DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1-2 and 9-11 is/are rejected under 35 U.S.C. 103 as being unpatentable over Grill et al. (hereinafter ‘Grill’, U. S. PGPub No. 2021/0128920) in view of Malik et al. (hereinafter ‘Malik’, U.S. PGPub No. 2012/0093376).
In regards to claim 1, Grill discloses a method for determining stimulation settings for a neurostimulation device, the method comprising (a) measuring neural activity in a subject using a recording electrode, the neural activity being in response to a neurostimulation delivered to a neural target in the subject by a neurostimulation device ([0037]: "In some embodiments, the device comprises an implantable pulse generator that is capable of generating and delivering non-regular patterns of stimulation while simultaneously recording neural activity.", [0047]: "The electrical stimulation contemplated can be any kind of stimulation, including, without limitation electrical stimulation through an implanted pulse generator, a pulse generator positioned ex-vivo with a lead and/or electrode positioned in-vivo, or any combination of such.", [0125]: "The present invention discloses a device that can deliver non-regular temporal patterns of stimulation and simultaneously record neural activity and mitigate the effects of the stimulus artifacts through amplifier blanking and/or stimulation relay...The application of ECAPs to DBS parameter optimization and closed-loop systems is disclosed. ECAPs indicate how neurons respond during stimulation, and so may better reveal the mechanisms of action of DBS and serve as a reliable feedback signal for brain stimulation."), (b) constructing an autoregressive model based on the measured neural activity ([0109]: "The present invention also describes an implantable pulse generator that is capable of generating and delivering non-regular patterns of stimulation while simultaneously recording neural activity. The implantable pulse generator uses an amplifier-blanking paradigm that briefly grounds the inputs during a short period encompassing the stimulation pulse. This prevents violating the input specifications of the amplifiers (railing the amplifiers), and the short gaps in the data can be overcome with real-time or posts processing analysis methods described below.", [0120]: "The other data processing approach is to work around the gaps and directly estimate the statistics or signal characteristics of interest. If the spectral content of the recorded signal is of interest, one could bypass reconstructing the data in the gaps entirely and instead train an AR model on the data around the gaps and calculate the power spectrum calculated directly from the model.", [0122]: "The recorded neural activity can be used to guide in vivo optimization or learning algorithm based modulation of the temporal pattern of stimulation."), (c) estimating a neural state of the subject based on the autoregressive model, wherein the neural state comprises a real-time state of the subject’s nervous system ([0034]: "In one embodiment, the present invention utilizes a model-based design to achieve a more optimal stimulation pattern for use in connection with a nervous system, one or more nerve cells, or nervous tissue (e.g., a human nervous system).", [0120]: "The other data processing approach is to work around the gaps and directly estimate the statistics or signal characteristics of interest. If the spectral content of the recorded signal is of interest, one could bypass reconstructing the data in the gaps entirely and instead train an AR model on the data around the gaps and calculate the power spectrum calculated directly from the model.", [0124]: "Non-regular temporal patterns of stimulation can be updated and optimized intermittently to meet stimulation objectives (e.g., suppress recorded pathological patterns of neural activity while minimizing energy usage) or changes in the patient or his or her treatment, e.g., different drug treatments. Further, the continuous recording of neural activity allows real-time optimization of the non-regular pattern of stimulation via an automated optimization algorithm incorporated into the implantable pulse generator or through communication (such as wireless communication) from another device (such as a computer, smart phone, etc.) directly to the implantable pulse generator.", see claims 5 and 15, wherein the parameters of stimulation which are constantly being estimated and optimized in real-time represent the "neural state" of the subject), (d) determining stimulation settings based on the estimated neural state ([0124]: "Non-regular temporal patterns of stimulation can be updated and optimized intermittently to meet stimulation objectives (e.g., suppress recorded pathological patterns of neural activity while minimizing energy usage) or changes in the patient or his or her treatment, e.g., different drug treatments. Further, the continuous recording of neural activity allows real-time optimization of the non-regular pattern of stimulation via an automated