, Wahlin, Winblad & Bäckman, 1995). These studies highlight the critical role of individual difference and task variables in episodic memory functioning in late life, consistent with a contextualist view of episodic memory (Bäckman, Mantyla, & Herlitz, 1990).
One issue that has received scant attention in studies of episodic memory in late adulthood is whether the oldest-old demonstrate pictorial superiority effects in retention. The pictorial superiority effect (PSE) refers to the finding that concrete items are better remembered when presented in a pictorial format than in a verbal format. Paivio (1971) advanced the dual-coding theory to explain the memorial advantage of pictures relative to their verbal referents. This theory holds that pictures can be dually represented in memory by visual and verbal codes, whereas words are represented primarily by verbal codes. Pictures are better remembered than words, on the assumption that two codes are better than one. An alternative explanation for the memorial efficacy of pictures is that the sensory codes for pictures are richer than those of words, leading to a more differentiated representation that is less susceptible to interference (Nelson, Reed, & Walling, 1976). It has also been suggested that pictures benefit from greater conceptual processing than do words (Stenberg, 2006). Whether code redundancy, a richer sensory representation of items in picture format, or enhanced conceptual processing of pictures accounts for the pictorial superiority effect is a matter of theoretical interest and debate (Mintzer & Snodgrass, 1999; Paivio, 1991). Differing theoretical perspectives notwithstanding, the PSE is widely recognized as a reliable and robust episodic memory phenomenon. Accordingly, researchers have used pictures and words as memory stimuli in many studies of implicit and explicit retrieval processes (e.g., Weldon & Roediger, 1987), false recognition (e.g., Dodson & Schacter, 2002) and in more recent neuropsychological studies using brain imaging techniques (e.g., Springer, McIntosh, Winocur, & Grady, 2005). The finding that pictures are better remembered than words is also a matter of practical importance, as pictures may be useful in educational contexts to aid in retention of written material (Cherry, Park, Frieske, & Smith, 1996).
Ample experimental evidence documents the reliability and generality of the PSE across numerous subject populations and memory measures. For instance, PSEs have been found for children and adolescents (Whitehouse, Maybery, & Durkin, 2006) and for college students in free recall (e.g., Toglia, Hinman, Dayton, Catalano, 1997), paired associate recall (e.g., Nelson et al., 1976), and recognition (e.g., Snodgrass & Asiaghi, 1977). Persons with mental retardation show a PSE, as do their normal intelligence counterparts (Cherry, Applegate, & Reese, 2002). Other evidence indicates that healthy older adults demonstrate PSEs as do younger adults in free recall (Rissenberg & Glazner, 1986; Winograd, Smith, & Simon, 1982) and recognition (Park, Puglisi, & Sovacool, 1983). To our knowledge, the present study is the first to examine evidence for pictorial superiority in the oldest-old, defined here as individuals age 90 years and older. Whether nonagenarians show a pictorial superiority advantage in long-term retention is an important issue, having both theoretical and practical implications. From a theoretical perspective, evidence of a PSE would suggest that the oldest-old form and utilize nonverbal (i.e., imaginal/visuospatial) memory codes as well as their younger counterparts do to support long-term retention. Evidence of a pictorial advantage in memory would also suggest a direction for the development of mnemonic aids to improve everyday retention in very old adults.
In this article, we examine episodic memory from a contextualist perspective, focusing on the role of individual differences, stimulus format, and retrieval task characteristics (Bäckman et al., 1990). To ensure an adequate representation of the age variable, the experimental design included comparison groups of middle age adults (ages 45-59 years), young-old adults (ages 60-74 years), old-old adults (ages 75-89 years) and the oldest-old adults (age 90 years and over). Participants were enrolled in the Louisiana Healthy Aging Study (LHAS), a multidisciplinary study of the determinants of human longevity and healthy aging. Our primary objective was to determine whether the oldest-old show a pictorial superiority effect in measures of longterm retention as do their younger counterparts. Participants studied and later remembered simple, black and white line drawings and their matching verbal referents from the Snodgrass and Vanderwart (1980) corpus. Free recall and recognition were tested. We expected that the oldest-old would perform more poorly than their younger counterparts on these long-term memory measures, consistent with prior research on episodic memory in late life. Further, the oldest-old should show a pictorial superiority advantage in free recall and recognition, in line with earlier studies with young-old adults (e.g., Park et al., 1983; Winograd et al., 1982). Based on expectations issuing from a contextualist perspective (Bäckman et al., 1990; Craik et al., 1987), we hypothesized that the oldest-old would benefit more than their younger counterparts from pictures as memory stimuli coupled with recognition as the retrieval task. Such a pattern of outcomes would confirm the reliability and generality of the pictorial superiority effect as well as demonstrate the context sensitivity of episodic remembering well into the ninetieth decade of life.
Our second objective was to examine the contributions of cognitive status and working memory capacity to episodic memory performance. The Mini Mental State Exam (MMSE, Folstein, Folstein, & McHugh, 1975) was selected as an index of cognitive status in the present research based on several considerations, including adequate psychometric properties (Tombaugh & McIntyre, 1992), frequency of use in other large scale studies of cognition in very old adults (cf. Luszcz, Bryan, & Kent, 1997; Wahlin et al., 1993) and ease of administration. The MMSE also offers a broad glimpse of cognitive functionality in that is comprised of 7 subscales (orientation, registration, attention and calculation, recall, language, repetition and comprehension). Prior research has shown that MMSE scores of 25 and above are predictive of episodic memory in very old age (e.g., Wahlin et al., 1993), implying that mild deficits in cognitive status may contribute to episodic memory impairment. We included the MMSE as a predictor of memory performance to shed further light on this issue. With respect to working memory as an individual difference variable, the digit span tests from the WAIS (Wechsler, 1981) and the size judgment span task (SJS, Cherry, Elliott & Reese, 2007) were included as indices of short-term retention and working memory capacity. The SJS has been shown to account for greater amounts of age-related variance than the backward digit span test in several different episodic memory tasks across a variety of subject groups (Cherry et al., 2007). As a result, SJS was selected to represent the working memory construct in this study. Regression analyses were conducted to determine the amount of variance accounted for by MMSE and SJS in word and picture free recall as criterion measures of episodic memory. We also examined the proportion of age-related variance that could be accounted for by SJS and MMSE to clarify the contribution of these individual ability differences to episodic memory performance in late adulthood.
