DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
Claims 1-19 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claims 1-19 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 1 recites:
(1) “blocking adverse influence of nonlinearity and non-stationarity of the brain signal to the [Kalman filter]”
The phrase “blocking adverse influence” is indefinite because it is a subjective, result-oriented limitation that does not provide an objective boundary for the scope of the claim. The claim does not specify what constitutes an “adverse influence” (e.g., instability, estimation error, divergence, variance increase, bias, overshoot), nor does it specify any measurable threshold or condition establishing when such influence is “blocked.” Absent an objective standard, a person of ordinary skill in the art cannot determine with reasonable certainty whether a given system “blocks adverse influence” as required. See MPEP § 2173.05(b) (relative/subjective terms); MPEP § 2173.05(g) (functional or result-oriented language).
(2) “utilizing the [Kalman filter] to provide smooth generation of the control signal”
The phrase “smooth generation” is indefinite as a term of degree without an accompanying objective measure, definition, or boundary in the claim. “Smooth” may refer to, for example, reduced variance, reduced jerk, bounded acceleration, low-pass characteristics, bandwidth limitation, or reduced oscillation; however, the claim provides no metric (e.g., spectral bound, derivative bound, variance threshold) to determine what level or type of smoothness is required. Accordingly, the scope of claim 1 is not reasonably certain. See MPEP § 2173.05(e) (terms of degree); MPEP § 2173.05(b).
(3) “allowing the nonlinear mapping to be adaptively and continuously updated to follow nonlinearity and non-stationarity of the brain signal”
The limitation “adaptively and continuously updated to follow nonlinearity and non-stationarity” is indefinite because it recites an intended result without specifying an objective manner of satisfying the limitation. The claim does not define what it means to “follow” the nonlinearity/0non-stationarity (e.g., minimize a specified error measure, track a specified drift model, maintain performance above a specified threshold), nor does it define “continuously” (e.g., update every sample, periodically, upon detecting drift, within a defined update interval). Without an objective criterion, the boundaries of the claim remain unclear. See MPEP § 2173.05(b) (subjective/relative terminology); MPEP § 2173.05(g) (purely functional/result-oriented language).
Claim 2 recites determining probabilities that a candidate movement action “is the movement action as intended by the brain signal.” While the claims generally relate to controlling a machine/prosthetic from brain signals, the phrase “intended” introduces a subjective mental state and an unclaimed ground truth. The claim does not specify any objective mechanism by which the “intended” prosthetic movement is determined (e.g., a presented target/cue, user confirmation input, measured kinematics of an intended limb trajectory, task success signal, or other external label). Consequently, a person of ordinary skill in the art cannot determine, with reasonable certainty, when a candidate action satisfies “as intended,” nor when the probability/score requirement is met. The scope of claim 2 is therefore not reasonably certain.
Claim 3 recites “determining a reward … according to whether or not the winner is actually the movement action as intended” and further recites “computing the plurality of weights according to at least the reward.” This language is indefinite because the claim does not provide an objective standard for determining whether a selected action is “actually” the user’s “intended” movement action. The claim does not specify how intention is established or verified (e.g., by an external cue/target, user confirmation, measured kinematics, task success, or other objectively determinable criterion). Absent such an objective mechanism, a person of ordinary skill in the art cannot determine with reasonable certainty when the reward condition is satisfied and, thus, cannot ascertain the scope of the claim. See MPEP § 2173.05(b), (g).
Claim 5 defines a binary reward signal (rt = 1 or 0), but the condition for setting rt still depends on whether the selected candidate movement action “is the movement action as intended by the brain signal.” Because the claim does not recite an objective standard or mechanism for determining “intended” movement, the binary formulation does not cure the indefiniteness. A person of ordinary skill in the art cannot determine with reasonable certainty when rt must be 1 versus 0. See MPEP § 2173.05(b), (g).
Allowable Subject Matter
Claims 4-7, and 14-17 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims, and if the 112 rejections were overcome.
Claims 4–7 and 14–17 are allowable over the prior art of record because the prior art, either alone or in combination, fails to teach or suggest the particular, expressly-recited mathematical formulations and update rules required by these claims.
