Deep learning and predictive modelling for generating normalised muscle function parameters from signal images of mandibular electromyography

Abstract

Challenges arise in accessing archived signal outputs due to proprietary software limitations. There is a notable lack of exploration in open-source mandibular EMG signal conversion for continuous access and analysis, hindering tasks such as pattern recognition and predictive modelling for temporomandibular joint complex function. To Develop a workflow to extract normalised signal parameters from images of mandibular muscle EMG and identify optimal clustering methods for quantifying signal intensity and activity durations. A workflow utilising OpenCV, variational encoders and Neurokit2 generated and augmented 866 unique EMG signals from jaw movement exercises. k-means, GMM and DBSCAN were employed for normalisation and cluster-centric signal processing. The workflow was validated with data collected from 66 participants, measuring temporalis, masseter and digastric muscles. DBSCAN (0.35 to 0.54) and GMM (0.09 to 0.24) exhibited lower silhouette scores for mouth opening, anterior protrusion and lateral excursions, while K-means performed best (0.10 to 0.11) for temporalis and masseter muscles during chewing activities. The current study successfully developed a deep learning workflow capable of extracting normalised signal data from EMG images and generating quantifiable parameters for muscle activity duration and general functional intensity. Graphical Abstract

Predictive modelling of freeway space utilising clinical history, normalised muscle activity, dental occlusion, and mandibular movement analysis

“…Standardisation across all 66 participants was achieved by normalising signal sweeps for each activity using an in-house deep learning-based software, which produced standardised quotients for both muscle intensity and activity duration 15 . The EMG images were standardised to range between 1604 × 579 pixels and 1617 × 590 pixels, with padded normalisation to ensure that the resulting signal had a standard length of max(M, 5)− min(M, 5) + 1 16 . These normalisation methods were implemented following previous evaluations 16 , that were subsequently repurposed to create an in-house, open-source signal processing tool, Dental Loop Signals v1.0 ( https://github.com/SoftwareImpacts/SIMPAC-2023-498 ) 15 .…”

“…The EMG images were standardised to range between 1604 × 579 pixels and 1617 × 590 pixels, with padded normalisation to ensure that the resulting signal had a standard length of max(M, 5)− min(M, 5) + 1 16 . These normalisation methods were implemented following previous evaluations 16 , that were subsequently repurposed to create an in-house, open-source signal processing tool, Dental Loop Signals v1.0 ( https://github.com/SoftwareImpacts/SIMPAC-2023-498 ) 15 .…”

Variables Associated with Jaw Clicking in a South Australian Population: A Cross-Sectional Study

“…Dynamic assessments involve evaluating jaw functions during the movement of the jaws and incorporate soft tissue parameters, such as muscle activity during mouth opening and chewing, along with mandibular hard tissue translation, most frequently during vertical displacements like mouth opening [15]. To quantitatively assess mandibular function during movement, certain parameters such as maximum mouth opening (MMO), masticatory muscle activities, and independent joint vibration integrals and amplitude values exceeding 300 Hz can be measured [16,17]. Clinically, while definitive diagnoses of the TMJ can only be made through magnetic imaging, these parameters can aid in the preliminary non-invasive guidance towards considering possible temporomandibular disorders (TMDs) using the Piper's classification system [16].…”