Method
Participants
A total of 160 individuals participated in this study. The four age groups included middle-age adults (M = 51.7 years, SD = 3.5, age range 45 to 58 years), young-old adults (M = 69.7 years, SD = 3.2, age range 61 to 74 years); old-old adults (M = 80.5 years, SD = 3.8, age range 75 to 89 years); and the oldest-old adults (M = 92.4 years, SD = 1.8, age range 90 to 97 years). All were enrolled in the LHAS, a multidisciplinary study of the determinants of longevity conducted in collaboration with LSU Health Sciences Center in New Orleans, the University of Alabama at Birmingham, and the Pennington Biomedical Research Center in Baton Rouge, LA. LHAS participants are sampled randomly from the Voters Registration 2000 files for those age 20 to 64 years old and from the Medicare Beneficiary Enrollment Data file of the Center for Medicare and Medicaid Services (CMS) for those age 65 years older and above for the eight parishes (counties) constituting the Greater Baton Rouge community. Corrected binocular visual acuity was assessed with a standard Snellen eye chart. All participants in this study were visually capable as and free of neurological impairment due to stroke or adult dementia. All scored at least a 25 or higher on the Mini-Mental State Exam (MMSE; Folstein et al., 1975). Table 1 presents a summary of the individual difference and self-reported health characteristics of the sample.
Table 1.
Summary of Individual Difference Characteristics
| Age group |
|||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Middle-age (n = 40) |
Young-old (n = 40) |
Old-old (n = 40) |
Oldest-old (n = 40) |
||||||||||
| % | M | SD | % | M | SD | % | M | SD | %> | M | SD | ||
| MMSEa | 29.7 | 0.69 | 29.1 | 0.82 | 28.5 | 1.28 | 27.8 | 1.52 | |||||
| Vocabularyb | 25.2 | 7.7 | 25.4 | 7.1 | 24.4 | 7.7 | 22.4 | 6.2 | |||||
| GDSc | 1.60 | 2.16 | 1.45 | 1.78 | 1.60 | 1.58 | 2.62 | 2.39 | |||||
| Years of Education | |||||||||||||
| < 7th grade | 0.00 | 0.00 | 0.00 | 2.50 | |||||||||
| 7th to 9th grade | 0.00 | 0.00 | 5.00 | 7.50 | |||||||||
| 10th to 11th grade | 0.00 | 2.50 | 7.50 | 10.50 | |||||||||
| High school | 37.50 | 40.00 | 17.50 | 22.50 | |||||||||
| Partial college or training | 30.00 | 25.00 | 32.50 | 25.00 | |||||||||
| College degree | 17.50 | 15.00 | 27.50 | 25.00 | |||||||||
| Graduate degree | 15.00 | 17.50 | 10.00 | 7.50 | |||||||||
| Health at the present time | |||||||||||||
| Excellent | 30.00 | 17.50 | 10.00 | 12.82 | |||||||||
| Good | 55.00 | 70.00 | 72.50 | 61.54 | |||||||||
| Fair | 12.50 | 12.50 | 17.50 | 25.64 | |||||||||
| Poor | 2.50 | 0.00 | 0.00 | 0.00 | |||||||||
| Health prevents activities | |||||||||||||
| A great deal | 5.13 | 10.00 | 7.50 | 23.08 | |||||||||
| A little/some | 39.90 | 30.00 | 60.00 | 46.15 | |||||||||
| Not at all | 58.97 | 60.00 | 32.50 | 30.77 | |||||||||
| Health compared to others | |||||||||||||
| Better than | 50.00 | 55.00 | 62.50 | 97.44 | |||||||||
| The same as | 45.00 | 42.50 | 37.50 | 2.56 | |||||||||
| Worse | 5.00 | 2.50 | 0.00 | 0.00 | |||||||||
| Number of clubs and social organizations | |||||||||||||
| None | 25.00 | 5.00 | 2.50 | 10.26 | |||||||||
| Between 1-3 | 65.00 | 77.50 | 65.00 | 72.92 | |||||||||
| Between 4-6 | 5.00 | 12.50 | 25.00 | 10.26 | |||||||||
| More than 6 | 5.00 | 5.00 | 7.50 | 2.56 | |||||||||
| Number of hours per week spent outside of home | |||||||||||||
| None | 0.00 | 0.00 | 2.50 | 2.56 | |||||||||
| Between 1-5 | 0.00 | 20.00 | 17.50 | 38.46 | |||||||||
| Between 6-12 | 7.50 | 30.00 | 20.00 | 15.38 | |||||||||
| Between 13-19 | 10.00 | 20.00 | 20.00 | 12.82 | |||||||||
| More than 19 | 82.50 | 30.00 | 40.00 | 30.77 | |||||||||
| Social support | |||||||||||||
| Very satisfied | 45.00 | 77.50 | 75.00 | 87.18 | |||||||||
| Fairly satisfied | 40.00 | 17.50 | 22.50 | 7.69 | |||||||||
| A little satisfied | 15.00 | 5.00 | 0.00 | 5.13 | |||||||||
| Not satisfied | 0.00 | 0.00 | 2.50 | 0.00 | |||||||||
| Confidant | 97.50 | 95.00 | 87.50 | 85.00 | |||||||||
Mini-Mental State Exam (Folstein, Folstein, & McHugh, 1975).
Vocabulary (Jastak & Jastak, 1965).
Geriatric Depression Scale (Sheikh & Yesavage, 1986).
We conducted one-way analyses of variance (ANOVAs) and chi square tests of independence (when indicated) on the individual difference data with age group as a between group factor. An ANOVA on the MMSE scores yielded a significant age group effect, F (3, 156) = 20.42, MSE = 1.28, p < 0.001. Pairwise comparisons (Tukey) confirmed that the oldest-old adults' mean MMSE score was lower than the three comparison groups' scores (p < 0.05 for each comparison, see Table 1). A short-form of the Wechsler Adult Intelligence Scale Vocabulary subtest (Jastak & Jastak, 1965) was given as a measure of verbal intelligence. Analyses of the vocabulary scores yielded a non-significant age group effect (p = 0.23). Participant's scores on the Geriatric Depression Scale (GDS; Sheikh & Yesavage, 1986) indicated that the majority (93%) were within the normal range at the time of testing. An ANOVA on the GDS scores yielded a significant age group effect, F(3, 155) = 2.83, p = 0.04, favoring the oldest-old adults whose mean GDS score was numerically higher than the other groups, yet well below the cutoff score of 6 representing mild depression.
