Detecting Fatigue From Voice in Generalised Myasthenia Gravis

ConditionMyasthenia Gravis Generalised

Trial statusNot yet recruiting

Trial phaseNot listed

Trial typeObservational

Biological sexAll

Age18+

SponsorThymia Limited

About this trial

The goal of this observational study is to learn if computer analysis of voice recordings can detect a type of exhaustion called "central fatigue" in adults with generalised myasthenia gravis.

The main questions it aims to answer are:

1. Can advanced voice analysis accurately tell when participants are experiencing deep exhaustion based on how they speak? 2. How easy and acceptable is voice-based fatigue monitoring for people with myasthenia gravis?

Participants will:

1. Record themselves reading short passages and answering questions out loud twice daily (morning and evening), twice a week, for 4 weeks. 2. Answer brief questionnaires about their energy levels, mood, and myasthenia gravis symptoms during each session. 3. Use their own devices (computer, tablet, or smartphone) to complete all study activities online from home.

Eligibility criteria

This trial does not accept healthy volunteers

Qualifiers

Adults ≥18 years old

Self-reported generalised Myasthenia Gravis diagnosis confirmed by healthcare provider for ≥6 months

Disease stability for ≥6 months (no hospitalisations, medication changes, or significant symptom worsening)

English as first language

Disqualifiers

Pure ocular Myasthenia Gravis

Diagnosed mild cognitive impairment or dyslexia

Speech or hearing impairments affecting voice recording

Unable to provide credible diagnostic information (healthcare provider diagnosis, antibody test results, current medications)

Trial population

The study population consists of a convenience sample of adults with generalised Myasthenia Gravis recruited through a decentralised, multi-channel approach across the United States and United Kingdom. This remote recruitment strategy leverages established myasthenia gravis patient networks, charities and digital platforms like social media groups to reach geographically dispersed participants. The decentralised methodology is particularly appropriate for studying a rare condition where patients may have limited mobility due to fatigue and muscle weakness.

Trial design

Design model

Cohort

Time perspective

Prospective

Treatments tested in this trial

Not listed

Trial groups

No trial groups listed

Trial outcomes

Primary outcomes

Accuracy of AI Model for Binary Central Fatigue Classification as Assessed by Voice Biomarker Analysis

Binary classification performance (presence vs. absence of central fatigue) of the artificial intelligence-based system using voice biomarker analysis, with the subjective fatigue scale serving as ground truth. Performance will be measured using sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) metrics through cross-validation methods.

Time frame

Across 16 assessment sessions over 4 weeks from enrolment

Secondary outcomes

Study Completion Rate Among Enrolled Participants

Percentage of enrolled participants who complete all 16 required assessment sessions out of the total number of participants who begin the study

Time frame

From enrolment through completion of final assessment session at 4 weeks

Individual Session Completion Rate Across All Participants

Percentage of individual assessment sessions completed across all enrolled participants out of the total possible sessions

Time frame

From enrolment through completion of final assessment session at 4 weeks

Adherence to Specified Assessment Time Windows

Percentage of completed sessions that occur within the designated time windows out of all completed sessions

Time frame

From enrolment through completion of final assessment session at 4 weeks

Participant Acceptability of Voice-Based Monitoring System

Self-reported acceptability scores including ease of use, satisfaction, and willingness for future use of the voice-based fatigue monitoring approach, assessed through three researcher developed 7-point Likert scales (1=lowest, 7=highest)

Time frame

At completion of final assessment session at 4 weeks

Other outcomes

Exploratory: Correlation Between Central and Peripheral Fatigue

Correlation between AI-detected central fatigue severity scores (0 = low to 1 = high) and self-reported peripheral fatigue symptoms measured by subscales of the MG-Symptoms PRO, namely "physical fatigue" and "muscle weakness fatigability" (Regnault et al., 2024). The score for each scale ranges from 0 to 100, with higher values indicating more severe symptoms. Analyses will employ Spearman's rank correlation to account for potential non-linear relationships.

Time frame

4 weeks

Exploratory: Fatigue-Mood Relationship

Correlation between central fatigue measures (both self-reported using the Modified Fatigue Impact Scale (MFIS) and AI-detected probability scores between 0 = low and 1 = high) and depression (PHQ-8) and anxiety (GAD-7) scores. The MFIS uses a 5-point scale from 0 = never to 4 = almost always with a total score range of 0-84. The PHQ-8 and GAD-7 use 4-point scales from 0 = not at all to 3 = nearly every day, with total score ranges of 0-24 for PHQ-8 and 0-21 for GAD-7.

Time frame

4 weeks

Locations

This trial has no locations.

Sponsors and contacts

Click on the lead sponsor to view all of their trials.

Thymia Limited

Lead sponsor

UCB Pharma

Collaborator

Source ClinicalTrials.gov

This trial is not recruiting at the moment. You can still explore other options: