Improved 3D real‐time MRI of speech production

Abstract

Purpose To provide 3D real‐time MRI of speech production with improved spatio‐temporal sharpness using randomized, variable‐density, stack‐of‐spiral sampling combined with a 3D spatio‐temporally constrained reconstruction. Methods We evaluated five candidate (k, t) sampling strategies using a previously proposed gradient‐echo stack‐of‐spiral sequence and a 3D constrained reconstruction with spatial and temporal penalties. Regularization parameters were chosen by expert readers based on qualitative assessment. We experimentally determined the effect of spiral angle increment and kz temporal order. The strategy yielding highest image quality was chosen as the proposed method. We evaluated the proposed and original 3D real‐time MRI methods in 2 healthy subjects performing speech production tasks that invoke rapid movements of articulators seen in multiple planes, using interleaved 2D real‐time MRI as the reference. We quantitatively evaluated tongue boundary sharpness in three locations at two speech rates. Results The proposed data‐sampling scheme uses a golden‐angle spiral increment in the kx−ky plane and variable‐density, randomized encoding along kz. It provided a statistically significant improvement in tongue boundary sharpness score (P < .001) in the blade, body, and root of the tongue during normal and 1.5‐times speeded speech. Qualitative improvements were substantial during natural speech tasks of alternating high, low tongue postures during vowels. The proposed method was also able to capture complex tongue shapes during fast alveolar consonant segments. Furthermore, the proposed scheme allows flexible retrospective selection of temporal resolution. Conclusion We have demonstrated improved 3D real‐time MRI of speech production using randomized, variable‐density, stack‐of‐spiral sampling with a 3D spatio‐temporally constrained reconstruction.

Multimodal dataset of real-time 2D and static 3D MRI of healthy French speakers

“…However, this method requires several identical repetitions of the same target utterance, which leads to artifacts in case of non-periodicity, and increases acquisition time. Real-time MRI allows high spatio-temporal resolution without repeating and is usually based on spoiled gradient echo sequences 8 , 13 , 14 . Acquisition can be sped-up by usage of non-cartesian (generally undersampled) schemes which ensure good coverage of the k-space centre.…”

“…A spiral encoding scheme was applied in 15 , 16 . and was thereafter combined with sparse-SENSE constrained reconstruction methods 8 , 17 . In 18 , 19 , a radial encoding scheme was used together with a compressed SENSE reconstruction.…”

Tongue Contour Tracking and Segmentation in Lingual Ultrasound for Speech Recognition: A Review

“…However, magnetic resonance imaging (MRI) has a better resolution, and it can provide more information about the soft tissues [ 11 ], vocal tract, and craniofacial structure [ 12 , 13 ]. MRI is used for real-time image acquisition [ 11 , 14 , 15 ] to visualize the vocal tract either in 2D or 3D orientation [ 16 , 17 ] and enhance the speech analysis. However, MRI is huge in size and very expensive compared to ultrasound.…”

Sub-millisecond 2D MRI of the vocal fold oscillation using single-point imaging with rapid encoding

“…MRI is already an established technique for the visualization of vocal tract dynamics in voice production [24][25][26] and singing [27][28][29]. Recently, dynamic 3D reconstructions of speech production with 15 fps have been presented [30].…”