Claim 4 requires computing, from the transformed brain signal components, a plurality of probabilities for candidate movement actions using the expressly-recited exponential normalization (softmax-type) relationship with a controlling parameter (α), and selecting a winning candidate movement action based on the computed probabilities (i.e., selection of k* from among the N probability values). The prior art of record does not disclose or render obvious this specific probability computation formula as recited, nor the resulting winner-selection constrained by that formula in the claimed brain-signal-to-movement-action context. Claim 14 incorporates these limitations via execution of the method of claim 4 and is allowable for the same reasons.
Claim 5 further limits the method by explicitly defining the reward due to selecting the k*th candidate movement action as rt = 1 if the selected candidate is the movement action as intended by the brain signal, and rt = 0 otherwise. The prior art of record does not teach or suggest this particular reward-definition constraint in combination with the claimed probability-based candidate selection framework of claims 2–4 as claimed.
Claim 6 recites a specific three-layer neural network architecture (input layer with Dz nodes, hidden layer with J hidden units, output layer with N nodes) and further requires computing and updating the plurality of weights according to the expressly-recited update relationships for vjk and wij, including (i) a defined error term δ based on the reward and the computed probability associated with the selected winner (Pat = k*), (ii) an error-expansive function f(δ), and (iii) a specified form for the hidden-unit value hj (logistic-type function of the weighted input). The prior art of record does not disclose or render obvious these particular, explicitly-recited weight update equations and associated variables/functions as claimed for reinforcement-learning based adaptation in the claimed brain-machine interface pipeline. Claim 16 incorporates these limitations via execution of the method of claim 6 and is allowable for the same reasons.
Claims 14–16 are directed to systems comprising a sensing device for capturing a brain signal, a machine for performing the movement action, and a computer configured to execute the corresponding methods of claims 14–16, respectively. Because the prior art of record fails to teach or suggest the specific computational constraints recited in claims 4–6 (and incorporated into the corresponding system claims 14–16), and further fails to render obvious the claimed configuration of the system to carry out these particular computations and updates in the recited brain-signal-driven control context, claims 14–16 are allowable over the prior art of record under 35 U.S.C. 103.
Claim Rejections - 35 USC § 103
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.
Claims 1-3, 8, 10-13, 18, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over US Pre-Grant 2019/0025917 (Francis et al; Francis) in view of “Bayesian Population Decoding of Motor Cortical Activity Using a Kalman Filter,” Neural Computation, January 2006, 18(1):80-118, Wu et al; Wu.
Regarding claim 1:
Francis teaches:
1. A computer-implemented method for controlling a machine to perform a movement action determined by a brain signal of a subject, the method comprising:
(Francis, ¶0106)
“According to some embodiments, methods and systems of the subject technology can improve a reinforcement learning agent of a BMI. Motor signals from a motor cortex of the subject's brain can be received to determine and intended action to be taken by an external device [i.e. A computer-implemented method for controlling a machine to perform a movement action determined by a brain signal of a subject, the method comprising:].”
2. processing the brain signal with a neural network (NN) for applying a nonlinear mapping defined by a plurality of weights of the NN to the brain signal to thereby yield a transformed brain signal;
(Francis, ¶0137)
“The firing rates of the simulated M1 in a given trial were provided as input to the Multi-Layer Perceptron (MLP). MLP had one hidden layer with 120 units wherein the output of each unit is a nonlinear function (tan h) of the weighted inputs. MLP consisted of 8 outputs whose outputs were the state-action value for each available action [i.e. processing the brain signal with a neural network (NN) for applying a nonlinear mapping by a plurality of weights of the NN to the brain signal].”
(Francis, ¶0138)
“Using the output of the simulated neural ensemble as the input to an artificial neural network the Q value for each potential action was determined. Specifically, a multilayer perceptron (MLP) with a single hidden layer consisting of, for example, 120 units was used to calculate the Q value given an input from the neural ensemble [i.e. to thereby yield a transformed brain signal;].”
