Leveraging Speech Analytics for Objective Measurement in Rare Neurologic Diseases

The search for sensitive, objective biomarkers in neurodegenerative diseases has intensified as clinical trials increasingly aim to detect subtle changes in function earlier and more reliably. Traditional outcome measures, while foundational, often rely on intermittent, clinician-reported assessments that may lack sensitivity to small but meaningful changes in disease trajectory, particularly in rare conditions with heterogeneous presentations.

In this context, speech is gaining attention as a potential digital biomarker that integrates multiple neurologic domains, including motor control, cognition, coordination, and language. A newly facilitated study involving Redenlab and the Translational Centre for Speech Disorders at Murdoch Children’s Research Institute is evaluating the role of speech analytics in hypomyelination with atrophy of the basal ganglia and cerebellum (H-ABC), a rare neurodegenerative disorder characterized by motor dysfunction, ataxia, and communication impairment. The study leverages advanced acoustic and linguistic analyses to capture multidimensional aspects of speech, with the goal of improving clinical assessment and supporting future therapeutic development.

To better understand the scientific and clinical rationale behind this approach, NeurologyLive^® spoke with Adam Vogel, PhD, Chief Science Officer at Redenlab and Professor of Speech Neuroscience at The University of Melbourne. In this Q&A, Vogel discusses how speech reflects underlying neurologic function, its potential advantages over traditional endpoints, and what is needed for broader integration into clinical trial design.

NeurologyLive: Can you describe the rationale for using speech as a biomarker in neurodegenerative diseases, particularly in a condition like H-ABC?

Adam Vogel, PhD: Speech is one of the most important human behaviors. It requires integration of motor and cognitive functions and depends on precise coordination of respiratory control, phonation, articulation, prosody, and language. When something goes wrong with the brain, the complex nature of speech makes effective communication challenging. This is particularly relevant in neurodegenerative and neurodevelopmental disorders where changes in motor control, coordination, or cognition emerge.

In H-ABC, this rationale is strong because the condition affects brain systems that are central to movement, coordination, and cognition. H-ABC is hypomyelination with atrophy of the basal ganglia and cerebellum, and clinical features can include motor impairment, dystonia, ataxia, swallowing problems, and speech difficulties. Speech offers a practical way to measure functional neurological change in a domain that is meaningful to patients and families: the ability to communicate.

What specific speech parameters (e.g., intelligibility, pause duration, articulation) are being evaluated in this study, and how might they correlate with underlying neurologic function?

We are interested in speech measures that are interpretable, clinically meaningful, and technically robust. These include intelligibility, or how well a listener can understand the speaker; speech naturalness; timing measures like speaking rate; voice quality; and acoustic measures that reflect articulation. Together, these parameters provide a multidimensional profile of communication.

The neurological interpretation depends on the parameter. Reduced intelligibility may reflect impaired articulatory precision, abnormal prosody, reduced respiratory support, or broader motor coordination difficulties. Slower articulation rate and increased pausing can reflect motor slowing, impaired planning, fatigue, or reduced coordination. Voice and resonance features may reflect laryngeal, respiratory, or oral motor control. In H-ABC, where basal ganglia and cerebellar systems are affected, timing, rhythm, articulatory control, and coordination are particularly relevant domains to measure.

How does speech-based analysis compare with traditional clinical outcome measures in terms of sensitivity, objectivity, and scalability for use in trials?

Traditional clinical outcome measures remain important, but they often rely on broad ordinal ratings, infrequent clinic visits, are typically clinician derived (and therefore subjective) and relatively coarse categories of function. Speech-based measures can complement those tools by providing more frequent, objective, and quantitative measurement of a function that is directly relevant to daily life. A short speech sample can generate multiple acoustic and linguistic measures, which can increase sensitivity to change over time.

Scalability is another important advantage. Speech can be collected in clinic or remotely, using standardized tasks and quality-managed digital workflows. FDA guidance on digital health technologies specifically recognizes the role of digital tools in remote data acquisition for clinical investigations, including the potential to improve trial efficiency and reduce participant burden. For rare diseases such as H-ABC, where participants are geographically dispersed and clinic-based assessment can be difficult, this is important.

Given the heterogeneity in communication methods in H-ABC, how is the study accounting for non-verbal or alternative communication strategies in its analysis?

In conditions like H-ABC, not every participant will be able to complete the same spoken tasks, with some individuals minimally verbal. Redenlab has developed methods for interpreting even small elements of vocalization, examining sound inventories and maturation trajectories. It is important to capture speech in ways that are natural, representative and meaningful. For clinical trials, it’s also important to remember we are measuring change within the individual, using the participant as their own control.

Looking ahead, what steps are needed for speech biomarkers to gain broader regulatory acceptance and integration into clinical trial endpoints?

Regulatory acceptance will require the same evidentiary discipline expected of any clinical trial endpoint. Speech markers need a clearly defined context of use, standardized collection procedures, evidence of analytical validity, test–retest reliability, sensitivity to change, and clinically meaningful interpretation. The field needs to show that the measure is accurate, reproducible, relevant to patient function, and interpretable in the context of treatment benefit.

For speech measures to be integrated into clinical trial endpoints, we also need strong links between digital speech measures and clinical outcome assessment frameworks. That includes demonstrating how changes in speech relate to communication function, caregiver burden, daily participation, and patient- or family-prioritized outcomes. FDA’s digital health technology guidance and clinical outcome assessment frameworks provide a useful direction. The technology must be fit for purpose, the measure must be meaningful, and the endpoint must be interpretable for regulatory decision-making.

Transcript edited for clarity.

REFERENCE
1. Speech could be a key clinical marker in neurodegenerative disease drug trials, says leading biotech. News release. Redenlab. April 14, 2026. Accessed May 6, 2026. https://redenlab.com/speech-could-be-a-key-clinical-marker-in-neurodegenerative-disease-drug-trials-says-leading-biotech/