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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on October 22, 2025 has been entered.
Response to Arguments
This Office Action is in response to the amendment filed on October 22, 2025. As directed by the amendment, Claims 1, 36, and 58 have been amended. Claims 63 and 64 are new claims. Claims 1-12, 14, 36, 55-58, 63, and 64 are pending in the instant application.
Regarding the Office Action mailed July 22, 2025:
Applicant has resolved all objections to the claims. Therefore, the objections are withdrawn.
Applicant’s arguments regarding the 35 USC 103 rejections have been fully considered but they are not persuasive.
Regarding the prior art, Applicant argues that the use of machine learning algorithms/models do not teach Claims 1 and 36 and that Kaemmerer teaches away from determining rules since it’s evaluating the accuracy of existing machine learning models (Remarks: Page 7).
Examiner respectfully disagrees with this argument. The prior art Wong already has a precedence of determining rules and determining therapy outcomes based on these rules. The device of Wong is already capable of determining different characteristics of tremor and using data as feedback to modify, adjust, and set various stimulation parameters (Wong: paragraph 0222). The device of Wong is basically figuring out what kinds of settings or parameters need to be implemented to produce a desired therapeutic outcome. Wong also does teach the use of predictive adaptation in which the device can “learn” or predict when tremors will occur and apply treatment ahead of time (Wong: paragraphs 0240-0242).
Kaemmerer merely further elaborates on details involved in predictive learning or adaptation, including the use of machine learning. Kaemmerer establishes the use of a cross-validation scheme to make comparisons (Kaemmerer: paragraph 0103). Machine learning does involve the determination of rules since that is the algorithm’s way of determining of the rules are accurate or correct enough to produce the desired outcome. This is done by comparing training sets with test sets. The instant specification establishes that determining the rules is related to predicting stimulation settings for a given patient (based on their tremor features) that will result in the best therapeutic effect (Specification: paragraph 0081) and that the rule generation engine can employ machine learning modeling to determine rules (Specification: paragraph 0085). This is no different from what has been previously established in the prior art Wong and Kaemmerer. Applicant has not provided any further details in the claims on how their rule determination techniques are drastically distinct from the adaptation techniques and machine learning techniques taught by Wong and Kaemmerer.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-3, 6-12, 14, 36, 55-58, 63, and 64 are rejected under 35 U.S.C. 103 as being unpatentable over Wong et al. (US 2017/0157398 A1) in view of Bibian et al. (US 8,538,512 B1) and Kaemmerer et al. (US 2016/0144186 A1).
Regarding Claim 1, Wong discloses a wearable neurostimulation device (Device 10 designed to be worn on wrist or arm, paragraph 0171) for transcutaneously stimulating one or more peripheral nerves of a user (a peripheral nerve stimulator, Abstract; 10 allows customization and optimization of transcutaneous electrical treatment, paragraph 0170), the device comprising: one or more electrodes (16, Fig 1D; electrical contacts in band 14 and/or housing 12 transmit stimulation waveform to disposable electrodes 16, paragraph 0171) configured to generate electric stimulation signals (electrodes being spaced on band to deliver electrical stimuli, paragraph 0008); one or more sensors (20, Fig 1E) configured to detect motion signals (motion sensor configured to measure motion, paragraph 0021; detecting tremor characteristics by processing one or more motion sensors, paragraph 0254), wherein the one or more sensors are operably connected to the wearable neurostimulation device (20 is connected to 22 and is inside 12, Fig 1E); and one or more hardware processors (22, Fig 1E; controller or processor 22, paragraph 0172) configured to: receive raw signals in time domain from the one or more sensors (controller programmed to determine characteristics of tremor based on signal generated by motion sensor, paragraph 0023; may require detecting tremor characteristics by processing one or more motion sensors, paragraph 0254); separate the raw signals into a plurality of frames (a 3 axis gyroscope can be used to measure tremor, each axis is individually windowed, paragraph 0254; multiple sensors and axes are individually windowed and processed to detect tremor characteristics); for each of the plurality of frames: transform the raw signals into a frequency domain (each axis is individually windowed, Fourier transform is applied, paragraph 0254); calculate a first energy in a first frequency band of the transformed signal for a respective frame (each axis is individually windowed, Fourier transform is applied, magnitude of each axis calculated, square root of sum of squares of axes are calculated as a function of frequency, paragraph 0254; first energy in first frequency band is the leftmost section of the frequency band, see annotated Fig 31A below); calculate a second energy in a second frequency band of the transformed signal for the respective frame, wherein the second frequency band includes a first frequency corresponding to a tremor (each axis is individually windowed, Fourier transform is applied, magnitude of each axis calculated, square root of sum of squares of axes are calculated as a function of frequency, paragraph 0254; second energy in second frequency band is the middle section of the frequency band, band is also called the tremor band which is where a tremor is expected, see annotated Fig 31A below); extract