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 .
Response to Amendment
Claims 1-20 were previously pending in this application. The amendment filed 17 February 2026 has been entered and the following has occurred: Claims 1, 6, 10, & 15-16 have been amended. No claims have been added or cancelled.
Claims 1-20 remain pending in the application.
Claim Objections
Claim 15 is objected to because of the following informalities:
Regarding Claim 15, the limitation “a group of patients who share a similarity in at least one of a physiological feature, and a symptom diagnosis, a neuromodulator device type” rather than “a group of patients who share a similarity in at least one of a physiological feature, a symptom diagnosis, and a neuromodulator device type”
Appropriate correction is required.
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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Choi et al. (U.S. Patent Publication No. 2023/0123383), hereinafter “Choi”, in view of Bartlett et al. (U.S. Patent Publication No. 2023/0191130), hereinafter “Bartlett”, further in view of Annoni et al. (U.S. Patent Publication No. 2023/0103448), hereinafter “Annoni”.
Claim 1 –
Regarding Claim 1, Choi discloses a computer-implemented method comprising:
collecting, by a data collection module (See Choi Par [0049] which discloses the use of one or more electric elements, components or modules, for performing a function if the element may be programmed for performing or otherwise structurally arranged for to perform that action, and is understood to read on each instance of the various “units” recited hereinafter), a first set of patient data and a first set of treatment data associated with a patient population treated with neuromodulation (See Choi Par [0131] which discloses captured data including context data and sensor data, the context data including data relating to neurostimulation therapy settings for each patient and timestamps for when the therapy was provided, i.e. treatment data, and the sensor data including biometric or other types of data captured by varying devices including a “smart” watch, a health monitoring device, or other types of consumer electronic devices, i.e. patient data);
clustering according to a set of vectors of features corresponding to the first set of patient data, by a clustering module, the patient population into a plurality of cohorts, wherein the at least one cohort is clustered according to an implant location, the implant location being a feature in the set of vector of features (See Choi Par [0134] which discloses ML/AI processing of the captured data to develop a predictive disorder model, such that the algorithm classifies patients by disorder, i.e. cohorts, and includes classification techniques, such as k-means clustering, which is understood to include clustering and/or a clustering module; See Choi Par [0196] which discloses features of two population groups, i.e. cohorts, to control neurostimulation; See Choi Par [0194] which discloses classifying being defined for classification of the state of a patient to control a neurostimulation therapy, such that spatial locations in the EEG data that correspond to cortical activity data can be analyzed and compared to healthy/control groups, however not specifically described as clustered according to said implant locations);
generating, by a state discovery module, a plurality of states using a second set of patient data associated with a cohort in the plurality of cohorts (See Choi Par [0134]-[0135] which discloses ML/AI processing of the captured data to develop a predictive disorder model, such that the algorithm classifies patients by disorder (e.g. health state, disorder state, disorder level, etc.), such that the predictive disorder model may include various disorder states, i.e. a plurality of states, such as in the case of Parkinson’s Disease, the predictive disorder model may include model components such as tremor, rigidity, facial drop, balance, hallucinations, and/or any other relevant disorder symptoms to create an overall disorder classification, i.e. state/class, that reflects an ensemble of the selected model components);
generating, by an action discovery module, a plurality of actions using a second set of treatment data associated with the cohort (See Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while each neurostimulation stimulation therapy session (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of neurostimulation parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient, i.e. plurality of actions for each subsequent set/instance of treatment);
generating, by a recommendation module, and based on a current setting of a neuromodulation device corresponding to the patient using time-series treatment data associated with the patient a neuromodulator action policy for a patient in the cohort (BRI of “neuromodulator action policy” in light of Applicant’s Specification is understood to include any course of action recommended for the neuromodulator that is tailored to the patient, therefore see Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient; See Choi Par [0094] & [0129] which discloses the therapy program parameters including one or more sets of stimulation parameters corresponding to different lead/electrode combinations, pulse amplitude, stimulation level, pulse width, pulse frequency or inter-pulse period, pulse repetition parameter (e.g., number of times for a given pulse to be repeated for respective stimulation sets or “stimsets” during the execution of a program), etc., electrode configuration data for delivery of electrical pulses (e.g., as cathodic nodes, anodic nodes, or configured as inactive nodes, etc.), stimulation pattern identification (e.g., tonic stimulation, burst stimulation, noise stimulation, biphasic stimulation, monophasic stimulation, and/or the like), etc.; See Choi Par [0113] which discloses “big data” being received and used by the data analytics platform for purposes of determining neurostimulation parameters, including patients’ health data, physiological/behavioral data, sensor data, daily activity data, therapy settings (i.e. electrode configuration data for delivery of electrical pulses, stimulation pattern identification, etc. as mentioned in Choi Par [0094]), and further discloses in Choi Par [0178] that the system can be used to make small adjustments to programmed parameters to explore the therapeutic state space when patient performs various exercises, and would therefore constitute adjustments from a “current” programmed setting to an adjusted setting; See Choi Par [0131] which specifically mentions timestamps according to the time of day when said neurostimulation therapy and associated settings were applied for each patient), the neuromodulator action policy including a sequence of adjustments to the neuromodulation device (See Choi Par [0131] which specifically mentions timestamps according to the time of day when said neurostimulation therapy and associated settings were applied for each patient; See Choi Par [0133] which discloses that the neurostimulation therapy can include stimulation pulses (by the neurostimulation device) that may be delivered continuously, or can be delivered to target tissue of the patient for a period of time and not delivered during another period of time, constituting a sequence of adjustments to the neurostimulation device; Choi Par [0178] that the system can be used to make small adjustments to programmed parameters over time to titrate the patient’s neurostimulation therapy during performance of tasks or activities and exploring the therapeutic state space when patient performs various exercises, and would therefore constitute adjustments from a “current” programmed setting to an adjusted setting; See Choi Par [0131] which specifically mentions timestamps according to the time of day when said neurostimulation therapy and associated settings were applied for each patient, i.e. also sequence of adjustments from a timestamp aspect; See Choi Par [0228] which discloses that the automated changes, i.e. adjustments, to the neurostimulator device parameters may be implemented to apply neurostimulation according to the modified parameters and additional iterations of the process, i.e. a sequence of adjustments, may be performed if deemed appropriate by the clinician);
transmitting, by the recommendation module, neuromodulator action policy to the neuromodulation device (see Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient, i.e. transmitted to neurostimulator; See Choi Par [0094] & [0129] which discloses the therapy program parameters including one or more sets of stimulation parameters corresponding to different lead/electrode combinations, pulse amplitude, stimulation level, pulse width, pulse frequency or inter-pulse period, pulse repetition parameter (e.g., number of times for a given pulse to be repeated for respective stimulation sets or “stimsets” during the execution of a program), etc., electrode configuration data for delivery of electrical pulses (e.g., as cathodic nodes, anodic nodes, or configured as inactive nodes, etc.), stimulation pattern identification (e.g., tonic stimulation, burst stimulation, noise stimulation, biphasic stimulation, monophasic stimulation, and/or the like), etc.; See Choi Par [0073] which specifically discloses “stimulation therapy settings for secure transmission to the patient device, which may be securely transmitted via encrypted communications”, constituting transmission of the neuromodulator action policy to the patient neuromodulation device); and
adjusting, by the neuromodulation device, a neuromodulator setting of the neuromodulation device, the adjusting the neuromodulation setting based on the neuromodulator action policy (see Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient; See Choi Par [0094] & [0129] which discloses the therapy program parameters including one or more sets of stimulation parameters corresponding to different lead/electrode combinations, pulse amplitude, stimulation level, pulse width, pulse frequency or inter-pulse period, pulse repetition parameter (e.g., number of times for a given pulse to be repeated for respective stimulation sets or “stimsets” during the execution of a program), etc., electrode configuration data for delivery of electrical pulses (e.g., as cathodic nodes, anodic nodes, or configured as inactive nodes, etc.), stimulation pattern identification (e.g., tonic stimulation, burst stimulation, noise stimulation, biphasic stimulation, monophasic stimulation, and/or the like), etc., which is understood to constitute adjustment of a neuromodulator setting).
While Choi discloses collecting patient data from a population, clustering said population into cohorts, generating one or more states and/or actions according to varying patient data from said cohorts, and subsequently implementing said actions in the form of a neurostimulation treatment plan and modifying parameters of said plan based on a current setting, and conducting a sequence of adjustments to the device, as shown above, and Choi specifically discloses the use of Markov decision processes as a classification technique to classify disorders of patients, Choi does not explicitly mention determining a plurality of probabilities associated with a transition from a first state to a second state in the plurality of states, and using said probabilities for generating the neuromodulator action policy, as given by the following bolded portions of the limitations below:
determining, by a decision-making module based on the plurality of actions, a plurality of probabilities associated with a transition from a first state to a second state in the plurality of states; and
generating, by a recommendation module based on the plurality of probabilities, and based on a current setting of a neuromodulation device corresponding to the patient using time-series treatment data associated with the patient a neuromodulator action policy for a patient in the cohort, the neuromodulator action policy including a sequence of adjustments to the neuromodulation device.
Therefore, Bartlett discloses determining, by a decision-making module based on the plurality of actions, a plurality of probabilities associated with a transition from a first state to a second state in the plurality of states (See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process in which transitions between states are stochastic, marked by transition probabilities, which is the probability of moving from one state to another state by taking an action, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states); and generating, by a recommendation module based on the plurality of probabilities, and based on a current setting of a neuromodulation device corresponding to the patient using time-series treatment data associated with the patient a neuromodulator action policy for a patient in the cohort, the neuromodulator action policy including a sequence of adjustments to the neuromodulation device (See Bartlett Par [0048]-[0049] & Fig. 2 which discloses the generation of a policy function for a neuromodulation device/system, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states, and further describes that the Markov decision process modeling can be used to generate state-reward relationships to build up a policy function for said neuromodulation device/system, as shown in FIG. 2, which is used to model mappings between current neural firing patterns, i.e. initial state, and applied stimuli, i.e. action, and series of (neuro)stimulation actions to take to move to desired states).
