‘A IS FOR APPEL?’: IMPROVING WRITTEN WORD RETRIEVAL IN SEMANTIC DEMENTIA THROUGH WORD RETRAINING

Abstract

Semantic dementia (SD), the semantic variant of primary progressive aphasia, is a dementia syndrome which significantly impacts word knowledge. This results in marked deficits in naming and understanding words, but also can affect written word production or spelling. While improvements in verbal word retrieval have been demonstrated through word retraining, research studies have not yet focused on written word retrieval, despite this having been identified as a possible area of remediation. The aim of this study was to investigate potential benefits of word retraining on written word retrieval in patients with SD. Data was analysed from 8 patients with SD who completed an individually tailored 2-month word training program. Participants engaged in a “Look, Listen, Repeat” practice, viewing photographs of everyday objects and hearing and seeing word labels on their home computer. Performance on both verbal and written word retrieval was assessed at baseline, immediate post-intervention, and at a 2-month follow up. Written word accuracy was scored either as correct if the word provided was spelled correctly, or otherwise as incorrect. Incorrect responses were then classified by error type (‘don't know/ no written response’, ‘correct word retrieved but spelt incorrectly’, ‘wrong word retrieved – semantic error’, ‘wrong word retrieved – other error’, ‘neologistic response’). All patients improved both in their ability to retrieve words verbally and in written form immediately post-intervention (p<0.001). For the majority of participants, this change reflected an improvement from no written response to the correct response. In patients with milder semantic impairments, up to 100% of newly relearned words were spelt correctly at immediate post-intervention and were well maintained at the 2-month follow up. Written retrieval was less accurate for those with severe semantic impairments, however, the majority of newly relearned words were spelt correctly by these patients (57% accuracy). Words untrained did not change over the period of the study. Patients with semantic impairments can show successful relearning of written words. While spelling errors are more likely to occur for patients with severe deficits, accurate relearning of written labels is still possible. Future word retraining programs should include both verbal and written practice.