Prosthetic Categories

Combining Brain-Computer Interfaces and AI in Neuroprosthetics

    The development of brain-computer interfaces (BCI) is thought to be the most important technological breakthrough in recent years for people experiencing limb loss. BCIs, which allow the brain and prosthesis to communicate directly, play a huge role in developing several brain-controlled prosthetic limbs.

     Brain-computer interfaces and artificial intelligence mean better neuroprosthetics.

    However, the technology itself is flawed, and experts say BCIs would work better when paired with artificial intelligence (AI) systems. In this article, we take a peek into how combining both systems means better neuroprosthetics.  

    How does BCI work?  

    BCI attempts to mimic how a brain would communicate with a biological limb to move and perform tasks. To bring the same dynamic into a brain and prosthesis setup, a method to detect signals is required.

    There are various ways to detect brain signals, and these can be classified into two categories: invasive and non-invasive.

    Invasive methods include electrocorticography, which uses electrodes placed directly on an exposed surface of the brain, and microelectrode arrays, which can record and stimulate neurons at multiple sites simultaneously, among others. 

    Non-invasive systems include electroencephalography, which uses small, metal discs attached to the scalp; functional magnetic resonance imaging, which uses imaging technology to measure brain activity by detecting changes associated with blood flow; functional near-infrared spectroscopy, which uses optical imaging to measure changes in hemoglobin concentrations within the brain; and magnetoencephalography, which maps brain activity by recording magnetic fields produced by electrical currents occurring naturally in the brain.

    Many prefer the latter because non-invasive systems don't carry the risk of tissue damage, plus it can be implemented quickly.  

    Using these electrophysiological techniques, BCIs are equipped to record the brain's activity, decode its meaning, and influence activity in specific regions to affect function.

    Limitations of BCI  

    However, BCIs have their limitations. Neuroscientists cannot clearly discern a person's intentions based on the brain's background electrical activity and match it to the actions of, say, a robotic arm.

    It's a good thing that recent developments in AI methodologies show that it's capable of outperforming humans in encoding and decoding neural signals. AI is primarily seen as the perfect helper in processing signals from the brain before they reach the prosthesis.

    How does AI enhance BCI?  

    AI is a set of general methodologies that uses a computer to model intelligent behavior with minimal human involvement. AI can be taught to match and even surpass human performance in specific tasks.

    For AI to work with BCIs, there needs to be a constant provision of internal parameters to the algorithms. These internal parameters may include electrical properties of the neural tissues, the device's energy consumption, and stimulation frequencies.

    The AI algorithms then use the information to identify which data is useful and the logic in the data while concurrently producing the desired functions.

    BCI and AI in neuroprosthetics and limb rehabilitation  

    BCI in limb rehabilitation and neuroprosthetics has evolved to allow users to perform essential tasks, such as arm movements, self-feeding, and reaching and grasping. Both invasive and non-invasive BCI systems are capable of neural control of a robotic arm in humans. This was seen in 2012 and 2013 when two separate studies sought the capability of people with tetraplegia to reach and grasp using a brain-controlled robotic arm.  

    However, researchers have found that approaches using non-invasive systems provide limited control, and most complex movements rely on AI. Researchers are looking into the development of AI to provide a better way to decode neural signals faster.

    The state of the technology  

    As of this writing, the use of BCIs and AI for cognitive training is still in its infancy, clinical BCI applications are currently limited, and the algorithms produced by machine learning systems are rarely predictable and understandable in the real world. Until these issues are addressed, and researchers find a way to understand what happens between a person's thoughts and the technology acting on their behalf, combining BCI and AI in neuroprosthetics and rehabilitation may only lead to problems.

    But one thing's for sure: BCI based on AI (artificial intelligence) is a rapidly advancing field. Many in the scientific community want to refine the technology further to improve the quality of life of people living with limb loss.