The interface segments the hand into three groups-the thumb, the index and middle fingers, and the ring and small fingers. Each group can move independently in two dimensions, allowing simultaneous control of multiple parts. By mentally commanding these movements, the participant guided the quadcopter through a virtual obstacle course.
"This is a greater degree of functionality than anything previously based on finger movements," said Matthew Willsey, U-M assistant professor of neurosurgery and biomedical engineering, and the lead author of the study published in Nature Medicine. The research was conducted during Willsey's tenure at Stanford University, alongside his collaborators.
Noninvasive techniques, such as electroencephalography (EEG), capture brain signals from the scalp, but they lack the precision needed for fine motor control. This study demonstrated a sixfold improvement in quadcopter flight performance when signals were captured directly from motor neurons using implanted electrodes rather than EEG.
To implement the interface, surgeons placed electrodes in the participant's motor cortex, connecting them to a pedestal on the skull that interfaces with a computer. "It takes the signals created in the motor cortex when the participant tries to move their fingers and uses an artificial neural network to interpret those intentions for controlling virtual fingers," Willsey explained. "We then send a signal to control a virtual quadcopter."
Part of the BrainGate2 clinical trials, this research explores the fusion of neural signals and machine learning for external device control by individuals with neurological impairments. The participant, who sustained a spinal cord injury years earlier, began collaborating with the research team in 2016 and expressed a strong interest in flying.
"The quadcopter simulation was not an arbitrary choice; the participant had a passion for flying," noted Donald Avansino, a co-author and computer scientist at Stanford University. "The platform fulfilled his desire for flight while demonstrating control of multiple fingers."
Co-author Nishal Shah, soon to join Rice University as a professor of electrical and computer engineering, emphasized the broader implications of the study: "Controlling fingers is a stepping stone; the ultimate goal is whole body movement restoration."
Stanford neurosurgery professor Jaimie Henderson highlighted the societal impact of the research: "People often focus on restoring essential functions like eating, dressing, and mobility. But recreation and social connection are equally important. People want to play games and interact with their friends."
Henderson believes the technology's potential extends far beyond games: "With brain control of multiple virtual fingers, we can develop multifactor control schemes for diverse applications, from operating CAD software to composing music."
Research Report:Investigational Device. Limited by Federal law to investigational use
Related Links
University of Michigan
UAV News - Suppliers and Technology
Subscribe Free To Our Daily Newsletters |
Subscribe Free To Our Daily Newsletters |