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 disclosure of Choi, which discloses collecting patient data from a population, clustering said population into cohorts, generating one or more states and/or actions according to varying patient data from said cohorts, and subsequently implementing said actions in the form of a neurostimulation treatment plan and modifying parameters of said plan, to further include determining a plurality of probabilities associated with a transition from a first state to a second state in the plurality of states and generating, by a recommendation module based on the plurality of probabilities, a neuromodulator action policy for a patient in the cohort, as disclosed by Bartlett, because this allows for modeling the states as a Markov decision process, in which transitions between states are stochastic and marked by transition probabilities, and therefore allows the relationship between neural firing patterns and applied stimuli to drive neural firing patterns and brain state towards a desired target to be modeled for subsequent input/use in reinforcement learning techniques (See Bartlett Par [0048]-[0049]).
While Choi and Bartlett effectively disclose generating a neuromodulator action policy based on current settings and modifying the device to include a sequence of adjustments to the various neuromodulation/treatment device parameters based on the determined neuromodulator action policy, Choi and Bartlett do not seem to effectively disclose the changing neuromodulation/treatment device parameters specifically relating to a wave shape of the waveform to be delivered by the neuromodulation/treatment device, as given by the following limitation:
the neuromodulator setting comprising waveform parameter, the waveform parameter comprising a wave shape.
However, Annoni discloses the neuromodulator setting comprising waveform parameter, the waveform parameter comprising a wave shape (See Annoni Par [0073] which discloses the use of a neuromodulator, including an implantable neuromodulator control generation and delivery of electrostimulation using the information about electrostimulation parameters and electrode configuration from the learning system; See Annoni Par [0074] which discloses a controller circuit controlling the generation and delivery of electrostimulation in a closed-loop fashion by adaptively adjusting one or more stimulation parameters or stimulation electrode configuration based on the pain score, and the electrostimulations may differ in at least one of the stimulation energy, pulse amplitude, pulse width, stimulation frequency, duration, on-off cycle, pulse shape or waveform). The disclosure of Annoni is directly applicable to the combined disclosure of Choi and Bartlett because the disclosures share limitations and capabilities, such as being directed towards the development of neurostimulation/neuromodulation therapy plans.
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 combined disclosure of Choi and Bartlett discloses generating a neuromodulator action policy, including various neuromodulation/treatment device parameters being changed based on the determined neuromodulator action policy to specifically include the neuromodulator setting comprising a waveform parameter, the waveform parameter comprising a wave shape, as disclosed by Annoni, because this allows for more individualized electrostimulation being delivered based on varying pain parameters experienced by the patient (See Annoni Par [0073]-[0074]).
While the combination of Choi, Bartlett and Annoni discloses clustering patients into varying cohorts based on feature vectors and/or classification of the state of a patient to control a neurostimulation therapy, such that spatial locations in the EEG data that correspond to cortical activity data can be analyzed and compared to healthy/control groups, this combination of references does not specifically describe clustering patient data into cohorts according to said implant locations as required by the following bolded portions of the limitation found below:
clustering according to a set of vectors of features corresponding to the first set of patient data, by a clustering module, the patient population into a plurality of cohorts, wherein the at least one cohort is clustered according to an implant location, the implant location being a feature in the set of vector of features.
However, Schulte discloses clustering patient data into cohorts according to said implant locations (See Schulte Par [0139] & [0245] which discloses collecting data from feature extraction engines of a plurality of users or patients, and the aggregation or collection of data can be taken from multiple different patients may be taken from subpopulations of patients with similar characteristics, such as neurostimulation devices distributed across geographic locations, and further discloses in Schulte Par [0140] & [0246] that for example, extracted features can be weighted based on, for example, geospatial data and/or temporal data indicating more or less closeness to a location and/or time specified in health related determination; See Schulte Par [0147] which discloses the use of aggregation learning algorithms and/or the learning algorithm develop rules between features and one or more parameters of one or more response profiles that correspond to therapy outcomes and or treatment parameters, such as those features disclosed in Schulte Par [0139]-[0140] & [0245]-[0246] regarding geographic location of neurostimulation devices). The disclosure of Schulte is directly applicable to the combined disclosure of Choi, Bartlett, and Annoni, because the disclosures share limitations and capabilities, such as being directed towards the development of neurostimulation/neuromodulation therapy plans.
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 disclosure of Choi, Bartlett, and Annoni, which already discloses clustering patients into varying cohorts based on feature vectors and/or classification of the state of a patient to control a neurostimulation therapy to further include clustering patient data into cohorts according to said implant locations, as disclosed by Schulte, because this allows for developing rules between features and one or more shared parameters of one or more response profiles that correspond to neurostimulation therapy outcomes, such as based on geographic locations of neurostimulation devices (See Schulte Par [0139], [0147], & [0245]).
Claim 2 –
Regarding Claim 2, Choi, Bartlett, Annoni, and Schulte disclose the method of claim 1 in its entirety. Choi and Bartlett further disclose a method, further comprising:
determining a current state of the patient using time-series patient data associated with the patient (See Choi Par [0134]-[0135] which discloses ML/AI processing of the captured data to develop a predictive disorder model, such that the algorithm classifies patients by disorder (e.g. health state, disorder state, disorder level, etc.), such that the predictive disorder model may include various disorder states, i.e. a plurality of states, such as in the case of Parkinson’s Disease, the predictive disorder model may include model components such as tremor, rigidity, facial drop, balance, hallucinations, and/or any other relevant disorder symptoms to create an overall disorder classification, i.e. state/class, that reflects an ensemble of the selected model components); and
determining the neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state (See Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient, albeit not specifically recited for highest probability to transition from one state to another; See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process in which transitions between states are stochastic, marked by transition probabilities, which is the probability of moving from one state to another state by taking an action, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states, and further discloses in Bartlett Par [0050]-[0052] that the policy function can map state-to-action spaces and effectively chooses the actions, such as the stimulus, by evaluating how successful a given action taken by the actor was and how it should adjust to minimize current and future errors, and thereby maximize probability of transitioning and Bartlett further discloses that a reward function fundamentally dictates and quantifies (neuro)stimulation goals, such that Bartlett utilizes a mean-square error loss function and that the function is asymptotically the maximum likelihood estimator, i.e. maximum probability, for the process of transitioning from an observed neural state to a new/desired state based on a given action/process, constituting determining a neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state).
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 disclosure of Choi, Bartlett, Annoni, and Schulte to further include additional embodiments found in Bartlett regarding determining the neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state, because by choosing actions that maximize said probability to transition from a current state to another, the system can effectively adjust parameters of the (neuro)stimulation action policy to minimize current and future errors of said transition (see Bartlett Par [0050]).
Claim 3 –
Regarding Claim 3, Choi, Bartlett, Annoni, and Schulte disclose the method of claim 2 in its entirety. Choi and Bartlett further disclose a method, further comprising:
wherein the sequence of adjustments have the highest probability to transition from the current state to another state (See Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient, albeit not specifically recited for highest probability to transition from one state to another; See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process in which transitions between states are stochastic, marked by transition probabilities, which is the probability of moving from one state to another state by taking an action, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states, and further discloses in Bartlett Par [0050]-[0052] that the policy function can map state-to-action spaces and effectively chooses the actions, such as the stimulus, by evaluating how successful a given action taken by the actor was and how it should adjust to minimize current and future errors, and thereby maximize probability of transitioning and Bartlett further discloses that a reward function fundamentally dictates and quantifies (neuro)stimulation goals, such that Bartlett utilizes a mean-square error loss function and that the function is asymptotically the maximum likelihood estimator, i.e. maximum probability, for the process of transitioning from an observed neural state to a new/desired state based on a given action/process, constituting determining a neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state).
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 disclosure of Choi, Bartlett, Annoni, and Schulte, to further include additional embodiments found in Bartlett regarding determining the neuromodulator action policy by selecting one or more actions of the neuromodulator associated with a highest probability to transition from the current state to another state, because by choosing actions that maximize said probability to transition from a current state to another, the system can effectively adjust parameters of the (neuro)stimulation action policy to minimize current and future errors of said transition (see Bartlett Par [0050]).
Claim 4 –
Regarding Claim 4, Choi, Bartlett, Annoni, and Schulte disclose the method of claim 1 in its entirety. Choi does not disclose, but Bartlett further discloses a method, further comprising:
generating a Markov decision process using time-series patient data and time-series treatment data associated with the cohort (See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process in which transitions between states are stochastic, marked by transition probabilities, which is the probability of moving from one state to another state by taking an action, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states, and further discloses in Bartlett Par [0050]-[0052] that the policy function can map state-to-action spaces and effectively chooses the actions, such as the stimulus, by evaluating how successful a given action taken by the actor was and how it should adjust to minimize current and future errors, and thereby maximize probability of transitioning and Bartlett further discloses that a reward function fundamentally dictates and quantifies (neuro)stimulation goals, such that Bartlett utilizes a mean-square error loss function and that the function is asymptotically the maximum likelihood estimator, i.e. maximum probability, for the process of transitioning from an observed neural state to a new/desired state based on a given action/process, constituting determining a neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state); and
determining the plurality of probabilities using the Markov decision process (See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process in which transitions between states are stochastic, marked by transition probabilities, which is the probability of moving from one state to another state by taking an action, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states, and further discloses in Bartlett Par [0050]-[0052] that the policy function can map state-to-action spaces and effectively chooses the actions, such as the stimulus, by evaluating how successful a given action taken by the actor was and how it should adjust to minimize current and future errors, and thereby maximize probability of transitioning and Bartlett further discloses that a reward function fundamentally dictates and quantifies (neuro)stimulation goals, such that Bartlett utilizes a mean-square error loss function and that the function is asymptotically the maximum likelihood estimator, i.e. maximum probability, for the process of transitioning from an observed neural state to a new/desired state based on a given action/process, constituting determining a neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state).
