Slide Slam A12
Online spaced-repetition training for treating word-finding difficulties in aphasia
John de Grosbois1, Lathushikka Canthiya1,2, Aaron Phillipp-Muller1,2, Natasha Hickey1, Benazir Hodzic-Santor1,2, Jed Meltzer1,2; 1Baycrest Health Sciences, 2University of Toronto
Aphasia is a language processing deficit associated with left-hemisphere damage, which consistently results in an inability to produce known words. A common treatment approach is to repeatedly drill items in a picture-naming task. Although such training is relatively easy to administer and score, it is time-consuming and perhaps best suited for self-administration outside of therapy hours. Accordingly, efforts have been made to improve the efficiency of this type of training by making use of specific, item-scheduling strategies such as spaced-repetition, in which the time between presentations of the same item is manipulated. The purpose of the current study was two-fold. First, this study evaluated the potential utility of online picture-naming training performed without a therapist present. And second, the relative effectiveness of three unique spaced-repetition training schedules were contrasted both to each other, and to an untrained control condition. Twenty-two individuals, 18 of whom suffered from stroke-induced aphasia (four with primary progressive aphasia) were recruited. Each participant first completed a pretest over ZOOM in which 80 candidate pictures were selected for training from a pool of 292 pictures using a picture-naming task. The 80 pictures were then subdivided into four conditions, controlling for both word-frequency and the number of syllables. Sixty of the pictures (i.e., 3 conditions) were used in the training portion of the study, and the remaining twenty served as a control condition. The training portion of the study was conducted over two weeks for each participant. That is, each participant trained for 10-days (i.e., Monday-to-Friday). Each day, participants completed between 30 and 60 minutes of picture naming practice via an online learning portal. On each training trial, participants attempted to name a presented picture aloud, were given the answer, and self-scored. Each day, all three training conditions were practiced in a randomized order. The first training condition was a traditional, randomized “large-deck shuffling” of 20 items. The second training conditions was a “small-deck shuffling” approach (i.e., 5 items each, in 4 groupings) that resulted in a relatively closer average spacing of item repeats. And the third training condition was an adaptive algorithm, in which items correctly identified were repeated less frequently over time. Post-training teleconference evaluations were completed for all 80 selected pictures both at the end of the training period, and four weeks later. Performance was quantified on both response-accuracy and response-time. Overall, the training was successful at alleviating word-finding difficulties in the participants, with trained pictures showing higher accuracy and a faster verbal reaction time relative to the untrained control items. Also, the ‘small-deck’ condition under-performed on response accuracy relative to the two other training conditions (i.e., ‘big-deck’ and ‘adaptive’). Further, the ‘adaptive’ condition showed a modest decease in response-time, relative to the ‘big-deck’ condition overall. Thus, remote picture naming training is effective for treating word-finding difficulties in aphasia, and continued use of adaptive scheduling algorithms may maximize both response accuracy and efficiency.