features from the combined frames in a frequency domain (peak frequency in 4-12 Hz range is identified, frequency detected by determining frequency at maximum value in 4-12 Hz range, paragraph 0254; each axis is individually windowed, Fourier transform is applied, magnitude of each axis calculated, square root of sum of squares of axes are calculated as a function of frequency, paragraph 0254); determine rules based on the extracted features (determine various characteristics of tremor and using data as feedback to modify, adjust, and set various stimulation parameters, paragraph 0222; tremor frequency can be measured at all times and then used to update stimulation in real time, paragraph 0259; the “rules” are the same as the stimulation parameters involved as these “rules” would be dependent on the tremor characteristics determined); and determine neurostimulation therapy outcomes based on an application of the determined rules on operational data (determine various characteristics of tremor and using data as feedback to modify, adjust, and set various stimulation parameters, paragraph 0222; tremor frequency can be measured at all times and then used to update stimulation in real time, paragraph 0259; when stimulation is applied to user, the stimulation would inherently determine the therapy outcome in which the user experiences less tremors, more comfort, less pain, etc.).
Wong also discloses different ways to resolve boundary artifacts from being falsely interpreted as a signal maximum (paragraph 0253) and a need to differentiate tremor movement from non-tremor (or voluntary) movements by segregating voluntary band (0.1-3 Hz) and tremor band (4-12 Hz) (paragraph 0260).
Wong fails to explicitly disclose determine a motion artifact in the respective frame based on a comparison of the first energy with the second energy; combine each of the respective frames from the plurality of frames based on the determination of the motion artifact; wherein determining the neurostimulation therapy outcomes comprises comparing an examined output calculated by an examined rule with a potential output calculated by a potential rule based on cross-validation accuracy; and wherein both the examined rule and the potential rule are selected from a set of potential rules.
However, Bibian, of the same field of endeavor and reasonably pertinent to the problem of detecting motion artifacts, teaches a brain dysfunction and seizure detector monitor and system (Abstract) including determine a motion artifact in the respective frame based on a comparison of the first energy with the second energy (invention possesses ability to more completely remove motion and other artifacts by firmware and/or software correction that utilizes information collected preferably from a sensor or device to detect body motion, 3D accelerometer connected to microprocessing unit, microprocessor applies particular tests and algorithms comparing the two signal sets to correct any motion artifacts, processor may apply more complicated frequency analysis, frequency analysis preferably in the form of wavelet analysis can be applied to yield artifact detection, Column 14, Lines 25-62; motion artifacts would obviously be detected by comparing different energies or frequencies involved in each signal); combine each of the respective frames based on the determination of the motion artifact (invention possesses ability to more completely remove motion and other artifacts by firmware and/or software correction that utilizes information collected preferably from a sensor or device to detect body motion, 3D accelerometer connected to microprocessing unit, microprocessor applies particular tests and algorithms comparing the two signal sets to correct any motion artifacts, processor may apply more complicated frequency analysis, frequency analysis preferably in the form of wavelet analysis can be applied to yield artifact detection, Column 14, Lines 25-62; signals would obviously need to be combined to allow for comparison and for producing the signal without the motion artifacts) since comparisons of signals are a known way to remove motion artifacts in signals (Column 14, Lines 25-62).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the processor and motion sensors to allow comparison of signals for detection and removal of motion artifacts, as taught by Bibian, since comparisons of signals are a known way to remove motion artifacts in signals (Bibian: Column 14, Lines 25-62). The determination of motion artifacts within a signal is obviously a well-known consideration done by one of ordinary skill in the art. Wong already discusses the importance of determining tremor movements from non-tremor movements (since the device is a wearable device on the wrist or arm of the user) and knowing that a non-tremor, voluntary activity falls within the band of 0.1-3 Hz (Wong: paragraph 0260). Bibian simply further supports the idea of making this determination through the comparison of signals and identification/detection of motion artifacts. Wong already discusses that tremors are to be found in the 4-12 Hz range (Wong: paragraphs 0228 and 0254). Thus, it is obvious for one of ordinary skill in the art to discern that any frequencies outside of that range would be considered noise or artifacts in the signal, which would obviously include motion artifacts. Additionally, signal processing techniques for removing noise and artifacts are well known as signal comparisons through correlation or cross-correlation are known to one of ordinary skill in the art. Other techniques within the frequency domain, like applying low-pass or high-pass or band-pass filters, are well-known to filter out noise/artifacts expected at frequency ranges outside of the desired frequency range. Thus, one of ordinary skill in the art would be motivated to remove noise/artifacts through various well-known signal processing techniques.
