A longitudinal study of the performance of the elderly and young on the Tower of Hanoi puzzle and Rey recall.

Normal performance on the Tower of Hanoi puzzle by amnesic patients has been taken as support for viewing this problem solving task as having a nondeclarative memory component. Individuals in each decade of life between the 20s and 80s were asked to solve this puzzle four times in four sessions with intersession intervals from 1 to 7 days (Davis & Keller, 1998). Participants in their 70s and 80s were significantly impaired compared to participants in their 20s and 30s. The elderly were also significantly impaired on five immediate trials of a 15 words verbal recall test. Participants were readministered these tests an average of 6.6 years later for the elderly (n = 12) and 7.7 years later for the young (n = 11). For the Tower of Hanoi, the performance of the elderly, but not the young individuals, was significantly poorer than their original performance. For the verbal recall test, no significant change over time was detected for the young or elderly participants. These findings support the view that some nondeclarative and/or problem solving tasks demonstrate as great or greater decline with age than some declarative tasks.

A ten-year longitudinal examination of repetition priming, incidental recall, free recall, and recognition in young and elderly.

The effects of age and time on nondeclarative and declarative memory in young and elderly were examined in a 10-year longitudinal study using tests of word-stem priming, incidental recall, free recall, and recognition. The elderly were significantly impaired on all tests, but no reliable longitudinal decrement by the elderly was detected for priming, incidental recall,or recognition. The elderly demonstrated a significant longitudinal decline in declarative memory as assessed by a test of free recall. While nondeclarative memory declines with age, the longitudinal findings are consistent with the view that declarative memory is more susceptible to the effects of aging.

A comparison of computerized and standard versions of the Wisconsin Card Sorting Test.

The manual version of the Wisconsin Card Sorting Test (WCST) is the most frequently used neuropsychological instrument for differentiating normal from populations with frontal-lobe impairments. A number of computerized versions have become available, and their ability to accurately detect frontal-lobe dysfunction is based on the assumption of equivalence with the standard manual version. Accordingly, comparisons of the distribution properties of central tendency, variability, and shape between the manual version with four computerized versions of the WCST were performed (n = 22 per condition). None of the computerized versions were found to be equivalent to the manual version on all assessment measures. Given the discrepancies between the manual and computerized versions, it is concluded that the norms provided for the standard manual presentation method should not be used for the computerized versions, and for continued use of the computerized versions, new norms for computer versions need to be established. Thus, clinicians and experimenters must use caution when basing conclusions on scores from computer versions of the WCST.

The effect of age on the learning of a nondeclarative category classification task.

A category classification task was administered to participants in their twenties through their eighties. Participants studied a set of high distortions of a prototype dot pattern and were then asked to choose whether or not a new set of dot patterns (random patterns, high distortions, low distortions, and the prototype) belonged to the same category of dot patterns as studied. Participants were also administered a recognition test after studying a second set of dot patterns. There were no significant differences for age groups on the pattern recognition test. In 2 of the 3 analyses of the category classification task, there were no significant age effects. However, there was a small age effect in one analysis with the young making more accurate classifications on two aspects of the task. The results are consistent with the view that small age-related effects may exist for some tests of nondeclarative memory.

The detection of simulated malingering using a computerized category classification test.

A category classification test was used to differentiate between normal student control participants, students instructed to malinger a memory deficit, and amnesic patients. Controls (N = 44) and amnesic patients (N = 10) were instructed to do their best, while simulators of malingering (N = 43) were instructed to fake a memory deficit for credit and possible financial compensation. Participants studied a list of high distortions of a prototype dot pattern and were then asked to choose whether or not a new set of dot patterns (random patterns, high distortions, low distortions, and the prototype) belonged to the same category of dot patterns as studied. Malingerers performed significantly worse than normal controls and amnesic patients. A discriminant function analysis showed that the classification test can be used to correctly classify participants as simulated malingerers, controls, or amnesic patients significantly higher than chance. These results indicate that a category classification test can be used in the detection of simulated malingering and that some tests of implicit memory provide a potential supplement to standard forced choice tests in the detection of malingering.

The detection of simulated malingering using a computerized priming test.

A word completion priming test was used to differentiate between normal student control subjects and students instructed to malinger. Controls (n = 60) were instructed to do their best, while malingering subjects (n = 60) were instructed to fake a memory deficit for credit and possible financial compensation. Subjects initially rated and completed stems for words that had at least 10 possible completions. Thirty minutes later, subjects rated and completed stems for words that were either uniquely defined by the stem or could only be completed with a variation of the word. Simulated malingerers and controls differed significantly on response latencies (time to produce rated words-time to produce baseline words, 10 second time limit) and priming scores. Discriminant function analyses showed that as high as 92% of the controls could be correctly identified, and 73% of the malingerers could be correctly identified. These results indicate that priming tests can be used in the detection of malingering.