Abstract
Previous studies have suggested that presenting to-be-memorised lyrics in a singing mode, instead of a speaking mode, may facilitate learning and retention in normal adults. In this study, seven healthy older adults and eight participants with mild Alzheimer's disease (AD) learned and memorised lyrics that were either sung or spoken. We measured the percentage of words recalled from these lyrics immediately and after 10 minutes. Moreover, in AD participants, we tested the effect of successive learning episodes for one spoken and one sung excerpt, as well as long-term retention after a four week delay. Sung conditions did not influence lyrics recall in immediate recall but increased delayed recall for both groups. In AD, learning slopes for sung and spoken lyrics did not show a significant difference across successive learning episodes. However, sung lyrics showed a slight advantage over spoken ones after a four week delay. These results suggest that singing may increase the load of initial learning but improve long-term retention of newly acquired verbal information. We further propose some recommendations on how to maximise these effects and make them relevant for therapeutic applications.
Notes
1We also analysed the size of chunks of words recalled. There was no clear influence of conditions in initial learning. In repeated learning, AD produced larger chunks in the Sung condition compared to the Spoken one and in delayed recall only (Z = 2.41, p < .05).
2Because of the adaptive procedure, the task stops or goes on depending on the number of words recalled. A problem resulting from this procedure is that the three other measures (sung words, size of chunk, and intrusion errors) depend on the number of words recalled. For example, when a participant learns more lines, the task becomes more difficult and he has more chances to make mistakes with word position; in that case, his score of intrusions will be high because overall, he recalled more words and had more lines to learn than others. To avoid this contamination of measures and control for performance level, we only adjusted performance for the first three trials (the learning of the first four lines) that were always completed by all participants.
3Note that results of S1 (initial learning) can differ slightly from the previous results of initial learning because scores were recalculated with the 6/8 AD participants who took part in repeated learning.