Participants rated their level of education on a 7-point scale (see Table 1). A χ2 test of independence revealed no significant association between educational attainment and age group (p = 0.2). Participants' responses to a demographic questionnaire that contained a subset of self-perceived health questions from the Older American Resources and Services Multidimensional Functional Assessment Questionnaire (Duke University Center for the Study of Aging and Human Development, 1975) indicated that most were generally in good health. Analyses of the ratings for health at the present time and age group yielded a non-significant association by a χ2 test (p = 0.21). For health prevents activities, the χ2 test for independence related to age groups showed a significant association (p = .006), owing to the oldest-old adults who rated their health as standing in the way of doing things they want to do more often than the other age groups. The χ2 test for independence of health compared to others and age groups showed a significant association (p = 0.0002). The oldest-old adults rated their health as better than their age mates more often than did the younger groups. Analyses of the social activity ratings measured by the number of clubs and social organizations yielded a significant association to age groups by χ2 test (p = 0.02) due to the middle-age group who reported fewer clubs and social organizations compared to their older counterpart groups. Analyses of the number of hours per week spent outside of the home also yielded a significant association to age groups by χ2 test (p < 0.001), favoring the middle-age group who reported more hours per week spent outside of their home compared to the other groups.
Participants rated their satisfaction with social support they receive for dealing with day to day problems. Analyses of the social support ratings yielded a significant association to age groups by χ2 test (p = 0.002), due to the middle-age adults whose social support ratings were lower than the other groups' ratings. Participants indicated whether they had a confidant, described as someone they can talk to about issues that concern them. Analyses of the confidant ratings yielded a non-significant association to age groups (p = 0.15).
Individual Differences Measures
Three measures of working memory span were administered that required participants to simultaneously hold and process auditorily-presented information. These measures included the Forward Digit Span (FDS) and Backward Digit Span (BDS) tests from the WAIS (Wechsler, 1981). Digit span tests were scored by giving full (set size) credit for sequences where both of the two trials were correct and half credit if only one trial per set size was correct. The Size Judgment Span (SJS) test (Cherry et al., 2007) required the manipulation of visuospatial information. Participants heard progressively longer sequences of individual words whose referents could be easily visualized (e.g., frog-piano-hairpin) and repeated the sequence of words in order of the referents' relative physical size, from the smallest to the largest item (e.g., hairpin-frog-piano). Two practice trials were given (two item sequences). After practice, participants were given three trials of two, three trials of three, and so forth, until they missed three consecutive trials within a sequence length. The SJS test was scored by giving full credit to sequence levels where at least two out of three trials were correct and half credit if only one of three trials was correct.
Materials
Stimulus items were simple, black and white line drawings and matching words. The drawings were taken from a standard corpus and matched for image and name agreement (Snodgrass & Vanderwart, 1980). The total stimulus pool consisted of 64 pictures and matching words, representing 8 taxonomic categories with 8 exemplars per category. Four acquisition lists of 16 items were created. Two lists contained target items presented for study (one picture list, one word list). For each participant, 16 study items were presented in picture format and the other 16 in word format. The non-presented lists served as distracter items in the picture and word recognition tests. Acquisition lists were counterbalanced so that each item appeared as a study item and a non-presented item equally often across participants and stimulus formats. All study and test stimuli were presented individually on 6 × 9 inch index cards.
Design
The design was a 4 × 2 mixed factorial with age group (middle-age, young-old, old-old, oldest-old) as a between groups variable and stimulus format (words, pictures) as a repeated measures factor. For half of the participants, the study/test materials were pictures in block 1 and words in block 2. For the other half, this order was reversed to control for presentation order effects. Forty persons were tested in each between group condition.
Procedure
Participants were tested individually. At the beginning of the session, an example item was shown (a picture or a word, depending on presentation order assignment) to familiarize the participants with the stimuli. A 3-item practice task followed. Participants named each item aloud as it was shown to ensure that all stimuli were encoded and that we had a record of possible unique verbal labels assigned to pictures (e.g., naming the “sofa” “a couch” or “davenport”). They orally recalled the study items and the experimenter recorded their responses. A practice recognition task followed, where 6 items were presented in turn (3 target items and 3 foils interspersed). Participants made a yes/no judgment for each item.
On the first block of the experiment proper, 16 items were presented for study (5 s rate). A two-minute distractor task followed, where they described their favorite foods (block 1) and favorite holidays (block 2). Following the distracter task, participants orally recalled as many of the studied items as possible. Recognition followed, using a mixed list of 16 studied items and 16 foils randomly interspersed presented in turn (8 s rate). Participants made yes/no judgments for each item. On the second trial, the stimulus format was changed relative to the first trial (i.e., from pictures to words or viceversa). The working memory measures were administered next. Debriefing followed.
Results
Overview of Scoring and Analyses
For each participant, the working memory measures were scored as discussed previously. Free recall was scored as the proportion of items correctly recalled (out of 16). Recognition was scored by calculating hit and false alarm rates that were used to derive a measure of corrected recognition (hits minus false alarms). Separate analyses of variance (ANOVAs) were conducted on the dependent measures as a function of age group and stimulus format accounting for within subject variability. Correlations among these variables were calculated, followed by sequential regression analyses that examined the independent contributions of age, cognitive status, and working memory span to word and picture free recall performance.
Mixed Model Analyses of Variance
Free Recall
Mean proportion correct by age group and stimulus format appears in Table 2 (upper panel). A mixed model type ANOVA with random subject effect, age group and stimulus format main effects, and age group × stimulus format interaction was conducted. On the free recall scores, a significant main effect of age group, F (3, 156) = 31.73, MSE = 0.04, p < 0.001 was found. Mean recall performance was highest for the middle-age adults (0.58), followed by the young-old (0.47), old-old (0.44) and the oldest-old adults (0.27), as expected. The main effect of stimulus format was also significant, F (1, 156) = 117.91, MSE = 0.02, p < 0.001. Recall of pictures exceeded that of words, with means of 0.52 and 0.36, respectively. This is an important finding, which confirms the pictorial superiority effect in free recall with the present materials and procedures. The age group × stimulus format interaction was non-significant, thus, all groups showed a pictorial advantage of comparable magnitude in free recall.
Table 2.