3. updating the plurality of weights by a reinforcement learning (RL) process such that the NN learns the nonlinear mapping by RL, allowing the nonlinear mapping to be adaptively and continuously updated to follow nonlinearity and non-stationarity of the brain signal;
(Francis, ¶0137)
“The firing rates of the simulated M1 in a given trial were provided as input to the Multi-Layer Perceptron (MLP). MLP had one hidden layer with 120 units wherein the output of each unit is a nonlinear function (tan h) of the weighted inputs. MLP consisted of 8 outputs whose outputs were the state-action value for each available action. An action was executed by the RL agent based on the ε-greedy policy resulting in either a correct (reward=1) or an incorrect movement (reward=−1) of the cursor [i.e. [i.e. updating the plurality of weights by a reinforcement learning (RL) process such that the NN learns the nonlinear mapping by RL].” The temporal difference error, which utilizes the scalar reward value, is used to update the weights of the MLP through backpropagation [i.e. allowing the nonlinear mapping to be adaptively and continuously updated to follow nonlinearity and non-stationarity of the brain signal;].”
Francis does not teach:
1. and processing the transformed brain signal with a Kalman filter (KF) to yield a control signal for controlling the machine to perform the movement action, thereby utilizing the KF to provide smooth generation of the control signal while blocking adverse influence of nonlinearity and non-stationarity of the brain signal to the KF in generating the control signal.
Wu teaches:
1. and processing the transformed brain signal with a Kalman filter (KF) to yield a control signal for controlling the machine to perform the movement action, thereby utilizing the KF to provide smooth generation of the control signal while blocking adverse influence of nonlinearity and non-stationarity of the brain signal to the KF in generating the control signal.
(Wu, pg. 86, Sect. 2.2)
“To that end, we developed a Kalman filtering method (Gelb, 1974; Welch & Bishop, 2001) that provides a rigorous and well-understood framework that addresses these issues. This approach provides a control-theoretic model for the encoding of hand movement in motor cortex and for inferring, or decoding, this movement from the firing rates of a population of neurons [i.e. and processing the transformed brain signal with a Kalman filter (KF) to yield a control signal for controlling the machine to perform the movement action,].”
(Wu, pg. 82)
“The Kalman filter has a number of desirable properties for motor cortical decoding. The inclusion of prior information about the system state enables an efficient recursive formulation of the decoding algorithm and effectively smooths noisy estimates in a mathematically principled way; this is particularly important for decoding complex, natural hand motions required for neural motor prostheses [i.e. thereby utilizing the KF to provide smooth generation of the control signal while blocking adverse influence of nonlinearity and non-stationarity of the brain signal to the KF in generating the control signal].”
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to combine Francis and Wu because Francis teaches using brain-derived evaluative (reward-expectation) feedback to autonomously update a BMI policy in novel/unstable, drifting neural environments, while Wu teaches a well-established Kalman-filter decoder that produces continuous real-time control signals for movement. Using Wu’s Kalman decoder as the actor/policy implementation in Francis’s actor-critic RL-BMI predictably yields stable continuous control with reduced recalibration burden by updating decoder/policy parameters when performance degrades. The combination is a routine integration of a standard continuous state estimator/decoder with an online reinforcement-learning adaptation signal to improve robustness under nonstationary neural inputs, as “The Kalman filter has a number of desirable properties for motor cortical decoding (Wu, pg. 82).”
Regarding claim 2:
Francis and Wu teach the method of claim 1.
Francis teaches:
1. equating the transformed brain signal to be a plurality of scores respectively associated with a plurality of candidate movement actions competing to be the movement action,
(Francis, ¶0137)
“MLP consisted of 8 outputs whose outputs were the state-action value for each available action [i.e. equating the transformed brain signal to be a plurality of scores respectively associated with a plurality of candidate movement actions].”
(Francis, ¶0138)
“Using the output of the simulated neural ensemble as the input to an artificial neural network the Q value for each potential action was determined. Specifically, a multilayer perceptron (MLP) with a single hidden layer consisting of, for example, 120 units was used to calculate the Q value given an input from the neural ensemble. 99% of the time the action with the highest Q value was executed (the “greedy” part of the e-greedy policy), and the other 1% of the time a random action was taken (the exploratory rate, the ‘ε’ part of the ε-greedy policy) [i.e. competing to be the movement action,].”