optimization algorithm incorporated into the implantable pulse generator or through communication (such as wireless communication) from another device (such as a computer, smart phone, etc.) directly to the implantable pulse generator.", see claims 5 and 15, wherein the parameters of stimulation which are constantly being estimated and optimized represent the "neural state" of the subject), and (e) delivering neurostimulation to the neural target in the subject with the neurostimulation device using the stimulation settings which were updated during a feedback loop during delivery of the neurostimulation ([0069]: "The present invention's optimized temporal pattern of stimulation that produces symptom relief at a lower average frequency has advantages over conventional high frequency DBS (typically 130 to 185 Hz). IPGs (or external pulse generators) delivering the optimized low frequency pattern of stimulation consume less energy, and reduced energy consumption translates into longer battery life and less frequent IPG replacement.", [0125]: "The present invention discloses a device that can deliver non-regular temporal patterns of stimulation and simultaneously record neural activity and mitigate the effects of the stimulus artifacts through amplifier blanking and/or stimulation relay...Third, novel applications for using the recorded neural activity as feedback to modulate the stimulation parameters are described. The temporal pattern of stimulation can be a stimulation parameter that is modified, optimized, or otherwise learned based on the recorded neural activity.").
However, Grill does not explicitly disclose that the autoregressive model is adaptively updated based on prediction error feedback during delivery of the neurostimulation. However, Grill does disclose using the Burg algorithm for the autoregressive model (Grill, [0119]: "Note that any method for calculating the AR model (least squares, Burg's algorithm, etc.) could be used, and other types of models could be used to fill in the gaps with data.").
Malik teaches an autoregressive model which uses the Burg algorithm ([0053]: "The coefficient and variance estimates of AR models up to order p are computed by the Burg algorithm during AR(p) model parameter estimation"). Malik further goes to explain that the Burg algorithm uses prediction error minimization ([0042]: "The Burg algorithm for autoregressive (AR) coefficient estimation uses least squares forward-backward prediction error minimization and is constrained to satisfy Levinson-Durbin recursions (LDR)[34,35]."). Therefore, Malik teaches the Burg algorithm providing prediction error minimization and Grill discloses the Burg algorithm being used for the autoregression model. Thus, Grill/Malik combination discloses an autoregressive model which minimizes prediction error during delivery of the neurostimulation. Therefore, it would be obvious to one of ordinary skill in the art before the effective filing date of the neurostimulation system to modify the autoregression model such that the Burg algorithm - and thus a prediction error minimization algorithm - is used, as taught by Malik, as doing so would provide a more efficient computation model to better estimate the model’s parameters ([0005]: "Preferred embodiments utilize a computationally efficient cyclic descent algorithm for maximum likelihood estimation of the model parameters.").
In regards to claim 2, Grill/Malik combination discloses the invention substantially as described above in claim 1.
Malik further teaches an autoregressive model which uses the Burg algorithm ([0053]: "The coefficient and variance estimates of AR models up to order p are computed by the Burg algorithm during AR(p) model parameter estimation"). Malik further goes to explain that the Burg algorithm uses prediction error minimization ([0042]: "The Burg algorithm for autoregressive (AR) coefficient estimation uses least squares forward-backward prediction error minimization and is constrained to satisfy Levinson-Durbin recursions (LDR)[34,35]."). Therefore, Malik teaches the Burg algorithm providing prediction error minimization and Grill discloses the Burg algorithm being used for the autoregression model, which is constructed based on measured neural activity (Grill, [0109], [0120], [0122]). Thus, Grill/Malik combination discloses that the autoregressive model is constructed based on the measured neural activity to minimize prediction error of a response to the neurostimulation delivered to the neural target. Therefore, it would be obvious to one of ordinary skill in the art before the effective filing date of the neurostimulation system to modify the autoregression model such that the Burg algorithm - and thus a prediction error minimization algorithm - is used, as taught by Malik, as doing so would provide a more efficient computation model to better estimate the model’s parameters ([0005]: "Preferred embodiments utilize a computationally efficient cyclic descent algorithm for maximum likelihood estimation of the model parameters.").
