- Email: [email protected]
A fifteen-item modification of the fuld object-memory evaluation: Preliminary data from healthy middle-aged adults
Archives of Clinrcal Neumpsychologv, Vol. 3. pp. 345-349. 1988 Pnntcd I” the USA All nghts reserved Copyright
@ 1988 Nawnrl
0887-5177/88 13.00 + .oO Academy of NeUrOpsychOloglsls
A Fifteen-kern Modification of the Fuld Object-Memory Evaluation: Preliminary Data From Healthy Middle-Aged Adults Linda Davenport,
Fiona F. Brown, George kin,
and Craig Van Dyke
Kterans Administratmn Medrcal Center and Universrty of Calrfornra,San Francrsco
A modification of the Fuld Object-Memory Evaluation (FOME) test was developed for use with middle-aged adults. To avoid the ceiling t$fet which occurs when the IO-item FOME is used with adults in this age group, we made the test more dtfficuultby increasing the number of items to 15. We obtained preliminary data on the I5-item version of the test by administering it to 32 healthy, welleducated adults aged 35 to 55. This version of the test was difficult enough so that even high functioning adults needed multiple trials to memorize all items. No sex differences, and no significant correlations with education or age for any aspect of test performance were found. Wepresent mean scores for recall on each trial, as well as averaged scores for storage, retrieval, consistent retrieval, and recall failure across five trrals.
The selective reminding (SR) procedure for assessment of memory was designed to differentiate storage, retrieval, and retention aspects of memory processes (Buschke & Fuld, 1974). The procedure entails multiple opportunities to learn a list of items; after each free recall attempt, the subject is selectively reminded of only those items which were not recalled on that attempt. This minimizes overlearning of previously recalled items. The use of verbal distracters between consecutive trials precludes retention in imme-
This research was supported by NINCDS Grant No. ROI-NS22029-01 and by Veterans’Administration General Medical Research Funds. Requests for reprints should be sent to George Fein, Ph.D., SFVAMC (116R), 4150 Clement Street, San Francisco, CA 94121, U.S.A. 345
346
L. Davenport et al.
diate memory; therefore subsequent recall reflects processing in long-term storage. In the Fuld Object-Memory Evaluation (FOME), which uses the SR procedure, 10 stimulus objects are initially presented both tactually and visually rather than as an oral word list (Fuld, 1977). This ensures that subjects actually process the stimuli prior to recall trials, mimimizing attention deficits as a confounding factor in interpreting FOME performance results. The FOME was designed for use with and normed on two groups of elderly subjects: community residents in good health, and patients in nursing homes. We wanted to use its procedures to test memory in individuals during middle adulthood, when memory processes can be impaired as a consequence of disease or as a side-effect of pharmacologic agents. To this end, we needed reference data for a comparable sample of healthy adults within this age group. Our early work with the lo-item FOME indicated that many average- to high-functioning adults could remember all ten items on the first and on most subsequent trials. To avoid this ceiling effect, we made the memory task more difficult by increasing the number of objects from 10 to 15. During our preliminary testing, we determined that 15 items made the task sufficiently difficult that even high functioning adults needed multiple trials to memorize the entire list, but not so difficult that it overloaded impaired individuals. In this paper, we present data from well-functioning middleaged adults for our 15-item modification of the FOME (FOME-15). METHOD Subjects
Sixteen female and sixteen male subjects, aged 35-55 years, were recruited from the staff of the San Francisco Veterans’ Administration Medical Center and from the community, Candidates were interviewed by telephone to obtain medical and psychiatric histories. We excluded those who reported a history of stroke, head injury, epilepsy, diabetes, hypertension, drug abuse, alcoholism, or psychiatric disorders. Twenty-eight of the 32 subjects were Caucasian, and the remaining four were black. All were high school graduates; the mean educational level was 15.8 years, with a range of 12 to 21 years. Distributions of educational levels were similar for women and men. We used the New Adult Reading Test (NART) (Nelson, 1982) to estimate IQ. This has a high correlation with WARS-R Full Scale IQ (Nelson & McKenna, 1975), and we have found a very high correlation with IQ estimates based on the WAIS-R Vocabulary subtest in a sample of patients with multiple sclerosis. The NART takes only a few minutes to administer, and
