Aims: To detect differences in speech fluency in separate primaryprogressive aphasia syndromes (PPA) using automated analysistechniques. The resulting linguistic features are evaluated for theiruse in a predictive model to identify common patterns in speakerswith PPA. As fluency is observable in audio recordings, its quantifi-cation may provide a low-cost instrument that augments sponta-neous speech analyses in clinical practice.Methods and Procedures: Speech was recorded in 14 controls, 7nonfluent variant (nfvPPA) and 8 semantic variant (svPPA) speakers.The recordings were annotated for speech and non-speech withKaldi, a common toolkit for speech processing software. Variablesrelating to fluency (pause rate, number of pauses, length of pauses)were analyzed.Outcomes and Results: The best fitting distribution of pause dura-tion was a combination of two Gaussian distributions, correspond-ing with pause categories short vs. long.Group level differences were found in the rate of pauses andproportion of silence: nfvPPA speakers use more short pausesrelative to long pauses than control speakers, and the duration ofshort and long pauses is longerĶ¾ svPPA speakers use more longerpauses relative to short pauses. Their short pauses are significantlyshorter than those from control speakers.Participants in both PPA groups pause more frequently. SvPPAspeakers are typically perceived as fluent. However, our analysisshows their fluency patterns to be distinct from control speakers, ifthe long-short distinction is observed.Conclusions: Automatic measurements of pause duration showmeaningful distinctions across the groups and might provide futureaid in clinical assessment.
