A Randomized Comparative Effectiveness Trial of Many-Features and Few-Features Variants of Semantic Feature Analysis...

Semantic feature analysis (SFA) is a well-established and well-studied treatment for naming impairment in aphasia. Both theory and correlational evidence suggest that feature generation is a key active ingredient of the treatment. The purpose of this study was to test whether the number of features generated by participants in SFA is causally related to outcomes.

Forty-one individuals with aphasia due to left-hemisphere stroke of at least 6 months postonset were randomized in a 1:1 ratio to receive either a many-features variant of SFA in which they were asked to generate 11 features per trial or a few-features variant in which they were asked to generate five features per trial. The entire study team, except for three treating clinicians, was blind to group assignment. Random group assignments were stratified by aphasia severity, blocked in groups of four, generated a priori by a team-external consultant, and managed by a treating clinician. One participant dropped out of the study following randomization, and the remaining participants were assessed twice prior to receiving 42 hr of SFA over 3 weeks, at treatment exit, and at 2- to 3-month follow-up. Primary outcomes were naming accuracy for treated and untreated, semantically related objects.

On average, participants demonstrated large and statistically reliable increases in accuracy for treated items and reliable and small increases for untreated items. Assignment to many-features or few-features treatment did not robustly affect any outcomes.

Active manipulation of the number of features per trial that SFA participants are required to generate, within the range of 5-11, does not causally influence treatment outcomes.

https://doi.org/10.23641/asha.33936256.