Those with dysfunctional vocal cords can speak using AI-supported adhesive neck patch

The device is attached over the larynx or voicebox of a person, and can interpret the movements of the muscles as speech, using machine learning techniques. The researchers have demonstrated an accuracy of 95 per cent for the device.

The device has two components, a self-powered sensor that detects the movements and translates them into electrical signals. A machine learning algorithm then translates the signals into speech. There is also an actuation component to convert the speech signals into the desired vocal expression. Each of the two components have a layer of copper induction coils forming a magnetic induction layer, as well as a layer made up of a soft, biocompatible silicone compound called polydimethylsiloxane or PDMS. Sandwiched between the two components is a layer of micromagnets that generates a magnetic field.

When mechanical forces from the larynx induces changes in the magnetic field, the high fidelity magnetoelastic layers in each of the two components generate the necessary signals for sensing purposes. The research has been presented in Nature Communications.

The device will be tested on those with speech disorders

The current patches were tested by volunteers who spoke out preset sentences aloud and silently, to evaluate the performance of the neck patch. Lead author of the paper, Jun Chensays, “Existing solutions such as handheld electro-larynx devices and tracheoesophageal- puncture procedures can be inconvenient, invasive or uncomfortable. This new device presents a wearable, non-invasive option capable of assisting patients in communicating during the period before treatment and during the post-treatment recovery period for voice disorders.” The researchers intend to improve the vocabulary of the device, and test it on people with speech disorders.