I5-Item FOME for Middle-Aged Adults
347
can be completed within the five-minute distractor period which is part of the FOME-15 procedure. Estimated IQ in this sample ranged from 96 to 126, with a mean of 114. This sample was well above average in both IQ and education level; however, it was comparable on these measures to several clinical samples we had recruited over the past few years. While not representative of the population as a whole, it may nevertheless be an appropriate comparison group for clinical samples in many research settings. Procedure
In the original FOME, 10 common household objects (e.g., toothbrush, matchbook) are presented to the subject for identification: first tactually, then visually. In the FOME-15, we increased the number of test items from 10 to 15. Eight items were taken from Form I of the FOME and seven from Form II. We eliminated one of each pair of objects with similar names or related functions. For example, we used “nail” and excluded “nail file”; we used “lock” and eliminated “key.” The 15 items are listed in Table 1. After the presentation, the objects were removed from sight, and the subject was distracted by a 60-second verbal fluency test which involved naming items from a specific semantic category as rapidly as possible. The test-taker was then asked to recall as many stimulus objects as possible, and was selectively reminded of omitted items. To ensure adequate time for recall of the additional items, we increased the recall time limit from 60 to 90 seconds. The sequence (verbal distractor- recall attempt-selective reminding) was repeated a total of five times. The distracters before trials 2 through 5 were additional semantic category, verbal fluency tests, each 30 seconds long. The distractor after the fifth trial lasted for five minutes, during which we administered the NART. A final, unexpected recall attempt then took place, and any items not recalled were presented to the subject in an oral multiple choice format to assess recognition. Further details of test administration are described elsewhere (Fuld, 1977, 1980). We calculated mean scores across trials for each of several measures: retrieval (number of items recalled), storage (cumulative retrieval),’ consistent retrieval (items recalled on two consecutive trials), and recall failure (omission of items on two consecutive trials, despite reminder) (cf. Fuld, tlt should be noted that, in order to retain comparability with the original FOME, we have maintained the original terminology. However, the term “storage,” which refers to all items recalled on any of the first five trials, is actually a misnomer. It is possible for a subject to omit some item on every trial, thus not to be credited with having “stored” it, and yet to recognize it from the multiple choice offered after the delayed recall. We consider that such recognition is evidence of prior storage. Thus, what is scored as storage on this test may in certain cases be an underestimate of actual storage.
348
L. Davenport
et al.
TABLE 1 FOME-15 Test Items Card Nail Matches
Rubber Pencil Bottle
Band
Scissors Can Spoon
Ball Rmg Comb
Toothbrush Button Lock
1977; La Rue, D’Elia, Clark, Spar, & Jarvik, 1986). We also calculated average scores for retrieval on each of the six recall trials. We compared men and women on the summary variables, using multivariate analysis of variance. In addition, we tested for effects of age and education by computing the multiple-correlation between age and the set of summary variables.
RESULTS There were no sex differences and no significant multiple-correlations of the FOME-15 summary variables with either age or educational level. Therefore, we treated the data as coming from a single population and averaged scores from all 32 subjects. Table 2 presents average recall scores for each trial and average scores across all trials for retrieval, storage, consistent retrieval, and recall failure. One-sided 95% confidence intervals (x1.64 SD) are provided as a cut-off score indicative of a deficit relative to this reference group. TABLE 2 All Subjects: Summary Statistics Across Trials, and Recall on Each Trial Separately
Mean
Retrieval Storage Consistent retrieval Recall failweb
Recall Trtal Trial Trial Trial Trral Trial
1 2 3 4 5 6
SD
Lower limit of confidence intervala
Averaged across Trials 12.6 1.2 13.9 0.6 10.8 1.9 0.4 0.6 lndrvidual 10.3 12.5 13.2 13.1 13.8 14.0
Trials 1.7 1.3 1.5 1.8 1.5 1.3
“x-1.64 (SD) bRecall failure is an error score, thus the confidence an upper hnut.