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 disclosure of Choi, Bartlett, Annoni, and Schulte, to further include additional embodiments found in Bartlett regarding generating a Markov decision process using time-series patient data and time-series treatment data associated with the cohort and determining the plurality of probabilities using the Markov decision process, because by modeling the states as a Markov decision process, in which transitions between states are stochastic and marked by transition probabilities, this allows the relationship between neural firing patterns and applied stimuli to drive neural firing patterns and brain state towards a desired target to be modeled for subsequent input/use in reinforcement learning techniques (See Bartlett Par [0048]-[0049]).
Claim 5 –
Regarding Claim 5, Choi, Bartlett, Annoni, and Schulte disclose the method of claim 4 in its entirety. Choi and Bartlett further disclose a method, further comprising:
receiving a user feedback on the neuromodulator action policy (See Choi Par [0097] which discloses the system facilitating patient input/feedback with respect to a trial therapy or treatment, by providing an interface where input/feedback can be augmented with one or more data labeling buttons, icons, pictograms, etc., and wherein the patient input/feedback data may be provided to a network-based AI/ML model for facilitating intelligent decision-making with respect to whether the IMD/NIMI device should be deployed in a more permanent manner (e.g., implantation) and/or whether a particular therapy setting or a set of settings, including context-sensitive therapy program selection, may need to be optimized or otherwise reconfigured; See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process for use in reinforcement learning techniques, i.e. the AI/ML model of Choi being modified to be a Markov decision process, as disclosed by Bartlett); and
adjusting the Markov decision process based on the user feedback (See Choi Par [0097] which discloses the system facilitating patient input/feedback with respect to a trial therapy or treatment, by providing an interface where input/feedback can be augmented with one or more data labeling buttons, icons, pictograms, etc., and wherein the patient input/feedback data may be provided to a network-based AI/ML model for facilitating intelligent decision-making with respect to whether the IMD/NIMI device should be deployed in a more permanent manner (e.g., implantation) and/or whether a particular therapy setting or a set of settings, including context-sensitive therapy program selection, may need to be optimized or otherwise reconfigured; See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process for use in reinforcement learning techniques, i.e. the AI/ML model of Choi being modified to be a Markov decision process, as disclosed by Bartlett, and therefore by Choi stating the patient input/feedback data may be provided to a network-based AI/ML model and Bartlett specifying the model as a Markov decision process for use in reinforcement learning techniques, this limitation is considered to be met).
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 disclosure of Choi, Bartlett, Annoni, and Schulte to further include additional embodiments found in Bartlett regarding generating a Markov decision process using time-series patient data and time-series treatment data associated with the cohort and determining the plurality of probabilities using the Markov decision process, because by modeling the states as a Markov decision process, in which transitions between states are stochastic and marked by transition probabilities, this allows the relationship between neural firing patterns and applied stimuli to drive neural firing patterns and brain state towards a desired target to be modeled for subsequent input/use in reinforcement learning techniques (See Bartlett Par [0048]-[0049]).
Claim 6 –
Regarding Claim 6, Choi, Bartlett, Annoni, and Schulte disclose the method of claim 1 in its entirety. Choi further discloses a method, wherein:
a cohort represents a group of patients who share a similarity in at least one of a physiological feature, a symptom diagnosis, a neuromodulator device type (See Choi Par [0134] which discloses ML/AI processing of the captured data to develop a predictive disorder model, such that the algorithm classifies patients by disorder, i.e. cohorts, and includes classification techniques, such as k-means clustering, which is understood to include clustering and/or a clustering module; See Choi Par [0134]-[0135] which discloses ML/AI processing of the captured data to develop a predictive disorder model, such that the algorithm classifies patients by disorder (e.g. health state, disorder state, disorder level, etc.), such that the predictive disorder model may include various disorder states, i.e. a plurality of states, such as in the case of Parkinson’s Disease, the predictive disorder model may include model components such as tremor, rigidity, facial drop, balance, hallucinations, and/or any other relevant disorder symptoms to create an overall disorder classification, i.e. state/class, that reflects an ensemble of the selected model components).
Claim 7 –
Regarding Claim 7, Choi, Bartlett, Annoni, and Schulte disclose the method of claim 1 in its entirety. Choi further discloses a method, further comprising:
presenting the neuromodulator action policy to the patient (See Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient; See Choi Par [0065] which discloses therapy instructions, i.e. neurostimulation therapy instructions, may be mediated, proxied or otherwise relayed by way of a controller device associated with the patient, i.e. presented to the patient); and
automatically adjusting a neuromodulator setting based on the neuromodulator action policy (See Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient).
Claim 8 –
Regarding Claim 8, Choi, Bartlett, Annoni, and Schulte disclose the method of claim 1 in its entirety. Choi further discloses a method, wherein:
the patient data includes at least one of a symptom level, an activity level, a sleep quality, and a mood level (See Choi Par [0131] which discloses captured data including context data and sensor data, the context data including data relating to neurostimulation therapy settings for each patient and timestamps for when the therapy was provided, i.e. treatment data, and the sensor data including biometric or other types of data captured by varying devices including a “smart” watch, a health monitoring device, or other types of consumer electronic devices, i.e. patient data; See Choi Par [0184] which discloses bio signals including heart rate, sleep quality, time standing, heart rate variability, flights of stairs climbed, distance of walking and running, count of step, location information using GPS, and posture information using accelerometer and gyroscope, a pain score; See Choi Fig. 11H which includes querying the patient for rating their mood level in the last seven days and Choi Par [0108]-[0109] which further discloses querying the patient for emotional states, well-being scores, i.e. mood level, reported levels of pain, activity levels, etc.; See Choi Par [0134]-[0135] which discloses ML/AI processing of the captured data to develop a predictive disorder model, such that the algorithm classifies patients by disorder (e.g. health state, disorder state, disorder level, etc.), such that the predictive disorder model may include various disorder states, i.e. a plurality of states, such as in the case of Parkinson’s Disease, the predictive disorder model may include model components such as tremor, rigidity, facial drop, balance, hallucinations, and/or any other relevant disorder symptoms to create an overall disorder classification, i.e. state/class, that reflects an ensemble of the selected model components); and
a state in the plurality of states includes a combination of at least one of the symptom level, the activity level, the sleep quality, and the mood level (See Choi Par [0134]-[0135] which discloses ML/AI processing of the captured data, e.g. biosignal data, patient-reported data, etc., to develop a predictive disorder model, such that the algorithm classifies patients by disorder (e.g. health state, disorder state, disorder level, etc.), such that the predictive disorder model may include various disorder states, i.e. a plurality of states, such as in the case of Parkinson’s Disease, the predictive disorder model may include model components such as tremor, rigidity, facial drop, balance, hallucinations, and/or any other relevant disorder symptoms to create an overall disorder classification, i.e. state/class, that reflects an ensemble of the selected model components).
Claim 9 –
Regarding Claim 9, Choi, Bartlett, Annoni, and Schulte disclose the method of claim 1 in its entirety. Choi further discloses a method, wherein:
the treatment data includes at least one of a neuromodulator frequency, a neuromodulator current, a neuromodulator voltage, and a neuromodulator intensity (See Choi Par [0131] which discloses captured data including context data and sensor data, the context data including data relating to neurostimulation therapy settings for each patient and timestamps for when the therapy was provided; See Choi Par [0094] & [0129] which discloses the therapy program parameters including one or more sets of stimulation parameters corresponding to different lead/electrode combinations, pulse amplitude, stimulation level, pulse width, pulse frequency or inter-pulse period, pulse repetition parameter (e.g., number of times for a given pulse to be repeated for respective stimulation sets or “stimsets” during the execution of a program), etc., electrode configuration data for delivery of electrical pulses (e.g., as cathodic nodes, anodic nodes, or configured as inactive nodes, etc.), stimulation pattern identification (e.g., tonic stimulation, burst stimulation, noise stimulation, biphasic stimulation, monophasic stimulation, and/or the like), etc.); and
the neuromodulator action policy includes adjusting at least one of the neuromodulator frequency, the neuromodulator current, the neuromodulator voltage, and the neuromodulator intensity (See Choi Par [0131] which discloses captured data including context data and sensor data, the context data including data relating to neurostimulation therapy settings for each patient and timestamps for when the therapy was provided; See Choi Par [0094] & [0129] which discloses the therapy program parameters including one or more sets of stimulation parameters corresponding to different lead/electrode combinations, pulse amplitude, stimulation level, pulse width, pulse frequency or inter-pulse period, pulse repetition parameter (e.g., number of times for a given pulse to be repeated for respective stimulation sets or “stimsets” during the execution of a program), etc., electrode configuration data for delivery of electrical pulses (e.g., as cathodic nodes, anodic nodes, or configured as inactive nodes, etc.), stimulation pattern identification (e.g., tonic stimulation, burst stimulation, noise stimulation, biphasic stimulation, monophasic stimulation, and/or the like), etc.).