Wong-Bibian combination teaches the use of predictive adaptation and using predictive algorithms to predict when tremors will increase (Wong: paragraphs 0240-0242). Wong-Bibian combination fails to teach wherein determining neurostimulation therapy outcomes comprises comparing an examined output calculated by an examined rule with a potential output calculated by a potential rule based on cross-validation accuracy; and wherein both the examined rule and the potential rule are selected from the set of potential rules.
However, Kaemmerer, of the same field of endeavor and reasonably pertinent to the problem of neurostimulation, teaches a method for selecting a combination of electrodes (Abstract) including determining the neurostimulation therapy outcomes comprises comparing an examined output calculated by an examined rule with a potential output calculated by a potential rule based on cross-validation accuracy; and wherein both the examined rule and the potential rule are selected from a set of potential rules (machine learning algorithms/models may be utilized to enable device to select combination of electrodes for patient, classifier performance may be evaluated based on number of errors produced in a leave-one-out cross-validation scheme, features may be standardized across training observations before being used to train classifier, in order to predict group of new observation, score representing likelihood of being in group is calculated for each of the possible group, and observation is classified as being in group with largest score, paragraph 0103; leave-one-out cross-validation would compare examined output or test set with potential rule or training set to see how accurate or how well the classifier performs; the set of potential rules is merely the series of training sets) to evaluate the performance based on number of errors produced in cross-validation scheme (paragraph 0103).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize machine learning with cross-validation capabilities to verify and select the most effective stimulation parameters, as taught by Kaemmerer, to evaluate the performance based on number of errors produced in cross-validation scheme (Kaemmerer: paragraph 0103). Utilizing machine learning to improve a device’s ability to adapt to a patient’s particular needs is well-known in the art. Cross-validation is a well-known type of machine learning in which training sets and test sets are compared to confirm or verify if the performance of the algorithm is correct or effective. Kaemmerer shows that utilizing cross-validation is obvious as it would allow one of ordinary skill in the art to obtain further feedback and ensuring that the device is treating the tremors effectively. It also trains the device to select and choose the most effective combination of electrodes and stimulation parameters for particular tremors. Since Wong already obtains feedback to adjust stimulation parameters, having machine learning would simply further improve upon the existing feedback capabilities. Applicant has not claimed any particular features within the algorithm that are significantly distinct from well-known machine learning methods.
Regarding Claim 2, Wong-Bibian-Kaemmerer combination teaches the sensors are operably attached to the wearable neurostimulation device (Wong: 20 is within housing 12, Fig 1E; housing 12 containing sensors 20, paragraph 0172).
Regarding Claim 3, Wong-Bibian-Kaemmerer combination teaches the raw signals relate to tremor activity of the user (Wong: detecting tremor characteristics by processing one or more motion sensors, paragraph 0254).
Regarding Claim 6, Wong-Bibian-Kaemmerer combination teaches the sensors comprise one or more of a gyroscope, accelerometer, and magnetometer (Wong: motion sensor selected from group consisting of accelerometer, a gyroscope, a magnetometer, and a bend sensor, paragraph 0070).
Regarding Claim 7, Wong-Bibian-Kaemmerer combination teaches first frequency band is between about 0 Hz and about 2.5 Hz (Wong: voluntary band is 0.1-3 Hz, non-tremor or voluntary movements, paragraph 0260).
Regarding Claim 8, Wong-Bibian-Kaemmerer combination teaches the second frequency band is between about 4 Hz and about 12 Hz (Wong: typical tremor frequencies are 4-12 Hz, paragraph 0228; tremor band of 4-12 Hz, paragraph 0260).
Regarding Claim 9, Wong-Bibian-Kaemmerer combination does not explicitly teach the second frequency band is between about 3 Hz and about 8 Hz.
