Using Generative Artificial Intelligence to Classify Primary Progressive Aphasia from Connected Speech

Abstract

Neurodegenerative dementia syndromes, such as Primary Progressive Aphasias (PPA), have traditionally been diagnosed based in part on verbal and nonverbal cognitive profiles. Debate continues about whether PPA is best subdivided into three variants and also regarding the most distinctive linguistic features for classifying PPA variants. In this study, we harnessed the capabilities of artificial intelligence (AI) and natural language processing (NLP) to first perform unsupervised classification of concise, connected speech samples from 78 PPA patients. Large Language Models discerned three distinct PPA clusters, with 88.5% agreement with independent clinical diagnoses. Patterns of cortical atrophy of three data-driven clusters corresponded to the localization in the clinical diagnostic criteria. We then used NLP to identify linguistic features that best dissociate the three PPA variants. Seventeen features emerged as most valuable for this purpose, including the observation that separating verbs into high and low-frequency types significantly improves classification accuracy. Using these linguistic features derived from the analysis of brief connected speech samples, we developed a classifier that achieved 97.9% accuracy in predicting PPA subtypes and healthy controls. Our findings provide pivotal insights for refining early-stage dementia diagnosis, deepening our understanding of the characteristics of these neurodegenerative phenotypes and the neurobiology of language processing, and enhancing diagnostic evaluation accuracy.One sentence summaryComputational linguistic analyses of naturalistic speech samples can classify the aphasic variant of patients similarly to expert clinicians and identify well-established and novel linguistic features crucial for classification.

Using explainable artificial intelligence to identify linguistic biomarkers of amyloid pathology in primary progressive aphasia

“…This interpretation is consistent with the finding by Slegers et al, 2021 who reported verb frequency as a distinctive feature between amyloid positive and negative classes. 8 In our recent work, we also showed that verb frequency is a variable that significantly increases classification accuracy among the three variants of PPA 44 .…”

Language Abnormalities in Alzheimer’s Disease Arise from Reduced Informativeness: A Cross-Linguistic Study in English and Persian

“…We have previously provided a list of definitions and examples for features extracted from Stanza. 37 In addition to the Stanza-derived feature set, we included word length, sentence length, average log frequency of all words and content words, total number of sentences, total number of words, and syntax frequency. 35…”

Impaired language in Alzheimer’s disease: A comparison between English and Persian implicates content-word frequency rather than the noun-verb distinction

“…English features a group of especially high-frequency verbs, such as "take" and "get," which may emerge as a distinct subset within the broader category of verbs upon cluster analyses based on frequency. 53 These high-frequency verbs possess abstract and flexible meanings as illustrated by phrases like "take a walk," "take a break," and "take a bath," which convey very distinct actions. 54,55 Conversely, in Persian, light verbs are more commonly integrated into complex verb structures rather than combined with independent noun phrases as in the English examples just mentioned.…”