Text
E-book Models and Analysis of Vocal Emissions for Biomedical Applications : 11th International Workshop, December, 17-19, 2019
The effects of neurodegenerative diseases on speech are well known. Phonation, articulation, prosody and fluency are speech characteristics strongly affected by Parkinson’s Disease (PD) [1]. On the other hand, it is also well known that neuroacustical stimulation based on binaural beats may modify the brain activity measured in the cortex by electroencephalography (EEG) [2]. Binaural beats stimulation is a perceptual phenomenon where two sine waves are presented to a subject separately through each ear [3]. This study is intended to test if neuroacoustical stimulation is able to modify the phonation stability of PD patients by the evaluation of phonation features derived from the acoustic analysis of PD patients phonation before (pre) and after (post) stimulation. Relevant aspects as the signal combinations used in active and inert stimulation are presented and discussed, as well as the order in which stimuli are applied. The statistical analysis of data is also presented and discussed. Results seem to avail the feasibility of this methodology in improving neuromotor stability in PD patients. Each patient was submitted to neuroacoustical stimulations and speech recordings. The stimuli included active (supposedly exerting a modification in the functional performance of the patient to approach normat ivit y) and inert or neuter (supposedly not exerting such positive modification of functional performance). The active stimulations consisted of pink noise and a pair of sine waves of 154 and 168 Hz, applied to the left and right ears during 10 minutes to generate a perceptual beat in the upper auditory pathways of 14 Hz, as this frequency is supposed to be associated to the EEG activity decay detected in PD [4]. The inert stimulations were based on the application of simple pink noise during the same amount of time. The specific protocol is shown in Figure 1. In a first session (top) patients were randomly selected to receive an active or an inert stimulation. In a second session separated at least seven days from the first one, the patient received an inert or an active stimulation (the reverse than the one in the first session). The stimulation session (bottom) consisted ofa voice recording task (pre-stimulus), followed by a binaural stimulation task lasting 10 minutes, to end with another voice recording task (post-stimulus). Recording tasks of at least two-seconds of the vowel [a:] were conducted immediately before (pre) and after (post) each stimulation session. These were analyzed following the phonation feature extraction described in [5]. The who le study was approved by the Ethical Committee of Universidad Politécnica de Madrid, and it adhered to the Declaration of Helsinki. All the volunteer participants signed an informed consent document.
Tidak tersedia versi lain