Claim 10 –
Regarding Claim 10, Choi discloses a computer program product comprising one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions executable by a processor to cause the processor to perform operations (See Choi Par [0052]) comprising:
collecting, by a data collection module (See Choi Par [0049] which discloses the use of one or more electric elements, components or modules, for performing a function if the element may be programmed for performing or otherwise structurally arranged for to perform that action, and is understood to read on each instance of the various “units” recited hereinafter), a first set of patient data and a first set of treatment data associated with a patient population treated with neuromodulation (See Choi Par [0131] which discloses captured data including context data and sensor data, the context data including data relating to neurostimulation therapy settings for each patient and timestamps for when the therapy was provided, i.e. treatment data, and the sensor data including biometric or other types of data captured by varying devices including a “smart” watch, a health monitoring device, or other types of consumer electronic devices, i.e. patient data);
clustering according to a set of vectors of features corresponding to the first set of patient data, by a clustering module, the patient population into a plurality of cohorts, wherein at least one cohort is clustered according to an implant location, the implant location being a feature in the set of vectors of features (See Choi Par [0134] which discloses ML/AI processing of the captured data to develop a predictive disorder model, such that the algorithm classifies patients by disorder, i.e. cohorts, and includes classification techniques, such as k-means clustering, which is understood to include clustering and/or a clustering module; See Choi Par [0196] which discloses features of two population groups, i.e. cohorts, to control neurostimulation; See Choi Par [0194] which discloses classifying being defined for classification of the state of a patient to control a neurostimulation therapy, such that spatial locations in the EEG data that correspond to cortical activity data can be analyzed and compared to healthy/control groups, however not specifically described as clustered according to said implant locations);
generating, by a state discovery module, a plurality of states using a second set of patient data associated with a cohort in the plurality of cohorts (See Choi Par [0134]-[0135] which discloses ML/AI processing of the captured data to develop a predictive disorder model, such that the algorithm classifies patients by disorder (e.g. health state, disorder state, disorder level, etc.), such that the predictive disorder model may include various disorder states, i.e. a plurality of states, such as in the case of Parkinson’s Disease, the predictive disorder model may include model components such as tremor, rigidity, facial drop, balance, hallucinations, and/or any other relevant disorder symptoms to create an overall disorder classification, i.e. state/class, that reflects an ensemble of the selected model components);
generating, by an action discovery module, a plurality of actions using a second set of treatment data associated with the cohort (See Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while each neurostimulation stimulation therapy session (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of neurostimulation parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient, i.e. plurality of actions for each subsequent set/instance of treatment);
generating, by a recommendation module based on the plurality of probabilities, and based on a current setting of a neuromodulation device corresponding to the patient using time-series treatment data associated with the patient a neuromodulator action policy for a patient in the cohort (BRI of “neuromodulator action policy” in light of Applicant’s Specification is understood to include any course of action recommended for the neuromodulator that is tailored to the patient, therefore see Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient; See Choi Par [0094] & [0129] which discloses the therapy program parameters including one or more sets of stimulation parameters corresponding to different lead/electrode combinations, pulse amplitude, stimulation level, pulse width, pulse frequency or inter-pulse period, pulse repetition parameter (e.g., number of times for a given pulse to be repeated for respective stimulation sets or “stimsets” during the execution of a program), etc., electrode configuration data for delivery of electrical pulses (e.g., as cathodic nodes, anodic nodes, or configured as inactive nodes, etc.), stimulation pattern identification (e.g., tonic stimulation, burst stimulation, noise stimulation, biphasic stimulation, monophasic stimulation, and/or the like), etc.; See Choi Par [0113] which discloses “big data” being received and used by the data analytics platform for purposes of determining neurostimulation parameters, including patients’ health data, physiological/behavioral data, sensor data, daily activity data, therapy settings (i.e. electrode configuration data for delivery of electrical pulses, stimulation pattern identification, etc. as mentioned in Choi Par [0094]), and further discloses in Choi Par [0178] that the system can be used to make small adjustments to programmed parameters to explore the therapeutic state space when patient performs various exercises, and would therefore constitute adjustments from a “current” programmed setting to an adjusted setting; See Choi Par [0131] which specifically mentions timestamps according to the time of day when said neurostimulation therapy and associated settings were applied for each patient), the neuromodulator action policy including a sequence of adjustments to the neuromodulation device (See Choi Par [0131] which specifically mentions timestamps according to the time of day when said neurostimulation therapy and associated settings were applied for each patient; See Choi Par [0133] which discloses that the neurostimulation therapy can include stimulation pulses (by the neurostimulation device) that may be delivered continuously, or can be delivered to target tissue of the patient for a period of time and not delivered during another period of time, constituting a sequence of adjustments to the neurostimulation device; Choi Par [0178] that the system can be used to make small adjustments to programmed parameters over time to titrate the patient’s neurostimulation therapy during performance of tasks or activities and exploring the therapeutic state space when patient performs various exercises, and would therefore constitute adjustments from a “current” programmed setting to an adjusted setting; See Choi Par [0131] which specifically mentions timestamps according to the time of day when said neurostimulation therapy and associated settings were applied for each patient, i.e. also sequence of adjustments from a timestamp aspect; See Choi Par [0228] which discloses that the automated changes, i.e. adjustments, to the neurostimulator device parameters may be implemented to apply neurostimulation according to the modified parameters and additional iterations of the process, i.e. a sequence of adjustments, may be performed if deemed appropriate by the clinician);
transmitting, by the recommendation module, neuromodulator action policy to a neuromodulation device (see Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient, i.e. transmitted to neurostimulator; See Choi Par [0094] & [0129] which discloses the therapy program parameters including one or more sets of stimulation parameters corresponding to different lead/electrode combinations, pulse amplitude, stimulation level, pulse width, pulse frequency or inter-pulse period, pulse repetition parameter (e.g., number of times for a given pulse to be repeated for respective stimulation sets or “stimsets” during the execution of a program), etc., electrode configuration data for delivery of electrical pulses (e.g., as cathodic nodes, anodic nodes, or configured as inactive nodes, etc.), stimulation pattern identification (e.g., tonic stimulation, burst stimulation, noise stimulation, biphasic stimulation, monophasic stimulation, and/or the like), etc.; See Choi Par [0073] which specifically discloses “stimulation therapy settings for secure transmission to the patient device, which may be securely transmitted via encrypted communications”, constituting transmission of the neuromodulator action policy to the patient neuromodulation device); and
adjusting, by the neuromodulation device, a neuromodulator setting of the neuromodulation device, the adjusting the neuromodulation setting based on the neuromodulator action policy (see Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient; See Choi Par [0094] & [0129] which discloses the therapy program parameters including one or more sets of stimulation parameters corresponding to different lead/electrode combinations, pulse amplitude, stimulation level, pulse width, pulse frequency or inter-pulse period, pulse repetition parameter (e.g., number of times for a given pulse to be repeated for respective stimulation sets or “stimsets” during the execution of a program), etc., electrode configuration data for delivery of electrical pulses (e.g., as cathodic nodes, anodic nodes, or configured as inactive nodes, etc.), stimulation pattern identification (e.g., tonic stimulation, burst stimulation, noise stimulation, biphasic stimulation, monophasic stimulation, and/or the like), etc., which is understood to constitute adjustment of a neuromodulator setting).
While Choi discloses collecting patient data from a population, clustering said population into cohorts, generating one or more states and/or actions according to varying patient data from said cohorts, and subsequently implementing said actions in the form of a neurostimulation treatment plan and modifying parameters of said plan based on a current setting, and conducting a sequence of adjustments to the device, as shown above, and Choi specifically discloses the use of Markov decision processes as a classification technique to classify disorders of patients, Choi does not explicitly mention determining a plurality of probabilities associated with a transition from a first state to a second state in the plurality of states, and using said probabilities for generating the neuromodulator action policy, as given by the following bolded portions of the limitations below:
determining, by a decision-making module based on the plurality of actions, a plurality of probabilities associated with a transition from a first state to a second state in the plurality of states; and
generating, by a recommendation module based on the plurality of probabilities, and based on a current setting of a neuromodulation device corresponding to the patient using time-series treatment data associated with the patient a neuromodulator action policy for a patient in the cohort, the neuromodulator action policy including a sequence of adjustments to the neuromodulation device.
Therefore, Bartlett discloses determining, by a decision-making module based on the plurality of actions, a plurality of probabilities associated with a transition from a first state to a second state in the plurality of states (See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process in which transitions between states are stochastic, marked by transition probabilities, which is the probability of moving from one state to another state by taking an action, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states); and generating, by a recommendation module based on the plurality of probabilities, and based on a current setting of a neuromodulation device corresponding to the patient using time-series treatment data associated with the patient a neuromodulator action policy for a patient in the cohort, the neuromodulator action policy including a sequence of adjustments to the neuromodulation device (See Bartlett Par [0048]-[0049] & Fig. 2 which discloses the generation of a policy function for a neuromodulation device/system, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states, and further describes that the Markov decision process modeling can be used to generate state-reward relationships to build up a policy function for said neuromodulation device/system, as shown in FIG. 2, which is used to model mappings between current neural firing patterns, i.e. initial state, and applied stimuli, i.e. action, and series of (neuro)stimulation actions to take to move to desired states).
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 disclosure of Choi, which discloses collecting patient data from a population, clustering said population into cohorts, generating one or more states and/or actions according to varying patient data from said cohorts, and subsequently implementing said actions in the form of a neurostimulation treatment plan and modifying parameters of said plan, to further include determining a plurality of probabilities associated with a transition from a first state to a second state in the plurality of states and generating, by a recommendation module based on the plurality of probabilities, a neuromodulator action policy for a patient in the cohort, as disclosed by Bartlett, because this allows for modeling the states as a Markov decision process, in which transitions between states are stochastic and marked by transition probabilities, and therefore allows the relationship between neural firing patterns and applied stimuli to drive neural firing patterns and brain state towards a desired target to be modeled for subsequent input/use in reinforcement learning techniques (See Bartlett Par [0048]-[0049]).
