A quantitative approach to analysing conversational turn-taking in Alzheimer's disease and frontotemporal dementia

Abstract

Communicating and understanding words require a wide range of cognitive resources involving phonological, lexical, semantic and perceptual processing. Disturbance in these domains is associated with pathology in temporal and/or frontal cortical areas. The subsequent abnormal breakdown in fluent speech can be observed in frontotemporal dementia or Alzheimer’s disease, in which brain atrophy in these areas is common. Organization and interaction during a conversation depend, additionally, on the ability to plan and exchange information. Such executive resources are required for structuring, planning, monitoring and organizing speech output, and are impaired when critical networks in dorsolateral prefrontal brain regions become disrupted.

The aim of this study was to identify speech parameters, to characterise and describe the subgroups of dementia which could be used to improve the correct identification of dementia types at an early stage of the disease. Our analysis was based on the interactional organization of conversations and overlapping behaviour. As a new approach in speech analysis of dementia, we aimed to develop a method to quantify traditionally qualitative measurements.

We carried out two studies: In the first we recorded conversations elicited using a map task in healthy subjects paired with a familiar or an unfamiliar partner. We aimed to identify candidate parameters for characterising conversational behaviour and to investigate the existence of a Familiarity effect. Twenty-four conversations were recorded: twelve between familiar pairs and twelve between unfamiliar pairs. Conversations were analysed for overlap behaviour in the categories Confirmations, Predictions, Full Turn-Taking, Failed Turn-Taking, Failed Turn-Taking Completed and Others.

In the second we recorded nine Alzheimer’s disease patients and three frontotemporal dementia patients using the same methodology of pairing, conversation stimulation and analysis as for the healthy subjects. The familiar partner was a family member or carer and the unfamiliar was a research assistant. For the patients recording sessions were repeated every three to eight months over a period of a year.

Analysis showed that the chosen experimental design and the parameters describing planning and executive abilities in speech are promising in terms of supporting our main research question on the subject of finding characteristic conversational behaviour in subtypes of dementia. We also found indicators for the Familiarity effect in healthy and dementia speech. The case studies, which we conducted for three FTD patients, allowed providing an insight in to certain turn-taking behaviour by using examples of the transcripts. Again, we found that Familiarity is a promising indicator in conjunction with the categorical analysis for characterising FTD speech.

We demonstrated changes in the underlying structure of Alzheimer’s disease and frontotemporal dementia speech with respect to the frequency of overlaps and outlined speech strategies of the patients to bypass cognitive limitations over time.

More information

Identifiers

ORCID for Anna Barney: orcid.org/0000-0002-6034-1478

Catalogue record

Export record