Eduardo Fernández
Director of the Institute of Bioengineering at the Miguel Hernandez University of Elche and director of the Biomedical Neuroengineering group at the Center for Biomedical Research Network on Bioengineering, Biomaterials and Nanomedicine (CIBER-BBN)
Patients with severe paralysis caused by conditions such as amyotrophic lateral sclerosis (ALS) or other conditions, such as spinal cord injuries, often experience significant communication and mobility problems. These problems vary depending on the level and severity of the injury or disease and make any task a major challenge. However, despite their physical limitations, most of these patients' brain activity remains intact. This means they are able to continue thinking, feeling, and making decisions. To improve their quality of life, brain-computer interfaces (BCIs) are being developed. These devices record the electrical activity of these patients' brains and transform these signals into commands or instructions that can be used to control, in real time, external devices such as robotic arms, computer cursors, or wheelchairs.
Traditionally, these brain-computer interfaces (BCIs) rely solely on recording brain activity to control external devices. However, many of the activities performed with the aid of these devices are goal-oriented. Consider, for example, searching for a glass of water or a specific icon on a computer screen. Once the goal is known, subsequent human actions are often quite stereotypical and can be assisted with artificial intelligence techniques. The challenge is identifying the specific goal the user is pursuing. In this context, the group led by Dr. Jonathan Kao at the University of California, Los Angeles (UCLA) has investigated the possibility of using other sources of information, such as previous movements and context, along with advanced artificial intelligence techniques to infer the user's goal and assist in their movements. In this way, AI helps decode the user's intentions from their brain signals, even when these are noisy or incomplete. This reduces cognitive load and improves the experience, without taking control away from the user.
This is a preliminary proof-of-concept study. The results of their experiments in three healthy controls and one patient with a spinal cord injury suggest that this approach can significantly improve control of the movement of a computer cursor and the actions of a robotic arm. For example, the paralyzed participant in these studies performed almost four times better in cursor control, thanks to the assistance of the AI assistant. Furthermore, this participant was able to control a robotic arm to move a series of blocks to random locations, a task they could not perform without the assistance of the AI assistant.
This technology offers a more intuitive and functional framework for the development of new brain-computer interfaces. However, it should be noted that this research was conducted on a single patient, so there is no control group nor is it a randomized trial. Furthermore, the tasks used in the study were not designed to simulate activities of daily living. Therefore, we must be aware that further clinical studies with a sufficient number of patients and in larger settings are still needed.
The future is bright, and we must be prepared to use the results of this research to improve the quality of life of patients with communication and mobility problems. However, many issues affecting the performance and widespread use of these systems still need to be resolved.