While Choi and Bartlett effectively disclose generating a neuromodulator action policy based on current settings and modifying the device to include a sequence of adjustments to the various neuromodulation/treatment device parameters based on the determined neuromodulator action policy, Choi and Bartlett do not seem to effectively disclose the changing neuromodulation/treatment device parameters specifically relating to a wave shape of the waveform to be delivered by the neuromodulation/treatment device, as given by the following limitation:
the neuromodulator setting comprising waveform parameter, the waveform parameter comprising a wave shape.
However, Annoni discloses the neuromodulator setting comprising waveform parameter, the waveform parameter comprising a wave shape (See Annoni Par [0073] which discloses the use of a neuromodulator, including an implantable neuromodulator control generation and delivery of electrostimulation using the information about electrostimulation parameters and electrode configuration from the learning system; See Annoni Par [0074] which discloses a controller circuit controlling the generation and delivery of electrostimulation in a closed-loop fashion by adaptively adjusting one or more stimulation parameters or stimulation electrode configuration based on the pain score, and the electrostimulations may differ in at least one of the stimulation energy, pulse amplitude, pulse width, stimulation frequency, duration, on-off cycle, pulse shape or waveform).
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 combined disclosure of Choi and Bartlett discloses generating a neuromodulator action policy, including various neuromodulation/treatment device parameters being changed based on the determined neuromodulator action policy to specifically include the neuromodulator setting comprising a waveform parameter, the waveform parameter comprising a wave shape, as disclosed by Annoni, because this allows for more individualized electrostimulation being delivered based on varying pain parameters experienced by the patient (See Annoni Par [0073]-[0074]).
While the combination of Choi, Bartlett and Annoni discloses clustering patients into varying cohorts based on feature vectors and/or classification of the state of a patient to control a neurostimulation therapy, such that spatial locations in the EEG data that correspond to cortical activity data can be analyzed and compared to healthy/control groups, this combination of references does not specifically describe clustering patient data into cohorts according to said implant locations as required by the following bolded portions of the limitation found below:
clustering according to a set of vectors of features corresponding to the first set of patient data, by a clustering module, the patient population into a plurality of cohorts, wherein the at least one cohort is clustered according to an implant location, the implant location being a feature in the set of vector of features.
However, Schulte discloses clustering patient data into cohorts according to said implant locations (See Schulte Par [0139] & [0245] which discloses collecting data from feature extraction engines of a plurality of users or patients, and the aggregation or collection of data can be taken from multiple different patients may be taken from subpopulations of patients with similar characteristics, such as neurostimulation devices distributed across geographic locations, and further discloses in Schulte Par [0140] & [0246] that for example, extracted features can be weighted based on, for example, geospatial data and/or temporal data indicating more or less closeness to a location and/or time specified in health related determination; See Schulte Par [0147] which discloses the use of aggregation learning algorithms and/or the learning algorithm develop rules between features and one or more parameters of one or more response profiles that correspond to therapy outcomes and or treatment parameters, such as those features disclosed in Schulte Par [0139]-[0140] & [0245]-[0246] regarding geographic location of neurostimulation devices).
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 disclosure of Choi, Bartlett, and Annoni, which already discloses clustering patients into varying cohorts based on feature vectors and/or classification of the state of a patient to control a neurostimulation therapy to further include clustering patient data into cohorts according to said implant locations, as disclosed by Schulte, because this allows for developing rules between features and one or more shared parameters of one or more response profiles that correspond to neurostimulation therapy outcomes, such as based on geographic locations of neurostimulation devices (See Schulte Par [0139], [0147], & [0245]).
Claim 11 –
Regarding Claim 11, Choi, Bartlett, Annoni, and Schulte disclose the computer program product of claim 10 in its entirety. Choi and Bartlett further disclose a computer program product, further comprising:
determining a current state of the patient using time-series patient data associated with the patient (See Choi Par [0134]-[0135] which discloses ML/AI processing of the captured data to develop a predictive disorder model, such that the algorithm classifies patients by disorder (e.g. health state, disorder state, disorder level, etc.), such that the predictive disorder model may include various disorder states, i.e. a plurality of states, such as in the case of Parkinson’s Disease, the predictive disorder model may include model components such as tremor, rigidity, facial drop, balance, hallucinations, and/or any other relevant disorder symptoms to create an overall disorder classification, i.e. state/class, that reflects an ensemble of the selected model components); and
determining the neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state (See Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient, albeit not specifically recited for highest probability to transition from one state to another; See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process in which transitions between states are stochastic, marked by transition probabilities, which is the probability of moving from one state to another state by taking an action, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states, and further discloses in Bartlett Par [0050]-[0052] that the policy function can map state-to-action spaces and effectively chooses the actions, such as the stimulus, by evaluating how successful a given action taken by the actor was and how it should adjust to minimize current and future errors, and thereby maximize probability of transitioning and Bartlett further discloses that a reward function fundamentally dictates and quantifies (neuro)stimulation goals, such that Bartlett utilizes a mean-square error loss function and that the function is asymptotically the maximum likelihood estimator, i.e. maximum probability, for the process of transitioning from an observed neural state to a new/desired state based on a given action/process, constituting determining a neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state).
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 disclosure of Choi, Bartlett, Annoni, and Schulte, to further include additional embodiments found in Bartlett regarding determining the neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state, because by choosing actions that maximize said probability to transition from a current state to another, the system can effectively adjust parameters of the (neuro)stimulation action policy to minimize current and future errors of said transition (see Bartlett Par [0050]).
Claim 12 –
Regarding Claim 12, Choi, Bartlett, Annoni, and Schulte disclose the computer program product of claim 11 in its entirety. Choi and Bartlett further disclose a computer program product, further comprising:
wherein the sequence of adjustments have the highest probability to transition from the current state to another state (See Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient, albeit not specifically recited for highest probability to transition from one state to another; See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process in which transitions between states are stochastic, marked by transition probabilities, which is the probability of moving from one state to another state by taking an action, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states, and further discloses in Bartlett Par [0050]-[0052] that the policy function can map state-to-action spaces and effectively chooses the actions, such as the stimulus, by evaluating how successful a given action taken by the actor was and how it should adjust to minimize current and future errors, and thereby maximize probability of transitioning and Bartlett further discloses that a reward function fundamentally dictates and quantifies (neuro)stimulation goals, such that Bartlett utilizes a mean-square error loss function and that the function is asymptotically the maximum likelihood estimator, i.e. maximum probability, for the process of transitioning from an observed neural state to a new/desired state based on a given action/process, constituting determining a neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state).
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 disclosure of Choi, Bartlett, Annoni, and Schulte, to further include additional embodiments found in Bartlett regarding determining the neuromodulator action policy by selecting one or more actions of the neuromodulator associated with a highest probability to transition from the current state to another state, because by choosing actions that maximize said probability to transition from a current state to another, the system can effectively adjust parameters of the (neuro)stimulation action policy to minimize current and future errors of said transition (see Bartlett Par [0050]).
Claim 13 –
Regarding Claim 13, Choi, Bartlett, Annoni, and Schulte disclose the computer program product of claim 10 in its entirety. Choi does not disclose, but Bartlett further discloses a computer program product, further comprising:
generating a Markov decision process using time-series patient data and time-series treatment data associated with the cohort (See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process in which transitions between states are stochastic, marked by transition probabilities, which is the probability of moving from one state to another state by taking an action, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states, and further discloses in Bartlett Par [0050]-[0052] that the policy function can map state-to-action spaces and effectively chooses the actions, such as the stimulus, by evaluating how successful a given action taken by the actor was and how it should adjust to minimize current and future errors, and thereby maximize probability of transitioning and Bartlett further discloses that a reward function fundamentally dictates and quantifies (neuro)stimulation goals, such that Bartlett utilizes a mean-square error loss function and that the function is asymptotically the maximum likelihood estimator, i.e. maximum probability, for the process of transitioning from an observed neural state to a new/desired state based on a given action/process, constituting determining a neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state); and
determining the plurality of probabilities using the Markov decision process (See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process in which transitions between states are stochastic, marked by transition probabilities, which is the probability of moving from one state to another state by taking an action, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states, and further discloses in Bartlett Par [0050]-[0052] that the policy function can map state-to-action spaces and effectively chooses the actions, such as the stimulus, by evaluating how successful a given action taken by the actor was and how it should adjust to minimize current and future errors, and thereby maximize probability of transitioning and Bartlett further discloses that a reward function fundamentally dictates and quantifies (neuro)stimulation goals, such that Bartlett utilizes a mean-square error loss function and that the function is asymptotically the maximum likelihood estimator, i.e. maximum probability, for the process of transitioning from an observed neural state to a new/desired state based on a given action/process, constituting determining a neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state).
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 disclosure of Choi, Bartlett, Annoni, and Schulte, to further include additional embodiments found in Bartlett regarding generating a Markov decision process using time-series patient data and time-series treatment data associated with the cohort and determining the plurality of probabilities using the Markov decision process, because by modeling the states as a Markov decision process, in which transitions between states are stochastic and marked by transition probabilities, this allows the relationship between neural firing patterns and applied stimuli to drive neural firing patterns and brain state towards a desired target to be modeled for subsequent input/use in reinforcement learning techniques (See Bartlett Par [0048]-[0049]).
