Typicality-based semantic treatment for anomia results in multiple levels of generalisation.

The study did not randomly assign participants to receive therapy or no therapy, so the natural history control comparison cannot fully rule out time-related change or repeated testing effects. Some participants did not reach the response benchmark, and the responder and nonresponder analyses were done after looking at treatment response. Five participants received a home exercise program because they were not responding well, so the separate effect of that program cannot be separated from the main therapy. The control comparisons found few significant differences between responders and natural history controls, especially for untrained naming and standardized language tests. The participants had chronic aphasia after left-hemisphere stroke and were selected for stable, low naming accuracy, so the results may not apply to acute recovery or very mild naming problems. The design did not test whether training typical items would produce the same generalization pattern as training atypical items.

The only funding statement supplied says the project was completed under NIH/NIDCD grant 1P50DC012283, with Cynthia Thompson as principal investigator. No author conflicts of interest are reported in the supplied text.

Do not conclude that typicality-based semantic feature analysis produces broad language recovery for all people with aphasia. The gains were strongest in participants who responded to treatment, and the natural history control comparison did not show significant differences for untrained naming or standardized language tests.