Acquisition and generalization responses in aphasia treatment: Evidence from sentence-production treatment

Dickey, Michael Walsh and Hula, William and Yoo, HyunSoo (2014) Acquisition and generalization responses in aphasia treatment: Evidence from sentence-production treatment. [Clinical Aphasiology Paper]

PDF 536-1014-1-RV(Dickey-Hula-Yoo).pdf Download (106kB)

Abstract

Treatment of Underlying Forms (TUF) promotes not only acquisition of treated sentence types but also generalization to related but untreated sentences (e.g., Thompson, Shapiro, Kiran & Sobecks, 2003). In a meta-analysis examining TUF treatment outcomes, Dickey and Yoo (2010) found evidence that the factors governing TUF acquisition and generalization may be different. They found that general auditory comprehension ability but not overall aphasia severity or sentence-comprehension impairment predicted participants’ acquisition of treated sentences. In contrast, none of these factors were related to participants’ generalization to related but untreated sentences. Interestingly, Meinzer and colleagues (2010) found similar results for naming treatment: brain areas that were positively related to acquiring treated items were not associated with generalization to untreated words. These findings suggest that the mechanisms responsible for acquisition and generalization responses to aphasia treatment may be distinct. The current study examined this question further by testing the dose-response relationships for TUF, for both acquisition and generalization. It analyzed existing TUF treatment studies by using multilevel generalized linear regression to model changes in probe accuracy over the course of treatment. One model estimated the slope and intercept of acquisition and generalization curves in response to increasing amounts of treatment. A second set of models tested whether these dose-response relationships were moderated by aphasia severity (viz. Dickey & Yoo, 2010). Determining whether acquisition and generalization curves exhibit similar slopes and intercepts, and whether they are moderated by the same factors, will help establish how similar the two treatment responses are. Comparing the slopes and intercepts of these curves can also shed light on whether similar amounts of treatment are needed to promote acquisition and generalization.

Actions (login required)

View Item