However, Wong further teaches typical tremor frequencies are 4-12 Hz and tremor bands are 4-12 Hz (Wong: paragraph 0228; paragraph 0260).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to expect a second frequency band between about 3 Hz and about 8 Hz, as taught by Wong, since typical tremor frequencies would fall within this frequency range. Since patients would have tremor frequencies that vary over time, some of those tremor frequencies would fall within the claimed frequency range.
Regarding Claim 10, Wong-Bibian-Kaemmerer combination teaches wherein the features comprise at least one or more of: amplitude (Wong: tremor amplitude, paragraph 0232), bandwidth (Wong: tremor band of 4-12 Hz, paragraph 0260), power (Wong: energy under the curve, paragraph 0233; spectral power at a frequency or spectral energy in the 4-12 Hz band, paragraph 0232), peak frequency (Wong: peak frequency in 4-12 Hz range, paragraph 0254; peak frequency in spectral density curve, paragraph 0238).
Regarding Claim 11, Wong-Bibian-Kaemmerer combination teaches the features comprise at least one or more of kinematic features (Wong: measuring motion of patient’s arm or wrist during specific task, determining characteristics of tremor, paragraph 0077; specific task is kinetic, paragraph 0078), wherein the kinematic features include regularity, amplitude and shape of the signal of each of the respective frames from the plurality of frames (Wong: Figs 24A-24D showcase the tremors over time including regularity (or period), amplitude, and shape of tremor signal; tremor amplitude and frequency can have daily patterns, paragraph 0240; match certain tremor characteristics including phase, frequency, and amplitude, paragraph 0206; Fig 25B depict disease segmentation separating kinetic tremor characteristics from resting tremor characteristics, paragraph 0244).
Regarding Claim 12, Wong-Bibian-Kaemmerer combination teaches wherein the features comprise at least one or more of: amplitude or power spectral density ("PSD") at peak tremor frequency (Wong: spectral power at a frequency or spectral energy in the 4-12 Hz band, paragraph 0232; peak frequency in 4-12 Hz range, paragraph 0254; peak frequency in spectral density curve, paragraph 0238), summed amplitude or PSD in an approximately 4 to 12 Hz band (Wong: magnitude of each axis calculated and square root of sum of squares of the axes calculated as function of frequency, sum spectrum is filtered and peak frequency in 4-12 Hz range is identified, paragraph 0254).
Regarding Claim 14, Wong-Bibian-Kaemmerer combination teaches the features comprise at least one of displacement (Wong: position and orientation can be determined by integrating accelerometer or gyroscope signals, combining positions in one or more axes, paragraph 0227; gyroscope used to measure tremor from wrist, paragraph 0254; sensors used provide position or displacement data), functional principal component analysis (PCA) (Wong: technique implemented using principal components analysis, paragraph 0244), filtering (Wong: filter may be required to eliminate noise oscillations, paragraph 0228; box car filter or other low pass filter, band pass filter, paragraph 0254), mean (Wong: frequency can be updated sporadically instead of continuously, tremor frequency does not vary dramatically, mean frequency, paragraph 0255; mean frequency capable of being utilized since frequency shifts happen over long periods of time).
Regarding Claim 63, Wong-Bibian-Kaemmerer combination teaches the neurostimulation therapy outcomes comprise predicting patient tremor severity at a given point (Wong: understanding historical tremor measurements and the time therapy was applied can inform therapy needed on successive days, neural networks, Kalman filters, and other such predictive algorithms can be used to predict when tremor will increase and apply pre-emptive treatment, paragraph 0240).