In regards to claim 9, Grill/Malik combination discloses the invention substantially as described above in claim 1.
Grill further discloses that the neural activity is measured using one or more recording electrodes ([0037]: "In some embodiments, the device comprises an implantable pulse generator that is capable of generating and delivering non-regular patterns of stimulation while simultaneously recording neural activity.", [0047]: "The electrical stimulation contemplated can be any kind of stimulation, including, without limitation electrical stimulation through an implanted pulse generator, a pulse generator positioned ex-vivo with a lead and/or electrode positioned in-vivo, or any combination of such.").
In regards to claim 10, Grill/Malik combination discloses the invention substantially as described above in claim 9.
Grill further discloses that the one or more recording electrodes are part of the neurostimulation device ([0037]: "In some embodiments, the device comprises an implantable pulse generator that is capable of generating and delivering non-regular patterns of stimulation while simultaneously recording neural activity.", [0047]: "The electrical stimulation contemplated can be any kind of stimulation, including, without limitation electrical stimulation through an implanted pulse generator, a pulse generator positioned ex-vivo with a lead and/or electrode positioned in-vivo, or any combination of such.").
In regards to claim 11, Grill/Malik combination discloses the invention substantially as described above in claim 9.
Grill further discloses that the stimulation settings comprise a neurostimulation waveform ([0091]: "Although the stimulation waveforms and patterns are the same across all subjects, the duration of stimulation artifacts are variable, apparently as a result of differences in the tissue properties around the electrodes.").
Claim(s) 3-4, 12, and 19-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Grill in view of Malik and in further view of Bortolami et al. (hereinafter ‘Bortolami’, U.S. PGPub 2018/0066943).
In regards to claim 3, Grill/Malik combination discloses the invention substantially as described above in claim 1.
However, Grill/Malik combination does not disclose that the neural state is estimated using a Kalman filter constructed based on the autoregressive model.
Bortolami teaches using a Kalman filter constructed from an autoregressive model ([0032]: "The individual instrument biases are modelled as autoregressive-moving-average (ARMA) processes in order to construct the Kalman filter 302."). Therefore, it would be obvious to one of ordinary skill in the art before the effective filing date of the neurostimulation system to construct a Kalman filter from the autoregressive model, as taught by Bortolami, as doing so would remove random events or biases from the detected signals that are fed into the autoregressive model ([0006]: "Using the plurality of stochastic models and the plurality of readings, for each motion sensor/temperature/rotation rate combination, a Kalman filter is built. The Kalman filter identifies bias and random walk states as a function of time.", Kalman filter can remove random events or biases.), wherein walk states may be analogous to neural states.
In regards to claim 4, Grill/Malik/Bortolami combination discloses the invention substantially as described above in claim 3.
Bortolami further teaches extracting model coefficients from the autoregressive model and constructing the Kalman filter using the model coefficients (see [0033]). Therefore, it would be obvious to one of ordinary skill in the art before the effective filing date of the neurostimulation system to construct a Kalman filter from coefficients extracted from the autoregressive model, as taught by Bortolami, as constructing a Kalman filter is possible after coefficients from the autoregression model are extracted ([0077]: "Once the AR models for the coefficient processes and the noise parameters are identified for each of the sensors, a Kalman filter for producing an estimate for the VG model parameter errors can be built.").
