Systems and methods for VOA model generation and use

• US 9,792,412 B2
• Filed: 08/22/2013
• Issued: 10/17/2017
• Est. Priority Date: 11/01/2012
• Status: Expired due to Fees

• Current Assignee: Boston Scientific Neuromodulation Corporation (Boston Scientific Corporation)
• Original Assignee: Boston Scientific Neuromodulation Corporation (Boston Scientific Corporation)
• Inventors: Moffitt, Michael A., Steinke, G. Karl
• Primary Examiner(s): Sitiriche, Luis
• Application Number:
  US13/973,113
• Publication Number:
  US 20140122379A1
• Time in Patent Office: 1,517 Days
• Field of Search: None
• CPC Class Codes: A61N 1/36128 Control systems
• CPC Class Codes: A61N 1/37247 User interfaces, e.g. input or presentation means
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 3/084 Backpropagation, e.g. using gradient descent
• CPC Class Codes: G06N 3/086 using evolutionary algorithms, e.g. genetic algorithms or genetic programming
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G16H 20/40 relating to mechanical, radiation or invasive therapies, e.g. surgery, laser therapy, dialysis or acupuncture
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 50/50 for simulation or modelling of medical disorders

Abstract

A computer implemented system and method provides a volume of activation (VOA) estimation model that receives as input two or more electric field values of a same or different data type at respective two or more positions of a neural element and determines based on such input an activation status of the neural element. A computer implemented system and method provides a machine learning system that automatically generates a computationally inexpensive VOA estimation model based on output of a computationally expensive system.

Claims

1. A computer-implemented method, comprising:

obtaining, by a computer processor, electric field data corresponding to settings of a non-cylindrically symmetrical implanted leadwire that is adapted for stimulating anatomical tissue, the electric field data including for each of a plurality of neural elements a respective plurality of electric values for a same electric field parameter;

determining in a first determining step, by the processor and for each of the neural elements, a respective activation status based on the respective plurality of electric values associated with the respective neural element; and

the processor determining in a second determining step, and outputting an indication of, an estimated activated tissue region corresponding to a combination of points surrounding the leadwire corresponding to those of the neural elements for which an active status is determined;

wherein the determining of the second determining step is performed by executing a first module that at least one of;

does not base the determining of the second determining step on input of different sets of values of an electric field at different points in time;

does not use more than one differential equation;

or

is generated based on observed functioning of a second module that uses differential equations, wherein the first module uses only linear equations.

2. The method of claim 1, wherein the first module is automatically generated by execution of a machine learning module based on analysis by the machine learning module of functioning of the second module.

3. The method of claim 2, wherein the machine learning module is an Artificial Neural Network learning module.

4. The method of claim 1, wherein the neural elements are axons.

5. The method of claim 4, wherein the axons are all copies of a same model axon, each plotted at a respective point surrounding the leadwire.

6. The method of claim 1, wherein the electric field parameter is a second difference activating function.

7. The method of claim 6, further comprising:

selecting, by the processor, for each of the plurality of neural elements, and as the respective plurality of electric values for the neural element, the second difference activating function values located at predefined locations of the respective neural element.

8. The method of claim 7, wherein the predefined locations are nodes of Ranvié

r of the respective neural element.

9. A computer-implemented method, comprising:

responsive to receipt of user input settings of an implanted leadwire that is adapted for stimulating anatomical tissue, wherein the implanted leadwire is non-cylindrically symmetrical;

determining in a first determining step, by a computer processor, electric field data corresponding to the user input settings, the electric field data including for each of a plurality of neural elements a respective plurality of electric values for a same electric field parameter;

determining in a second determining step, by the processor and for each of the neural elements, a respective activation status based on the respective plurality of electric values associated with the respective neural element;

determining in a third determining step, by the processor, an estimated activated tissue region corresponding to a combination of points surrounding the leadwire corresponding to those of the neural elements for which an active status is determined; and

generating and displaying, by the processor, a graphical representation of a volume relative to at least one of a graphical representation of the leadwire and a graphical representation of anatomical structures, the volume corresponding to, and being based on, the estimated activated tissue region,

wherein one or more of the determining in the second determining step or the determining in the third determining step is performed by executing a first module that at least one of;

does not base the one or more of the determining in the second determining step or the determining in the third determining step on input of different sets of values of an electric field at different points in time;

does not use more than one differential equation;

or

is generated based on observed functioning of a second module that uses differential equations, wherein the first module uses only linear equations.

