AI-Based Speech Intelligibility Assessment: Efficacy & Reliability (AI Solution)

Clinical Study on the Efficacy and Reliability of Speech Intelligibility Assessment for Post-stroke Patients Using an AI Solution

compare the auditory-perceptual assessment conducted by speech-language pathologists using the U-TAP with the assessment performed via the AI solution (Blings) program in patients with speech disorders, and to evaluate the efficacy and reliability of these two methods for speech intelligibility assessment.

Study Overview

• Status: Not yet recruiting
• Conditions: Patients With Speech Disorders

• Study Type: Observational
• Enrollment (Estimated): 40

Contacts and Locations

• Study Contact: Name: Jisoo Baik; Phone Number: 082+055-360-4159; Email: zisoo@pusan.ac.kr

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: No

• Sampling Method: Non-Probability Sample

Study Population

The study population consists of adult patients aged 18 years or older with neurologic disorders-including amyotrophic lateral sclerosis (ALS), poliomyelitis, cerebellar-related disorders (such as cerebellar atrophy or degeneration and cerebellar lesions), and stroke-who present with concomitant speech disorders, such as dysarthria, motor speech disorders, and voice disorders.

Participants must have sufficient cognitive function, defined as a Korean Mini-Mental State Examination (K-MMSE) score of 16 or higher, and must be capable of understanding and following the examiner's instructions. Individuals with visual, perceptual, or auditory impairments that interfere with participation in speech assessment, as well as those deemed unsuitable for study participation by the investigator for other clinical or practical reasons, are excluded.

Description

Inclusion Criteria:

• Adults aged 18 years or older.
• Patients with neurologic disorders, including amyotrophic lateral sclerosis, poliomyelitis, cerebellar disorders, cerebellar lesions, and stroke.
• Patients presenting with concomitant speech disorders, such as dysarthria, motor speech disorders, and voice disorders.

Exclusion Criteria:

• Subjects with a score of less than 16 points on the Korean Mini-Mental State Examination (K-MMSE).
• Subjects with visual, perceptual, or auditory impairments that prevent them from following the examiner's instructions.
• Subjects judged by the researcher as unsuitable for participation in this clinical study due to other reasons.

Study Plan

How is the study designed?

Number of groups / cohorts

1

Cohorts and Interventions

Group / Cohort
Observational study(Patients with speech disorders)

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Word-level consonant accuracy (%) derived from speech samples obtained during the U-TAP assessment	The agreement between consonant accuracy analyzed by the AI-based speech disorder analysis program (Blings) and the evaluations conducted independently by two certified speech-language pathologists (Rater 1 and Rater 2). Consonant accuracy is calculated at the word level and expressed as a percentage score ranging from 0% to 100%. The mean value of consonant accuracy scores from Rater 1 and Rater 2 is used as the reference human evaluation.	Visit 1 (Baseline)

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Test-retest reliability of word-level consonant accuracy (%) calculated by Blings	The consistency of consonant accuracy scores generated by Blings from the same participant across two visits. Comparison of Blings-derived consonant accuracy scores (%) between Visit 1 and Visit 2. Scores range from 0% to 100%.	Visit 1 and Visit 2 (7 ± 5 days interval)
Inter-rater agreement of word-level consonant accuracy (%) among Blings, Rater 1, and Rater 2	Agreement in consonant accuracy evaluations between the AI-based analysis and human raters. Pairwise comparisons: Rater 1 vs. Rater 2, Blings vs. Rater 1, Blings vs. Rater 2. Scores expressed as percentages (0-100%)	Visit 1 and Visit 2 (7 ± 5 days interval)
Time required to complete speech analysis (seconds)	Comparison of analysis time between the AI-based system (Blings) and human raters. Total analysis time measured in seconds for: Rater 1, Rater 2, Blings	Visit 1 and Visit 2 (7 ± 5 days interval)

Collaborators and Investigators

• Sponsor: Pusan National University Yangsan Hospital

Study record dates

Study Major Dates

• Study Start (Estimated): February 1, 2026
• Primary Completion (Estimated): December 31, 2026
• Study Completion (Estimated): December 31, 2026

Study Registration Dates

• First Submitted: December 22, 2025
• First Submitted That Met QC Criteria: January 8, 2026
• First Posted (Estimated): January 16, 2026

Study Record Updates

• Last Update Posted (Estimated): January 16, 2026
• Last Update Submitted That Met QC Criteria: January 8, 2026
• Last Verified: January 1, 2026

More Information

Terms related to this study

• Other Study ID Numbers: 11-2025-086

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: UNDECIDED

Drug and device information, study documents

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