In regards to claim 12, Grill discloses a controller for controlling a neurostimulation device, comprising: an input that receives neural activity signals measured from a subject ([0037]: "In some embodiments, the device comprises an implantable pulse generator that is capable of generating and delivering non-regular patterns of stimulation while simultaneously recording neural activity.", [0047]: "The electrical stimulation contemplated can be any kind of stimulation, including, without limitation electrical stimulation through an implanted pulse generator, a pulse generator positioned ex-vivo with a lead and/or electrode positioned in-vivo, or any combination of such.", [0125]: "The present invention discloses a device that can deliver non-regular temporal patterns of stimulation and simultaneously record neural activity and mitigate the effects of the stimulus artifacts through amplifier blanking and/or stimulation relay...The application of ECAPs to DBS parameter optimization and closed-loop systems is disclosed. ECAPs indicate how neurons respond during stimulation, and so may better reveal the mechanisms of action of DBS and serve as a reliable feedback signal for brain stimulation."), a processor to: receive the neural activity signals from the input ([0046]: "The pulse generator may include a processor to which applies a non-regular pulse train."), construct an autoregressive model based on the neural activity signals ([0109]: "The present invention also describes an implantable pulse generator that is capable of generating and delivering non-regular patterns of stimulation while simultaneously recording neural activity. The implantable pulse generator uses an amplifier-blanking paradigm that briefly grounds the inputs during a short period encompassing the stimulation pulse. This prevents violating the input specifications of the amplifiers (railing the amplifiers), and the short gaps in the data can be overcome with real-time or posts processing analysis methods described below.", [0120]: "The other data processing approach is to work around the gaps and directly estimate the statistics or signal characteristics of interest. If the spectral content of the recorded signal is of interest, one could bypass reconstructing the data in the gaps entirely and instead train an AR model on the data around the gaps and calculate the power spectrum calculated directly from the model.", [0122]: "The recorded neural activity can be used to guide in vivo optimization or learning algorithm based modulation of the temporal pattern of stimulation."), update the autoregressive model during delivery of the stimulation ([0069]: "The present invention's optimized temporal pattern of stimulation that produces symptom relief at a lower average frequency has advantages over conventional high frequency DBS (typically 130 to 185 Hz). IPGs (or external pulse generators) delivering the optimized low frequency pattern of stimulation consume less energy, and reduced energy consumption translates into longer battery life and less frequent IPG replacement.", [0125]: "The present invention discloses a device that can deliver non-regular temporal patterns of stimulation and simultaneously record neural activity and mitigate the effects of the stimulus artifacts through amplifier blanking and/or stimulation relay...Third, novel applications for using the recorded neural activity as feedback to modulate the stimulation parameters are described. The temporal pattern of stimulation can be a stimulation parameter that is modified, optimized, or otherwise learned based on the recorded neural activity."), estimate a neural state which comprises a real-time state of the subject’s nervous system ([0034]: "In one embodiment, the present invention utilizes a model-based design to achieve a more optimal stimulation pattern for use in connection with a nervous system, one or more nerve cells, or nervous tissue (e.g., a human nervous system).", [0120]: "The other data processing approach is to work around the gaps and directly estimate the statistics or signal characteristics of interest. If the spectral content of the recorded signal is of interest, one could bypass reconstructing the data in the gaps entirely and instead train an AR model on the data around the gaps and calculate the power spectrum calculated directly from the model.", [0124]: "Non-regular temporal patterns of stimulation can be updated and optimized intermittently to meet stimulation objectives (e.g., suppress recorded pathological patterns of neural activity while minimizing energy usage) or changes in the patient or his or her treatment, e.g., different drug treatments. Further, the continuous recording of neural activity allows real-time optimization of the non-regular pattern of stimulation via an automated optimization algorithm incorporated into the implantable pulse generator or through communication (such as wireless communication) from another device (such as a computer, smart phone, etc.) directly to the implantable pulse generator.", see claims 5 and 15, wherein the parameters of stimulation which are constantly being estimated and optimized in real-time represent the "neural state" of the subject), determine stimulation settings based on the estimated neural state ([0124]: "Non-regular temporal patterns of stimulation can be updated and optimized intermittently to meet stimulation objectives (e.g., suppress recorded pathological patterns of neural activity while minimizing energy usage) or changes in the patient or his or her treatment, e.g., different drug treatments. Further, the continuous recording of neural activity allows real-time optimization of the non-regular pattern of stimulation via an automated optimization algorithm incorporated into the implantable pulse generator or through communication (such as wireless communication) from another device (such as a computer, smart phone, etc.) directly to the implantable pulse generator.", see claims 5 and 15, wherein the parameters of stimulation which are constantly being estimated and optimized represent the "neural state" of the subject), and an output that receives the stimulation settings from the processor and communicates the stimulation settings to a neurostimulation device ([0069]: "The present invention's optimized temporal pattern of stimulation that produces symptom relief at a lower average frequency has advantages over conventional high frequency DBS (typically 130 to 185 Hz). IPGs (or external pulse generators) delivering the optimized low frequency pattern of stimulation consume less energy, and reduced energy consumption translates into longer battery life and less frequent IPG replacement.", [0125]: "The present invention discloses a device that can deliver non-regular temporal patterns of stimulation and simultaneously record neural activity and mitigate the effects of the stimulus artifacts through amplifier blanking and/or stimulation relay...Third, novel applications for using the recorded neural activity as feedback to modulate the stimulation parameters are described. The temporal pattern of stimulation can be a stimulation parameter that is modified, optimized, or otherwise learned based on the recorded neural activity.").
