[Expressive prosodic patterns in individuals with Alzheimer's disease].

This paper describes a useful technique for quantifying the degree of speech deficits in dementia of GDS 4 Alzheimer's type (DAT). Production of prosodic speech in DAT patients and healthy older controls was analysed using variation in fundamental frequency (F0) measures on a reading task. The prosogram computational model was used to analyze the prosodic contours of the speech samples, using melodic styling of F0 based on perceptual principles and prominence detection of spectral and amplitude fluctuations in the speech signal. Results revealed significant differences in most of these prosodic parameters among the DAT group. Normal pitch variation in speech and variations in syllable timing were reduced in the DAT group, these features cause "flat" speech prosody in these patients. These speech parameters may have diagnostic and prognostic value for Alzheimer's disease and therefore could be a useful aid in clinical trials.