Acoustic analysis of the effects of sustained wakefulness on speech

Abstract

Exposing healthy adults to extended periods of wakefulness is known to induce changes in psychomotor functioning [Maruff et al. (2005). J. Sleep Res. 14, 21-27]. The effect of fatigue on speech is less well understood. To date, no studies have examined the pitch and timing of neurologically healthy individuals over 24 h of sustained wakefulness. Therefore, speech samples were systematically acquired (e.g., every 4 h) from 18 healthy adults over 24 h. Stimuli included automated and extemporaneous speech tasks, sustained vowel, and a read passage. Measures of timing, frequency and spectral energy were derived acoustically using PRAAT and significant changes were observed on all tasks. The effect of fatigue on speech was found to be strongest just before dawn (after 22 h). Specifically, total speech time, mean pause length, and total signal time all increased as a function of increasing levels of fatigue on the reading tasks; percentage pause and mean pause length decreased on the counting task; F4 variation decreased on the sustained vowel tasks /a:/; and alpha ratio increased on the extemporaneous speech tasks. These findings suggest that acoustic methodologies provide objective data on central nervous system functioning and that changes in speech production occur in healthy adults after just 24 h of sustained wakefulness.

Logopenic and Nonfluent Variants of Primary Progressive Aphasia Are Differentiated by Acoustic Measures of Speech Production

“…The reading samples were used to examine syllable segregation by studying the features of hesitations and within and between word silences using three measures [39], [40]:…”

The Prediction of Fatigue Using Speech as a Biosignal

“…For example in [7] it was demonstrated that when subjects were kept awake for a period of 24 hours, the duration of their pauses gradually increased for read speech, and the variation in the 4th formant decreased for sustained vowel sounds.…”

Analysis of the effects of physical task stress on the speech signal

“…Solutions for speech recognition systems were explored in that study, and later studies also considered stress classification systems (Cairns and Hansen, 1994;Womack and Hansen, 1996;Bou-Ghazale and Hansen, 2000). Contemporary study of acoustic variability of speech focuses on one type of speech variability at a time, and groups emotions into one category, and situational speaker stressors into another, including cognitive task stress (Lindstrom et al, 2008), Lombard effect (Boril and Hansen, 2010), fatigue (Vogel et al, 2010), and physical task stress. Godin and Hansen (2008) showed that fundamental frequency, and the percent of frames voiced in an utterance, are affected by physical task stress, but that the standard deviation of fundamental frequency in an utterance is not affected.…”