PNG
media_image1.png
460
471
media_image1.png
Greyscale
Regarding Claim 36, Wong discloses a neuromodulation device (Device 10 designed to be worn on wrist or arm, paragraph 0171) for modulating one or more nerves of a user (a peripheral nerve stimulator, Abstract; 10 allows customization and optimization of transcutaneous electrical treatment, paragraph 0170), the device comprising: one or more electrodes (16, Fig 1D; electrical contacts in band 14 and/or housing 12 transmit stimulation waveform to disposable electrodes 16, paragraph 0171) configured to generate neuromodulation signals (electrodes being spaced on band to deliver electrical stimuli, paragraph 0008); one or more sensors (20, Fig 1E) configured to detect motion signals (motion sensor configured to measure motion, paragraph 0021; detecting tremor characteristics by processing one or more motion sensors, paragraph 0254); and one or more hardware processors (22, Fig 1E; controller or processor 22, paragraph 0172) configured to: receive raw signals in time domain from the one or more sensors (controller programmed to determine characteristics of tremor based on signal generated by motion sensor, paragraph 0023; may require detecting tremor characteristics by processing one or more motion sensors, paragraph 0254); separate the raw signals into a plurality of frames (a 3 axis gyroscope can be used to measure tremor, each axis is individually windowed, paragraph 0254; multiple sensors and axes are individually windowed and processed to detect tremor characteristics); for each of the plurality of frames: transform the raw signals into a frequency domain (each axis is individually windowed, Fourier transform is applied, paragraph 0254); calculate a first parameter in a first frequency band of the transformed signal for a respective frame (each axis is individually windowed, Fourier transform is applied, magnitude of each axis calculated, square root of sum of squares of axes are calculated as a function of frequency, paragraph 0254; first parameter in first frequency band is the leftmost section of the frequency band, see annotated Fig 31A below); calculate a second parameter in a second frequency band of the transformed signal for the respective frame, wherein the second frequency band includes a first frequency corresponding to a physiological characteristic of the user (each axis is individually windowed, Fourier transform is applied, magnitude of each axis calculated, square root of sum of squares of axes are calculated as a function of frequency, paragraph 0254; second parameter in second frequency band is the middle section of the frequency band, band is also called the tremor band which is where a tremor is expected, see annotated Fig 31A below); extract features from the combined frames in a frequency domain (peak frequency in 4-12 Hz range is identified, frequency detected by determining frequency at maximum value in 4-12 Hz range, paragraph 0254; each axis is individually windowed, Fourier transform is applied, magnitude of each axis calculated, square root of sum of squares of axes are calculated as a function of frequency, paragraph 0254); determine rules based on the extracted features (determine various characteristics of tremor and using data as feedback to modify, adjust, and set various stimulation parameters, paragraph 0222; tremor frequency can be measured at all times and then used to update stimulation in real time, paragraph 0259; the “rules” are the same as the stimulation parameters involved as these “rules” would be dependent on the tremor characteristics determined); and determine neuromodulation therapy outcomes based on an application of the determined rules on operational data (determine various characteristics of tremor and using data as feedback to modify, adjust, and set various stimulation parameters, paragraph 0222; tremor frequency can be measured at all times and then used to update stimulation in real time, paragraph 0259; when stimulation is applied to user, the stimulation would inherently determine the therapy outcome in which the user experiences less tremors, more comfort, less pain, etc.).
Wong also discloses different ways to resolve boundary artifacts from being falsely interpreted as a signal maximum (paragraph 0253) and a need to differentiate tremor movement from non-tremor (or voluntary) movements by segregating voluntary band (0.1-3 Hz) and tremor band (4-12 Hz) (paragraph 0260).
Wong fails to explicitly disclose determine a motion artifact in the respective frame based on a comparison of the first parameter with the second parameter; combine each of the respective frames from the plurality of frames based on the determination of motion artifact; wherein determining the neurostimulation therapy outcomes comprises comparing an examined output calculated by an examined rule with a potential output calculated by a potential rule based on cross-validation accuracy; and wherein both the examined rule and the potential rule are selected from a set of potential rules.