10.7 12.9 1.7 1.3
1.6 10.4 10.7 10.2 11.4 11.8
interval
is
15-Item FOME for Middle-Aged
349
Adults
No subject recalled all 15 items on either of the first two trials. Twentyone subjects (66%) recalled all 15 items at least once on Trials 3 through 5. Only four of the 13 subjects who initially recalled 15 items on Trial 3 or 4 recalled all items on two consecutive trials. Failure to remember all items on two consecutive trials, therefore, should not be interpreted as evidence of impairment. Most subjects (85 X) had stored all items by Trial 3; only one had failed to store every item by Trial 5. Recall failures were unusual; only three subjects (9%) failed to recall as many as three items on two consecutive trials after the SR procedure. On the delayed recall trial, in which subjects were given a multiple choice of items they had omitted, all subjects correctly recognized all items. Failure at this recognition task should, therefore, alert the examiner to the possibility of an impairment in retention over a brief time interval. DISCUSSION
We have adapted the FOME test for use with middle-aged adults by increasing the length of the list to be learned from the standard 10 items to 15 items. This modified version, the FOME-15, eliminates ceiling effects even in healthy adults of average to above-average ability. We present preliminary data on a well-characterized sample of subjects, which will be an appropriate comparison group for many clinical samples. The cut-off scores suggested by our data may be biased toward classifying individuals of average or below-average IQ and educational level as impaired, and therefore should be used cautiously with such individuals. Clearly, it will be important to supplement these preliminary data with scores obtained from a larger sample of subjects who are more representative of the general population. With such a normative data base, the FOME15 may prove to be a valuable tool for investigating pathological and iatrogenie memory changes during middle adulthood. REFERENCES Buschke, H., & Fuld, P. A. (1974). Evaluating storage, retention, memory and learning. Neurology, 24, 1019-1025. Fuld, P. A. (1977). The Fuld ObJecf-memory Evaluation. Chicago:
and retrieval Stoelting
in disordered
Instrument
Com-
pany. Fuld, P. A. (1980). Guaranteed stimulus processing in the evaluation of memory and learning. Cortex, 16, 255-271. La Rue, A., D’Elia, L. F., Clark, E. 0.. Spar, J. E., & Jarvik, L. F. (1986). Clinical tests of memory in dementia, depression, and healthy aging. Journal of Psychology and Agmg, 1, 69-77. Nelson, H. E. (1982). Test Mama/for the Adult Reading Test. Windsor, England: NFER Wilson Publishing Co. Nelson, H. E., & McKenna, P. (1975). The use of current reading ability in the assessment of dementia. British Journal of Social and Clinrcal Psychology, 14, 259-267.
@ 1988 Nawnrl
0887-5177/88 13.00 + .oO Academy of NeUrOpsychOloglsls
A Fifteen-kern Modification of the Fuld Object-Memory Evaluation: Preliminary Data From Healthy Middle-Aged Adults Linda Davenport,
Fiona F. Brown, George kin,
and Craig Van Dyke
Kterans Administratmn Medrcal Center and Universrty of Calrfornra,San Francrsco
A modification of the Fuld Object-Memory Evaluation (FOME) test was developed for use with middle-aged adults. To avoid the ceiling t$fet which occurs when the IO-item FOME is used with adults in this age group, we made the test more dtfficuultby increasing the number of items to 15. We obtained preliminary data on the I5-item version of the test by administering it to 32 healthy, welleducated adults aged 35 to 55. This version of the test was difficult enough so that even high functioning adults needed multiple trials to memorize all items. No sex differences, and no significant correlations with education or age for any aspect of test performance were found. Wepresent mean scores for recall on each trial, as well as averaged scores for storage, retrieval, consistent retrieval, and recall failure across five trrals.
The selective reminding (SR) procedure for assessment of memory was designed to differentiate storage, retrieval, and retention aspects of memory processes (Buschke & Fuld, 1974). The procedure entails multiple opportunities to learn a list of items; after each free recall attempt, the subject is selectively reminded of only those items which were not recalled on that attempt. This minimizes overlearning of previously recalled items. The use of verbal distracters between consecutive trials precludes retention in imme-
This research was supported by NINCDS Grant No. ROI-NS22029-01 and by Veterans’Administration General Medical Research Funds. Requests for reprints should be sent to George Fein, Ph.D., SFVAMC (116R), 4150 Clement Street, San Francisco, CA 94121, U.S.A. 345
