Parallel Decoders for Neural Prosthetic Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current neural prosthetics are limited in their functionality as they can only perform continuous actions, lacking the ability to decode discrete action states necessary for more complex tasks like selecting or grasping, which restricts their clinical utility.
Innovation Solution
A brain-machine interface system that combines a continuous decoder for kinematics with a discrete action state decoder, learning distribution models and transition models for states like velocity, idle, and task states, enabling the control of prosthetic devices to perform discrete actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If only a continuous decoder is used to control neural prosthetics, then the device can perform continuous kinematic actions, but it cannot perform discrete actions such as selecting or grasping
Solution Approach 1:
The control system is segmented into two independent parallel decoders: a continuous decoder for kinematic control and a discrete action state decoder for discrete actions. This segmentation allows each decoder to specialize in its specific function while working together to provide comprehensive control capability.
Solution Approach 2:
The dual-decoder architecture provides multi-functionality by enabling the prosthetic device to perform both continuous movements and discrete actions through the same neural interface system. The continuous decoder handles position and velocity control while the discrete decoder handles selection and grasping, making the system universally applicable to various control tasks.
2Adaptability or versatility
If a discrete action state decoder is added in parallel with the continuous decoder, then the prosthetic device can perform discrete actions like clicking or grasping, but the system complexity increases
Solution Approach 1:
The control system is segmented into two independent parallel decoders: a continuous decoder for kinematic control and a discrete action state decoder for discrete actions. This segmentation allows each decoder to specialize in its specific function while working together to provide comprehensive control capability.
Solution Approach 2:
The continuous decoder and discrete action state decoder are merged into a unified control framework that processes neural signals simultaneously. Both decoders receive the same neural input and their outputs are combined to control the prosthetic device, achieving synergistic functionality.
3Reliability
If distribution models and transition models are learned for multiple discrete states, then reliable decoding of discrete action states is achieved, but the learning and computational requirements increase
Solution Approach 1:
Distribution models and transition models are learned in advance during a training phase before actual use. This preliminary learning allows the system to store pre-computed statistical models that can be rapidly applied during operation, improving real-time decoding reliability without adding computational burden during execution.
Solution Approach 2:
The system uses learned distribution models to evaluate the likelihood of different discrete states and uses transition models to predict state changes. This feedback mechanism continuously refines state estimation by comparing predicted transitions with actual observed states, improving decoding accuracy over time.
Data Source
AI summary
A brain machine interface for control of prosthetic devices is provided. In its control, the interface utilizes parallel control of a continuous decoder and a discrete action state decoder. In the discrete decoding, we not only learn states affiliated with the task, but also states related to the velocity of the prosthetic device and the engagement of the user. Moreover, we not only learn the distributions of the neural signals in these states, but we also learn the interactions/transitions between the states, which is crucial to enabling a relatively higher level of performance of the prosthetic device. Embodiments according to this parallel control system enable us to reliably decode not just task-related states, but any “discrete action state,” in parallel with a neural prosthetic “continuous decoder,” to achieve new state-of-the-art levels of performance in brain-machine interfaces.