However, Bibian, of the same field of endeavor and reasonably pertinent to the problem of detecting motion artifacts, teaches a brain dysfunction and seizure detector monitor and system (Abstract) including determine a motion artifact in the respective frame based on a comparison of the first parameter with the second parameter (invention possesses ability to more completely remove motion and other artifacts by firmware and/or software correction that utilizes information collected preferably from a sensor or device to detect body motion, 3D accelerometer connected to microprocessing unit, microprocessor applies particular tests and algorithms comparing the two signal sets to correct any motion artifacts, processor may apply more complicated frequency analysis, frequency analysis preferably in the form of wavelet analysis can be applied to yield artifact detection, Column 14, Lines 25-62; motion artifacts would obviously be detected by comparing different parameters or frequencies involved in each signal); combine each of the respective frames from the plurality of frames based on the determination of motion artifact (invention possesses ability to more completely remove motion and other artifacts by firmware and/or software correction that utilizes information collected preferably from a sensor or device to detect body motion, 3D accelerometer connected to microprocessing unit, microprocessor applies particular tests and algorithms comparing the two signal sets to correct any motion artifacts, processor may apply more complicated frequency analysis, frequency analysis preferably in the form of wavelet analysis can be applied to yield artifact detection, Column 14, Lines 25-62; signals would obviously need to be combined to allow for comparison and for producing the signal without the motion artifacts) since comparisons of signals are a known way to remove motion artifacts in signals (Column 14, Lines 25-62).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the processor and motion sensors to allow comparison of signals for detection and removal of motion artifacts, as taught by Bibian, since comparisons of signals are a known way to remove motion artifacts in signals (Bibian: Column 14, Lines 25-62). The determination of motion artifacts within a signal is obviously a well-known consideration done by one of ordinary skill in the art. Wong already discusses the importance of determining tremor movements from non-tremor movements (since the device is a wearable device on the wrist or arm of the user) and knowing that a non-tremor, voluntary activity falls within the band of 0.1-3 Hz (Wong: paragraph 0260). Bibian simply further supports the idea of making this determination through the comparison of signals and identification/detection of motion artifacts. Wong already discusses that tremors are to be found in the 4-12 Hz range (Wong: paragraphs 0228 and 0254). Thus, it is obvious for one of ordinary skill in the art to discern that any frequencies outside of that range would be considered noise or artifacts in the signal, which would obviously include motion artifacts. Additionally, signal processing techniques for removing noise and artifacts are well known as signal comparisons through correlation or cross-correlation are known to one of ordinary skill in the art. Other techniques within the frequency domain, like applying low-pass or high-pass or band-pass filters, are well-known to filter out noise/artifacts expected at frequency ranges outside of the desired frequency range. Thus, one of ordinary skill in the art would be motivated to remove noise/artifacts through various well-known signal processing techniques.
Wong-Bibian combination teaches the use of predictive adaptation and using predictive algorithms to predict when tremors will increase (Wong: paragraphs 0240-0242). Wong-Bibian combination fails to teach wherein determining neurostimulation therapy outcomes comprises comparing an examined output calculated by an examined rule with a potential output calculated by a potential rule based on cross-validation accuracy; and wherein both the examined rule and the potential rule are selected from the set of potential rules.
However, Kaemmerer, of the same field of endeavor and reasonably pertinent to the problem of neurostimulation, teaches a method for selecting a combination of electrodes (Abstract) including determining the neurostimulation therapy outcomes comprises comparing an examined output calculated by an examined rule with a potential output calculated by a potential rule based on cross-validation accuracy; and wherein both the examined rule and the potential rule are selected from a set of potential rules (machine learning algorithms/models may be utilized to enable device to select combination of electrodes for patient, classifier performance may be evaluated based on number of errors produced in a leave-one-out cross-validation scheme, features may be standardized across training observations before being used to train classifier, in order to predict group of new observation, score representing likelihood of being in group is calculated for each of the possible group, and observation is classified as being in group with largest score, paragraph 0103; leave-one-out cross-validation would compare examined output or test set with potential rule or training set to see how accurate or how well the classifier performs; the set of potential rules is merely the series of training sets) to evaluate the performance based on number of errors produced in cross-validation scheme (paragraph 0103).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize machine learning with cross-validation capabilities to verify and select the most effective stimulation parameters, as taught by Kaemmerer, to evaluate the performance based on number of errors produced in cross-validation scheme (Kaemmerer: paragraph 0103). Utilizing machine learning to improve a device’s ability to adapt to a patient’s particular needs is well-known in the art. Cross-validation is a well-known type of machine learning in which training sets and test sets are compared to confirm or verify if the performance of the algorithm is correct or effective. Kaemmerer shows that utilizing cross-validation is obvious as it would allow one of ordinary skill in the art to obtain further feedback and ensuring that the device is treating the tremors effectively. It also trains the device to select and choose the most effective combination of electrodes and stimulation parameters for particular tremors. Since Wong already obtains feedback to adjust stimulation parameters, having machine learning would simply further improve upon the existing feedback capabilities. Applicant has not claimed any particular features within the algorithm that are significantly distinct from well-known machine learning methods.
Regarding Claim 55, Wong-Bibian-Kaemmerer combination teaches the first frequency band is between about 0 Hz and about 2.5 Hz (Wong: voluntary band is 0.1-3 Hz, non-tremor or voluntary movements, paragraph 0260).
