Progressive apraxia of speech: delays to diagnosis and rates of alternative diagnoses

Responsible development of clinical speech AI: Bridging the gap between clinical research and technology

This perspective article explores the challenges and potential of using speech as a biomarker in clinical settings, particularly when constrained by the small clinical datasets typically available in such contexts. We contend that by integrating insights from speech science and clinical research, we can reduce sample complexity in clinical speech AI models with the potential to decrease timelines to translation. Most existing models are based on high-dimensional feature representations trained with limited sample sizes and often do not leverage insights from speech science and clinical research. This approach can lead to overfitting, where the models perform exceptionally well on training data but fail to generalize to new, unseen data. Additionally, without incorporating theoretical knowledge, these models may lack interpretability and robustness, making them challenging to troubleshoot or improve post-deployment. We propose a framework for organizing health conditions based on their impact on speech and promote the use of speech analytics in diverse clinical contexts beyond cross-sectional classification. For high-stakes clinical use cases, we advocate for a focus on explainable and individually-validated measures and stress the importance of rigorous validation frameworks and ethical considerations for responsible deployment. Bridging the gap between AI research and clinical speech research presents new opportunities for more efficient translation of speech-based AI tools and advancement of scientific discoveries in this interdisciplinary space, particularly if limited to small or retrospective datasets.

Rate Modulation Abilities in Acquired Motor Speech Disorders

PURPOSE

The purpose of this study was to describe, compare, and understand speech modulation capabilities of patients with varying motor speech disorders (MSDs) in a paradigm in which patients made highly cued attempts to speak faster or slower.

METHOD

Twenty-nine patients, 12 with apraxia of speech (AOS; four phonetic and eight prosodic subtype), eight with dysarthria (six hypokinetic and two spastic subtype), and nine patients without any neurogenic MSD completed a standard motor speech evaluation where they were asked to repeat words and sentences, which served as their "natural" speaking rate. They were then asked to repeat lower complexity (counting 1-5; repeating "cat" and "catnip" 3 times each) and higher complexity stimuli (repeating "catastrophe" and "stethoscope" 3 times each and "My physician wrote out a prescription" once) as fast/slow as possible. Word durations and interword intervals were measured. Linear mixed-effects models were used to assess differences related to MSD subtype and stimuli complexity on bidirectional rate modulation capacity as indexed by word duration and interword interval. Articulatory accuracy was also judged and compared.

RESULTS

Patients with prosodic AOS demonstrated a reduced ability to go faster; while they performed similarly to patients with spastic dysarthria when counting, patients with spastic dysarthria were able to increase rate similar to controls during sentence repetition; patients with prosodic AOS could not and made increased articulatory errors attempting to increase rate. AOS patients made more articulatory errors relative to other groups, regardless of condition; however, their percentage of errors reduced with an intentionally slowed speaking rate.

CONCLUSIONS

The findings suggest comparative rate modulation abilities in conjunction with their impact on articulatory accuracy may support differential diagnosis between healthy and abnormal speech and among subtypes of MSDs (i.e., type of dysarthria or AOS). Findings need to be validated in a larger, more representative cohort encompassing several types of MSDs.

SUPPLEMENTAL MATERIAL

https://doi.org/10.23641/asha.22044632.