Background Surgical resection of tongue cancer may impair swallowing and speech. Knowledge of tongue muscle architecture affected by the resection could aid in patient counseling. Diffusion tensor imaging (DTI) enables reconstructions of muscle architecture in vivo. Reconstructing crossing fibers in the tongue requires a higher‐order diffusion model. Purpose To develop a clinically feasible diffusion imaging protocol, which facilitates both DTI and constrained spherical deconvolution (CSD) reconstructions of tongue muscle architecture in vivo. Study Type Cross‐sectional study. Subjects/Specimen One ex vivo bovine tongue resected en bloc from mandible to hyoid bone. Ten healthy volunteers (mean age 25.5 years; range 21–34 years; four female). Field Strength/Sequence Diffusion‐weighted echo planar imaging at 3 T using a high‐angular resolution diffusion imaging scheme acquired twice with opposing phase‐encoding for B 0 ‐field inhomogeneity correction. The scan of the healthy volunteers was divided into four parts, in between which the volunteers were allowed to swallow, resulting in a total acquisition time of 10 minutes. Assessment The ability of resolving crossing muscle fibers using CSD was determined on the bovine tongue specimen. A reproducible response function was estimated and the optimal peak threshold was determined for the in vivo tongue. The quality of tractography of the in vivo tongue was graded by three experts. Statistical Tests The within‐subject coefficient of variance was calculated for the response function. The qualitative results of the grading of DTI and CSD tractography were analyzed using a multilevel proportional odds model. Results Fiber orientation distributions in the bovine tongue specimen showed that CSD was able to resolve crossing muscle fibers. The response function could be determined reproducibly in vivo. CSD tractography displayed significantly improved tractography compared with DTI tractography ( P = 0.015). Data Conclusion The 10‐minute diffusion imaging protocol facilitates CSD fiber tracking with improved reconstructions of crossing tongue muscle fibers compared with DTI. Level of Evidence: 2 Technical Efficacy: Stage 1 J. Magn. Reson. Imaging 2019;50:96–105.
Recent advances in real-time magnetic resonance imaging (rtMRI) of the vocal tract provides opportunities for studying human speech. This modality together with acquired speech may enable the mapping of articulatory configurations to acoustic features. In this study, we take the first step by training a deep learning model to classify 27 different phonemes from midsagittal MR images of the vocal tract. An American English database was used to train a convolutional neural network for classifying vowels (13 classes), consonants (14 classes) and all phonemes (27 classes) of 17 subjects. Classification top-1 accuracy of the test set for all phonemes was 57%. Error analysis showed voiced and unvoiced sounds often being confused. Moreover, we performed principal component analysis on the network's embedding and observed topological similarities between the network learned representation and the vowel diagram. Saliency maps gave insight into the anatomical regions most important for classification and show congruence with known regions of articulatory importance. We demonstrate the feasibility for deep learning to distinguish between phonemes from MRI. Network analysis can be used to improve understanding of normal articulation and speech and, in the future, impaired speech. This study brings us a step closer to the articulatory-to-acoustic mapping from rtMRI.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.