Mean Free Recall and Recognition as a Function of Age Group and Stimulus Format
| Stimulus format |
||||||
|---|---|---|---|---|---|---|
| Free Recall | Words | Pictures | ||||
| Age group | M | SD | M | SD | ||
| Middle-age | 0.52 | 0.17 | 0.64 | 0.18 | ||
| Young-old | 0.39 | 0.19 | 0.56 | 0.18 | ||
| Old-old | 0.35 | 0.19 | 0.53 | 0.17 | ||
| Oldest-old | 0.19 |
0.14 |
0.34 |
0.16 |
||
| Group mean | 0.36 | 0.52 | ||||
| Recognition |
Words |
Pictures |
||||
| Age group | M | SD | M | SD | ||
| Middle-age | 0.87 | 0.14 | 0.95 | 0.07 | ||
| Young-old | 0.83 | 0.12 | 0.95 | 0.07 | ||
| Old-old | 0.83 | 0.14 | 0.95 | 0.06 | ||
| Oldest-old | 0.65 |
0.24 |
0.89 |
0.10 |
||
| Group mean | 0.80 | 0.94 | ||||
Notes. Free recall scores reflect mean proportion correct. Recognition scores reflect corrected recognition (hits minus false alarms).
To provide insight into participants' organizational strategies on the free recall task, we conducted clustering analyses using the Scoring Options for Recall Tests (SORT) program, version 2.0 (Elie & Payne, 1999). Two dependent measures were of particular interest. That is, we examined the number of taxonomic categories accessed, which is assumed to reflect participants' retrieval plan, and the number of items recalled per category, thought to reflect participants' encoding strategy (Bäckman & Wahlin, 1995). Table 3 presents the age group means for the number of categories accessed and the number of items recalled per category. For the number of categories accessed, a significant effect of age group occurred, F (3, 156) = 33.68, MSE = 0.80, p < 0.0001. Means for the middle-age (3.64), young-old (3.34) and old-old adults (3.17) were greater than the mean for the oldest-old (2.29). The effect of stimulus format was significant, F (1, 156) = 47.78, MSE = 0.53, p < 0.0001, favoring pictures compared to words. Means, in order, were 3.39 and 2.83 for pictures and words. The Age Group × Stimulus Format interaction was significant, F (3, 156) = 3.17, MSE = 0.53, p = 0.03. As can be seen in Table 3, all four groups accessed a greater number of categories with pictures compared to words. The significance of the interaction is attributable to the middle age adults whose performance was near ceiling for the number of categories accessed. For the number of items recalled per category, a significant effect of age group, F (3, 156) = 18.52, MSE = 0.52, p < 0.0001 was found. Means for the middle-age (2.53), young-old (2.17) and old-old (2.15) exceeded the mean for the oldest-old adults (1.69). The effect of stimulus format was significant, F (1, 156) = 84.53, MSE = 0.25, p < 0.0001, favoring pictures compared to words. Means, in order, were 2.39 and 1.88 for pictures and words. Taken together, the results of these analyses are suggestive of deficits in both encoding and retrieval processes for the oldest-old, compared to their younger counterparts (see Bäckman & Larsson, 1992; Bäckman & Wahlin, 1995, for similar outcomes). These data also imply that the PSE observed for all age groups in this study may be mediated at least in part by participants' organizational strategies applied at both encoding and retrieval.
Table 3.
Analyses of Clustering in Free Recall
| Number of categories accessed |
Items per category |
||||
|---|---|---|---|---|---|
| Age group | Words | Pictures | Words | Pictures | |
| Middle-age | 3.55 | 3.73 | 2.34 | 2.72 * | |
| Young-old | 3.03 | 3.65 ** | 1.91 | 2.42 ** | |
| Old-old | 2.88 | 3.45 ** | 1.87 | 2.43 ** | |
| Oldest-old | 1.85 |
2.73 ** |
1.39 |
1.98 ** |
|
| Mean | 2.83 | 2.83 | 1.88 | 2.39 | |
Note: Stimulus lists were comprised of 4 taxonomic categories with 4 exemplars per category.
p ≤ .01
p ≤ .001.
Recognition
The analysis of corrected recognition scores yielded a significant effect of age group, F (3, 156) = 15.56, MSE = 0.02, p < 0.0001. Means for the middle-age (0.91), young-old (0.89) and old-old (0.89) exceeded the mean for the oldest-old adults (0.77). The main effect of stimulus format was significant, F (1, 156) = 112.88, MSE = 0.01, p < 0.0001 due to the superior recognition of pictures compared to words. Means, in order, were 0.94 and 0.80 for pictures and words. Importantly, the Age Group × Stimulus Format interaction was significant, F (3, 156) = 6.49, MSE = 0.01, p < 0.001 owing to the disproportionate memorial benefit of pictures compared to words for the oldest-old as the means in Table 2 indicate. Interpretative caution is warranted, however, as means for the three younger comparison groups are approaching ceiling in the picture condition. As a result, our analytic strategy presented in the sections that follow will be confined to word and picture free recall as the principle measures of episodic memory, neither of which is hampered by ceiling or floor effects.
Working Memory Analyses
Span estimates for the forward and backward digit span, and size judgment span tests appear in Table 4. An ANOVA on the forward digit span scores yielded a significant effect of age group, F (3, 156) = 6.16, MSE = 1.00, p = 0.0006. Pairwise comparisons confirmed that the span estimate for the middle-age adults was significantly higher than that of the young-old, old-old and oldest-old adults who did not differ from each other. An ANOVA on the backward digit span scores yielded a significant age group effect, F (3, 156) = 4.11, MSE = 0.96, p = 0.008. The middle-age adults' mean span estimate was significantly greater than the oldest-old adults' span estimate (p = 0.004). An ANOVA on the size judgment span scores yielded a significant effect of age group, F (3, 156) = 14.44, MSE = 0.71, p < 0.0001. Pairwise comparisons confirmed that the middle-aged adults' mean span estimate was significantly greater than the old-old and oldest-old adults' span estimate (p ≤ 0.001 for all). Similarly, the young-old adults' mean span exceeded that of the oldest-old adults (p < 0.001). Together, the results of these analyses confirm that the size judgment span test reliably discriminated among the comparison groups more effectively than did the forward and backward digit span measures (see Cherry et al., 2007, for a similar result).
Table 4.
