Machine-learning assisted swallowing assessment: a deep learning-based quality improvement tool to screen for post-stroke dysphagia.

Introduction: Post-stroke dysphagia is common and associated with significant morbidity and mortality, rendering bedside screening of significant clinical importance. Using voice as a biomarker coupled with deep learning has the potential to improve patient access to screening and mitigate the subjectivity associated with detecting voice change, a component of several validated screening protocols. Methods: In this single-center study, we developed a proof-of-concept model for automated dysphagia screening and evaluated the performance of this model on training and testing cohorts. Patients were admitted to a comprehensive stroke center, where primary English speakers could follow commands without significant aphasia and participated on a rolling basis. The primary outcome was classification either as a pass or fail equivalent using a dysphagia screening test as a label. Voice data was recorded from patients who spoke a standardized set of vowels, words, and sentences from the National Institute of Health Stroke Scale. Seventy patients were recruited and 68 were included in the analysis, with 40 in training and 28 in testing cohorts, respectively. Speech from patients was segmented into 1,579 audio clips, from which 6,655 Mel-spectrogram images were computed and used as inputs for deep-learning models (DenseNet and ConvNext, separately and together). Clip-level and participant-level swallowing status predictions were obtained through a voting method. Results: The models demonstrated clip-level dysphagia screening sensitivity of 71% and specificity of 77% (F1 = 0.73, AUC = 0.80 [95% CI: 0.78-0.82]). At the participant level, the sensitivity and specificity were 89 and 79%, respectively (F1 = 0.81, AUC = 0.91 [95% CI: 0.77-1.05]). Discussion: This study is the first to demonstrate the feasibility of applying deep learning to classify vocalizations to detect post-stroke dysphagia. Our findings suggest potential for enhancing dysphagia screening in clinical settings. https://github.com/UofTNeurology/masa-open-source.

Oxidative stress-dependent MMP-13 activity underlies glucose neurotoxicity.

BACKGROUND: A complication of diabetes is neuropathy, a condition of sensory axon degeneration that originates in the epidermis. The mechanisms remain unknown but reactive oxygen species (ROS) have been implicated in this condition. In this study, we assessed the role of ROS and a candidate downstream target, MMP-13 in glucose-induced sensory axon degeneration in zebrafish and mice. METHODS: The effects of glucose on metabolism and sensory axon degeneration were assessed using qPCR and live imaging. ROS were analyzed using pentafluorobenzene-sulfonyl fluorescein and activation of the NF-κB stress response was determined using Tg(NF-κB:GFP) zebrafish. The role of MMP-13 and ROS in glucose-dependent axon degeneration was determined in zebrafish following treatment with the antioxidant, N-acetylcysteine and the MMP-13 inhibitor, DB04760. Neuropathic mice fed on a high-fat/high-sugar diet were treated with the MMP-13 inhibitor, CL-82198 to assess sensory recovery. RESULTS: Glucose treatment of zebrafish induced metabolic changes that resemble diabetes. Sensory axon degeneration was mediated by ROS-induced MMP-13 and prevented upon antioxidant treatment or MMP-13 inhibition. MMP-13 inhibition also reversed neuropathy in diabetic mice. CONCLUSION: We demonstrate that zebrafish are suitable to study glucose-induced neurotoxicity. Given the effects in zebrafish and mice, MMP-13 inhibition may be beneficial in the treatment of human diabetic neuropathy.

Correction: A single Danio rerio hars gene encodes both cytoplasmic and mitochondrial histidyl-tRNA synthetases.

[This corrects the article DOI: 10.1371/journal.pone.0185317.].

A single Danio rerio hars gene encodes both cytoplasmic and mitochondrial histidyl-tRNA synthetases.

Histidyl tRNA Synthetase (HARS) is a member of the aminoacyl tRNA synthetase (ARS) family of enzymes. This family of 20 enzymes is responsible for attaching specific amino acids to their cognate tRNA molecules, a critical step in protein synthesis. However, recent work highlighting a growing number of associations between ARS genes and diverse human diseases raises the possibility of new and unexpected functions in this ancient enzyme family. For example, mutations in HARS have been linked to two different neurological disorders, Usher Syndrome Type IIIB and Charcot Marie Tooth peripheral neuropathy. These connections raise the possibility of previously undiscovered roles for HARS in metazoan development, with alterations in these functions leading to complex diseases. In an attempt to establish Danio rerio as a model for studying HARS functions in human disease, we characterized the Danio rerio hars gene and compared it to that of human HARS. Using a combination of bioinformatics, molecular biology, and cellular approaches, we found that while the human genome encodes separate genes for cytoplasmic and mitochondrial HARS protein, the Danio rerio genome encodes a single hars gene which undergoes alternative splicing to produce the respective cytoplasmic and mitochondrial versions of Hars. Nevertheless, while the HARS genes of humans and Danio differ significantly at the genomic level, we found that they are still highly conserved at the amino acid level, underscoring the potential utility of Danio rerio as a model organism for investigating HARS function and its link to human diseases in vivo.

Identification of target genes downstream of semaphorin6A/PlexinA2 signaling in zebrafish.

BACKGROUND: Semaphorin (Sema)/Plexin (Plxn) signaling is important for many aspects of neuronal development, however, the transcriptional regulation imposed by this signaling pathway is unknown. Previously, we identified an essential role for Sema6A/PlxnA2 signaling in regulating proliferation and cohesion of retinal precursor cells (RPCs) during early eye development. This study used RNA isolated from control, Sema6A-deficient and PlxnA2-deficient zebrafish embryos in a microarray analysis to identify genes that were differentially expressed when this signaling pathway was disrupted. RESULTS: We uncovered a set of 58 transcripts, and all but 1 were up-regulated in both sema6A and plxnA2 morphants. We validated gene expression changes in subset of candidates that are suggested to be involved in proliferation, migration or neuronal positioning. We further functionally evaluated one gene, rasl11b, as contributing to disrupted proliferation in sema6A and plxna2 morphants. Our results suggest rasl11b negatively regulates proliferation of RPCs in the developing zebrafish eye. CONCLUSIONS: Microarray analysis has generated a resource of target genes downstream of Sema6A/PlxnA2 signaling, which can be further investigated to elucidate the downstream effects of this well-studied neuronal and vascular guidance signaling pathway. Developmental Dynamics 246:539-549, 2017. © 2017 Wiley Periodicals, Inc.