Automated analysis of written language in the three variants of primary progressive aphasia

2023

Preprint

Self Cite

Recently we proposed and tested a novel framework based on information theory about the features of agrammatism in patients with the nonfluent variant of primary progressive aphasia (nfvPPA). These features include short sentences, simplified structures, the omission of function words, reduced use of verbs, and increased use of heavy verbs. After distinguishing the syntactic and lexical features of agrammatism, we proposed that the lexical features are not defects but rather a compensatory response to structural deficits. We showed that although patients with nfvPPA have difficulty using complex structures, they use more informative words to facilitate the transfer of their intended message. In this work, we demonstrated the generalizability of our findings to patients with Broca's aphasia who share similar neuroanatomical involvements with patients with nfvPPA. We found that patients with Broca's aphasia use simpler syntactic structures as measured by higher syntax frequency. We also found that the use of simpler syntax predicts the use of lower-frequency words. Furthermore, our study showed that producing sentences of lower word frequency is likely achieved through the canonical features of agrammatism, i.e., higher proportions of content words over function words, nouns over verbs, and heavy verbs over light verbs.

Section: Discussionsupporting

confidence: 86%

An information-theoretic analysis of agrammatism in Broca’s aphasia

2023

Preprint

Self Cite

“…As a result, language production in nfvPPA patients might already be at the maximum capacity for using low frequency words. 50 Lastly, the comparison between speaking and writing in nfvPPA makes it possible to further probe the functional locus of the bottleneck in language production. Our findings rule out the possibility for the articulatory stage to be the limiting factor as the same lexical and syntactic properties were found in the written samples.…”

Section: Discussionmentioning

confidence: 99%

A computational approach for measuring sentence information via surprisal: theoretical implications in nonfluent primary progressive aphasia

Michaelov

Josephy‐Hernandez

et al. 2022

Preprint

Self Cite

Nonfluent aphasia is a language disorder characterized by simplified sentence structures as well as word-level abnormalities such as a reduced use of verbs and function words. According to the predominant account of the disorder, both structural and word-level features are caused by a core deficit in the processing of syntax. Under this account, however, it remains unclear why nonfluent patients choose semantically richer verbs and may have an intact comprehension of verbs and function words. Here, we propose and test the hypothesis that the word-level features of nonfluency reflect a process that selects lexically richer words to increase the information content of sentences. We use a computational linguistic method to measure the information content of sentences in the language of patients with nonfluent primary progressive aphasia (nfvPPA) (n = 36) and healthy controls (n = 133). We measure sentence information using surprisal, a metric calculated by the average probability of occurrence of words in a sentence given their preceding context. We found that by packaging their structurally simple sentences with lower frequency words, nfvPPA patients produce sentences with similar surprisal as that of healthy speakers. Furthermore, we found that higher sentence surprisal in nfvPPA correlates with a lower function-to-all-word ratio, a lower verb-to-noun ratio, and a higher heavy-to-all-verb ratio. Surprisal is an effective quantitative index of sentence information. Using surprisal allows for testing an account of nonfluent aphasia that regards word-level features of nonfluency as adaptive rather than defective symptoms, a finding that may entail revisions in therapeutic approaches to nonfluent speech.

“…The goal of developing this package is to increase the precision and objectivity of language assessments while reducing human labor (as outlined in (63)). The toolbox uses a number of natural language processing toolkits and software such as Stanford Parser (64), spaCy (65), and NLTK as well as text analysis libraries in R. Quantitext receives transcribed language samples as input and generates as outputs a number of metrics such as sentence length, log word frequency, log syntax frequency (63), content units (as in (66)), efficiency of lexical and syntactic items (as in (67)), part of speech tags, and the distinction of heavy and light verbs.…”

Section: Methodsmentioning

confidence: 99%

Less is more in language production: Shorter sentences contain more informative words

Ren

Quimby

et al. 2022

Preprint

Self Cite

Agrammatism is characterized by short sentences, the omission of function words, a higher ratio of heavy to light verbs, and a decreased use of verbs relative to nouns. Despite the observation of these phenomena more than two centuries ago, there has been no unifying theory to explain all features of agrammatism. Here, by first examining the language of patients with primary progressive aphasia, we show that the seemingly heterogeneous features of agrammatism can be explained by a process that selects lower frequency words over their higher frequency alternatives in the context of a limitation in sentence production, likely to increase the informational content of sentences. We further show that when healthy speakers are constrained to produce short sentences, features of agrammatism emerge in their language. Finally, we show that these findings instantiate a general property in healthy language production in which shorter sentences are constructed by selecting lower frequency words.