346
L. Davenport et al.
diate memory; therefore subsequent recall reflects processing in long-term storage. In the Fuld Object-Memory Evaluation (FOME), which uses the SR procedure, 10 stimulus objects are initially presented both tactually and visually rather than as an oral word list (Fuld, 1977). This ensures that subjects actually process the stimuli prior to recall trials, mimimizing attention deficits as a confounding factor in interpreting FOME performance results. The FOME was designed for use with and normed on two groups of elderly subjects: community residents in good health, and patients in nursing homes. We wanted to use its procedures to test memory in individuals during middle adulthood, when memory processes can be impaired as a consequence of disease or as a side-effect of pharmacologic agents. To this end, we needed reference data for a comparable sample of healthy adults within this age group. Our early work with the lo-item FOME indicated that many average- to high-functioning adults could remember all ten items on the first and on most subsequent trials. To avoid this ceiling effect, we made the memory task more difficult by increasing the number of objects from 10 to 15. During our preliminary testing, we determined that 15 items made the task sufficiently difficult that even high functioning adults needed multiple trials to memorize the entire list, but not so difficult that it overloaded impaired individuals. In this paper, we present data from well-functioning middleaged adults for our 15-item modification of the FOME (FOME-15). METHOD Subjects
Sixteen female and sixteen male subjects, aged 35-55 years, were recruited from the staff of the San Francisco Veterans’ Administration Medical Center and from the community, Candidates were interviewed by telephone to obtain medical and psychiatric histories. We excluded those who reported a history of stroke, head injury, epilepsy, diabetes, hypertension, drug abuse, alcoholism, or psychiatric disorders. Twenty-eight of the 32 subjects were Caucasian, and the remaining four were black. All were high school graduates; the mean educational level was 15.8 years, with a range of 12 to 21 years. Distributions of educational levels were similar for women and men. We used the New Adult Reading Test (NART) (Nelson, 1982) to estimate IQ. This has a high correlation with WARS-R Full Scale IQ (Nelson & McKenna, 1975), and we have found a very high correlation with IQ estimates based on the WAIS-R Vocabulary subtest in a sample of patients with multiple sclerosis. The NART takes only a few minutes to administer, and
I5-Item FOME for Middle-Aged Adults
347
can be completed within the five-minute distractor period which is part of the FOME-15 procedure. Estimated IQ in this sample ranged from 96 to 126, with a mean of 114. This sample was well above average in both IQ and education level; however, it was comparable on these measures to several clinical samples we had recruited over the past few years. While not representative of the population as a whole, it may nevertheless be an appropriate comparison group for clinical samples in many research settings. Procedure
In the original FOME, 10 common household objects (e.g., toothbrush, matchbook) are presented to the subject for identification: first tactually, then visually. In the FOME-15, we increased the number of test items from 10 to 15. Eight items were taken from Form I of the FOME and seven from Form II. We eliminated one of each pair of objects with similar names or related functions. For example, we used “nail” and excluded “nail file”; we used “lock” and eliminated “key.” The 15 items are listed in Table 1. After the presentation, the objects were removed from sight, and the subject was distracted by a 60-second verbal fluency test which involved naming items from a specific semantic category as rapidly as possible. The test-taker was then asked to recall as many stimulus objects as possible, and was selectively reminded of omitted items. To ensure adequate time for recall of the additional items, we increased the recall time limit from 60 to 90 seconds. The sequence (verbal distractor- recall attempt-selective reminding) was repeated a total of five times. The distracters before trials 2 through 5 were additional semantic category, verbal fluency tests, each 30 seconds long. The distractor after the fifth trial lasted for five minutes, during which we administered the NART. A final, unexpected recall attempt then took place, and any items not recalled were presented to the subject in an oral multiple choice format to assess recognition. Further details of test administration are described elsewhere (Fuld, 1977, 1980). We calculated mean scores across trials for each of several measures: retrieval (number of items recalled), storage (cumulative retrieval),’ consistent retrieval (items recalled on two consecutive trials), and recall failure (omission of items on two consecutive trials, despite reminder) (cf. Fuld, tlt should be noted that, in order to retain comparability with the original FOME, we have maintained the original terminology. However, the term “storage,” which refers to all items recalled on any of the first five trials, is actually a misnomer. It is possible for a subject to omit some item on every trial, thus not to be credited with having “stored” it, and yet to recognize it from the multiple choice offered after the delayed recall. We consider that such recognition is evidence of prior storage. Thus, what is scored as storage on this test may in certain cases be an underestimate of actual storage.