However, Grill does not disclose explicitly using prediction error feedback, extracting coefficients from the autoregressive model, or estimating a neural state based on the extracted autoregressive model coefficients.
While Grill does not explicitly disclose that the autoregressive model is adaptively updated based on prediction error feedback during delivery of the neurostimulation, Grill does disclose using the Burg algorithm for the autoregressive model (Grill, [0119]: "Note that any method for calculating the AR model (least squares, Burg's algorithm, etc.) could be used, and other types of models could be used to fill in the gaps with data.").
Malik teaches an autoregressive model which uses the Burg algorithm ([0053]: "The coefficient and variance estimates of AR models up to order p are computed by the Burg algorithm during AR(p) model parameter estimation"). Malik further goes to explain that the Burg algorithm uses prediction error minimization ([0042]: "The Burg algorithm for autoregressive (AR) coefficient estimation uses least squares forward-backward prediction error minimization and is constrained to satisfy Levinson-Durbin recursions (LDR)[34,35]."). Therefore, Malik teaches the Burg algorithm providing prediction error minimization and Grill discloses the Burg algorithm being used for the autoregression model. Thus, Grill/Malik combination discloses an autoregressive model which minimizes prediction error during delivery of the neurostimulation. Therefore, it would be obvious to one of ordinary skill in the art before the effective filing date of the neurostimulation system to modify the autoregression model such that the Burg algorithm - and thus a prediction error minimization algorithm - is used, as taught by Malik, as doing so would provide a more efficient computation model to better estimate the model’s parameters ([0005]: "Preferred embodiments utilize a computationally efficient cyclic descent algorithm for maximum likelihood estimation of the model parameters.").
Bortolami further teaches extracting model coefficients from the autoregressive model and constructing a Kalman filter using the model coefficients, the Kalman filter further used to estimate the subject’s neural state (see [0033], [0032]: "The individual instrument biases are modelled as autoregressive-moving-average (ARMA) processes in order to construct the Kalman filter 302. "). Therefore, it would be obvious to one of ordinary skill in the art before the effective filing date of the neurostimulation system to construct a Kalman filter from coefficients extracted from the autoregressive model, as taught by Bortolami, as constructing a Kalman filter is possible after coefficients from the autoregression model are extracted and doing so would remove random events or biases from the detected signals that are fed into the autoregressive model ([0077]: "Once the AR models for the coefficient processes and the noise parameters are identified for each of the sensors, a Kalman filter for producing an estimate for the VG model parameter errors can be built.", [0006]: "Using the plurality of stochastic models and the plurality of readings, for each motion sensor/temperature/rotation rate combination, a Kalman filter is built. The Kalman filter identifies bias and random walk states as a function of time.", Kalman filter can remove random events or biases.).
In regards to claim 13, Grill/Malik/Bortolami combination discloses the invention substantially as described above in claim 12.
