Childhood Apraxia of Speech is a highly controversial disorder due to a lack of consensus on the features that define it and the etiologic conditions that explain its origin. The term Suspected Apraxia of Speech (sAOS) has been proposed as an interim term for this putative clinical entity. The point prevalence of sAOS in young children has been estimated at approximately 0.1%. The long-term objective of this proposal is to develop a valid, reliable, and efficient means to classify children as positive for sAOS. In addition to the contributions to theoretical explication of AOS, the software-based diagnostic tools resulting from this work will allow any certified speech-language pathologist to determine if a child's speech includes prosodic features that fall within a 95% confidence interval supporting the diagnosis of sAOS.
The aim for this first period of planned programmatic research is to develop automated diagnostic markers for sAOS with clinically adequate sensitivity and specificity (> 90% positive and negative likelihood ratios). The four specific aims are: (a) to automate and improve the sensitivity and specificity of two existing (manually derived) prosodic markers, (b) to develop four additional automatic, prosody-based diagnostic markers, (c) to derive a single diagnostic index based on a statistical derivative from the six individual markers, and (d) to validate the composite diagnostic marker using classification data obtained from expert clinical researchers. Procedures are divided into four phases. In Year 1, automated versions of existing markers will be developed that determine speech-event locations using automatic speech recognition (ASR). Based on two pilot studies, this technique is expected to yield results equivalent to published data. The sensitivity of the markers will be improved by methods including normalizing by speaking rate and vowel identity. In Year 2, new automated markers will be created based on ASR and speech-signal processing techniques. These markers will measure variation in interstress timing, linguistic rhythm, speaking rate, and glottal-source characteristics. In the first part of Year 3, results from all six markers will be combined into a single diagnostic index using multi-layer perceptrons. In the latter part of Year 3, per-child errors will be evaluated to determine relationships between specific prosodic factors and the diagnosis of sAOS, providing insight into the features and definition of sAOS.
| Hosom, John-Paul (2009) Speaker-Independent Phoneme Alignment Using Transition-Dependent States. Speech Commun 51:352-368 |
| Hosom, John-Paul; Shriberg, Lawrence; Green, Jordan R (2004) Diagnostic Assessment of Childhood Apraxia of Speech Using Automatic Speech Recognition (ASR) Methods. J Med Speech Lang Pathol 12:167-171 |
