Bayesian semiparametric Markov renewal mixed models for vocalization syntax.

Speech and language play an important role in human vocal communication. Studies have shown that vocal disorders can result from genetic factors. In the absence of high-quality data on humans, mouse vocalization experiments in laboratory settings have been proven useful in providing valuable insights into mammalian vocal development, including especially the impact of certain genetic mutations. Such data sets usually consist of categorical syllable sequences along with continuous intersyllable interval (ISI) times for mice of different genotypes vocalizing under different contexts. ISIs are of particular importance as increased ISIs can be an indication of possible vocal impairment. Statistical methods for properly analyzing ISIs along with the transition probabilities have however been lacking. In this article, we propose a class of novel Markov renewal mixed models that capture the stochastic dynamics of both state transitions and ISI lengths. Specifically, we model the transition dynamics and the ISIs using Dirichlet and gamma mixtures, respectively, allowing the mixture probabilities in both cases to vary flexibly with fixed covariate effects as well as random individual-specific effects. We apply our model to analyze the impact of a mutation in the Foxp2 gene on mouse vocal behavior. We find that genotypes and social contexts significantly affect the length of ISIs but, compared to previous analyses, the influences of genotype and social context on the syllable transition dynamics are weaker.

Human GNPTAB stuttering mutations engineered into mice cause vocalization deficits and astrocyte pathology in the corpus callosum.

Stuttering is a common neurodevelopmental disorder that has been associated with mutations in genes involved in intracellular trafficking. However, the cellular mechanisms leading to stuttering remain unknown. Engineering a mutation in N-acetylglucosamine-1-phosphate transferase subunits α and ß (GNPTAB) found in humans who stutter into the mouse Gnptab gene resulted in deficits in the flow of ultrasonic vocalizations similar to speech deficits of humans who stutter. Here we show that other human stuttering mutations introduced into this mouse gene, Gnptab Ser321Gly and Ala455Ser, produce the same vocalization deficit in 8-day-old pup isolation calls and do not affect other nonvocal behaviors. Immunohistochemistry showed a marked decrease in staining of astrocytes, particularly in the corpus callosum of the Gnptab Ser321Gly homozygote mice compared to wild-type littermates, while the staining of cerebellar Purkinje cells, oligodendrocytes, microglial cells, and dopaminergic neurons was not significantly different. Diffusion tensor imaging also detected deficits in the corpus callosum of the Gnptab Ser321Gly mice. Using a range of cell type-specific Cre-drivers and a Gnptab conditional knockout line, we found that only astrocyte-specific Gnptab-deficient mice displayed a similar vocalization deficit. These data suggest that vocalization defects in mice carrying human stuttering mutations in Gnptab derive from abnormalities in astrocytes, particularly in the corpus callosum, and provide support for hypotheses that focus on deficits in interhemispheric communication in stuttering.