2. an individual score of a respective candidate movement action being indicative to a probability that the respective candidate movement action is the movement action as intended by the brain signal;
(Francis, ¶0137)
“MLP consisted of 8 outputs whose outputs were the state-action value for each available action [i.e. an individual score of a respective candidate movement action].”
(Francis, ¶0138)
“Using the output of the simulated neural ensemble as the input to an artificial neural network the Q value for each potential action was determined. Specifically, a multilayer perceptron (MLP) with a single hidden layer consisting of, for example, 120 units was used to calculate the Q value given an input from the neural ensemble [i.e. being indicative to a probability that the respective candidate movement action is the movement action as intended by the brain signal;].”
3. computing the plurality of weights according to at least the plurality of scores;
(Francis, ¶0138)
“The temporal difference error, which utilizes the scalar reward value, is used to update the weights of the MLP through backpropagation.”
4. and updating the NN with the computed plurality of weights for configuring the nonlinear mapping, whereby the equating of the transformed brain signal to be the plurality of scores in computing the plurality of weights
(Francis, ¶0137)
“MLP had one hidden layer with 120 units wherein the output of each unit is a nonlinear function (tan h) of the weighted inputs [i.e. and updating the NN with the computed plurality of weights for configuring the nonlinear mapping,]. MLP consisted of 8 outputs whose outputs were the state-action value for each available action. An action was executed by the RL agent based on the ε-greedy policy resulting in either a correct (reward=1) or an incorrect movement (reward=−1) of the cursor. The temporal difference error, which utilizes the scalar reward value, is used to update the weights of the MLP through backpropagation [i.e. whereby the equating of the transformed brain signal to be the plurality of scores in computing the plurality of weights].”
5. and updating the NN with the computed plurality of weights guides the nonlinear mapping to follow nonlinearity and non-stationarity of the brain signal while allowing RL to be applied to NN learning.
(Francis, ¶0138)
“Exploratory rate, defined as the percentage of steps in which an action is executed randomly irrespective of its optimality at a given state, was set at 1% (‘ε’ part of 6-greedy policy). The random exploration allows for discovery of new solutions by the RL agent, useful especially in an altering environment.”
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to modify Francis with Wu. The motivation is the same as claim 1.
Regarding claim 3:
Francis and Wu teach the method of claim 1.
Francis teaches:
1. computing, from the plurality of scores, a plurality of probabilities associated with the plurality of candidate movement actions,
(Francis, ¶0135)
“The action with the highest Q value is selected 1−ε percent of the time (exploitation) whereas a random action is performed c percent of the time (exploration) under the ε-greedy policy. There are also ways to change c given the systems performance as appreciated by persons skilled in the art [i.e. computing, from the plurality of scores, a plurality of probabilities associated with the plurality of candidate movement actions,].”
2. wherein an individual probability associated with the respective candidate movement action is the probability that the respective candidate movement action is the movement action as intended by the brain signal;
(Francis, ¶0106)
“Command signals can be provided to the external device based on (1) the motor signals and (2) a policy of the reinforcement learning agent. As used herein, a policy can refer to one or more operating parameters of an RL-BMI architecture that governs how detected motor signals are translated into action by a device [i.e. wherein an individual probability associated with the respective candidate movement action is the probability that the respective candidate movement action is the movement action as intended by the brain signal;].”
3. selecting, from the plurality of candidate movement actions, a winner in competing to be the movement action as intended according to the plurality of probabilities; determining a reward due to selecting the winner as the movement action according to whether or not the winner is actually the movement action as intended;
(Francis, ¶0135)
“Specifically, we use an ε-greedy policy as the actor and the Q learning paradigm, augmented with Eligibility Trace Q(λ), as the actor's update rule. An eligibility trace is extremely useful in dealing with the credit assignment problem. The action with the highest Q value is selected 1−ε percent of the time (exploitation) whereas a random action is performed c percent of the time (exploration) under the ε-greedy policy [i.e. selecting, from the plurality of candidate movement actions, a winner in competing to be the movement action as intended according to the plurality of probabilities;].”
