Speech Analysis Algorithm for Dysarthria Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current speech evaluation methods for dysarthria rely heavily on subjective assessments by speech-language pathologists, which are inconsistent, costly, and prone to bias, lacking objective measures that can provide reliable and repeatable results for diagnosing disease onset, progression, and treatment efficacy.
Innovation Solution
Development of novel speech analysis algorithms that combine subjective ratings with objective laboratory-implemented features to create a predictive software model, using envelope modulation spectrum, long-term average spectrum, spatio-temporal features, and dysphonia features to generate unbiased, perception-calibrated metrics for evaluating speech intelligibility and motor speech disorders.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If subjective assessments by speech-language pathologists are used, then face validity for characterizing speech deficits is improved, but consistency and repeatability deteriorate
Solution Approach 1:
The patent introduces an automated speech analysis system as an intermediary between the speech signal and the assessment outcome. This system processes speech signals through multiple objective measures (acoustic, spectral, temporal features) and combines them with machine learning algorithms to generate assessments that maintain face validity while improving consistency and repeatability across different evaluators and time points.
Solution Approach 2:
The patent transforms the assessment from subjective perceptual judgments to objective quantitative measurements by changing the parameters from qualitative ratings to multiple acoustic and spectral features (e.g., fundamental frequency, formant frequencies, spectral tilt, jitter, shimmer). This parameter transformation enables consistent and repeatable measurements while preserving clinical relevance.
2Measurement precision
If subjective assessments by speech-language pathologists are used, then face validity is improved, but cost and time consumption worsen
Solution Approach 1:
The patent replaces the mechanical system of human perceptual assessment with an automated computational system that processes speech signals through acoustic analysis algorithms. This substitution eliminates the time-consuming nature of manual subjective assessment while maintaining face validity through multiple objective measures that capture relevant speech characteristics.
Solution Approach 2:
The patent creates a computational model that copies and formalizes the assessment process, allowing it to be executed automatically without requiring speech-language pathologist time for each assessment. The system captures the essential features of speech deficits through objective measures that can be processed rapidly and consistently.
3Reliability
If existing objective measures are used, then repeatability is improved, but clinical interpretability worsens
Solution Approach 1:
The patent merges multiple objective measures (acoustic features, spectral features, temporal features) with machine learning algorithms to create a comprehensive assessment system. This combination maintains the repeatability of objective measurements while improving clinical interpretability by integrating multiple sources of information that collectively provide a more complete and clinically relevant picture of speech deficits.
Solution Approach 2:
The patent creates a composite assessment approach that combines multiple types of objective measures (acoustic, spectral, temporal) rather than relying on a single measure. This composite approach maintains the repeatability of each individual measure while improving overall clinical interpretability through the integration of complementary information from different measurement domains.
Data Source
AI summary
Systems and methods use patient speech samples as inputs, use subjective multi-point ratings by speech-language pathologists of multiple perceptual dimensions of patient speech samples as further inputs, and extract laboratory-implemented features from the patient speech samples. A predictive software model learns the relationship between speech acoustics and the subjective ratings of such speech obtained from speech-language pathologists, and is configured to apply this information to evaluate new speech samples. Outputs may include objective evaluation of the plurality of perceptual dimensions for new speech samples and/or evaluation of disease onset, disease progression, or disease treatment efficacy for a condition involving dysarthria as a symptom, utilizing the new speech samples.