10. A computer-implemented method, comprising:

obtaining, by a computer processor executing a machine learning module, output data of a first module that determines an activation status for each of a plurality of neural elements based on input characterizing stimulation settings of one or more non-cylindrically symmetrical implanted leadwires adapted for stimulating anatomical tissue;

obtaining, by the processor executing the machine learning module, at least a portion of the input processed by the first module to produce the obtained output data;

analyzing, by the processor executing the machine learning module, the obtained output data and input; and

based on the analysis, automatically generating, by the processor executing the machine learning module, at least one second module that determines an activation status for each of a plurality of neural elements based on input that (a) is different than the input and (b) characterizes stimulation settings of one or more non-cylindrically symmetrical implanted leadwires adapted for stimulating anatomical tissue.

11. The method of claim 10, wherein the machine learning module is an Artificial Neural Network (ANN) module.

12. The method of claim 11, wherein the ANN module implements pattern recognition with back-propagation of errors.

13. The method of claim 10, wherein the machine learning module implements at least one of decision tree analysis, association rules, genetic algorithms, and support vector machines.

14. The method of claim 10, wherein the machine learning module implements supervised learning based on user input rules.

15. The method of claim 10, wherein the input includes for the each of the plurality of neurons a respective activating function profile.

16. The method of claim 15, wherein the activating function profile includes a plurality of values of second central different activating function, each of the values being at a respective one of a plurality of predetermined positions of the respective neural element.

17. The method of claim 16, wherein the neural elements are axons and the plurality of predetermined positions are nodes of Ranvié

r.

18. The method of claim 10, wherein each of the plurality of neural elements is one of an axon, a dendrite, a cell body, and a glial cell.

19. The method of claim 10, wherein the input includes at least one of a potential field, an electric tensor field, and a current tensor field.

20. The method of claim 10, wherein the at least one second module includes a plurality of second modules, each adapted for performing the determining of the activation status for the each of the plurality of neural elements for a respective pulse width category.

21. The method of claim 10, wherein the at least one second module includes a first version module for determining activation statuses for neural elements associated with cathodically shaped activation function profiles, a second version module for determining activation statuses for neural elements associated with anodically shaped activation function profiles, and a third version module for determining activation statuses for neural elements associated with mixed anodic and cathodic activation function profiles.

22. The method of claim 10, wherein the one or more implanted leadwires for which the one or more second modules are adapted to determine activation statuses are different than the one or more implanted leadwires for the output data of the first module was provided.

23. The method of claim 10, further comprising:

providing, by the processor executing the machine learning nodule, a same input to each of the first and second modules;

comparing, by the processor executing the machine learning module, to each other respective output produced by each of the first and second modules in response to the same input; and

responsive to determining that the compared output differ, modifying, by the processor executing, the machine learning module, the second module.

24. A computer-implemented method, comprising:

obtaining, by a computer processor, electric field data corresponding to settings of non-cylindrically symmetrical implanted leadwire that is adapted for stimulating anatomical tissue, the electric field data including for each of a plurality of neural elements a respective plurality of electric values for a same electric field parameter; and

determining and outputting, by the processor and for each of the neural elements, a respective activation threshold based on the respective plurality of electric values associated with the respective neural element;

wherein the determining is performed by executing a first module that at least one of;

does not base the determining on input of different sets of values of an electric field at different points in time;

does not use more than one differential equation;

or

is generated based on observed functioning of a second module that uses differential equations, wherein the first module uses only linear equations.

25. A computer-implemented method, comprising:

obtaining, by a computer processor executing a machine learning module, output data of a first module that determines an activation threshold for each of a plurality of neural elements based on input characterizing stimulation settings of one or more non-cylindrically symmetrical implanted leadwires adapted for stimulating anatomical tissue;

obtaining, by the processor executing the machine learning module, at least a portion of the input processed by the first module to produce the obtained output;

analyzing, by the processor executing the machine learning module, the obtained output data and input; and

based on the analysis, automatically generating, by the processor executing the machine learning module, at least one second module that determines an activation threshold for each of a plurality of neural elements based on input that (a) is different than the input and (b) characterizes stimulation settings of one or more non-cylindrically symmetrical implanted leadwires adapted for stimulating anatomical tissue.