EgypDisSpeech-9

A Pilot Dataset and Feasibility Study of Egyptian Arabic Non-Standard Speech with Motor Speech Disorders

Abstract

Speech technologies rarely include recordings from Arabic speakers with motor speech disorders, which limits both clinical research and accessibility tools for Arabic speakers. We present EgypDisSpeech-9, a pilot feasibility study describing the collection protocol, metadata, and initial descriptive statistics for a small Egyptian Arabic dataset recorded from nine volunteer participants with non-standard speech (e.g., dysarthria related to ALS and other etiologies). The goal is not to release raw audio (IRB not currently obtained) but to report our methods, ethical safeguards, and lessons learned to guide future scaled collections and to encourage the community to develop Arabic clinical speech resources. We hypothesize that even a small, welldocumented pilot will reveal practical barriers (recruitment, annotation alignment, recording variability) that must be addressed before large-scale data release. We conclude with recommendations and a plan to obtain formal ethics approval for an expanded, shareable dataset.