348
L. Davenport
et al.
TABLE 1 FOME-15 Test Items Card Nail Matches
Rubber Pencil Bottle
Band
Scissors Can Spoon
Ball Rmg Comb
Toothbrush Button Lock
1977; La Rue, D’Elia, Clark, Spar, & Jarvik, 1986). We also calculated average scores for retrieval on each of the six recall trials. We compared men and women on the summary variables, using multivariate analysis of variance. In addition, we tested for effects of age and education by computing the multiple-correlation between age and the set of summary variables.
RESULTS There were no sex differences and no significant multiple-correlations of the FOME-15 summary variables with either age or educational level. Therefore, we treated the data as coming from a single population and averaged scores from all 32 subjects. Table 2 presents average recall scores for each trial and average scores across all trials for retrieval, storage, consistent retrieval, and recall failure. One-sided 95% confidence intervals (x1.64 SD) are provided as a cut-off score indicative of a deficit relative to this reference group. TABLE 2 All Subjects: Summary Statistics Across Trials, and Recall on Each Trial Separately
Mean
Retrieval Storage Consistent retrieval Recall failweb
Recall Trtal Trial Trial Trial Trral Trial
1 2 3 4 5 6
SD
Lower limit of confidence intervala
Averaged across Trials 12.6 1.2 13.9 0.6 10.8 1.9 0.4 0.6 lndrvidual 10.3 12.5 13.2 13.1 13.8 14.0
Trials 1.7 1.3 1.5 1.8 1.5 1.3
“x-1.64 (SD) bRecall failure is an error score, thus the confidence an upper hnut.
10.7 12.9 1.7 1.3
1.6 10.4 10.7 10.2 11.4 11.8
interval
is
15-Item FOME for Middle-Aged
349
Adults
No subject recalled all 15 items on either of the first two trials. Twentyone subjects (66%) recalled all 15 items at least once on Trials 3 through 5. Only four of the 13 subjects who initially recalled 15 items on Trial 3 or 4 recalled all items on two consecutive trials. Failure to remember all items on two consecutive trials, therefore, should not be interpreted as evidence of impairment. Most subjects (85 X) had stored all items by Trial 3; only one had failed to store every item by Trial 5. Recall failures were unusual; only three subjects (9%) failed to recall as many as three items on two consecutive trials after the SR procedure. On the delayed recall trial, in which subjects were given a multiple choice of items they had omitted, all subjects correctly recognized all items. Failure at this recognition task should, therefore, alert the examiner to the possibility of an impairment in retention over a brief time interval. DISCUSSION
We have adapted the FOME test for use with middle-aged adults by increasing the length of the list to be learned from the standard 10 items to 15 items. This modified version, the FOME-15, eliminates ceiling effects even in healthy adults of average to above-average ability. We present preliminary data on a well-characterized sample of subjects, which will be an appropriate comparison group for many clinical samples. The cut-off scores suggested by our data may be biased toward classifying individuals of average or below-average IQ and educational level as impaired, and therefore should be used cautiously with such individuals. Clearly, it will be important to supplement these preliminary data with scores obtained from a larger sample of subjects who are more representative of the general population. With such a normative data base, the FOME15 may prove to be a valuable tool for investigating pathological and iatrogenie memory changes during middle adulthood. REFERENCES Buschke, H., & Fuld, P. A. (1974). Evaluating storage, retention, memory and learning. Neurology, 24, 1019-1025. Fuld, P. A. (1977). The Fuld ObJecf-memory Evaluation. Chicago:
and retrieval Stoelting
in disordered
Instrument
Com-
pany. Fuld, P. A. (1980). Guaranteed stimulus processing in the evaluation of memory and learning. Cortex, 16, 255-271. La Rue, A., D’Elia, L. F., Clark, E. 0.. Spar, J. E., & Jarvik, L. F. (1986). Clinical tests of memory in dementia, depression, and healthy aging. Journal of Psychology and Agmg, 1, 69-77. Nelson, H. E. (1982). Test Mama/for the Adult Reading Test. Windsor, England: NFER Wilson Publishing Co. Nelson, H. E., & McKenna, P. (1975). The use of current reading ability in the assessment of dementia. British Journal of Social and Clinrcal Psychology, 14, 259-267.