Building Brain–Machine Interfaces to Restore Neurological Functions

Excerpt

Modern research on brain–machine interfaces (BMI) is a highly multidisciplinary field that has been developing at a stunning pace since the first experiment conducted 8 years ago that demonstrated direct control of a robotic manipulator by ensembles of neurons recorded in cortical and subcortical areas in awake, behaving rats (Chapin, Moxon et al. 1999). Since this pioneering study, an exponentially growing stream of research publications has provoked an enormous interest in BMIs among scientists from different fields and the lay public. This level of interest stems from both the use of BMIs to investigate the way large and distributed neural circuits operate in behaving animals and the perceived potential that BMI technology can realize for restoration of motor behaviors and other functions in patients suffering from devastating neurological conditions.

In theory, the group of patients that can benefit from BMI systems includes people who lost mobility as a consequence of neurodegenerative disorders, such as amyotrophic lateral sclerosis (ALS), severe trauma and irreversible spinal cord injuries, stroke, and cerebral palsy. As the risk–benefit factor of invasive BMIs improves, it is conceivable that the same technology may be accepted by patients with less severe degrees of body paralysis or even by amputees.

Future BMI technologies will not be limited to systems for restoration of mobility. We expect that systems for restoration of speech and restoration of communication between brain areas will likely emerge. These future neuroprostheses are expected to be seamlessly integrated with the human body as much as possible and to use the most advanced developments in robotics, material science, computational algorithms, and electrical engineering.

Notwithstanding these high expectations, much work has to be done to develop solutions for numerous issues that preclude an immediate translation of laboratory demonstrations into practical clinical applications. Most of BMI research has been conducted in monkeys and rats, and clinical trials in humans are only starting. In this chapter, we highlight the major obstacles faced by BMI research and lay out a roadmap that can transform experimental advances into clinical applications that will benefit millions of people worldwide in the next decade. This roadmap is based on a critical analysis of previous studies conducted in both experimental animals and human subjects. The milestones that we propose take into account the experience accumulated during the last 5 years by a multiuniversity consortium led by the Duke University Center for Neuroengineering.