Episodic versus semantic memory impairments and deficits in discourse production

Abstract

Event Abstract Back to Event Episodic versus semantic memory impairments and deficits in discourse production Bruna Seixas Lima1, 2*, Kelly Murphy3, 4, Angela Troyer3, 4, Brian Levine3, 5, Naida Graham1, 2, Carol Leonard1, 6, David Tang-Wai7, Sandra E. Black8, 9, 10, 11 and Elizabeth Rochon1, 2 1 University of Toronto, Rehabilitation Sciences Institute, Canada 2 Toronto Rehabilitaion Institute, Canada 3 University of Toronto, Department of Psychology, Canada 4 Baycrest Health Sciences, Neuropsychology and Cognitive Health, Canada 5 Baycrest Centre, Rotman Research Institute, Canada 6 University of Ottawa, Audiology and Speech-Language Pathology Program, Canada 7 University Health Network Memory Clinic, Department of Medicine, Canada 8 Sunnybrook Health Sciences, Canada 9 University of Toronto, Institute of Medical Science, Canada 10 Sunnybrook Health Sciences, L.C. Campbell Cognitive Neurology Research Unit, Canada 11 Heart and Stroke Foundation, Partnership for Stroke Recovery, Canada Discourse production requires organization, integration and coordination of linguistic units. These processes involve cognitive skills other than language, such as memory (Glosser & Deser, 1990). Neurodegenerative syndromes commonly affect autobiographical memory recall and often are accompanied by linguistic disruption. Semantic variant primary progressive aphasia (svPPA) is a disorder which affects the semantic memory system and whose most prominent symptom is aphasia (Gorno-Tempini et al., 2011). In mild cognitive impairment (MCI) there is greater than expected cognitive decline in comparison to healthy-aging, with high rates of progression to Alzheimer’s disease (AD) (Artero et al., 2006). Language impairment observed in MCI and AD patients includes impairments in verbal fluency, naming and discourse production (Murphy et al., 2006; Henry et al., 2004; Harris et al., 2008). Semantic memory (general knowledge) is known to be affected in svPPA but appears intact in amnestic MCI (aMCI). By contrast, episodic memory (memories of specific events) is impaired in aMCI, but spared in svPPA (Murphy et al., 2008, Moss et al., 2003). These two types of memory are necessary for autobiographical memory recall. This study aims to investigate discourse coherence and autobiographical memory recall in svPPA and aMCI patients. Given the impairments observed in these populations and the intrinsic relationship between language and memory, we predicted the discourse production of these groups would be differentially affected when recalling information related to the type of memory affected in each syndrome. To date, we have analysed the speech of 18 svPPA patients and 18 healthy controls matched for age and education, and the speech of five aMCI patients and five matched controls. Discourse was elicited by asking participants to recall autobiographical events. Memory recall was evaluated using the Autobiographical Interview (Levine et al., 2002). Narratives were transcribed and segmented into utterances which were classified as semantic or episodic details. In addition, a coherence rating scale was used to analyse the relevance and informativeness of these details to the topic of discourse. Details received a score from 0 to 3. The weighted average score was calculated for the whole narrative and for the semantic and episodic composites separately. Independent samples (corrected for multiple comparisons) t-tests showed that aMCI patients produced shorter narratives than controls (p<.001), while the linguistic output of svPPA patients was comparable to that of the controls. Both svPPA (p<.001) and aMCI groups (p<.001) produced fewer episodic details than controls, while producing a comparable number of semantic details. Interestingly, svPPA patients presented with deficits in overall coherence (p<.001), which were not observed in aMCI patients. Additionally, the deficit in coherence of the svPPA patients was observed when producing semantic (p<.001) but not episodic details. These preliminary results suggest that impairment in the semantic memory system may be linked to reduced discourse coherence. Ongoing analyses will incorporate more aMCI participants. We expect this study will aid in the understanding of the relationship between memory and linguistic processes and may support the development of new diagnostic tools for cognitively impaired individuals. Acknowledgements This work was supported by the Canadian Institutes of Health Research (grant numbers 82744 and 130462).