Rhetorical structure and Alzheimer’s disease

ABSTRACT

Background: Language is one of the first faculties afflicted by Alzheimer’s disease (AD). A growing body of work has focussed on leveraging automated analysis of speech to accurately predict the onset of AD. Previous work, however, did not address the effects of AD on the structure of discourse in spontaneous speech and literature.

Aims: Our goal is to identify the effects of AD on the structure of discourse, both in spontaneous speech and in literature.

Methods & Procedures: We use two existing data sets, DementiaBank and the Carolina Conversations Collection, to explore how AD manifests itself in spontaneous speech. This is done by automatically extracting discourse relations according to Rhetorical Structure Theory. We also study written novels, comparing authors with and without dementia using the same tools.

Outcomes & Results: Several discourse relations, especially those involving elaboration and attribution, are significant indicators of AD in speech. Indicators of the disease in written text, by contrast, involve relations of logical contingency.

Conclusions: Our work highlights how AD can alter discourse structures in both spontaneous speech and written text. Future work should combine discourse analysis with previously studied lexico-syntactic features.

KEYWORDS:

• Alzheimer’s disease
• Rhetorical Structure Theory
• computational linguistics
• spontaneous speech
• longitudinal analysis

Acknowledgements

Frank Rudzicz’s contribution is funded by an NSERC Discovery grant (RGPIN 435874) and by a Young Investigator award by the Alzheimer Society of Canada. Graeme Hirst’s contribution is funded by an NSERC Discovery grant. The original acquisition of the DementiaBank data was supported by NIH grants AG005133 and AG003705 to the University of Pittsburgh, and maintenance of the data archive is supported by NIH-NIDCD grant R01-DC008524 to Carnegie Mellon University. The authors thank Vanessa Wei Feng for providing the software tool used in the analysis of RST relations. We acknowledge and thank Professor Ian Lancashire for providing the Ross Macdonald data.

Disclosure statement

Frank Rudzicz is a co-founder of a software company, called WinterLigtht Labs Inc, that commercializes automatic speech-based assessment of cognitive disorder. No intellectual property of that company has been used in this research.

Notes

1. Bottom-up approaches build more abstract structures from smaller ones, starting with the atomic units themselves.

2. http://verbs.colorado.edu/ mpalmer/projects/verbnet.

3. Murdoch and Macdonald constitute the AD group, while James is the CT group in further statistical group analysis. Literary work can involve considerable review and editing, with assistants and editors changing the author’s original writing. However, there is no evidence that this is the case for the authors we consider. Le et al. (2011), drawing on Lancashire (2010), reviews the writing and editing processes of Christie, Murdoch, and James, and conclude that the later novels of each author do not deviate from the author’s earlier practices; in particular, Murdoch allowed no one else to edit her writing at all.

Additional information

Funding

Frank Rudzicz’s contribution is funded by an Natural Sciences and Engineering Research Council of Canada grant [RGPIN 435874] and by a Young Investigator award by the Alzheimer Society of Canada. Graeme Hirst’s contribution is funded by an NSERC Discovery grant. The original acquisition of the DementiaBank data was supported by NIH grants [AG005133 and AG003705] to the University of Pittsburgh, and maintenance of the data archive is supported by NIH-NIDCD grant [R01-DC008524] to Carnegie Mellon University.