Span Estimates for the Working Memory Measures
| Age group |
|||||
|---|---|---|---|---|---|
| Measure | Middle-age | Young-old | Old-old | Oldest-old | |
| Forward digit spana | |||||
| M | 6.20 | 5.49 | 5.54 | 5.30 | |
| SD | 1.11 | 0.96 | 1.05 | 0.88 | |
| Backward digit spanb | |||||
| M | 4.54 | 4.24 | 4.19 | 3.78 | |
| SD | 1.03 | 0.97 | 0.97 | 0.94 | |
| Size judgment spanc | |||||
| M | 4.70 | 4.36 | 3.99 | 3.53 | |
| SD | 1.00 | 0.70 | 0.73 | 0.90 | |
From the Wechsler Adult Intelligence Scale (Wechsler, 1981).
From the Wechsler Adult Intelligence Scale (Wechsler, 1981).
From Cherry, Elliott and Reese (2007).
Relationships Among Age, Individual Difference and Memory Measures
Correlation Analyses
Table 5 presents the correlations among chronological age, MMSE, FDS, BDS, SJS, and word and picture free recall1. Three aspects of these data warrant mention. First, age was inversely correlated with all of the memory measures, as expected. Second, for word recall, correlations with FDS, BDS and SJS were all significant and the correlation coefficients are larger in comparison to those obtained for picture recall, implying that picture recall may be less resource demanding than word recall. Third, of the three span estimates, SJS showed the strongest relationships to word and picture free recall, implying that SJS is the optimal variable to include in the sequential regression analyses reported next. In addition, a variable selection procedure was conducted using multiple regression and the all possible submodels selection method (SAS System for Linear Models, 1991). The best model was selected based on maximal R2 and minimal MSE values yielding a model that contains age, MMSE and SJS for inclusion in the regression analyses with free recall as the criterion response variable presented next.
Table 5.
Correlations Among Age, Cognitive Status and Memory Measures
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | |
|---|---|---|---|---|---|---|---|
| 1. Age | -- | ||||||
| 2 MMSE | - 0.54*** | -- | |||||
| 3. FDS | - 0.32*** | 0.31*** | -- | ||||
| 4. BDS | -0.26** | 0.29*** | 0.53*** | -- | |||
| 5. SJS | - 0.46*** | 0.44*** | 0.49*** | 0.45*** | -- | ||
| 6. Word recall | - 0.57*** | 0.54*** | 0.34*** | 0.29*** | 0.54*** | -- | |
| 7. Picture recall | - 0.51*** | 0.49*** | 0.17* | 0.21** | 0.39*** | 0.62*** | -- |
Note: MMSE indicates mini-mental state exam; FDS indicates forward digit span; BDS indicates backward digit span; SJS indicates size judgment span. N = 160.
p < 0.05
p < 0.01
p < 0.001.
Regression Analyses
Sequential regression analyses were conducted to assess the independent contributions of age, MMSE, and SJS to episodic memory using two criterion measures of performance (i.e., free recall of words and pictures). In all, four models were conducted separately for each criterion measure. In the first model, age alone was entered. In the second and third models, we wanted to determine the amount of variance in memory performance accounted for by age after adjusting for MMSE (Model 2) and SJS (Model 3). The fourth model was conducted to determine whether the MMSE and SJS were still predictive of memory performance after age had been statistically controlled for.
Results of the regression analyses appear in Tables 6 and 7. The cumulative R2 associated with the addition of subsequent variables entered into the regression equation is given in the first column. The increment in R2 associated with each additional variable appears in the second column. The beta weight associated with each variable upon entry into the equation appears in the third column. The F statistic and p value that show the statistical significance of R2 for the first variable or the increment in R2 for each additional variable, appear in the fourth and fifth columns2.
Table 6.
Summary of Sequential Regression Analyses for Word Free Recall
| R2 | Increment in R2 | β | F | P | |
|---|---|---|---|---|---|
| Model 1 | |||||
| Age | 0.326 | - 0.571 | 77.14 | < 0.0001 | |
| Model 2 | |||||
| MMSE | 0.295 | 0.543 | 77.61 | < 0.0001 | |
| Age | 0.403 | 0.108 | - 0.391 | 28.58 | < 0.0001 |
| Model 3 | |||||
| SJS | 0.288 | 0.537 | 77.9 | < 0.0001 | |
| Age | 0.420 | 0.132 | - 0.412 | 35.72 | < 0.0001 |
| Model 4 | |||||
| Age | 0.326 | - 0.571 | 94.6 | < 0.0001 | |
| MMSE | 0.403 | 0.077 | 0.33 | 22.59 | < 0.0001 |
| SJS | 0.463 | 0.060 | 0.286 | 17.25 | < 0.0001 |
Note. The increment in R2 is associated with the inclusion of additional variables entered into the regression equation. The F statistic denotes the statistical significance of R2 for the first variable or the increment in R2 associated with each additional variable entered into the regression equation. For all F statistics, the degrees of freedom were 1 and 156.
Table 7.
Summary of Sequential Regression Analyses for Picture Free Recall
| R2 | Increment in R2 | β | F | P | |
|---|---|---|---|---|---|
| Model 1 | |||||
| Age | 0.260 | - 0.510 | 55.55 | < 0.0001 | |
| Model 2 | |||||
| MMSE | 0.237 | 0.487 | 55.04 | < 0.0001 | |
| Age | 0.323 | 0.086 | - 0.349 | 19.97 | < 0.0001 |
| Model 3 | |||||
| SJS | 0.150 | 0.387 | 33.07 | < 0.0001 | |
| Age | 0.289 | 0.139 | - 0.421 | 30.74 | < 0.0001 |
| Model 4 | |||||
| Age | 0.260 | - 0.510 | 61.05 | < 0.0001 | |
| MMSE | 0.324 | 0.064 | 0.299 | 14.83 | < 0.001 |
| SJS | 0.335 | 0.011 | 0.128 | 2.80 | < 0.1 |
Note. The increment in R2 is associated with the inclusion of additional variables entered into the regression equation. The F statistic denotes the statistical significance of R2 for the first variable or the increment in R2 associated with each additional variable entered into the regression equation. For all F statistics, the degrees of freedom were 1 and 156.
