This paper proposes a methodological framework for validating tri-modal acquisition systems that integrate acoustic, articulatory kinematic, and neurophysiological signals for assessing dysarthric and aphasic populations. The framework includes a multi-modal architecture, validation protocols, evaluation metrics, and risk analysis for clinical deployment.
Key findings
Proposes a tri-modal acquisition system for comprehensive speech disorder assessment.
Includes a modular architecture with acoustic, EMA kinematic, and EEG signals.
Details validation protocols using benchmark dysarthric and aphasic speech corpora.
Presents evaluation metrics for intelligibility, severity classification, and diagnostic accuracy.
Limitations & open questions
Does not yet include results from experimental validation.
Requires further assessment of cross-corpus generalization and clinical deployment risks.