This study evaluated the accuracy of personalized automatic speech recognition (ASR) for recognizing disordered speech from a large cohort of individuals with a wide range of underlying etiologies using an open vocabulary. The performance of these models was benchmarked relative to that of expert…