Malik further teaches an autoregressive model which uses the Burg algorithm ([0053]: "The coefficient and variance estimates of AR models up to order p are computed by the Burg algorithm during AR(p) model parameter estimation"). Malik further goes to explain that the Burg algorithm uses prediction error minimization ([0042]: "The Burg algorithm for autoregressive (AR) coefficient estimation uses least squares forward-backward prediction error minimization and is constrained to satisfy Levinson-Durbin recursions (LDR)[34,35]."). Therefore, Malik teaches the Burg algorithm providing prediction error minimization and Grill discloses the Burg algorithm being used for the autoregression model, which is constructed based on measured neural activity (Grill, [0109], [0120], [0122]). Thus, Grill/Malik combination discloses that the autoregressive model is constructed based on the measured neural activity to minimize prediction error of a response to the neurostimulation delivered to the neural target. Therefore, it would be obvious to one of ordinary skill in the art before the effective filing date of the neurostimulation system to modify the autoregression model such that the Burg algorithm - and thus a prediction error minimization algorithm - is used, as taught by Malik, as doing so would provide a more efficient computation model to better estimate the model’s parameters ([0005]: "Preferred embodiments utilize a computationally efficient cyclic descent algorithm for maximum likelihood estimation of the model parameters.").
In regards to claim 14, Grill/Malik/Bortolami combination discloses the invention substantially as described above in claim 12.
Bortolami further teaches extracting model coefficients from the autoregressive model and constructing a Kalman filter using the model coefficients, the Kalman filter further used to estimate the subject’s neural state (see [0033], [0032]: "The individual instrument biases are modelled as autoregressive-moving-average (ARMA) processes in order to construct the Kalman filter 302. "). Therefore, it would be obvious to one of ordinary skill in the art before the effective filing date of the neurostimulation system to construct a Kalman filter from coefficients extracted from the autoregressive model, as taught by Bortolami, as constructing a Kalman filter is possible after coefficients from the autoregression model are extracted and doing so would remove random events or biases from the detected signals that are fed into the autoregressive model ([0077]: "Once the AR models for the coefficient processes and the noise parameters are identified for each of the sensors, a Kalman filter for producing an estimate for the VG model parameter errors can be built.", [0006]: "Using the plurality of stochastic models and the plurality of readings, for each motion sensor/temperature/rotation rate combination, a Kalman filter is built. The Kalman filter identifies bias and random walk states as a function of time.", Kalman filter can remove random events or biases.).
In regards to claim 19, Grill/Malik/Bortolami combination discloses the invention substantially as described above in claim 12.
Grill further discloses that the neural activity is measured using one or more recording electrodes ([0037]: "In some embodiments, the device comprises an implantable pulse generator that is capable of generating and delivering non-regular patterns of stimulation while simultaneously recording neural activity.", [0047]: "The electrical stimulation contemplated can be any kind of stimulation, including, without limitation electrical stimulation through an implanted pulse generator, a pulse generator positioned ex-vivo with a lead and/or electrode positioned in-vivo, or any combination of such.").
In regards to claim 20, Grill/Malik/Bortolami combination discloses the invention substantially as described above in claim 12.
Grill further discloses that the stimulation settings comprise a neurostimulation waveform ([0091]: "Although the stimulation waveforms and patterns are the same across all subjects, the duration of stimulation artifacts are variable, apparently as a result of differences in the tissue properties around the electrodes.").
Claim(s) 5 and 15 is/are rejected under 35 U.S.C. 103 as being unpatentable over Grill in view of Malik and Bortolami and in further view of Goodzeit et al. (hereinafter ‘Goodzeit’, U.S. Patent No. 8583371).
In regards to claim 5, Grill/Malik/Bortolami combination discloses the invention substantially as described above in claim 4.
However, Grill/Malik/Bortolami combination does not disclose that the Kalman filter is constructed using the model coefficients to generate Kalman filter coefficient matrices and Kalman gain for the Kalman filter.
