Is it time to stop “fishing”? A review of generalisation following aphasia intervention

Abstract

Background: The study of generalisation is integral to both demonstrating and maximising therapy effectiveness. With aphasia therapy aiming to promote change in everyday communication, there is a temptation to “fish” for any evidence of improvement; multiple tasks, elicitation methods, and measures may be used to identify any change that might then be attributed to therapy. Examining the evidence about generalisation will allow us to develop our theoretical understanding of change following aphasia intervention and strengthen our ability to form robust predictions for generalisation that can be tested. Aims: We explore what is meant by generalisation, considering it within the broader context of impact of intervention. We propose a framework focusing on linguistic generalisation at the levels of word, sentence, and connected speech. The discussion draws on the therapy literature for spoken production, exploring single-word therapies for nouns and verbs, sentence production therapies, and discourse therapies. Main Contribution: This article introduces a framework to conceptualise and describe generalisation within level, i.e., change to untreated stimuli within the same linguistic level as the focus of treatment, and across level, i.e., change at a different linguistic level to the focus of treatment. The existing evidence base for spoken production is reviewed using the framework, considering our current knowledge, the predictions we are able to make, and areas for future focus. Conclusions: There is an urgent need for further research in this area. While within-level generalisation has often been addressed systematically, there have been more limited attempts to systematically define, predict, and measure generalisation across linguistic levels. This has resulted in a limited evidence base from which to make decisions regarding when and how to best facilitate generalisation and where and how to best measure generalisation. We propose it is time to stop fishing for change in an opportunistic way and to develop a greater understanding of the relationship between change at the different linguistic levels of words, sentences, and connected speech. We need to use theory and evidence to predict change (selecting the right “pond” within which to fish) and identify valid and reliable ways to measure both targeted and generalised therapy effects (selecting the right “bait”).