Systems for noise-sensitive patient-specific adjustments of neuromodulation parameters
An example method for delivering neurostimulation energy may include performing a training procedure by delivering the neurostimulation energy to a neural target of the patient when the patient is at one or more postures. Electrical activity is sensed from the spinal cord, such as an electrospinogram (ESG). A relationship is determined between the sensed electrical activity and neurostimulation intensity that reduces influence of noise in the sensed electrical activity caused by dynamically changing posture of the patient using mathematical or statistical modeling of the extracted features. Stimulation parameters are modulated according to the determined relationship.
1. A method for delivering neurostimulation energy through a plurality of electrodes to a patient having a spinal cord, the method comprising:
performing a training procedure by delivering the neurostimulation energy at one or more neurostimulation intensity levels to a neural target of the patient when the patient is at a plurality of postures;
sensing electrical activity from the spinal cord, the electrical activity including the delivered neurostimulation energy;
extracting features from the sensed electrical activity when the patient is at the plurality of postures, including determining an influence of noise caused by dynamically changing posture of the patient;
determining a relationship between the sensed electrical activity and neurostimulation intensity that reduces the influence of the noise in the sensed electrical activity caused by the dynamically changing posture of the patient using mathematical or statistical modeling of the extracted features; and
modulating stimulation parameters according to the determined relationship.
2. The method of claim 1 , wherein sensing the electrical activity includes using an electrospinogram (ESG).
3. The method of claim 1 , wherein the mathematical modeling includes using a dynamical model incorporating one or more of a Kalman filter (KF) or a Hidden Markov Model (HMM).
4. The method of claim 1 , wherein the mathematical modeling includes using a risk-sensitive control strategy based on a selection of risk parameters specific for the patient.
5. The method of claim 4 , wherein the selection of risk parameters includes an indication of pain or discomfort tolerance of the patient.
6. The method of claim 1 , wherein the mathematical modeling includes using a probabilistic model implementing a neural network and a dynamical model incorporating one or more of a Kalman filter (KF) or a Hidden Markov Model (HMM).
7. The method of claim 1 , wherein the mathematical modeling includes using a probabilistic model implementing a neural network and a risk-sensitive control strategy based on a selection of risk parameters specific for the patient.
8. The method of claim 1 , wherein the mathematical modeling includes using a dynamical model incorporating one or more of a Kalman filter (KF) or a Hidden Markov Model (HMM) and a risk-sensitive control strategy based on a selection of risk parameters specific for the patient.
9. The method of claim 1 , wherein the mathematical modeling includes using a probabilistic model implementing a neural network, a dynamical model incorporating one or more of a Kalman filter (KF) or a Hidden Markov Model (HMM), and a risk-sensitive control strategy based on a selection of risk parameters specific for the patient.
10. The method of claim 9 , wherein the selection of risk parameters includes an indication of a discomfort threshold of the patient.
11. A non-transitory machine-readable medium including instructions, which when executed by a machine, cause the machine to perform a method for delivering neurostimulation energy through a plurality of electrodes to a patient having a spinal cord, the method comprises:
performing a training procedure by delivering the neurostimulation energy at one or more neurostimulation intensity levels to a neural target of the patient when the patient is at a plurality of postures;
sensing electrical activity from the spinal cord, the electrical activity including the delivered neurostimulation energy;
extracting features from the sensed electrical activity when the patient is at the plurality of postures, including determining an influence of noise caused by dynamically changing posture of the patient;
determining a relationship between the sensed electrical activity and neurostimulation intensity that reduces the influence of the noise in the sensed electrical activity caused by the dynamically changing posture of the patient using mathematical or statistical modeling of the extracted features; and
modulating stimulation parameters according to the determined relationship.
12. The machine-readable medium of claim 11 , wherein sensing the electrical activity includes using an electrospinogram (ESG).
13. The machine-readable medium of claim 11 , wherein the one or more neurostimulation intensity levels includes a discomfort threshold of the patient.
14. The machine-readable medium of claim 11 , wherein the one or more neurostimulation intensity levels includes a perception threshold of the patient.
15. The machine-readable medium of claim 11 , wherein the one or more neurostimulation intensity levels includes a therapeutic threshold of the patient.
16. A system for delivering neurostimulation energy through a plurality of electrodes, comprising:
an implantable pulse generator (IPG) configured for spinal neurostimulation therapy;
one or more leads configured to connect the IPG to the plurality of electrodes;
a stimulation control circuit configured to deliver the neurostimulation energy using the plurality of electrodes at one or more neurostimulation intensity levels to a neural target at or near a spinal cord of a patient when the patient is at a plurality of postures;
a sensing input configured to sense electrical activity from the spinal cord, the electrical activity including responses to the delivered neurostimulation energy;
a feature extraction module configured to extract features from the sensed electrical activity when the patient is at the plurality of postures, including determining an influence of noise caused by dynamically changing posture of the patient; and
a training module configured to determine a relationship between the sensed electrical activity and neurostimulation intensity that reduces the influence of the noise in the sensed electrical activity caused by the dynamically changing posture of the patient using mathematical or statistical modeling of the extracted features;
wherein the stimulation control circuit is further configured to modulate stimulation parameters according to the determined relationship.
17. The system of claim 16 , wherein the sensing input is configured to receive an electrospinogram (ESG).
18. The system of claim 16 , wherein the mathematical modeling includes using a dynamical model incorporating one or more of a Kalman filter (KF) or a Hidden Markov Model (HMM).
19. The system of claim 16 , wherein the mathematical modeling includes using a risk-sensitive control strategy based on a selection of risk parameters specific for the patient.
20. The system of claim 19 , wherein the selection of risk parameters includes an indication of pain or discomfort tolerance of the patient.