Systems and methods for optimizing programming and use of neuromodulation systems

• US 10,426,949 B2
• Filed: 10/26/2017
• Issued: 10/01/2019
• Est. Priority Date: 10/26/2016
• Status: Active Grant

• Current Assignee: Regents of The University of Minnesota (University of Minnesota)
• Original Assignee: Regents of The University of Minnesota (University of Minnesota)
• Inventors: Johnson, Matthew D., Pena, Edgar, Zhang, Simeng, Deyo, Steven, Xiao, YiZi
• Primary Examiner(s): Getzow, Scott M.
• Application Number:
  US15/794,333
• Publication Number:
  US 20180110973A1
• Time in Patent Office: 705 Days
• CPC Class Codes: A61B 5/055 involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging
• CPC Class Codes: A61B 5/4064 Evaluating the brain A61B5/031, A61B5/369, A61B5/14553 take precedence
• CPC Class Codes: A61B 5/4836 Diagnosis combined with treatment in closed-loop systems or methods A61B5/0036 takes precedence
• CPC Class Codes: A61N 1/0534 Electrodes for deep brain stimulation
• CPC Class Codes: A61N 1/36082 Cognitive or psychiatric applications, e.g. dementia or Alzheimer's disease
• CPC Class Codes: A61N 1/36128 Control systems
• CPC Class Codes: A61N 1/37247 User interfaces, e.g. input or presentation means

Abstract

Systems and methods for programming multi-electrode neuromodulation systems, with applications for at least deep brain stimulation (DBS) therapy. One or more configurations can be generated and presented to a user for selection. The selected configurations can be settings for programming a pulse generator. The one or more configurations can be Pareto optimal in terms of one or more objective values and can be generated using a particle swarm optimization. The generated configurations can be visualized on a Pareto front for user selection. Objective values can include minimizing power use, maximizing activation of neural pathways and/or regions of interest, minimizing activation of neural pathways and or regions of avoidance, and maximizing or minimizing the distance to sources of sensed functional data.

Claims

1. A method for programming a pulse generator with a configuration for a deep brain stimulation array including one or more electrodes, the configuration including a stimulation setting for each electrode in the deep brain stimulation array, the method comprising:

identifying one or more grid points representing a target tissue to be activated by stimulation through the one or more electrodes;

identifying one or more objective values to be balanced, the objective values being calculable based on the one or more grid points and a configuration;

generating one or more configurations, such that each configuration is Pareto optimal in terms of the objective values calculated based on the one or more grid points and the configuration;

presenting the one or more configurations to a user;

receiving from the user a selected one of the one or more configurations; and

providing, to a pulse generator capable of delivering therapy via each of the one or more electrodes, the selected configuration.

2. The method of claim 1, wherein generating the one or more configurations comprises:

in a memory, storing data elements representing one or more particles, each particle comprising a current configuration and a local leader configuration;

in the memory, allocating storage space for a set of dominant particles;

iteratively;

calculating the one or more objective values for each of the one or more particles based on the one or more grid points and the current configuration for the particle,

updating the set of dominant particles to contain only Pareto dominant particles, wherein a particle is Pareto dominant if there exist no other particles having at least one objective value that is better than a corresponding objective value of the particle and having no objective values that are worse than the corresponding objective values of the particle,

determining the local leader configuration for each particle,

determining a global leader configuration, and

modifying the current configuration for each particle based on the local leader configuration for the particle and the global leader configuration until a termination criterion has been met; and

providing the configurations of each particle in the set of dominant particles.

3. The method of claim 2, wherein the termination criterion is selected from the group consisting of:

a maximum number of iterations, convergence, and stalling.

4. The method of claim 2, wherein modifying the current configuration for each particle comprises modifying the current configuration for the particle based on an inertial component, a cognitive component, and a social component.

5. The method of claim 4, wherein:

each particle further comprises a set of previous configurations for the particle;

the inertial component is based on the current configuration for the particle and the set of previous configurations for the particle;

the cognitive component is based on the local leader configuration for the particle; and

the social component is based on the global leader configuration.

6. The method of claim 4, wherein modifying the current configuration for each particle further comprises modifying the current configuration for each particle based on one or more weighting factors.

7. The method of claim 6, wherein the functional data comprises spectral features of local field potential activity collected adjacent or distal to the stimulated electrodes.

8. The method of claim 2, further comprising:

identifying an objective function calculable based on the one or more objective values;

wherein determining the local leader configuration for each particle comprises

calculating the objective function for the current configuration for the particle, and

identifying the current configuration for the particle as the local leader configuration for the particle if the objective function for the current configuration is better than the objective function for a previous local leader configuration; and

wherein determining the global leader configuration comprises identifying the particle having the local leader configuration with the best objective function.