Goodzeit teaches using model coefficients to generate Kalman filter coefficient matrices and Kalman gain for the Kalman filter (Col. 4, lns. 31-34: "Next, the covariance propagation and Kalman gain module 150 calculates a Kalman gain matrix based on the AD model state covariance matrix using the corrected body rates and the measured gyro temperature.", lns. 36-39: "The AD dynamics model used by the AD logic, which is described in greater detail below, may include states representative of the inertial attitude error, the fixed/random bias error, and the temperature bias coefficients."). Therefore, it would be obvious to one of ordinary skill in the art before the effective filing date of the neurostimulation system to generate Kalman filter coefficient matrices and Kalman gain for the Kalman filter, as taught by Goodzeit, as generating a coefficient (or covariance) matrix and gain matrix are a standard part of Kalman filter formulation (Col. 4, lns. 34-36: "A covariance matrix and Kalman gain matrix are part of a standard Kalman filter formulation.").
In regards to claim 15, Grill/Malik/Bortolami combination discloses the invention substantially as described above in claim 14.
However, Grill/Malik/Bortolami combination does not disclose that the Kalman filter is constructed using the model coefficients to generate Kalman filter coefficient matrices and Kalman gain for the Kalman filter.
Goodzeit teaches using model coefficients to generate Kalman filter coefficient matrices and Kalman gain for the Kalman filter (Col. 4, lns. 31-34: "Next, the covariance propagation and Kalman gain module 150 calculates a Kalman gain matrix based on the AD model state covariance matrix using the corrected body rates and the measured gyro temperature.", lns. 36-39: "The AD dynamics model used by the AD logic, which is described in greater detail below, may include states representative of the inertial attitude error, the fixed/random bias error, and the temperature bias coefficients."). Therefore, it would be obvious to one of ordinary skill in the art before the effective filing date of the neurostimulation system to generate Kalman filter coefficient matrices and Kalman gain for the Kalman filter, as taught by Goodzeit, as generating a coefficient (or covariance) matrix and gain matrix are a standard part of Kalman filter formulation (Col. 4, lns. 34-36: "A covariance matrix and Kalman gain matrix are part of a standard Kalman filter formulation.").
Claim(s) 6-7 is/are rejected under 35 U.S.C. 103 as being unpatentable over Grill in view of Malik and in further view of Zehetleitner et al. (hereinafter ‘Zehetleitner’, U.S. PGPub No. 20200/096013).
In regards to claim 6-7, Grill/Malik combination discloses the invention substantially as described above in claim 1.
However, Grill/Malik combination does not disclose that the stimulation settings are generated using a linear quadratic regulator and that the linear quadratic regulator is constructed using the estimated neural state to define a cost function of the linear quadratic regulator.
Zehetleitner teaches generating parameters using a linear quadratic regulator which is constructed using a system with a feedback loop which defines a cost function ([0016]: "Thus, the control parameters of a PID controller, for example, as well as a linear-quadratic controller (LQR), can be determined. A linear-quadratic controller, also known as a Riccati controller, is a state controller for a linear dynamic system of which the feedback matrix is determined by minimizing a quadratic cost function, at which point the actual system pressure is controlled to the target pressure."). Therefore, it would be obvious to one of ordinary skill in the art before the effective filing date of the neurostimulation system to use a linear quadratic regulator with a defined cost function, as taught by Zehetleitner, as using a linear quadratic regulator for a system with control parameters or a linear dynamic system is well known in the art ([0017]: "Both the Fast Fourier Transform (FFT), the least squares method (LSQ), and the frequency characteristic design method, as well as linear-quadratic controllers (LQR), are well known…").
Claim(s) 8 is/are rejected under 35 U.S.C. 103 as being unpatentable over Grill in view of Malik and Zehetleitner and in further view of Nguyen (U.S. Patent No. 9227721).
In regards to claim 8, Grill/Malik/Zehetleitner combination discloses the invention substantially as described above in claim 7.
