Company: Tampere University of Technology
NI Product(s) Used: LabVIEW, myRIO
Industry: Biomedical, Research
Unilateral facial paralysis is a condition in which one side of the face is either partially or totally disabled due to paralysis of the facial nerve. There are a number of possible causes for the paralysis, including Bell’s palsy and physical trauma.
Unilateral facial paralysis is fairly common. Its most common and idiopathic form alone reportedly affects approximately one in every 60–70 persons. The paralysis affects daily activities such as blinking, eating, drinking, and social interaction. Problems with blinking can cause dryness of the eye and even loss of vision. A drooping mouth corner can make eating and drinking difficult. Smiling and other socially-important facial expressions become distorted. People suffering from paralysis may have concerns related to their appearance and may feel they cannot fully express themselves.
The goal of the treatment of facial paralysis is regaining lost functions to overcome the negative impact they have on the quality of life.
Permanent forms of the paralysis can be treated surgically. However, surgery involves risks, is expensive, and regaining lost functions is not guaranteed. In reality, some nerve repair surgery techniques have reported success rates as low as 20 percent.
Prosthetic technology is another approach to regain lost facial functions. Functional electrical stimulation can be used to activate facial muscles for reanimation. Facial pacing describes the act of reanimating the disabled side of the face based on the simultaneous measurement of the healthy side. For example, a smile detected on the healthy side of the face can be replicated on the disabled side using electrical stimulation. This type of pacing can provide a short-term solution to regain facial control while awaiting surgery or, in the case of temporary paralysis, while waiting for the paralysis to heal. For short-term facial pacing, we can use wearable devices that carry out measurement and stimulation via electrodes placed on the skin. In the future, prosthetic technology could be an effective alternative to costly repair surgeries whose success is never guaranteed. Miniaturized, implantable devices would be less obtrusive and more suitable for continuous use.
An effective facial pacing system has strict computational demands. Aside from fulfilling common medical device requirements, pacing requires determinism and true parallelism to achieve synchronous facial expressions. An acceptable delay between the sides of the face (detected versus stimulated movement) is tens of milliseconds.
Our research focuses on developing transcutaneous facial pacing beyond the state-of-theart. Current limitations include difficulties in activating the targeted muscles, difficulties in producing varying levels of muscle contraction, and problems related to the comfort of the stimulation. In order to study how these issues could be solved, we needed a research prototype system that allows us to fully customize the processing, from measurement to stimulation.
Currently, we have designed and constructed a prototype system that we are using for clinical trials under the supervision of the National Supervisory Authority for Welfare and Health (Valvira) in Finland. We have already carried out clinical trials at the local hospital to study the excitability of facial muscles of both healthy participants and patients with unilateral facial paralysis. We also have our first implementation of pacing for reanimating eye blinks. We are developing our system further to continue with more experimental research and clinical trials.
Our original idea was to make our own custom device with amplifiers for the measurement and stimulation, and combine it with a laptop to carry out the required processing. However, the connectivity to the computer would have created a bottleneck, leading to unacceptable processing delays. NI reconfigurable I/O (RIO) platform represented an ideal solution, offering reliable, deterministic computational power, whilst simplifying the integration of our custom electronics to form our facial pacing prototype system (Rantanen et al. 2016). Our device currently has four EMG measurement amplifiers with appropriate gain and bandwidth. Our four stimulation amplifiers produce constant-current signals to varying loads with ±48 mA maximum amplitude, and nominal maximum voltage of 100 V. The amplifier bandwidths allow arbitrary waveforms to be produced.
We chose the NI myRIO device as the embedded controller. We used it to measure the signals from the EMG amplifiers, carry out real-time processing on the data streams, and generate the signals for the stimulation amplifiers. A custom LabVIEW application, running on a standard laptop, wirelessly connects to the myRIO to provide a graphical user interface (GUI) and data logging. The GUI is required for configuring the processing and stimulation parameters, providing real-time visualisation of the signals, and setting-up the data logging. However, we use the myRIO integrated FPGA for time-critical tasks. Generation of accurate square wave stimulation pulses with pulse widths less than a millisecond wouldn’t be possible without an FPGA program running at tens of kilohertz. We also currently sample and process the EMG signals in the FPGA because a high sampling frequency allows fast detection of the onset of muscle activity from the background noise.
The myRIO greatly simplified the design of our prototype system, allowing us to implement a vast amount of functionality on a single device—from analog and digital I/O, to customized signal processing, to wireless connectivity. We use analog I/O to measure and generate signals. We need digital I/O to toggle our amplifiers and physical indicator LEDs and to read UI buttons and the status indicators of our stimulation amplifiers. The built-in wireless connectivity of the myRIO was important because we designed the prototype system to be a medical-grade device. Physically connecting it to a computer would have caused safety concerns. Choosing myRIO as the system controller was also a future-proof design choice. NI RIO architecture helps us easily reuse all of our current software IP, whilst upgrading from myRIO to a higher-end processing target (such as Single-Board RIO) with larger FPGA capacity, if required in the future.
Video 1. Demonstration of Real-Time Eye Blink Pacing
The NI platform accelerated the development of our first facial pacing prototype system. We achieved more than would have been possible using alternative technologies. We can take advantage of the extensive programmability of the prototype to develop facial pacing further, helping us discover more specific requirements for designing dedicated, wearable, lightweight devices in the future.
This work is a part of the Mimetic Interfaces project funded by the Academy of Finland (funding decision numbers 278529, 276567, and 278312). Mimetic Interfaces project is:
Ville Rantanen, Antti Vehkaoja, Jarmo Verho, Petr Veselý, Jani Lylykangas, Mirja Ilves, Eeva Mäkelä, Markus Rautiainen, Veikko Surakka, and Jukka Lekkala.
Prosthetic pacing device for unilateral facial paralysis.
In Proceedings of the XIV Mediterranean Conference on Medical and Biological Engineering and Computing 2016, volume 57 of IFMBE Proceedings, pages 647–652, Paphos, Cyprus, March-April 2016.
Tampere University of Technology