As can be seen in Table 6 (Model 1), age accounted for 32.6% of the variance in word recall. In Model 2, MMSE was entered first, accounting for a significant 29.5% of the variance. Following the statistical control of MMSE, the variance accounted for by age was reduced to 10.8%. This finding indicates that individual differences in MMSE account for some of the age-related variance in word recall. To be precise, the drop in variance accounted for by age, from 32.6% (Model 1) to 10.8% (Model 2) after MMSE was controlled for can be interpreted to indicate that MMSE accounts for 66.9% of the age-related variance in word recall (i.e., [.326 - .108]/.326). In Model 3, SJS accounted for a significant 28.8% of the variance and the contribution of age was reduced to 13.2%, confirming that SJS accounts for a sizeable portion of age-related variance in word recall. That is, the reduction in variance accounted for by age, from 32.6% (Model 1) to 13.2% (Model 3) after SJS was controlled indicates that SJS accounts for 59.5% of the agerelated variance in word recall (i.e., [.326 - .132]/.326). This is a noteworthy finding, indicating that SJS accounts for nearly as much age-related variance in word recall as does MMSE. Note, however, that cognitive status (MMSE) and working memory resource (SJS) influence word recall, but do not account for age effects completely because the 10.8% (Model 2) and the 13.2% (Model 3) of the variance accounted for by age in these models was significant (p < .0001 for all). In Model 4, following the addition of age, the R2 values associated with MMSE and SJS remained significant, but were reduced to 7.7% and 6.0% of the variance consecutively. Thus, MMSE and SJS accounted for unique variance in word recall beyond that already accounted for by age.
Inspection of Table 7 (Model 1) reveals that age accounted for a significant 26.0% of the variance in picture recall. In Model 2, MMSE accounted for a significant 23.7% of the variance in picture recall. Following the statistical control of MMSE, the variance accounted for by age was reduced to 8.6%. The drop in variance accounted for by age, from 26.0% (Model 1) to 8.6% (Model 2) after MMSE was controlled can be interpreted to indicate that MMSE accounts for 66.9% of the age-related variance in picture recall (i.e., [.260 - .086]/.260). In Model 3, SJS accounted for a significant 15.0% of the variance. The contribution of age was reduced to 13.9%, confirming that SJS accounts for age-related variance in picture recall. The drop in variance accounted for by age, from 26.0% (Model 1) to 13.9% (Model 3) after SJS was controlled indicates that SJS accounts for 46.5% of the age-related variance in picture recall (i.e., [.260 - .139]/.260). Note, however, that MMSE and SJS do not account for age effects in picture recall completely because the remaining 8.6% (Model 2) and the 13.9% (Model 3) of the variance accounted for by age in these models was significant (p < .0001 for all). In Model 4, following the addition of age, MMSE made a small but significant contribution (6.4%), although SJS did not account for additional variance in picture recall beyond that already accounted for by age and MMSE.
Discussion
The principle new findings of this study are as follows. First, we found clear evidence of a PSE for all age groups, implying that the oldest-old utilize nonverbal or visuospatial codes as well as their younger counterparts to support long-term retention. Follow-up analyses of clustering in free recall revealed that a greater number of categories were accessed and more items were recalled per category for pictures than words for all. Second, cognitive status and working memory were correlated with word and picture recall. Third, sequential regression analyses indicated that age, cognitive status and working memory predicted word and picture recall. Cognitive status and working memory also accounted for significant amounts of age-related variance. These results are discussed more fully next.
Pictorial Superiority Effect
The first finding was the PSE in free recall and recognition for all age groups. Our findings join others in the cognitive aging literature where a pictorial superiority advantage was observed for healthy young-old adults (Park et al., 1983; Rissenberg & Glazner, 1986; Winograd et al., 1982). To our knowledge, this study is the first to show a PSE for oldest-old, defined here as persons over 90 years of age. Nonagenarians showed a PSE in both free recall and recognition, confirming the reliability and generality of the finding. Consistent with a contextualist perspective, a significant age interaction occurred where the memorial benefit of pictures was the greatest in size for nonagenarians when retention was tested using a recognition task. Interpretative caution is warranted, as the younger age groups were approaching ceiling in picture recognition, accounting for the significance of the interaction (see Table 2, lower panel). Thus, we confine our discussion of episodic memory in healthy aging to word and picture free recall.
Follow-up analyses of clustering in free recall revealed a possible mechanism of the memorial benefit of pictures, namely, a greater number of categories were accessed and more items were recalled per category when pictures served as stimuli compared to words (see Table 3). These clustering indices were proportional for the nonagenarians compared to the other age groups, although number of categories accessed and number of items recalled were somewhat lower for the oldest-old. The results of the clustering analyses imply that nonagenarians engage in qualitatively similar organizational strategies in support of recall as do the younger age groups.
Overall, the finding that nonagenarians show a PSE is intriguing, having noteworthy theoretical and practical implications. Specifically, the results imply that the oldest-old utilize nonverbal or visuospatial codes as do younger age groups to support episodic memory. These findings also suggest that the oldest-old may draw on semantic memory knowledge (i.e., the taxonomic category membership of the to-be-remembered items) to enhance episodic memory in a manner similar to their younger counterparts (Bäckman & Wahlin, 1995). The finding that all showed greater use of categorical organization in free recall of pictures compared to words further implies that pictorial material may facilitate application of semantic memory skills, but more research is needed. From a practical perspective, our results imply that the oldest-old may benefit from pictorial illustrations to support retention of textual material, an exciting possibility that offers a direction for memory remediation in the everyday lives of very old adults.
Role of Individual Differences
Our second finding concerned the influence of cognitive status and working memory span estimates on episodic memory performance. Correlation analyses indicated that age was negatively related to performance on all memory measures, as expected. Of greater interest were the correlations among the cognitive status and working memory variables, word and picture free recall. To be precise, the r values were higher for MMSE, SJS and word recall, compared to the r values obtained for these variables and picture recall (see Table 5). This aspect of the data implies that word recall may be more resource demanding than picture recall, consistent with a contextualist perspective. Words may offer less cognitive support for remembering compared to pictures which are richer in stimulus properties important for retention, such as visual distinctiveness. Pictures may also invite a more elaborative or richer encoding than words by comparison, resulting in memorial gains such as those observed here.
The present findings, among others, attest to the powerful role of individual differences and cognitive support in episodic remembering in old age (Bäckman et al., 2000). One should note that preliminary analyses with educational attainment in this study revealed no significant age group differences in level of education, nor any significant relationships with secondary memory performance. This aspect of the data conflicts with earlier reports where educational attainment was related to memory performance (Bäckman & Wahlin, 1995; Hill et al., 1995; Wahlin, Bäckman, & Winblad, 1995). Community-dwelling older adults who self-select to participate in cognitive aging research typically vary in educational attainment, among other general intellectual and sociodemographic variables (Park & Cherry, 1989). The sampling strategy used in this study may have corrected for a potential selection bias. Alternatively, Hassing, Whalin, and Bäckman (1998) have since made the point that demographic variables may be less important to memory performance in very old age when participants have been extensively screened for health status. Further research to examine the dynamic relationship between health status and episodic memory performance seems warranted.