Claim 14 –
Regarding Claim 14, Choi, Bartlett, Annoni, and Schulte disclose the computer program product of claim 13 in its entirety. Choi and Bartlett further disclose a computer program product, further comprising:
receiving a user feedback on the neuromodulator action policy (See Choi Par [0097] which discloses the system facilitating patient input/feedback with respect to a trial therapy or treatment, by providing an interface where input/feedback can be augmented with one or more data labeling buttons, icons, pictograms, etc., and wherein the patient input/feedback data may be provided to a network-based AI/ML model for facilitating intelligent decision-making with respect to whether the IMD/NIMI device should be deployed in a more permanent manner (e.g., implantation) and/or whether a particular therapy setting or a set of settings, including context-sensitive therapy program selection, may need to be optimized or otherwise reconfigured; See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process for use in reinforcement learning techniques, i.e. the AI/ML model of Choi being modified to be a Markov decision process, as disclosed by Bartlett); and
adjusting the Markov decision process based on the user feedback (See Choi Par [0097] which discloses the system facilitating patient input/feedback with respect to a trial therapy or treatment, by providing an interface where input/feedback can be augmented with one or more data labeling buttons, icons, pictograms, etc., and wherein the patient input/feedback data may be provided to a network-based AI/ML model for facilitating intelligent decision-making with respect to whether the IMD/NIMI device should be deployed in a more permanent manner (e.g., implantation) and/or whether a particular therapy setting or a set of settings, including context-sensitive therapy program selection, may need to be optimized or otherwise reconfigured; See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process for use in reinforcement learning techniques, i.e. the AI/ML model of Choi being modified to be a Markov decision process, as disclosed by Bartlett, and therefore by Choi stating the patient input/feedback data may be provided to a network-based AI/ML model and Bartlett specifying the model as a Markov decision process for use in reinforcement learning techniques, this limitation is considered to be met).
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 disclosure of Choi, Bartlett, Annoni, and Schulte to further include additional embodiments found in Bartlett regarding generating a Markov decision process using time-series patient data and time-series treatment data associated with the cohort and determining the plurality of probabilities using the Markov decision process, because by modeling the states as a Markov decision process, in which transitions between states are stochastic and marked by transition probabilities, this allows the relationship between neural firing patterns and applied stimuli to drive neural firing patterns and brain state towards a desired target to be modeled for subsequent input/use in reinforcement learning techniques (See Bartlett Par [0048]-[0049]).
Claim 15 –
Regarding Claim 15, Choi, Bartlett, Annoni, and Schulte disclose the computer program product of claim 10 in its entirety. Choi further discloses a computer program product, wherein:
a cohort represents a group of patients who share a similarity in at least one of a physiological feature, a symptom diagnosis, and a neuromodulator device type (See Choi Par [0134] which discloses ML/AI processing of the captured data to develop a predictive disorder model, such that the algorithm classifies patients by disorder, i.e. cohorts, and includes classification techniques, such as k-means clustering, which is understood to include clustering and/or a clustering module; See Choi Par [0134]-[0135] which discloses ML/AI processing of the captured data to develop a predictive disorder model, such that the algorithm classifies patients by disorder (e.g. health state, disorder state, disorder level, etc.), such that the predictive disorder model may include various disorder states, i.e. a plurality of states, such as in the case of Parkinson’s Disease, the predictive disorder model may include model components such as tremor, rigidity, facial drop, balance, hallucinations, and/or any other relevant disorder symptoms to create an overall disorder classification, i.e. state/class, that reflects an ensemble of the selected model components).
Claim 16 –
Regarding Claim 16, Choi discloses a computer system comprising a processor and one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions executable by the processor to cause the processor to perform operations (See Choi Par [0052]) comprising:
collecting, by a data collection module (See Choi Par [0049] which discloses the use of one or more electric elements, components or modules, for performing a function if the element may be programmed for performing or otherwise structurally arranged for to perform that action, and is understood to read on each instance of the various “units” recited hereinafter), a first set of patient data and a first set of treatment data associated with a patient population treated with neuromodulation (See Choi Par [0131] which discloses captured data including context data and sensor data, the context data including data relating to neurostimulation therapy settings for each patient and timestamps for when the therapy was provided, i.e. treatment data, and the sensor data including biometric or other types of data captured by varying devices including a “smart” watch, a health monitoring device, or other types of consumer electronic devices, i.e. patient data);
clustering according to a set of vectors of features corresponding to the first set of patient data, by a clustering module, the patient population into a plurality of cohorts, wherein at least one cohort is clustered according to an implant location, the implant location being a feature in the set of vectors of features (See Choi Par [0134] which discloses ML/AI processing of the captured data to develop a predictive disorder model, such that the algorithm classifies patients by disorder, i.e. cohorts, and includes classification techniques, such as k-means clustering, which is understood to include clustering and/or a clustering module; See Choi Par [0196] which discloses features of two population groups, i.e. cohorts, to control neurostimulation; See Choi Par [0194] which discloses classifying being defined for classification of the state of a patient to control a neurostimulation therapy, such that spatial locations in the EEG data that correspond to cortical activity data can be analyzed and compared to healthy/control groups, however not specifically described as clustered according to said implant locations);
generating, by a state discovery module, a plurality of states using a second set of patient data associated with a cohort in the plurality of cohorts (See Choi Par [0134]-[0135] which discloses ML/AI processing of the captured data to develop a predictive disorder model, such that the algorithm classifies patients by disorder (e.g. health state, disorder state, disorder level, etc.), such that the predictive disorder model may include various disorder states, i.e. a plurality of states, such as in the case of Parkinson’s Disease, the predictive disorder model may include model components such as tremor, rigidity, facial drop, balance, hallucinations, and/or any other relevant disorder symptoms to create an overall disorder classification, i.e. state/class, that reflects an ensemble of the selected model components);
generating, by an action discovery module, a plurality of actions using a second set of treatment data associated with the cohort (See Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while each neurostimulation stimulation therapy session (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of neurostimulation parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient, i.e. plurality of actions for each subsequent set/instance of treatment);
generating, by a recommendation module, and based on a current setting of a neuromodulation device corresponding to the patient using time-series treatment data associated with the patient a neuromodulator action policy for a patient in the cohort (BRI of “neuromodulator action policy” in light of Applicant’s Specification is understood to include any course of action recommended for the neuromodulator that is tailored to the patient, therefore see Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient; See Choi Par [0094] & [0129] which discloses the therapy program parameters including one or more sets of stimulation parameters corresponding to different lead/electrode combinations, pulse amplitude, stimulation level, pulse width, pulse frequency or inter-pulse period, pulse repetition parameter (e.g., number of times for a given pulse to be repeated for respective stimulation sets or “stimsets” during the execution of a program), etc., electrode configuration data for delivery of electrical pulses (e.g., as cathodic nodes, anodic nodes, or configured as inactive nodes, etc.), stimulation pattern identification (e.g., tonic stimulation, burst stimulation, noise stimulation, biphasic stimulation, monophasic stimulation, and/or the like), etc.; See Choi Par [0113] which discloses “big data” being received and used by the data analytics platform for purposes of determining neurostimulation parameters, including patients’ health data, physiological/behavioral data, sensor data, daily activity data, therapy settings (i.e. electrode configuration data for delivery of electrical pulses, stimulation pattern identification, etc. as mentioned in Choi Par [0094]), and further discloses in Choi Par [0178] that the system can be used to make small adjustments to programmed parameters to explore the therapeutic state space when patient performs various exercises, and would therefore constitute adjustments from a “current” programmed setting to an adjusted setting; See Choi Par [0131] which specifically mentions timestamps according to the time of day when said neurostimulation therapy and associated settings were applied for each patient), the neuromodulator action policy including a sequence of adjustments to the neuromodulation device (See Choi Par [0131] which specifically mentions timestamps according to the time of day when said neurostimulation therapy and associated settings were applied for each patient; See Choi Par [0133] which discloses that the neurostimulation therapy can include stimulation pulses (by the neurostimulation device) that may be delivered continuously, or can be delivered to target tissue of the patient for a period of time and not delivered during another period of time, constituting a sequence of adjustments to the neurostimulation device; Choi Par [0178] that the system can be used to make small adjustments to programmed parameters over time to titrate the patient’s neurostimulation therapy during performance of tasks or activities and exploring the therapeutic state space when patient performs various exercises, and would therefore constitute adjustments from a “current” programmed setting to an adjusted setting; See Choi Par [0131] which specifically mentions timestamps according to the time of day when said neurostimulation therapy and associated settings were applied for each patient, i.e. also sequence of adjustments from a timestamp aspect; See Choi Par [0228] which discloses that the automated changes, i.e. adjustments, to the neurostimulator device parameters may be implemented to apply neurostimulation according to the modified parameters and additional iterations of the process, i.e. a sequence of adjustments, may be performed if deemed appropriate by the clinician);
transmitting, by the recommendation module, neuromodulator action policy to a neuromodulation device (see Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient, i.e. transmitted to neurostimulator; See Choi Par [0094] & [0129] which discloses the therapy program parameters including one or more sets of stimulation parameters corresponding to different lead/electrode combinations, pulse amplitude, stimulation level, pulse width, pulse frequency or inter-pulse period, pulse repetition parameter (e.g., number of times for a given pulse to be repeated for respective stimulation sets or “stimsets” during the execution of a program), etc., electrode configuration data for delivery of electrical pulses (e.g., as cathodic nodes, anodic nodes, or configured as inactive nodes, etc.), stimulation pattern identification (e.g., tonic stimulation, burst stimulation, noise stimulation, biphasic stimulation, monophasic stimulation, and/or the like), etc.; See Choi Par [0073] which specifically discloses “stimulation therapy settings for secure transmission to the patient device, which may be securely transmitted via encrypted communications”, constituting transmission of the neuromodulator action policy to the patient neuromodulation device); and
adjusting, by the neuromodulation device, a neuromodulator setting of the neuromodulation device, the adjusting the neuromodulation setting based on the neuromodulator action policy (see Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient; See Choi Par [0094] & [0129] which discloses the therapy program parameters including one or more sets of stimulation parameters corresponding to different lead/electrode combinations, pulse amplitude, stimulation level, pulse width, pulse frequency or inter-pulse period, pulse repetition parameter (e.g., number of times for a given pulse to be repeated for respective stimulation sets or “stimsets” during the execution of a program), etc., electrode configuration data for delivery of electrical pulses (e.g., as cathodic nodes, anodic nodes, or configured as inactive nodes, etc.), stimulation pattern identification (e.g., tonic stimulation, burst stimulation, noise stimulation, biphasic stimulation, monophasic stimulation, and/or the like), etc., which is understood to constitute adjustment of a neuromodulator setting).
