Cross-Modality Treatment Generalization in Sentence Comprehension in Aphasia: Evidence From Structural Priming and Webcam-Based Eye-Tracking.

Small samples: 24 people with aphasia and 16 control participants. Comprehension accuracy varied a lot between individuals, with 14 people with aphasia improving at one of the two post-tests and 10 showing little or no change; the authors say this variability is probably why the group-level accuracy result was null. Some improvement appeared on the control locative questions as well, which also got faster from baseline to the delayed post-test, so part of the speed-up may reflect repeated exposure to the task rather than the training. The authors argue the training-condition effect was larger on the target questions than on the controls. Participants were not randomly allocated to training condition; they were assigned to one of four balanced combinations of structure order and same-verb versus different-verb training. Only up to six training sessions were given, and the authors say this may not have been enough to change comprehension accuracy. The comprehension questions were hard: passive and double-object sentences are difficult for people with aphasia to understand, which the authors suggest may have held accuracy down. Webcam eye-tracking is less spatially accurate than laboratory equipment, calibration failed for some people with aphasia, and data were missing: 8 of 104 sessions for people with aphasia produced behavioural data only, no data at all were recorded in six sessions for people with aphasia and three sessions for control participants. Everyone had relatively preserved single-word and simple-sentence processing, so the results may not extend to people with more severe aphasia.

The authors declare no conflicts of interest. The work was funded by the National Institute on Deafness and Other Communication Disorders of the National Institutes of Health.

Do not read the faster response times and earlier eye movements as proof that this training improves how well people with aphasia understand sentences. The comprehension accuracy measure did not improve at either post-test, the control locative questions also got faster over the same period, and there was no untrained comparison group of people with aphasia, so part of the change could reflect practice with the task.