Episodic memory performance is hindered as a result of increased age and age related diseases, such as Alzheimer’s (Craik & Jennings, 1992; Nebes, 1992). Recently, the trend in studying memory and ageing has turned towards measures of veridical memory (i.e., memory for information that was presented) and measures of false memory (i.e., memory for information that was not presented).

The present study used three variations of DRM word lists: a pure semantic list, a pure phonological list, or a hybrid list of both semantically and phonologically related words. In addition, five different types of participants were analyzed. Three of those dimensions were on age: a young healthy group, a young-old healthy group, and a old-old healthy group. The remaining two categories of participants consisted of individuals with Alzheimer’s disease, varying on two levels: a very mild dementia group and a mild dementia group.

The purpose of the present study was to gain a better understanding of memory relating to age and memory performance of individuals with dementia of the Alzheimer type (DAT). Furthermore, memory performance on hybrid lists was examined and related to performance on semantic and phonological lists.

Participants

A total of 182 participants took part in the present study. Five categories of participant types were formed, three relating to age and two relating to Alzheimer’s. As divided by age group, the young healthy group consisted of 37 young adults (M=19.3 years; range 17-22 years), the young-old healthy group consisted of 46 older adults on the younger side of the scale (M=71.8 years; range 57-80 years), and the old-old healthy group consisted of 38 older adults on the older side of the scale (M=86.3 years; range 80-96 years). For individuals with DAT, a score of 0.5 on the CDR (Washington University Clinical Dementia Rating scale) classified them as having very mild dementia and a score of 1.0 classified them as having mild dementia.

Materials

In the present study six critical items were used: cold, dog, glass, gun, hard, and smoke. 12 items were presented per list. For example, the semantic list consisted of words such as hound, puppy, Lassie, mutt, pet, beware, etc. which all related to the critical non-presented item dog. The phonological list consisted of words such as log, doll, dug, frog, bog, dig, etc. which all related to the same critical non-presented item dog. Lastly, the hybrid list consisted of words such as hound, doll, Lassie, frog, pet, dig, etc. which all related to the same critical non-presented item dog. Materials were counterbalanced across participants.

Procedure

Participants were individually tested and were given a total of six lists: two pure semantic lists, two pure phonological lists, and two hybrid lists. Presentation order of lists was counterbalanced in blocks of three of each type.

Participants were told to verbally read the word in a list as it was presented to them on a computer monitor. At the beginning of each trail, a fixation point ( + ) appeared in the middle of the computer monitor and then the screen went blank. The blank screen lasted for 200-ms then a word in the list appeared in lowercase letters in the middle of the screen. The word was presented for 1,500 ms and then the screen went blank again for another 200-ms. This cycle was repeated for every word in a list. After presentation of all 12 words, the fixiation point ( + ) reappeared and the participants were asked to recall as many words as they could from the list that was just presented. The participants were given as much time as they needed and then the cycle started all over again for another word list until all 6 word lists had been presented.

All of the healthy older adults (young-old and old-old) and the individuals with DAT (very mild dementia and mild dementia) also participated in a 2-hr battery of psychometric tests designed to assess psychological functions including language, memory, and intelligence. These results showed clear changes across groups on these cognitive measures and supported the usefulness of categorizing different age groups. Correlations between veridical recall, false recall, and psychometric performance produced no systematic patterns of correlation (See Table 1, pg. 257).

Results

Several ANOVAs of different combinations of independent variables were conducted. It was found that overall, correct recall showed a serial curve as a function of group (See Figure 1, pg. 258). There was an effect of age and of DAT, which occurred most strongly for the primacy and middle serial positions. Furthermore, the semantic lists produced higher recall performance primarily at the primacy positions. In addition, both veridical and false critical recall as a function of group was analyzed (See Figure 2, pg. 260). The results indicated that there were differential effects of age and DAT on veridical recall and false recall.

In addition to these standard analyses, three other comparisons were conducted. One was an analysis of the recall position of the critical nonpresented items, such that this was calculated by dividing the output position of the critical nonpresented item by the total number of items recalled on that list (See Table 2, pg. 261). It was found that the relative output positions of the critical nonpresented words were about two thirds into the recalled items, independent of group, type of list, and number of items recalled. This pattern is consistent with the analyses reported by others (e.g. Balota, Cortese, et al., 1999; McEvoy, Nelson, & Komatsu, 1999; Roediger & McDermott, 1995).

The second additional analysis included a coding system for the possible intrusions of noncritical items. The results were coded on nine different types of intrusions: 1) semantically related but noncritical intrusions; 2) phonologically related but noncritical intrusions; 3) cross-list intrusions from previous semantic lists; 4) cross-list intrusions from previous phonological lists; 5) cross-list intrusions from previous hybrid lists; 6) cross-list critical item intrusions from previous semantic lists; 7) cross-list critical item intrusions from previous phonological lists; 8) cross-list critical item intrusions from previous hybrid lists; 9) other, apparently unrelated, intrusions (See Table 3, pg. 261).

It is important to note that participants produced very few noncritical intrusions per list and overall there was a very low level of intrusions. Hence, the figures on Table 3 should be interpreted with caution.

The third additional analysis was conducted on mean naming response latency during the encoding task. Words from the phonological lists took longer to name (868 ms) than words from either hybrid or semantic lists (843 ms and 840 ms, respectively). In addition, response latencies increased as a function of increased age and increasing DAT.

Discussion

This experiment yielded three main findings.

1. Veridical recall decreased in healthy older adults as compared with healthy young adults, whereas false recall increased as a function of age.
2. Veridical recall decreased in individuals with DAT as compared with age-matched healthy older adults, whereas false recall remained fairly stable.
3. Compared with pure semantic and pure phonological lists, hybrid lists produced higher false recall of the critical nonpresented word for all groups (See Figure 3, pg. 263).

The author argues that the present results support the idea of breakdowns in attentional control systems that select among activated pathways due to spreading activation and item-specific information due to memory encoding. In addition, the present study shows that there is a clear sensitivity to multiple phonological and semantic manipulations (even in individuals with mild DAT). Finally, the present study shows a superadditive effect (i.e. a higher increase of false memories as compared to a pure semantic or phonological list) regarding the hybrid lists.