Assessing speech dysfunction using BOLD and acoustic analysis in parkinsonism

摘要

Speech dysfunction is often associated with parkinsonism (Parkinson's disease (PD), Multiple System Atrophy (MSA), and Progressive Supranuclear Palsy (PSP)), along with characteristic motor features. Any or all of the following i.e. respiratory, phonatory, resonatory, or articulatory components of speech production may be affected. Articulatory imprecision, repetition of syllables (tachyphrenia), and tremor of oropharyngeal structures add to speech unintelligibility. We studied acoustics using spectrogram and its correlation with BOLD activation during voice/speech production across these subjects. Methods: BOLD studies were conducted on 108 subjects (29 PD, 20 MSA and 19 PSP and 40 controls) on 1.5T MR scanner using 130 dynamics. Active phase involved acquisition (10 volumes each) of audible reading of visually presented bi-syllabic meaningful Hindi simple words (5 types of non-nasal stop consonant categories, i.e. namely velars, palatals, retroflexes, dentals, bilabials and one nasal stop consonant) with interleaved silence during baseline. The subjects' voice samples were analyzed for acoustic parameters, namely formant frequencies of the adjoining vowels, voice onset time (VOT), and intensities using spectrogram. Correlation of BOLD activation in different brain areas with acoustic parameters was evaluated. Results: Voice intensity was significantly lowered, while VOTs were delayed in these patients as compared to healthy controls. All acoustic parameters were significantly affected for nasal consonants. BOLD activation correlated positively in primary motor cortex to VOTs, while F2 formants to activation of supplementary motor area. Conclusion: The differences in the acoustic quality of various stop consonants in patients may be helpful in differentiating these three parkinsonian disorders.