Analysis of spontaneous, conversational speech in dementia of Alzheimer type: Evaluation of an objective technique for analysing lexical performance

Romola S. Bucks, Sameer Singh, Joanne M. Cuerden, Gordon WilcockAphasiology 2000 · DOI

TL;DR

This paper introduces and validates a set of objective linguistic measures to analyze spontaneous speech in individuals with Alzheimer's disease (DAT). These measures effectively differentiate DAT patients from healthy older adults, showing promise as a diagnostic and prognostic tool for early detection and monitoring of the disease's progression.

What Problem Does This Paper Address?

Diagnosing and tracking Alzheimer's disease (DAT) can be challenging, especially in its early stages. While traditional cognitive tests exist, they can be insensitive to subtle communication deficits and may be difficult for severely affected individuals to complete. Spontaneous, conversational speech offers a more naturalistic window into language function, but analyzing it objectively has been difficult. This paper aims to bridge this gap by developing and evaluating a quantitative, objective method for analyzing lexical performance in spontaneous speech, which could lead to more sensitive diagnostic and prognostic tools.

Methods

The researchers developed a method for analyzing spontaneous conversational speech using eight linguistic measures. These measures fall into two categories:

  • Word Classes: Noun rate (N-rate), Pronoun rate (P-rate), Adjective rate (A-rate), and Verb rate (V-rate), all calculated per 100 words. These measures reflect the frequency of different word types used in speech.
  • Lexical Richness: Type Token Ratio (TTR), Brunét's Index (W), and Honoré's Statistic (R). These measures assess the diversity and sophistication of vocabulary used.

Additionally, a measure of semantic cohesion, the Clause-like Semantic Unit (CSU) rate, was included to assess the ability to form meaningful word clusters.

The speech samples were transcribed, and linguistic analyses were performed using computer programs. Statistical analyses, including Mann-Whitney U tests, Principal Components Analysis (PCA), and discriminant function analysis with cross-validation, were used to compare DAT patients and healthy older controls.

Datasets

The study involved two groups:

  • Participants with Dementia of Alzheimer Type (DAT): 8 individuals (4 female, 4 male) with mild to moderate-severe cognitive impairment, diagnosed according to DSM-III-R and NINCDS-ADRDA criteria.
  • Healthy Older Controls (HO): 16 individuals (7 male, 9 female) who were not suffering from any neurological or psychological illness.

The participants were interviewed using a semi-structured format, and their speech was recorded and transcribed to collect at least 1000 words of conversation per participant.

Key Insights and Ideas

  • Spontaneous speech is a rich source of information: The study highlights that analyzing natural conversation can reveal significant differences in language use between DAT patients and healthy individuals, which might be missed by traditional tests.
  • Lexical deficits are prominent in DAT: The findings strongly suggest that DAT is associated with significant changes in lexical performance, particularly a reduction in noun usage.
  • Objective measures are key: The development of objective, quantitative linguistic measures is crucial for reliable and sensitive assessment of language changes in DAT.
  • Differentiation is possible: The proposed linguistic measures can effectively discriminate between DAT patients and healthy older adults, paving the way for improved diagnostic tools.

Key Results

  • Significant differences found: DAT participants showed significantly lower noun rates (N-rate) and higher pronoun rates (P-rate), adjective rates (A-rate), and verb rates (V-rate) compared to healthy older controls.
  • Lexical richness: Healthy older adults generally exhibited richer speech across all three lexical richness measures (TTR, Brunét's W, and Honoré's R), indicating a less diverse vocabulary in DAT patients.
  • CSU-rate similarity: Interestingly, there was no significant difference in the Clause-like Semantic Unit (CSU) rate between the two groups, suggesting that DAT patients did not differ significantly in their ability to form noun and verb phrases.
  • High classification accuracy: Discriminant function analysis, even with cross-validation, achieved an 87.5% correct classification rate in distinguishing between DAT and healthy participants. The most important discriminators were N-rate, P-rate, and Brunét's Index (W).

Limitations and Future Work

  • Sample size: The study acknowledges that small sample sizes prevented statistical significance from being achieved for some correlations between demographic variables and linguistic measures in DAT participants.
  • Normative data: The authors emphasize the need for additional normative data from a larger sample of healthy older adults to explore the sensitivity of these measures to very early language symptoms in DAT and other dementing disorders.
  • Inter-rater reliability: While objective measures were used, the study did not assess inter-rater or intra-rater reliability, though it is anticipated to be minimal due to the objective nature of the measures.
  • Further exploration: Future work could investigate whether noun rate differences apply to all noun classes, explore the relationship between these measures and semantic memory, and apply these techniques to other types of dementia. The authors also suggest exploring gender differences in DAT language.

Summary updated

Abstract

Spontaneous, conversational speech in probable dementia of Alzheimer type (DAT) participants and healthy older controls was analysed using eight linguistic measures. These were evaluated for their usefulness in discriminating between healthy and demented individuals. The measures were; noun rate, pronoun rate, verb rate, adjective rate, clause-like semantic unit rate (all per 100 words), including three lexical richness measures; type token ratio (TTR), Brunet's Index (W) and Honore's statistic (R). Results suggest that these measures offer a sensitive method of assessing spontaneous speech output in DAT. Comparison between DAT and healthy older participants demonstrates that these measures discriminate well between these groups. This method shows promise as a diagnostic and prognostic tool, and as a measure for use in clinical trials. Further validation in a large sample of patient versus control 'norms' in addition to evaluation in other types of dementia is considered.
BibTeX
@inproceedings{bucks2000analysis,
  author = {Romola S. Bucks and Sameer Singh and Joanne M. Cuerden and Gordon Wilcock},
  title = {{Analysis of spontaneous, conversational speech in dementia of Alzheimer type: Evaluation of an objective technique for analysing lexical performance}},
  booktitle = {Aphasiology},
  year = {2000},
  doi = {10.1080/026870300401603},
}
Figure