However, Grill/Malik/Zehetleitner combination does not disclose that the linear quadratic regulator is further constructed based on model coefficients extracted from the autoregressive model.
Nguyen teaches using a coefficient extracted from a model to construct a linear quadratic regulator (Col. 30, lns. 11-14: "Hence, the first three terms in the cost function form the standard linear-quadratic regulator (LQR) cost function in these embodiments. The last term may be designed to minimize the drag coefficient with qD as a weighting constant and thus represents an enabling feature of the present invention.").Therefore, it would be obvious to one of ordinary skill in the art before the effective filing date of the neurostimulation system to use a linear quadratic regulator constructed from coefficients extracted from a model, as taught by Nguyen, as extracting a coefficient from a mathematical model is well known in the art (Col. 30, lns. 15-17: "It is assumed that the drag coefficient may be estimated accurately via a drag polar model for which the drag polar parameters are known.", wherein the drag polar model represents a mathematical model from which coefficients may be obtained, similar to the autoregressive model).
Claim(s) 16-17 is/are rejected under 35 U.S.C. 103 as being unpatentable over Grill in view of Malik and Bortolami and in further view of Zehetleitner.
In regards to claim 16-17, Grill/Malik/Bortolami combination discloses the invention substantially as described above in claim 12.
However, Grill/Malik/Bortolami combination does not disclose that the stimulation settings are generated using a linear quadratic regulator and that the linear quadratic regulator is constructed using the estimated neural state to define a cost function of the linear quadratic regulator.
Zehetleitner teaches generating parameters using a linear quadratic regulator which is constructed using a system with a feedback loop which defines a cost function ([0016]: "Thus, the control parameters of a PID controller, for example, as well as a linear-quadratic controller (LQR), can be determined. A linear-quadratic controller, also known as a Riccati controller, is a state controller for a linear dynamic system of which the feedback matrix is determined by minimizing a quadratic cost function, at which point the actual system pressure is controlled to the target pressure."). Therefore, it would be obvious to one of ordinary skill in the art before the effective filing date of the neurostimulation system to use a linear quadratic regulator with a defined cost function, as taught by Zehetleitner, as using a linear quadratic regulator for a system with control parameters or a linear dynamic system is well known in the art ([0017]: "Both the Fast Fourier Transform (FFT), the least squares method (LSQ), and the frequency characteristic design method, as well as linear-quadratic controllers (LQR), are well known…").
Claim(s) 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Grill in view of Malik, Bortolami, and Zehetleitner and in further view of Nguyen.
In regards to claim 18, Grill/Malik/Bortolami/Zehetleitner combination discloses the invention substantially as described above in claim 17.
However, Grill/Malik/Bortolami/Zehetleitner combination does not disclose that the linear quadratic regulator is further constructed based on model coefficients extracted from the autoregressive model.
Nguyen teaches using a coefficient extracted from a model to construct a linear quadratic regulator (Col. 30, lns. 11-14: "Hence, the first three terms in the cost function form the standard linear-quadratic regulator (LQR) cost function in these embodiments. The last term may be designed to minimize the drag coefficient with qD as a weighting constant and thus represents an enabling feature of the present invention.").Therefore, it would be obvious to one of ordinary skill in the art before the effective filing date of the neurostimulation system to use a linear quadratic regulator constructed from coefficients extracted from a model, as taught by Nguyen, as extracting a coefficient from a mathematical model is well known in the art (Col. 30, lns. 15-17: "It is assumed that the drag coefficient may be estimated accurately via a drag polar model for which the drag polar parameters are known.", wherein the drag polar model represents a mathematical model from which coefficients may be obtained, similar to the autoregressive model).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to BRYAN M LEE whose telephone number is (703)756-1789. The examiner can normally be reached 9:00 am - 6:00 pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Unsu Jung can be reached at (571)272-8506. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/B.M.L./Examiner, Art Unit 3796
/LYNSEY C Eiseman/Primary Examiner, Art Unit 3796