While Choi discloses collecting patient data from a population, clustering said population into cohorts, generating one or more states and/or actions according to varying patient data from said cohorts, and subsequently implementing said actions in the form of a neurostimulation treatment plan and modifying parameters of said plan based on a current setting, and conducting a sequence of adjustments to the device, as shown above, and Choi specifically discloses the use of Markov decision processes as a classification technique to classify disorders of patients, Choi does not explicitly mention determining a plurality of probabilities associated with a transition from a first state to a second state in the plurality of states, and using said probabilities for generating the neuromodulator action policy, as given by the following bolded portions of the limitations below:
determining, by a decision-making module based on the plurality of actions, a plurality of probabilities associated with a transition from a first state to a second state in the plurality of states; and
generating, by a recommendation module based on the plurality of probabilities, and based on a current setting of a neuromodulation device corresponding to the patient using time-series treatment data associated with the patient a neuromodulator action policy for a patient in the cohort, the neuromodulator action policy including a sequence of adjustments to the neuromodulation device.
Therefore, Bartlett discloses determining, by a decision-making module based on the plurality of actions, a plurality of probabilities associated with a transition from a first state to a second state in the plurality of states (See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process in which transitions between states are stochastic, marked by transition probabilities, which is the probability of moving from one state to another state by taking an action, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states); and generating, by a recommendation module based on the plurality of probabilities, and based on a current setting of a neuromodulation device corresponding to the patient using time-series treatment data associated with the patient a neuromodulator action policy for a patient in the cohort, the neuromodulator action policy including a sequence of adjustments to the neuromodulation device (See Bartlett Par [0048]-[0049] & Fig. 2 which discloses the generation of a policy function for a neuromodulation device/system, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states, and further describes that the Markov decision process modeling can be used to generate state-reward relationships to build up a policy function for said neuromodulation device/system, as shown in FIG. 2, which is used to model mappings between current neural firing patterns, i.e. initial state, and applied stimuli, i.e. action, and series of (neuro)stimulation actions to take to move to desired states).
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 disclosure of Choi, which discloses collecting patient data from a population, clustering said population into cohorts, generating one or more states and/or actions according to varying patient data from said cohorts, and subsequently implementing said actions in the form of a neurostimulation treatment plan and modifying parameters of said plan, to further include determining a plurality of probabilities associated with a transition from a first state to a second state in the plurality of states and generating, by a recommendation module based on the plurality of probabilities, a neuromodulator action policy for a patient in the cohort, as disclosed by Bartlett, because this allows for modeling the states as a Markov decision process, in which transitions between states are stochastic and marked by transition probabilities, and therefore allows the relationship between neural firing patterns and applied stimuli to drive neural firing patterns and brain state towards a desired target to be modeled for subsequent input/use in reinforcement learning techniques (See Bartlett Par [0048]-[0049]).
While Choi and Bartlett effectively disclose generating a neuromodulator action policy based on current settings and modifying the device to include a sequence of adjustments to the various neuromodulation/treatment device parameters based on the determined neuromodulator action policy, Choi and Bartlett do not seem to effectively disclose the changing neuromodulation/treatment device parameters specifically relating to a wave shape of the waveform to be delivered by the neuromodulation/treatment device, as given by the following limitation:
the neuromodulator setting comprising waveform parameter, the waveform parameter comprising a wave shape.
However, Annoni discloses the neuromodulator setting comprising waveform parameter, the waveform parameter comprising a wave shape (See Annoni Par [0073] which discloses the use of a neuromodulator, including an implantable neuromodulator control generation and delivery of electrostimulation using the information about electrostimulation parameters and electrode configuration from the learning system; See Annoni Par [0074] which discloses a controller circuit controlling the generation and delivery of electrostimulation in a closed-loop fashion by adaptively adjusting one or more stimulation parameters or stimulation electrode configuration based on the pain score, and the electrostimulations may differ in at least one of the stimulation energy, pulse amplitude, pulse width, stimulation frequency, duration, on-off cycle, pulse shape or waveform).
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 combined disclosure of Choi and Bartlett discloses generating a neuromodulator action policy, including various neuromodulation/treatment device parameters being changed based on the determined neuromodulator action policy to specifically include the neuromodulator setting comprising a waveform parameter, the waveform parameter comprising a wave shape, as disclosed by Annoni, because this allows for more individualized electrostimulation being delivered based on varying pain parameters experienced by the patient (See Annoni Par [0073]-[0074]).
While the combination of Choi, Bartlett and Annoni discloses clustering patients into varying cohorts based on feature vectors and/or classification of the state of a patient to control a neurostimulation therapy, such that spatial locations in the EEG data that correspond to cortical activity data can be analyzed and compared to healthy/control groups, this combination of references does not specifically describe clustering patient data into cohorts according to said implant locations as required by the following bolded portions of the limitation found below:
clustering according to a set of vectors of features corresponding to the first set of patient data, by a clustering module, the patient population into a plurality of cohorts, wherein the at least one cohort is clustered according to an implant location, the implant location being a feature in the set of vector of features.
However, Schulte discloses clustering patient data into cohorts according to said implant locations (See Schulte Par [0139] & [0245] which discloses collecting data from feature extraction engines of a plurality of users or patients, and the aggregation or collection of data can be taken from multiple different patients may be taken from subpopulations of patients with similar characteristics, such as neurostimulation devices distributed across geographic locations, and further discloses in Schulte Par [0140] & [0246] that for example, extracted features can be weighted based on, for example, geospatial data and/or temporal data indicating more or less closeness to a location and/or time specified in health related determination; See Schulte Par [0147] which discloses the use of aggregation learning algorithms and/or the learning algorithm develop rules between features and one or more parameters of one or more response profiles that correspond to therapy outcomes and or treatment parameters, such as those features disclosed in Schulte Par [0139]-[0140] & [0245]-[0246] regarding geographic location of neurostimulation devices).
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 disclosure of Choi, Bartlett, and Annoni, which already discloses clustering patients into varying cohorts based on feature vectors and/or classification of the state of a patient to control a neurostimulation therapy to further include clustering patient data into cohorts according to said implant locations, as disclosed by Schulte, because this allows for developing rules between features and one or more shared parameters of one or more response profiles that correspond to neurostimulation therapy outcomes, such as based on geographic locations of neurostimulation devices (See Schulte Par [0139], [0147], & [0245]).
Claim 17 –
Regarding Claim 17, Choi, Bartlett, Annoni, and Schulte disclose the computer system of claim 16 in its entirety. Choi and Bartlett further disclose a computer program product, further comprising:
determining a current state of the patient using time-series patient data associated with the patient (See Choi Par [0134]-[0135] which discloses ML/AI processing of the captured data to develop a predictive disorder model, such that the algorithm classifies patients by disorder (e.g. health state, disorder state, disorder level, etc.), such that the predictive disorder model may include various disorder states, i.e. a plurality of states, such as in the case of Parkinson’s Disease, the predictive disorder model may include model components such as tremor, rigidity, facial drop, balance, hallucinations, and/or any other relevant disorder symptoms to create an overall disorder classification, i.e. state/class, that reflects an ensemble of the selected model components); and
determining the neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state (See Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient, albeit not specifically recited for highest probability to transition from one state to another; See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process in which transitions between states are stochastic, marked by transition probabilities, which is the probability of moving from one state to another state by taking an action, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states, and further discloses in Bartlett Par [0050]-[0052] that the policy function can map state-to-action spaces and effectively chooses the actions, such as the stimulus, by evaluating how successful a given action taken by the actor was and how it should adjust to minimize current and future errors, and thereby maximize probability of transitioning and Bartlett further discloses that a reward function fundamentally dictates and quantifies (neuro)stimulation goals, such that Bartlett utilizes a mean-square error loss function and that the function is asymptotically the maximum likelihood estimator, i.e. maximum probability, for the process of transitioning from an observed neural state to a new/desired state based on a given action/process, constituting determining a neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state).
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 disclosure of Choi, Bartlett, Annoni, and Schulte to further include additional embodiments found in Bartlett regarding determining the neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state, because by choosing actions that maximize said probability to transition from a current state to another, the system can effectively adjust parameters of the (neuro)stimulation action policy to minimize current and future errors of said transition (see Bartlett Par [0050]).
Claim 18 –
Regarding Claim 18, Choi, Bartlett, Annoni, and Schulte disclose the computer system of claim 17 in its entirety. Choi and Bartlett further disclose a computer program product, further comprising:
wherein the sequence of adjustments have the highest probability to transition from the current state to another state (See Choi Par [0135]-[0136] which discloses the outputs of the predictive disorder model being used to evaluate the condition of a patient’s disorder or condition based on data captured while the neurostimulation stimulation therapy (or other type of therapy) is or is not being provided, such that the outputs can be utilized to assist with programming of therapy, e.g. neurostimulation, parameters (e.g., adjust the parameters, conditions for triggering delivery of neurostimulation, etc.) for said patient, albeit not specifically recited for highest probability to transition from one state to another; See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process in which transitions between states are stochastic, marked by transition probabilities, which is the probability of moving from one state to another state by taking an action, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states, and further discloses in Bartlett Par [0050]-[0052] that the policy function can map state-to-action spaces and effectively chooses the actions, such as the stimulus, by evaluating how successful a given action taken by the actor was and how it should adjust to minimize current and future errors, and thereby maximize probability of transitioning and Bartlett further discloses that a reward function fundamentally dictates and quantifies (neuro)stimulation goals, such that Bartlett utilizes a mean-square error loss function and that the function is asymptotically the maximum likelihood estimator, i.e. maximum probability, for the process of transitioning from an observed neural state to a new/desired state based on a given action/process, constituting determining a neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state).
