Efficient, generalizable brain-computer interface for neural feedback control

This technology is a brain-computer interface system for controlling motor commands that can generalize across different movements and contexts for an efficient brain-computer interface communication method.

Unmet Need: Unsupervised control of motor commands for brain-computer interface

Current brain-computer interfaces require supervised calibration for each individual while still having issues of low recognition accuracy. To communicate with an external prosthetic device, an individual must perform a series of supervised tasks to calibrate the motor commands. There are currently no brain-machine interfaces that can perform in a generalized manner without context or supervision.

The Technology: Generalizable brain-computer interface for neural motor control

This technology describes a method for a movement- and context-agnostic brain-computer interface that does not require supervised calibration. Using invariant dynamics and transition of neural activity across different movements, this interface can generate motor control commands in a generalizable manner. The motor commands can be learned without prior context and supervision to provide an alternative method for brain-computer interface communication.

This technology has been validated in animal models.

Applications:

  • Research tool for brain-machine interfaces
  • Software for motor control
  • Interface for prosthetics
  • Research tool for deciphering brain dynamic patterns during different activities

Advantages:

  • Unsupervised interface
  • Generalizable motor control commands
  • Automated and unbiased motor command modeling
  • Expandable set of information that can be encoded

Lead Inventor:

Rui Costa, DVM, PhD

Patent Information:

Patent Pending (WO/2024/196737)

Related Publications:

Tech Ventures Reference: