Voice Analysis in Patients With Neurologic Diseases

Advanced Voice Analysis With Machine Learning Algorithms in Patients With Neurologic Diseases

In this observational pilot study, the investigators will record and assess voice samples from healthy participants and those participants affected by neurologic diseases to evaluate possible differences in voice features.

Study Overview

• Status: Recruiting
• Conditions: Neurologic Disorder; Voice Disorders
• Intervention / Treatment: Other: Speech task

Detailed Description

In this study, the investigators will evaluate the clinical features of healthy participants and those participants with neurologic disorders by applying dedicated clinical scales. Also, the investigators will assess voice impairment by using perceptual examination tools. Then, the investigators will apply spectral analysis to assess the main frequency components of voice in healthy participants and in patients affected by neurologic disorders with a prominent voice impairment. To distinguish between healthy participants and patients affected by various neurologic diseases, the investigators will apply a voice analysis based on support vector machine (SVM) classifier that included a large number of features in addition to the main frequency components of voice.

For these purposes, the investigators will assess in detail the sensitivity, specificity, positive predictive value, and negative predictive value and accuracy of all diagnostic tests. Furthermore, the investigators will calculate the area under the receiver operating characteristic (ROC) curves to verify the optimal diagnostic threshold as reflected by the associated criterion (Ass. Crit.) and Youden Index (YI). To assess possible clinical-instrumental correlations, the investigators will also use a modified algorithm of SVM analysis to calculate a continuous numerical value (the likelihood ratio [LR]) providing a measure of voice impairment severity for each participant.

Voice recordings will be performed by asking participants to produce a specific speech task with their usual voice intensity, pitch, and quality. The speech task will consist of a sustained emission of a close mid-front unrounded vowel /e/ for at least 5 seconds. Voice recordings will be collected by using a high-definition audio-recorder placed at a distance of 5 cm from the mouth. Voice samples will be recorded in linear PCM format (.wav) at a sampling rate of 44.1 kHz, with 24-bit sample size. Voice analysis will consist of three separate processes: feature extraction, selection and classification. For feature extraction, the investigators will use the OpenSMILE (audEERING GmbH, Germany), dedicated software. Then, the investigators will select and classify voice feature by using SVM algorithm included in Weka.

• Study Type: Observational
• Enrollment (Anticipated): 100

Contacts and Locations

• Study Contact: Name: Antonio Suppa, MD, PhD; Phone Number: +0039 3494940365; Email: antonio.suppa@uniroma1.it

Study Locations

• Italy
  • Pozzilli, Italy, 86077
    • Recruiting
    • Antonio Suppa
    • Contact: Antonio Suppa; Phone Number: +00393494940365; Email: antonio.suppa@uniroma1.it

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Child; Adult; Older Adult
• Accepts Healthy Volunteers: Yes
• Genders Eligible for Study: All

• Sampling Method: Non-Probability Sample

Study Population

We will recruit neurologic patients not taking any oral medications or alcohol and any drugs acting on the central nervous system at the time of the study.

Description

Inclusion Criteria:

• Clinical diagnosis of neurologic disorders

Exclusion Criteria:

• smoking
• bilateral/unilateral hearing loss
• respiratory disorders
• conditions affecting the vocal cords, including nodules.

Study Plan

How is the study designed?

Number of groups / cohorts

1

Cohorts and Interventions

Group / Cohort	Intervention / Treatment
Patients; Patients affected by neurologic disorders showing a prominent voice impairment.	Other: Speech task; Speech task which consists of a sustained emission of the vowel /e/.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Voice analysis	Voice features obtained by using Support Vector Machine algorithm	Voice analysis with machine learning algorithms will be implemented immediately after voice recording, during the clinical evaluation of each participant.

Collaborators and Investigators

• Sponsor: Neuromed IRCCS

Study record dates

Study Major Dates

• Study Start (Actual): September 1, 2021
• Primary Completion (Actual): July 31, 2022
• Study Completion (Anticipated): July 31, 2023

Study Registration Dates

• First Submitted: April 12, 2021
• First Submitted That Met QC Criteria: April 12, 2021
• First Posted (Actual): April 15, 2021

Study Record Updates

• Last Update Posted (Actual): March 22, 2023
• Last Update Submitted That Met QC Criteria: March 21, 2023
• Last Verified: March 1, 2023

More Information

Terms related to this study

• Keywords: machine learning; neurologic disorders; voice analysis
• Additional Relevant MeSH Terms: Pathologic Processes; Respiratory Tract Diseases; Neurologic Manifestations; Otorhinolaryngologic Diseases; Laryngeal Diseases; Disease; Nervous System Diseases; Voice Disorders
• Other Study ID Numbers: DIPNEUROSCI_01

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No