Closed-loop Deep Brain Stimulation Using Neural Biomarkers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current closed-loop deep brain stimulation methods rely on generic, nonspecific biomarkers, which are inadequate for effectively addressing the distinct motor dysfunctions in Parkinson's disease, as they fail to differentiate between tremor and bradykinesia, leading to inadequate treatment outcomes.
Innovation Solution
The use of machine learning models to decode neural signals associated with tremor and bradykinesia from subthalamic and cortical recordings, allowing for precise identification and targeted deep brain stimulation based on specific motor features, with dorsal STN contacts sensing tremor signals and ventral contacts identifying bradykinesia signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If generic, nonspecific biomarkers are used to drive deep brain stimulation, then the stimulation can be implemented with simpler control mechanisms, but the treatment effectiveness is reduced because it fails to address distinct motor dysfunctions like tremor and bradykinesia separately
Solution Approach 1:
The patent segments the treatment by using separate control signals for different motor dysfunctions. Specifically, beta oscillation amplitude modulates stimulation amplitude to address bradykinesia, while beta oscillation frequency modulates stimulation frequency to address tremor. This segmentation allows each aspect of motor dysfunction to be treated independently with optimized parameters, improving treatment effectiveness without requiring overly complex integrated control mechanisms.
2Adaptability or versatility
If multiple control signals are used to drive location, type, and timing of stimulation, then distinct aspects of disease expression can be addressed, but the system complexity increases
Solution Approach 1:
The patent applies local quality by modulating different aspects of stimulation (amplitude and frequency) based on specific neural signals. The beta oscillation amplitude specifically controls stimulation amplitude for bradykinesia treatment, while beta oscillation frequency specifically controls stimulation frequency for tremor treatment. This localized control approach enables the system to address distinct motor features with appropriate parameters without requiring complete system redesign for each condition.
Solution Approach 2:
The patent implements dynamic control where stimulation parameters are continuously adjusted based on real-time neural signal characteristics. The beta oscillation amplitude and frequency dynamically modulate the stimulation amplitude and frequency respectively, allowing the system to adapt to changing motor symptoms. This dynamic adjustment provides versatility in treating different motor features while using a unified control framework that manages complexity through systematic parameter coupling.
3Measurement precision
If machine learning models are used to differentiate neural signals associated with tremor and bradykinesia, then precise identification of motor features is achieved, but the processing time and computational complexity increase
Solution Approach 1:
The patent replaces complex machine learning differentiation processes with a more efficient signal processing approach. Instead of using sophisticated ML models to differentiate between tremor and bradykinesia neural signals, the system directly utilizes beta oscillation characteristics (amplitude and frequency) as separate control signals. This substitution maintains measurement precision by leveraging the inherent discriminative properties of beta oscillations while significantly reducing computational complexity and processing time.
Data Source
AI summary
Closed-loop deep brain stimulation using neural and behavioral biomarkers of specific motor features is provided via training a machine learning model to differentiate first neural signals generated by a biological subject that are associated with tremor in the biological subject from second neural signals generated by the biological subject that are associated with bradykinesia in the biological subject; placing a first plurality and a second plurality of deep brain stimulation electrodes in a subthalamic nucleus (STN) of the biological subject; placing a plurality of microelectrodes and a plurality of macroelectrodes in the STN; in response to identifying, by the machine learning model, a given neural signal as one of the first or second neural signals, activating a corresponding one of the first or second plurality of electrodes with a therapeutically effective voltage and current to affect deep brain stimulation in the biological subject to treat a neuromotor disorder.