Neurorobotics represents a symbiosis between brain-computer interfaces (BCIs), artificial intelligence (AI) and robotics. It is comprised of embodied autonomous neurorobots whose “brains” are modeled after the neural networks in the human brain using AI and either prosthetic or orthopedic devices that are controlled with BCIs. Neurorobotics aims to create brains for bodies and bodies for brains, making robots with “human” brains that can better interact within real-world environments, and humans with “enhanced” brains and robotic body parts that can be intuitively controlled, in a manner similar to their biological counterparts. In this article, we will focus on the latter use cases.
Early neurorobots have been used in various industries such as the military, manufacturing, and entertainment, but revolutionary advances are expected for first come in the healthcare industry – upgrading standard methods for therapy and rehabilitation of patients suffering from a range of conditions including motor disabilities, poor cognition, hearing loss, attention-deficit hyperactivity disorder (ADHD), and autism spectrum disorder (ASD).
Artificial limbs and musculoskeletal disorders
Of note, efforts to improve the quality of life of patients suffering from amputations or musculoskeletal disorders have led to a boom in the field of neurorobotics. Neural prosthetics, orthoses and exoskeletons are currently in development, however clinical applications are likely still a few years out.
Importantly, neural prosthetics can be categorized as either invasive or non-invasive, depending on the type of signal receiver/recorder. Invasive neural prosthetics use microelectrodes implanted in the surface of the brain cortex (or at the nerve endings) that record neuronal signals while the user is performing motor imagination (MI). These signals are then decoded using mathematical software and transmitted to robotic arms or legs for task execution. A few such products are in advanced stages of development, one of which can be seen in the video below.
Non-invasive neural prosthetics use electroencephalograms (EEGs) as a way of recording neuronal activity upon MI. The neuronal signals are then processed in a similar manner to approaches that involve invasive prosthetics. Of note, since the received signal is lower in frequency and voltage and is less specific, it requires special noise filters, enhancers and continuous pursuit models that are stable enough to improve signal acquisition and also reliable enough for practical uses. Another example of a non-invasive BCI approach of treating musculoskeletal disorders such as paraplegia is the use of lower limb exoskeletons. While a product based on this approach is not yet on the market, pure exoskeletons (without BCI) are currently available, primarily for industrial purposes.
For now, which is better? While invasive approaches (using microelectrodes) can receive clearer signals that are strong enough to control a robotic arm, they cannot record signals from deep within the brain or from larger areas that include other limbs or impairments (which non-invasives can). Furthermore, the fact that they are invasive comes with many challenges; for example, that the implantable material needs to be biocompatible, safe and durable for years. A few attempts are on the horizon to remove these challenges.
Hearing impairment
According to the World Health Organization (WHO), roughly 3% of the world population has a disabling hearing impairment (ranging from moderate to complete deafness). Children with this disability have difficulties in speech development, language and improvement of cognitive skills, often performing poorly in school and later on in life. This is usually accompanied by stigmatization and social isolation. Until now, hearing impairment was treated using various enhancers and implants that needed surgery. An attempt to overcome this disability without surgery using a non-invasive approach with EEG and BCI entailed transforming musical information into different stimuli – such as vibration and tactile stimuli. In this study, different types of songs were played to the patients through a specially designed chair and arm bracelet that transmitted vibrations, while simultaneously recording brain activity with an EEG. Even though further development remains, the results obtained in this study were very promising.
Mental health and cognitive disturbances
Applications based on BCIs have been explored with the aim of improving attention span, executive functions and cognitive performance in patients of all ages. Traditionally, methods were based on suppressing and enhancing specific EEG waves during performance of a specific task. An example of such an application developed in 2014, for children with ASD, required that the patients modulate their brain activity and physiological responses while being active in social games, receiving an appropriate emotional feedback for a specific social encounter. Another application designed two years later, for cognitive improvement in the elderly, used an MI-based BCI, since the EEG-alpha and EEG-beta activities used for conducting motor functions are also associated with cognitive performance. Two birds with one stone! An example of a novel approach from 2019, in treating ADHD in children with a series of training games, was designed in a way that BCI can use individualized EEG patterns that represent optimal attention span. With training activities and motivation to proceed with the game, which can last for months, attention span was shown to be widened and maintained. All of these applications are still trials of the technologies, but it is expected that they will soon be eligible for commercial use.
Personalized medical treatment based on neurorobotics
With these types of devices, it is possible to gather tremendous amounts of healthcare data. By collecting various signals (from the brain, nerves, muscles, bloodstream, etc.), we could gain a proper insight into patients’ health status. How it could be done? If during EEG recording low transmission of brain signals happen (excluding technical issues), it could indicate sensorimotor lesions in the cortex that the patient has suffered and may not even know about (after trauma, minor stroke). With continuous recording of the brain activity, it may be even possible to predict epileptic seizures or visualize any other brain-wave irregularities (eg. insomnia, mood swings and disorders).
Another example of useful data received from implanted microelectrodes at the nerve endings, could be a timeline of patients’ regeneration process of the skin, muscles and nerves long after amputation or trauma. Based on these observations, personalized medical treatment could be developed for each patient, by specifying which therapy or rehabilitation method is needed, lowering the costs of delayed care or diagnosis and remarkably improving the quality of life.
How will neurorobotics disrupt healthcare?
The application of neurorobotics within the healthcare sector will require a multidisciplinary approach with collaboration between individuals from a variety of fields – neuroscientists, engineers, mathematicians, doctors, economists and sales marketing experts. Key technical challenges include smaller devices, lighter and more comfortable materials, and wireless and user-friendly aids, in addition to improved reliability and accuracy.
As these technologies reach stages of commercial viability, additional challenges in implementation must be overcome. How will the patients adapt to new aids, and what are potential problems created during use of these devices (skin, musculoskeletal, biomechanical and cardio-metabolic risks)? Can BCI systems be reliable and valid enough, having proper self-calibration, self-analysis and error-correction — without the need for continuous supervision? Will these aids be accessible and efficient in the long run? Finally, how will society react to the “upgraded” human body parts – will they be accepted, mocked, feared or admired (the “disabled or superhuman” question)?
Conclusion
Neurorobotics represents an exciting research area with enormous potential to impact the future of healthcare. Acknowledging that every organism is a complex yet beautiful machine of its own that can experience glitches that lead to irregularities and malfunctioning will require fixing and maintenance by experts. Nevertheless, not every glitch is the same, as no machinery is the same. Developing and upgrading personalized medical treatment with neurorobotics will allow to improve the quality of life for millions of individuals, bringing back motivation and self-dignity to those truly needing it to help us create healthy bodies for brains and even healthier brains for bodies.
Featured image shows SuitX’s PHOENIX Medical Exoskeleton (Source: SuitX).