Regarding Claim 56, Wong-Bibian-Kaemmerer combination teaches the second frequency band is between about 4 Hz and about 12 Hz (Wong: typical tremor frequencies are 4-12 Hz, paragraph 0228; tremor band of 4-12 Hz, paragraph 0260).
Regarding Claim 57, Wong-Bibian-Kaemmerer combination does not explicitly teach the second frequency band is between about 3 Hz and about 8 Hz.
However, Wong further teaches typical tremor frequencies are 4-12 Hz and tremor bands are 4-12 Hz (Wong: paragraph 0228; paragraph 0260).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to expect a second frequency band between about 3 Hz and about 8 Hz, as taught by Wong, since typical tremor frequencies would fall within this frequency range. Since patients would have tremor frequencies that vary over time, some of those tremor frequencies would fall within the claimed frequency range.
Regarding Claim 58, Wong-Bibian-Kaemmerer combination teaches wherein the features comprise at least one or more of: amplitude (Wong: tremor amplitude, paragraph 0232), bandwidth (Wong: tremor band of 4-12 Hz, paragraph 0260), power (Wong: energy under the curve, paragraph 0233; spectral power at a frequency or spectral energy in the 4-12 Hz band, paragraph 0232), peak frequency (Wong: peak frequency in 4-12 Hz range, paragraph 0254; peak frequency in spectral density curve, paragraph 0238).
Regarding Claim 64, Wong-Bibian-Kaemmerer combination teaches the neurostimulation therapy outcomes comprise predicting patient tremor severity at a given point (Wong: understanding historical tremor measurements and the time therapy was applied can inform therapy needed on successive days, neural networks, Kalman filters, and other such predictive algorithms can be used to predict when tremor will increase and apply pre-emptive treatment, paragraph 0240).
PNG
media_image1.png
460
471
media_image1.png
Greyscale
Claims 4-5 are rejected under 35 U.S.C. 103 as being unpatentable over Wong et al. (US 2017/0157398 A1), Bibian et al. (US 8,538,512 B1), and Kaemmerer et al. (US 2016/0144186 A1) as applied to Claim 1, and in further view of Rosenbluth et al. (US 2015/0321000 A1).
Regarding Claim 4, Wong-Bibian-Kaemmerer combination teaches the claimed invention of Claim 1. Wong-Bibian-Kaemmerer combination fail to teach one or more end effectors configured to generate stimulation signals other than electric stimulation signals.
However, Rosenbluth, of the same field of endeavor, teaches a peripheral nerve stimulator can be used to stimulate a peripheral nerve to treat tremor (Abstract) including one or more end effectors configured to generate stimulation signals other than electric stimulation signals (vibrotactile stimulation refers to excitation of proprioceptors by application of biomechanical load to soft tissue and nerves, paragraph 0117; effectors may be mechanical excitation of proprioceptors by means of vibrotactile or haptic sensation, might include force, vibration and/or motion, mechanical effectors include small motors, piezoelectrics, vibrotactile units comprised of mass and effector to move mass such that vibratory stimulus is applied, paragraph 0151) since this is a known way to reduce tremors and is capable of reducing tremors through several methods (paragraph 0153).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to add an end effector that generates vibrotactile stimulation, as taught by Rosenbluth, since this is a known way to reduce tremors and is capable of reducing tremors through several methods (Rosenbluth: paragraph 0153). The device already reduces tremors through electrical stimulations. By adding these effectors, the tremors would also be reduced through vibrotactile stimulation. This addition would provide multiple, different avenues in which tremors would be reduced and increase efficacy of treatment.
Regarding Claim 5, Wong-Bibian-Kaemmerer-Rosenbluth combination teaches the stimulation signals other than the electric stimulation signals are vibrational stimulation signals (Rosenbluth: vibrotactile stimulation refers to excitation of proprioceptors by application of biomechanical load to soft tissue and nerves, paragraph 0117; effectors may be mechanical excitation of proprioceptors by means of vibrotactile or haptic sensation, might include force, vibration and/or motion, mechanical effectors include small motors, piezoelectrics, vibrotactile units comprised of mass and effector to move mass such that vibratory stimulus is applied, paragraph 0151).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to BRIAN THAI-BINH KHONG whose telephone number is (571)272-1857. The examiner can normally be reached Monday to Thursday 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, Kendra Carter can be reached at (571) 272-9034. 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.
/BRIAN T KHONG/Examiner, Art Unit 3785
/KENDRA D CARTER/Supervisory Patent Examiner, Art Unit 3785