9. The method of claim 2, wherein each particle further comprises a local dominant set of configurations;

wherein calculating the one or more objective values for each of the one or more particles further comprises updating the local dominant set of configurations by

adding the current configuration to the local dominant set of configurations if there exist no other configurations in the local dominant set of configurations having at least one objective value that is better than the corresponding objective value of the current configuration and having no objective values that are worse than the corresponding objective values of the current configuration, and

removing any configuration from the local dominant set of configurations that have at least one objective value that is worse than the corresponding objective value of the current configuration and no objective values that are better than the corresponding objective values of the current configuration;

wherein determining the local leader configuration for each of the one or more particles comprises choosing a random configuration from the local dominant set of configurations; and

wherein determining the global leader configuration comprises choosing the local leader configuration for a random particle.

10. The method of claim 1 wherein the objective values are selected from the group consisting of:

a number of axons activated in one or more regions of interest,

a number of axons activated in one or more pathways of interest,

a number of axons activated in one or more regions of avoidance,

a number of axons activated in one or more pathways of avoidance, and

a total power required for stimulation.

11. The method of claim 1, further comprising:

receiving functional data related to the proximity of brain activity to each of the one or more electrodes;

wherein at least one of the objective values includes the proximity of brain activity to each of the electrodes activated by the configuration.

12. A system for stimulation of brain tissue via an implantable deep brain stimulation array including one or more electrodes, the system comprising:

a pulse generator programming device comprising;

a brain geometry engine configured to generate, from brain geometry data, one or more grid points representing a target tissue to be activated;

a discretization engine configured to determine and estimated modulation at each of the one or more grid points for a given configuration based on lead geometry data;

an optimization engine configured to generate one or more configurations, each configuration including a stimulation setting for each electrode, by

identifying one or more objective values to be balanced, the objective values being calculable based on the one or more grid points and a configuration; and

generating the one or more configurations such that each configuration is Pareto optimal in terms of the objective values calculated based on the one or more grid points and the configuration;

a user interface configured to present the generated one or configurations to the user and receive a selected one of the one or more configurations; and

a pulse generator in communication with the pulse generator programming device and electrically coupled to the implantable deep brain stimulation array and comprising a memory configured to store the selected configuration and a stimulation engine configured to deliver therapeutic pulses to each of the one or more electrodes according to the stimulation setting for the electrode of the configuration.

13. The system of claim 12, wherein the optimization engine comprises:

a memory adapted to store data elements representing one or more particles, each particle comprising a current configuration and a local leader configuration, and data elements representing a set of dominant particles; and

wherein the optimization engine is configured to generate the one or more configurations by iteratively;

calculating the one or more objective values for each of the one or more particles based on the one or more grid points and the current configuration for the particle,

updating the set of dominant particles to contain only Pareto dominant particles, wherein a particle is Pareto dominant if there exist no other particles having at least one objective value that is better than the corresponding objective value of the particle and having no objective values that are worse than the corresponding objective values of the particle,

determining the local leader configuration for each particle,

determining a global leader configuration, and

modifying the current configuration for each particle based on the local leader configuration for the particle and the global leader configuration until a termination criterion has been met;

wherein the one or more configurations are the configurations of each particle in the set of dominant particles.

14. The system of claim 13, wherein the optimization engine is further configured to:

identifying an objective function calculable based on the one or more objective values,

wherein determining the local leader configuration for each particle comprises;

calculating the objective function for the current configuration for the particle, and

identifying the current configuration for the particle as the local leader configuration for the particle if the objective function for the current configuration is better than the objective function for a previous local leader configuration, and

wherein determining the global leader configuration comprises identifying the particle having the local leader configuration with the best objective function.

15. The system of claim 13, wherein each particle further comprises a local dominant set of configurations;

wherein calculating the one or more objective values for each of the one or more particles further comprises updating the local dominant set of configurations by;

adding the current configuration to the local dominant set of configurations if there exist no other configurations in the local dominant set of configurations having at least one objective value that is better than the corresponding objective value of the current configuration and having no objective values that are worse than the corresponding objective values of the current configuration, and

removing any configuration from the local dominant set of configurations that have at least one objective value that is worse than the corresponding objective value of the current configuration and no objective values that are better than the corresponding objective values of the current configuration,

wherein determining the local leader configuration for each of the one or more particles comprises choosing a random configuration from the local dominant set of configurations, and

wherein determining the global leader configuration comprises choosing the local leader configuration for a random particle.

16. The system of claim 13, wherein modifying the current configuration for each particle comprises modifying the current configuration for the particle based on an inertial component, a cognitive component, and a social component.

17. The system of claim 16, wherein:

each particle further comprises a set of previous configurations for the particle;

the inertial component is based on the current configuration for the particle and the set of previous configurations for the particle;

the inertial component is based on the current configuration for the particle and any previous configurations for the particle;

the cognitive component is based on the local leader configuration for the particle; and

the social component is based on the global leader configuration.

18. The system of claim 16, wherein modifying the current configuration for each particle further comprises modifying the current configuration based on one or more weighting factors.

19. The system of claim 12, wherein the pulse generator further comprises:

a sensing engine configured to detect functional data related to the proximity of each of the one or more electrodes to brain activity;

and wherein at least one of the objective values includes the distance of brain activity to each of the electrodes activated by the configuration.

20. The system of claim 19, wherein the functional data comprises spectral features of local field potential activity collected adjacent or distal to the stimulated electrodes.