Machine-Learning-Based Detecting of Eyelid Closure and Smiling Using Surface Electromyography of Auricular Muscles in Patients with Postparalytic Facial Synkinesis: A Feasibility Study.

Surface electromyography (EMG) allows reliable detection of muscle activity in all nine intrinsic and extrinsic ear muscles during facial muscle movements. The ear muscles are affected by synkinetic EMG activity in patients with postparalytic facial synkinesis (PFS). The aim of the present work was to establish a machine-learning-based algorithm to detect eyelid closure and smiling in patients with PFS by recording sEMG using surface electromyography of the auricular muscles. Sixteen patients (10 female, 6 male) with PFS were included. EMG acquisition of the anterior auricular muscle, superior auricular muscle, posterior auricular muscle, tragicus muscle, orbicularis oculi muscle, and orbicularis oris muscle was performed on both sides of the face during standardized eye closure and smiling tasks. Machine-learning EMG classification with a support vector machine allowed for the reliable detection of eye closure or smiling from the ear muscle recordings with clear distinction to other mimic expressions. These results show that the EMG of the auricular muscles in patients with PFS may contain enough information to detect facial expressions to trigger a future implant in a closed-loop system for electrostimulation to improve insufficient eye closure and smiling in patients with PFS.

Exploring intrinsic triggers for functional facial electrostimulation based on intramuscular electromyography recordings.

Based on univariate intramuscular electromyography (EMG) recordings of facial muscles of patients suffering from chronic idiopathic facial palsy we propose a data-driven feature selection process for the discrimination of different mimic maneuvers. Following fundamental ideas of automatic EMG decompositions based on templates defined by motor unit action potentials, the proposed approach relies on a multiple template matching. Yet, the novel methodology utilizes templates derived from the intramuscular EMG signal itself without any supervisor interaction or a priori information by identifying abundant short signal sections (motifs). Focusing on motifs as individual, characteristical graphoelements of an EMG recording implies a high level of flexibility. In connection with facial palsy such a flexibility is necessary, since unique individual, also pathological, EMG patterns can be expected due to the high spatial variability of intramuscular recordings combined with random patterns of aberrant reinnervation. The proposed methodology is applied to EMG data of frontalis, zygomaticus, and orbicularis oculi muscle without patient- or muscle-specific adaptations.

Modelling of cortical and thalamic 600 Hz activity by means of oscillatory networks.

The purpose of this study is to investigate information processing in the primary somatosensory system with the help of oscillatory network modelling. Specifically, we consider interactions in the oscillatory 600Hz activity between the thalamus and the cortical Brodmann areas 3b and 1. This type of cortical activity occurs after electrical stimulation of peripheral nerves such as the median nerve. Our measurements consist of simultaneous 31-channel MEG and 32-channel EEG recordings and individual 3D MRI data. We perform source localization by means of a multi-dipole model. The dipole activation time courses are then modelled by a set of coupled oscillators, described by linear second-order ordinary delay differential equations (DDEs). In particular, a new model for the thalamic activity is included in the oscillatory network. The parameters of the DDE system are successfully fitted to the data by a nonlinear evolutionary optimization method. To activate the oscillatory network, an individual input function is used, based on measurements of the propagated stimulation signal at the biceps. A significant feedback from the cortex to the thalamus could be detected by comparing the network modelling with and without feedback connections. Our finding in humans is supported by earlier animal studies. We conclude that this type of rhythmic brain activity can be modelled by oscillatory networks in order to disentangle feed forward and feedback information transfer.

EMG analysis of the thenar muscles as a model for EMG-triggered larynx stimulation.

Paralysis of one or both sides of the larynx musculature compromises breathing and speech function. Currently there is no surgical remedy to restore adequate function of the larynx. A plausible alternative solution is triggered electrical stimulation of the paralysed larynx site using a laryngeal pacemaker. Triggering of the pacemaker succeeds via constant EMG measurement of the muscle activity of the healthy larynx side. The EMG data analysis described in this work is one possible approach for regulating pacemaker triggering. In this study we used EMG data from the thenar muscles as a model to calculate a trigger point.