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 disclosure of Choi, Bartlett, Annoni, and Schulte, to further include additional embodiments found in Bartlett regarding determining the neuromodulator action policy by selecting one or more actions of the neuromodulator associated with a highest probability to transition from the current state to another state, because by choosing actions that maximize said probability to transition from a current state to another, the system can effectively adjust parameters of the (neuro)stimulation action policy to minimize current and future errors of said transition (see Bartlett Par [0050]).
Claim 19 –
Regarding Claim 19, Choi, Bartlett, Annoni, and Schulte disclose the computer system of claim 16 in its entirety. Choi does not disclose, but Bartlett further discloses a computer program product, further comprising:
generating a Markov decision process using time-series patient data and time-series treatment data associated with the cohort (See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process in which transitions between states are stochastic, marked by transition probabilities, which is the probability of moving from one state to another state by taking an action, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states, and further discloses in Bartlett Par [0050]-[0052] that the policy function can map state-to-action spaces and effectively chooses the actions, such as the stimulus, by evaluating how successful a given action taken by the actor was and how it should adjust to minimize current and future errors, and thereby maximize probability of transitioning and Bartlett further discloses that a reward function fundamentally dictates and quantifies (neuro)stimulation goals, such that Bartlett utilizes a mean-square error loss function and that the function is asymptotically the maximum likelihood estimator, i.e. maximum probability, for the process of transitioning from an observed neural state to a new/desired state based on a given action/process, constituting determining a neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state); and
determining the plurality of probabilities using the Markov decision process (See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process in which transitions between states are stochastic, marked by transition probabilities, which is the probability of moving from one state to another state by taking an action, such that the processor quantifies the value of neural activity deficiency remaining in the current state versus transitioning to a new state which is dependent on measured error between current state and the desired state and current and past rewards for transitioning between states, and further discloses in Bartlett Par [0050]-[0052] that the policy function can map state-to-action spaces and effectively chooses the actions, such as the stimulus, by evaluating how successful a given action taken by the actor was and how it should adjust to minimize current and future errors, and thereby maximize probability of transitioning and Bartlett further discloses that a reward function fundamentally dictates and quantifies (neuro)stimulation goals, such that Bartlett utilizes a mean-square error loss function and that the function is asymptotically the maximum likelihood estimator, i.e. maximum probability, for the process of transitioning from an observed neural state to a new/desired state based on a given action/process, constituting determining a neuromodulator action policy by selecting one or more actions with a highest probability to transition from the current state to another state).
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 disclosure of Choi, Bartlett, Annoni, and Schulte to further include additional embodiments found in Bartlett regarding generating a Markov decision process using time-series patient data and time-series treatment data associated with the cohort and determining the plurality of probabilities using the Markov decision process, because by modeling the states as a Markov decision process, in which transitions between states are stochastic and marked by transition probabilities, this allows the relationship between neural firing patterns and applied stimuli to drive neural firing patterns and brain state towards a desired target to be modeled for subsequent input/use in reinforcement learning techniques (See Bartlett Par [0048]-[0049]).
Claim 20 –
Regarding Claim 20, Choi, Bartlett, Annoni, and Schulte disclose the computer system of claim 19 in its entirety. Choi and Bartlett further disclose a computer program product, further comprising:
receiving a user feedback on the neuromodulator action policy (See Choi Par [0097] which discloses the system facilitating patient input/feedback with respect to a trial therapy or treatment, by providing an interface where input/feedback can be augmented with one or more data labeling buttons, icons, pictograms, etc., and wherein the patient input/feedback data may be provided to a network-based AI/ML model for facilitating intelligent decision-making with respect to whether the IMD/NIMI device should be deployed in a more permanent manner (e.g., implantation) and/or whether a particular therapy setting or a set of settings, including context-sensitive therapy program selection, may need to be optimized or otherwise reconfigured; See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process for use in reinforcement learning techniques, i.e. the AI/ML model of Choi being modified to be a Markov decision process, as disclosed by Bartlett); and
adjusting the Markov decision process based on the user feedback (See Choi Par [0097] which discloses the system facilitating patient input/feedback with respect to a trial therapy or treatment, by providing an interface where input/feedback can be augmented with one or more data labeling buttons, icons, pictograms, etc., and wherein the patient input/feedback data may be provided to a network-based AI/ML model for facilitating intelligent decision-making with respect to whether the IMD/NIMI device should be deployed in a more permanent manner (e.g., implantation) and/or whether a particular therapy setting or a set of settings, including context-sensitive therapy program selection, may need to be optimized or otherwise reconfigured; See Bartlett Par [0048]-[0049] & Fig. 2 which disclose the states of the patients being modeled as a Markov decision process for use in reinforcement learning techniques, i.e. the AI/ML model of Choi being modified to be a Markov decision process, as disclosed by Bartlett, and therefore by Choi stating the patient input/feedback data may be provided to a network-based AI/ML model and Bartlett specifying the model as a Markov decision process for use in reinforcement learning techniques, this limitation is considered to be met).
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 disclosure of Choi, Bartlett, Annoni, and Schulte, to further include additional embodiments found in Bartlett regarding generating a Markov decision process using time-series patient data and time-series treatment data associated with the cohort and determining the plurality of probabilities using the Markov decision process, because by modeling the states as a Markov decision process, in which transitions between states are stochastic and marked by transition probabilities, this allows the relationship between neural firing patterns and applied stimuli to drive neural firing patterns and brain state towards a desired target to be modeled for subsequent input/use in reinforcement learning techniques (See Bartlett Par [0048]-[0049]).
Response to Arguments
Applicant's arguments filed 17 February 2026 have been fully considered but they are not persuasive:
Regarding 35 U.S.C. 103 rejections of claims 1-20, Applicant argues on p. 9-11 of Arguments/Remarks that Choi, Bartlett, and/or Annoni fail to disclose each and every element of independent claims 1, 10, & 16. More specifically, Applicant argues that Choi, Bartlett, and Annoni do not disclose the newly amended portions regarding “clustering according to a set of vectors of features corresponding to the first set of patient data, by a clustering module, the patient population into a plurality of cohorts, wherein at least one cohort is clustered according to an implant location, the implant location being a feature in the set of vectors of features” as found in independent claims 1, 10, & 16 (see p. 26 of Arguments/Remarks). Examiner agrees with Applicant’s arguments. Therefore, the rejections have been withdrawn. However, upon further consideration, a new ground of rejection is made under 35 U.S.C. 103 over Choi, in view of Bartlett, in view of Annoni, further in view of Schulte. Newly cited portions of Choi (Par [0194] & [0196]) disclose features of two population groups, i.e. cohorts, to control neurostimulation and classifying being defined for classification of the state of a patient to control a neurostimulation therapy, such that spatial locations in the EEG data that correspond to cortical activity data can be analyzed and compared to healthy/control groups, however not specifically described as clustered according to said implant locations. As such, Choi, Bartlett, and Annoni did not disclose the entirety of the limitation regarding clustering according to implant locations. However, Schulte is herein applied to cover said deficiencies, and read on clustering patient data according to said implant locations in one or more feature vectors. Therefore, the argued limitations are still met by Choi, Bartlett, Annoni, and Schulte, and as such, independent claims 1, 10, & 16 and claims dependent therefrom remain rejected under 35 U.S.C. 103 over Choi, in view of Bartlett, further in view of Annoni.
Regarding 35 U.S.C. 103 rejections of claims 1-20, Applicant argues on p. 11 of Arguments/Remarks that because independent claims 1, 10, & 16 are purportedly allowable over the prior art, dependent claims 2-9, 11-15, & 17-20, which depend from independent claims 1, 10, & 16 are also allowable over the prior art, by virtue of dependency. Examiner respectfully disagrees with Applicant’s arguments. As explained above, independent claims 1, 10, & 16 remain rejected under 35 U.S.C. 103 over Choi, in view of Bartlett, in view of Annoni, further in view of Schulte. Therefore, Applicant’s arguments in view of independent claims 1, 10, & 16 being purportedly allowable are rendered moot. As such, dependent claims 2-9, 11-15, & 17-20, which depend from independent claims 1, 10, & 16 also remain rejected under 35 U.S.C. 103 over Choi, in view of Bartlett, in view of Annoni, further in view of Schulte.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Mohammed et al. (NPL – “A Framework for Adapting Deep Brain Stimulation Using Parkinsonian State Estimates” – May 2020) disclose an approach that uses generative and discriminative machine learning models to more accurately estimate the symptom severity of patients and adjust deep brain stimulation therapy accordingly, such as by using a support vector machine;
Arcot Desai et al. (U.S. Patent Publication No. 2019/0117978) discloses a system for clinical decision making for a patient using data from the patient and data from other patients, and more particularly, to systems and methods that apply deep learning algorithms to data corresponding to electrical activity of the patient's brain, to identify information relevant to making clinical decisions for the patient, such as for neuromodulation therapy purposes;
Donovan et al. (U.S. Patent Publication No. 2023/0394668) discloses a system for determining likelihood of a patient favorably responding to a neuromodulation procedure based on a quantitative or objective score or determination based on a plurality of indicators of pain.
Applicant's amendment necessitated the new ground of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to HUNTER J RASNIC whose telephone number is (571)270-5801. The examiner can normally be reached M-F 8am-5:30pm.
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, Shahid Merchant can be reached on (571) 270-1360. 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.
/H.R./Examiner, Art Unit 3684
/Shahid Merchant/Supervisory Patent Examiner, Art Unit 3684
Prosecution Insights