AUTHOR=Suppa Antonio , Asci Francesco , Costantini Giovanni , Bove Francesco , Piano Carla , Pistoia Francesca , Cerroni Rocco , Brusa Livia , Cesarini Valerio , Pietracupa Sara , Modugno Nicola , Zampogna Alessandro , Sucapane Patrizia , Pierantozzi Mariangela , Tufo Tommaso , Pisani Antonio , Peppe Antonella , Stefani Alessandro , Calabresi Paolo , Bentivoglio Anna Rita , Saggio Giovanni ,  Lazio DBS Study Group , Altavista Maria Concetta , Calciulli Alessandra , Ciavarro Marco , Cortese Francesca , Daniele Antonio , Biase Alessandro De , D'Ercole Manuela , Biase Lazzaro Di , Giuda Daniela Di , Leo Pietro Di , Genovese Danilo , Imbimbo Isabella , Izzo Alessandro , Liperoti Rosa , Marano Giuseppe , Marano Massimo , Mazza Marianna , Monge Alessandra , Montano Nicola , Orsini Michela , Rigon Leonardo , Rizz Marina , Rocchi Camilla , Saporito Gennaro , Vacca Laura , Viselli Fabio 

TITLE=Effects of deep brain stimulation of the subthalamic nucleus on patients with Parkinson's disease: a machine-learning voice analysis

JOURNAL=Frontiers in Neurology

VOLUME=Volume 14 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2023.1267360

DOI=10.3389/fneur.2023.1267360

ISSN=1664-2295

ABSTRACT=Introduction: Deep brain stimulation of the subthalamic nucleus (STN-DBS) can exert relevant effects on voice in patients with Parkinson’s disease (PD). We here examined voice objectively in STN-DBS patients with PD, by using artificial intelligence. 
Materials and Methods: In a cross-sectional study, we enrolled 108 controls and 101 patients with PD. The cohort of PD consisted of a first group of 50 patients with STN-DBS and a second group of 51 patients under the best medical treatment. Voice was clinically evaluated using the Unified Parkinson’s Disease Rating Scale part-III subitem for voice (UPDRS-III-v). We recorded and then analysed voices using specific machine-learning algorithms. The likelihood ratio (LR) was also calculated as an objective measure for clinical-instrumental correlations.
Results: Clinically, voice impairment was greater in STN-DBS patients than in those under oral treatment. Machine-learning discriminated voices recorded from STN-DBS patients and those under oral treatments, objectively and with high accuracy. We also found significant clinical-instrumental correlations since the greater LRs, the higher UPDRS-III-v scores. 
Discussion: STN-DBS deteriorates speech in patients with PD as objectively demonstrated by machine-learning voice analysis.