Training flexible conceptual retrieval in post-stroke aphasia.

There was no control group, and randomisation or blinding are not reported, so practice, time, and other factors cannot be ruled out. The sample was small, with eleven patients, and the sample size was set by the maximum number available, so group effects may be unstable. Training was brief, six sessions over two to three weeks, and follow-up was immediate, so durability is unknown. Individual analyses had substantially reduced statistical power, and patient responses varied. The supplied text does not report the numerical results for the untrained semantic tasks, so generalisation beyond the training task is unclear. Patients were selected for semantic control deficits and left inferior frontal gyrus damage, so findings may not apply to other aphasia profiles. One patient was using a computerised naming therapy concurrently, but had used it regularly for many years. No part of the study procedures or analyses was pre-registered.

No funding source is stated in the supplied text. The authors declared no competing interests or financial interests.

Do not conclude that this two-week training improves everyday comprehension or untrained associations. The group improved mainly on repeatedly trained items, while novel and untrained items did not show reliable improvement, and individual results were inconsistent.