Identifiable Patterns of Trait, State, and Experience in Chronic Stroke Recovery.

Only 12 participants were analysed, and they differed in age, lesion size, time after stroke, and aphasia severity. There was no control group, so changes over time cannot be separated from normal variation, repeated scanning, or other influences. The therapy produced only small changes in language performance in the larger study, which may have made therapy-related brain differences harder to detect. The classifier used many brain connections and few scans, and the brain regions were large averages that could hide finer changes. The injured left hemisphere did not significantly separate pre- and post-therapy scans, so the finding does not show a simple left-brain change after therapy. The imitation task was underrepresented compared with the other tasks, and the number of scans varied by participant.

The authors declared no conflict of interest. The article's publication types list NIH and other government research support.

Do not read the scan classification results as proof that intensive imitation therapy changed the brain in a clinically meaningful way. The study had 12 participants, no control group, and the injured left hemisphere did not separate pre- versus post-therapy scans.