The third finding pertains to the sequential regression analyses. These results confirmed that cognitive status and working memory accounted for substantial amounts of age-related variance (see Tables 6 and 7), although age made an independent contribution to both word and picture free recall (see Bäckman & Wahlin, 1995, for a similar result). Further, the age-related variance accounted for by MMSE, the cognitive status measure, exceeded that of SJS, the working memory span measure, in both word and picture free recall. Specifically, MMSE accounted for 66.9% of the age-related variance in both word and picture recall. The same estimates for SJS were 59.5% and 46.5%, for word and picture recall, respectively. The MMSE provides a global index of cognitive status, perhaps having more executive function involvement than does SJS, where participants merely arrange a list of auditorily-presented items in order of physical size. Further research is necessary to understand the contributions of the executive component to performance on the present episodic memory task.
On a cautionary note, variance partitioning procedures such as the sequential regression analyses used here have been questioned on technical and substantive grounds (see Lindenberger & Pötter, 1998). With this caveat in mind, our data imply that MMSE may be preferable to SJS alone as an index of cognitive functionality. These two measures were significantly correlated (r = 0.44, p < 0.0001), although the correlation was reduced when age partialed out (r = 0.26, p = 0.0012). The differential predictive utility of the MMSE over the SJS suggests that it is the better choice when the analytic strategy is to account for age-related variance in episodic memory performance.
In closing, this study provides new evidence on the PSE as an episodic memory phenomenon that persists well into the ninetieth decade of life for healthy older adults. Two methodological limitations of the study warrant brief mention. First, self-reports of strategy use on the free recall task were not solicited, which may have provided valuable information on how participants approached the acquisition and retrieval tasks. Second, the younger groups were at ceiling in picture recognition, a typical problem in studies using pictorial stimuli and recognition tasks in the cognitive aging literature. Future research to address these limitations and explore the generality of the present findings seems is warranted.
Acknowledgments
This research was supported by grants from the Louisiana Board of Regents through the Millennium Trust Health Excellence Fund [HEF(2001-06)-02] and the National Institute on Aging P01 AG022064. This support is gratefully acknowledged.
Erin M Jackson is now at University of South Florida. We are grateful to Julieta Spears, Elizabeth Maggiore, Marla Erwin, and Travis Chidester for their assistance with participant testing. We also thank and Emily Boudreaux and Emily Smitherman for their help with data scoring.
Footnotes
Because level of education is a categorical variable, we conducted Spearman's correlations among the educational attainment scores (rated on a 7-point scale) and the memory measures reported here. As expected, level of education was positively correlated with MMSE, FDS, BDS, and SJS performance (r's between 0.19 and 0.39 and all p's < 0.02). However, level of education was not significantly correlated with picture free recall (r = 0.08, p = 0.32) and only marginally so with word free recall (r = 0.14, p = 0.07). Further, the age groups were no different in self-reported educational attainment. On the basis these outcomes, educational attainment was not included in the sequential regression analyses.
Type I Sum of Squares as well as the residual means square from the full model was used in calculation of the F and p values for the first variable and for each additional variable. The purpose of this procedure was to permit direct comparisons across regression models (Hocking, 1985). We used this procedure in the analyses of word free recall (Table 6) and picture free recall (Table 7).
References
- Bäckman L, Larsson M. Recall of organizable words and objects in adulthood: Influences of instructions, retention interval, and retrieval cues. Journal of Gerontology: Psychological Sciences. 1992;47:273–278. doi: 10.1093/geronj/47.4.p273. [DOI] [PubMed] [Google Scholar]
- Bäckman L, Mantyla T, Herlitz A. The optimization of episodic remembering in old age. In: Baltes PB, Baltes MM, editors. Successful aging: Perspectives for the behavioral sciences. NY: Cambridge University Press; 1990. pp. 118–163. [Google Scholar]
- Bäckman L, Small BJ, Wahlin A, Larsson M. Cognitive functioning in very old age. In: Craik FIM, Salthouse TA, editors. The handbook of aging and cognition. 2nd. Mahwah, New Jersey: Erlbaum; 2000. pp. 499–558. [Google Scholar]
- Bäckman L, Wahlin A. Influences of item organizability and semantic retrieval cues on word recall in very old age. Aging and Cognition. 1995;2:312–325. [Google Scholar]
- Cherry KE, Applegate H, Reese CM. Do adults with mental retardation show pictorial superiority effects in recall and recognition? Research in Developmental Disabilities. 2002;23:135–147. doi: 10.1016/s0891-4222(02)00091-4. [DOI] [PubMed] [Google Scholar]
- Cherry KE, Elliott EM, Reese CM. Age and individual differences in working memory: The size judgment span task. Journal of General Psychology. 2007;134:43–65. doi: 10.3200/GENP.134.1.43-65. [DOI] [PubMed] [Google Scholar]
- Cherry KE, Park DC, Frieske DA, Smith AD. Verbal and pictorial elaborations enhance memory in younger and older adults. Aging, Neuropsychology, and Cognition. 1996;3:15–29. [Google Scholar]
- Craik FIM. Age-related changes in human memory. In: Park D, Schwarz N, editors. Cognitive Aging: A Primer. Philadelphia: Psychology Press; 2000. pp. 75–91. [Google Scholar]
- Craik FIM, Byrd M, Swanson JM. Patterns of memory loss in three elderly samples. Psychology and Aging. 1987;2:79–86. doi: 10.1037/0882-7974.2.1.79. [DOI] [PubMed] [Google Scholar]
- Dodson CS, Schacter DL. When false recognition meets metacognition: The distinctiveness heuristic. Journal of Memory and Language. 2002;46:782–803. [Google Scholar]
