Drawing pictures at encoding enhances memory in healthy older adults and in individuals with probable dementia

ABSTRACT

We explored the efficacy of drawing pictures as an encoding strategy to enhance memory performance in healthy older adults and individuals with probable dementia. In an incidental encoding phase, participants were asked to either draw a picture or write out each word from a set of 30 common nouns for 40 seconds each. Episodic memory for the target words was compared in a group of healthy older adults to individuals with probable dementia (MMSE/MOCA range 4 to 25). In two experiments we showed that recall and recognition performance was higher for words that were drawn than written out during encoding, for both participant groups. We suggest that incorporating visuo-perceptual information into memory enhanced performance by increasing reliance on visual-sensory brain regions, which are relatively intact in these populations. Our findings demonstrate that drawing is a valuable technique leading to measurable gains in memory performance for individuals with probable dementia.

KEYWORDS:

• Episodic memory
• aging
• dementia
• drawing
• encoding tasks

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Four WRAP control participants scored below the established cutoff scores on the cognitive screening measures; their data were excluded and replaced with data from additional WRAP participants who scored in the cognitively healthy range.

2. Given that individuals with MCI or dementia typically have poorer overall memory than healthy older adults, it is difficult to determine the relative memorial benefit that drawing provides each age group. Given this, we also calculated the proportion of drawn and written words recalled by dividing the number of each by the total number of words recalled by each individual participant. For recognition, the number of words correctly recognized was divided by the total number of each word type (15 words for each type) in the recognition test. By taking this approach we could determine the relative benefit from each encoding trial type, and make comparisons across our groups. See Table 1 for means. As expected, the analysis revealed a main effect of Encoding trial type, F (1, 24) = 69.63, MSE = 5.00, p < 0.001, η² = 0.74, such that drawn words were better remembered than written, but no main effect of Group, F (1, 24) = 3.60, MSE = 0.17, p = 0.07, η² = 0.13, nor a Group X Encoding type interaction, F (1, 24) = 0.39, MSE = 0.03, p = 0.54, η² = 0.02.

3. We also analyzed the proportion of words recognized of Word Type in a 2 Group (Probable Dementia and Controls) X Word Type (Drawing, Writing, or Lure) mixed ANOVA (See Table 1 for means). As expected, the analysis revealed a main effect of Word Type, F (1, 48) = 53.77, MSE = 2.77, p < 0.001, η² = 0.69, such that a larger proportion of drawn then written words were recognized, but no main effect of Group, F (1, 24) = 0.43, MSE = 0.05, p = 0.52, η² = 0.02. Additionally, a Group X Word Type interaction was found, F (1, 48) = 8.48, MSE = 0.44, p = 0.001, η² = 0.26. To better understand the interaction, we performed two separate repeated measures ANOVAs, one for each of our participant groups, to compare memory for each word type. For the control group, the analysis revealed differences in the proportion of memory endorsements made between word types, F (1, 24) = 64.98, MSE = 2.68, p < 0.001, η² = 0.84. Paired samples t-tests revealed that a larger proportion of drawn words were recognized than written words, t (12) = 4.21, p = 0.001, and lures, t (12) = 11.71, p < 0.001, and a larger proportion of written words were recognized compared to lures, t (12) = 6.68, p < 0.001. For participants with probable dementia, the analysis revealed differences in the proportion of memory endorsements between word types, F (1, 24) = 8.53, MSE = 0.53, p = 0.002, η² = 0.42. Paired samples t-tests revealed no difference between the proportion of recognized drawn and written words, t (12) = 1.05, p = 0.32. However, a larger proportion of drawn words were recognized than lures, t (12) = 2.91, p = 0.01, and written words than lures, t (12) = 3.56, p = 0.004.

4. Three individuals with probable dementia scored above the established cutoff scores on the cognitive screening measures so were excluded from analysis and replaced by participants who scored in the cognitively impaired range.

5. Two WRAP control participants scored below the established cutoff scores on the cognitive screening measures; their data were excluded and replaced with data from additional WRAP participants who scored in the cognitively healthy range.

6. The proportion of words recalled from each encoding trial type was analyzed in a 2 Group (Probable Dementia and Controls) X 2 Encoding Trial Type (Drawing and Writing) mixed ANOVA (See Table 1 for means). As expected, the analysis revealed a main effect of Encoding trial type, F (1, 28) = 26.09, MSE = 2.48, p < 0.001, η² = 0.48, such that drawn words were better remembered than written, but no main effect of Group, F (1, 28) = 2.15, MSE = 0.07, p = 0.15, η² = 0.07, or Group X Encoding type interaction, F (1, 28) = 0.32, MSE = 0.03, p = 0.58, η² = 0.01.

7. The proportion of words recognized of each Word Type was analyzed in a 2 Group (Probable Dementia and Controls) X 2 Word Type (Drawing, Writing, or Lure) mixed ANOVA (See Table 1 for means). As expected, the analysis revealed a main effect of Word Type, F (1, 56) = 197.88, MSE = 4.53, p < 0.001, η² = 0.88, such that drawn words were better remembered than written, as well as a main effect of Group, F (1, 28) = 1.17, MSE = 0.08, p < 0.001, η² = 0.29, such that controls recognized a larger proportion of words than individuals with probable dementia, and a Group X Word Type interaction, F (1, 56) = 5.17, MSE = 0.12, p = 0.009, η² = 0.16. To better understand the interaction, we performed two separate repeated measures ANOVAs, one for each of our participant groups, to compare memory for each word type. For the control group, the analysis revealed differences in the proportion of memory endorsements made between word types, F (1, 28) = 204.44, MSE = 3.05, p < 0.001, η² = 0.94. Paired samples t-tests revealed that a larger proportion of drawn words were recognized than written words, t (14) = 7.55, p = 0.001, or lures, t (14) = 31.98, p < 0.001, and for written words compared to lures, t (14) = 10.03, p < 0.001. For participants with probable dementia, the analysis revealed differences in the proportion of memory endorsements between word types, F (1, 28) = 51.69, MSE = 1.59, p < 0.001, η² = 0.79. Paired samples t-tests revealed a larger proportion drawn words were recognized than written words, t (14) = 5.19, p < 0.001, or lures, t (14) = 8.17, p < 0.001, and a larger proportion written words were recognized than lures, t (14) = 6.48, p < 0.001.

Additional information

Funding

This work was supported by the Natural Sciences and Engineering Research Council of Canada [Discovery].