- Duke University Center for the Study of Aging and Human Development . OARS: Multidimensional Functional Assessment Questionnaire. Durham, NC: Author; 1975. [Google Scholar]
- Elie CJ, Payne DG. Scoring options for recall tests (SORT), version 2.0. Behavior Research Methods, Instruments, & Computers. 1999;31:380–383. [Google Scholar]
- Folstein MF, Folstein SE, McHugh PR. “Mini-Mental State”: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research. 1975;12:189–198. doi: 10.1016/0022-3956(75)90026-6. [DOI] [PubMed] [Google Scholar]
- Hassing L, Wahlin A, Bäckman L. Minimal influence of age, education and gender on episodic memory functioning in very old age: A population-based study of nonagenarians. Archives of Gerontology and Geriatrics. 1998;27:75–87. doi: 10.1016/s0167-4943(98)00101-0. [DOI] [PubMed] [Google Scholar]
- Hill RD, Wahlin A, Winblad B, Bäckman L. The role of demographic and life style variables in utilizing cognitive support for episodic remembering among very old adults. Journal of Gerontology: Psychological Sciences. 1995;50B:P219–227. doi: 10.1093/geronb/50b.4.p219. [DOI] [PubMed] [Google Scholar]
- Hocking RR. The Analysis of Linear Models. NY: Brooks/Cole; 1985. [Google Scholar]
- Jastak J, Jastak S. Short forms of the WAIS vocabulary subtest. Journal of Clinical Psychology. 1965;20:167–199. doi: 10.1002/1097-4679(196404)20:2<167::aid-jclp2270200202>3.0.co;2-9. [DOI] [PubMed] [Google Scholar]
- Lindenberger U, Pötter U. The complex nature of unique and shared effects in hierarchical linear regression: Implications for developmental psychology. Psychological Methods. 1998;3:218–230. [Google Scholar]
- Luszcz M, Bryan J, Kent P. Predicting episodic memory performance of very old men and women: Contributions from age, depression, activity, cognitive ability and speed. Psychology and Aging. 1997;12:340–351. doi: 10.1037//0882-7974.12.2.340. [DOI] [PubMed] [Google Scholar]
- Mintzer MZ, Snodgrass JG. The picture superiority effect: Support for the distinctiveness model. American Journal of Psychology. 1999;112:113–146. [PubMed] [Google Scholar]
- Nelson DL, Reed VS, Walling JR. Pictorial superiority effect. Journal of Experimental Psychology: Human Learning and Memory. 1976;2:523–528. [PubMed] [Google Scholar]
- Park DC, Puglisi JT, Sovacool M. Memory for pictures, words, and spatial location in older adults: evidence for pictorial superiority. Journal of Gerontology. 1983;38:582–588. doi: 10.1093/geronj/38.5.582. [DOI] [PubMed] [Google Scholar]
- Park DC, Cherry KE. Human subjects and cognitive aging research: A unique solution to a perennial problem. Educational Gerontology. 1989;15:563–571. [Google Scholar]
- Paivio A. Imagery and verbal processes. New York: Holt, Rinehart & Winston; 1971. [Google Scholar]
- Paivio A. Dual coding theory: Retrospect and current status. Canadian Journal of Psychology. 1991;43:255–287. [Google Scholar]
- Rissenberg M, Glanzer M. Picture superiority in free recall: the effects of normal aging and primary degenerative dementia. Journal of Gerontology. 1986;41:64–71. doi: 10.1093/geronj/41.1.64. [DOI] [PubMed] [Google Scholar]
- SAS System for Linear Models. Third Edition. Cary, NC: SAS Institute Inc.; 1991. pp. 20–22. [Google Scholar]
- Sheikh JI, Yesavage JA. Geriatric depression scale (GDS): Recent evidence and development of a shorter version. In: Brink TL, editor. Clinical Gerontology. NY: Haworth Press; 1986. pp. 165–173. [Google Scholar]
- Snodgrass JG, Asiaghi A. The pictorial superiority effect in recognition memory. Bulletin of the Psychometric Society. 1977;10:1–4. [Google Scholar]
- Snodgrass JG, Vanderwart MA. A standardized set of 260 pictures: Norms for name agreement, image agreement, familiarity, and visual complexity. Journal of Experimental Psychology: Human Learning and Memory. 1980;6:174–215. doi: 10.1037//0278-7393.6.2.174. [DOI] [PubMed] [Google Scholar]
- Springer MV, McIntosh AR, Winocur G, Grady CL. The relation between brain activity during memory tasks and years of education in young and older adults. Neuropsychology. 2005;19:181–192. doi: 10.1037/0894-4105.19.2.181. [DOI] [PubMed] [Google Scholar]
- Stenburg G. Conceptual and perceptual factors in the picture superiority effect. European Journal of Cognitive Psychology. 2006;18:813–847. [Google Scholar]
- Toglia MP, Hinman PJ, Dayton BS, Catalano J. The blocked-random effect in pictures and words. Perceptual and Motor Skills. 1997;84:976–978. doi: 10.2466/pms.1997.84.3.976. [DOI] [PubMed] [Google Scholar]
- Tombaugh TN, McIntyre NJ. The Mini-Mental State Examination: A comprehensive review. Journal of American Geriatrics Society. 1992;40:922–935. doi: 10.1111/j.1532-5415.1992.tb01992.x. [DOI] [PubMed] [Google Scholar]
- Wahlin A, Bäckman L, Mäntylä T, Herlitz A, Viitanen M, Winblad B. Prior knowledge and face recognition in a community-based sample of healthy, very old adults. Journal of Gerontology: Psychological Sciences. 1993;48:P54–61. doi: 10.1093/geronj/48.2.p54. [DOI] [PubMed] [Google Scholar]
- Wahlin A, Bäckman L, Winblad B. Free recall and recognition of slowly and rapidly presented words in very old age: A community-based study. Experimental Aging Research. 1995;21:251–271. doi: 10.1080/03610739508253984. [DOI] [PubMed] [Google Scholar]
- Wechsler D. Wechsler Adult Intelligence Scale-Revised. San Antonio, TX: The Psychological Corporation; 1981. [Google Scholar]
- Weldon MS, Roediger HL. Altering retrieval demands reverses the picture superiority effect. Memory & Cognition. 1987;15:269–280. doi: 10.3758/bf03197030. [DOI] [PubMed] [Google Scholar]
- Whitehouse AJO, Maybery MT, Durkin K. The development of the picture-superiority effect. British Journal of Developmental Psychology. 2006;24:767–773. [Google Scholar]
- Winograd E, Smith AD, Simon EW. Aging and the picture superiority effect in recall. Journal of Gerontology. 1982;37:70–75. doi: 10.1093/geronj/37.1.70. [DOI] [PubMed] [Google Scholar]
- Zacks RT, Hasher L, Li KZH. Human memory. In: Craik FIM, Salthouse TA, editors. The handbook of aging and cognition. Mahwah, NJ: Erlbaum; 2000. pp. 293–357. [Google Scholar]