(Francis, ¶0137)
“An action was executed by the RL agent based on the ε-greedy policy resulting in either a correct (reward=1) or an incorrect movement (reward=−1) of the cursor [i.e. determining a reward due to selecting the winner as the movement action according to whether or not the winner is actually the movement action as intended;].”
4. and computing the plurality of weights according to at least the reward.
(Francis, ¶0137)
“The temporal difference error, which utilizes the scalar reward value, is used to update the weights of the MLP through backpropagation.”
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to modify Francis with Wu. The motivation is the same as claim 1.
Regarding claim 8:
Francis and Wu teach the method of claim 1.
Francis teaches:
1. A system for capturing a brain signal of a subject and performing a movement action determined by the brain signal, the system comprising: a sensing device for capturing the brain signal from the subject;
(Francis, ¶0120)
“96 channel platinum microelectrode arrays (for example, from Blackrock Microsystems, Salt Lake City, Utah) were implanted in the contralateral and ipsilateral primary motor cortex (M1) of Monkey A and Monkey Z respectively [i.e. A system for capturing a brain signal of a subject and performing a movement action determined by the brain signal, the system comprising: a sensing device for capturing the brain signal from the subject;].”
2. a machine for performing the movement action;
(Francis, ¶0106)
“Motor signals from a motor cortex of the subject's brain can be received to determine and intended action to be taken by an external device [i.e. a machine for performing the movement action;] Command signals can be provided to the external device based on (1) the motor signals and (2) a policy of the reinforcement learning agent.”
3. and a computer configured to execute a computing process of processing the brain signal to determine the movement action and controlling the machine to perform the movement action according to the method of claim 1.
(Francis, ¶0110)
“The environment 100 includes an actor 102 (in some embodiments also referred to as an RL-BMI agent or a BMI agent), a critic 104, a brain 106, a target 108, a user 110, and an end effector 112. The components included in the environment 100 can be further broken down into more than one component and/or combined together in any suitable arrangement. Further, one or more components can be rearranged, changed, added, and/or removed. The actor 102 receives motor signal from the brain 106, and the critic 104 receives reward signal from the brain 106 [i.e. and a computer configured to execute a computing process of processing the brain signal to determine the movement action]. The RL-BMI agent (actor) 102 can be viewed as trying to map the neural activity in the brain 106 pertaining to the intended action to an appropriate action of the external actuator (end effector) 112 [i.e. and controlling the machine to perform the movement action according to the method of claim 1].”
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to modify Francis with Wu. The motivation is the same as claim 1.
Regarding claim 10:
Francis and Wu teach the method of claim 1.
Francis teaches:
1. wherein the machine is a prosthesis.
“The end effector 112 can include, without limitations, computer cursor, robotic arm, virtual arm, prosthetic limb, or speech generation device [i.e. wherein the machine is a prosthesis].”
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to modify Francis with Wu. The motivation is the same as claim 1.
Regarding claim 11:
Francis and Wu teach the method of claim 1.
Francis teaches:
1. wherein the machine is a second computer configured to generate the movement action on a virtual object for virtual-reality applications.
“As illustrated below in connections with FIGS. 6-8, reward expectation was indicated either via the color of the target or via the trajectory of the feedback cursor toward or away from the target [i.e. wherein the machine is a second computer configured to generate the movement action on a virtual object for virtual-reality applications].”
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to modify Francis with Wu. The motivation is the same as claim 1.
Regarding claim 12:
Francis and Wu teach the method of claim 1.
Francis teaches:
1. A system for capturing a brain signal of a subject and performing a movement action determined by the brain signal, the system comprising: a sensing device for capturing the brain signal from the subject;
(Francis, ¶0120)
“96 channel platinum microelectrode arrays (for example, from Blackrock Microsystems, Salt Lake City, Utah) were implanted in the contralateral and ipsilateral primary motor cortex (M1) of Monkey A and Monkey Z respectively [i.e. A system for capturing a brain signal of a subject and performing a movement action determined by the brain signal, the system comprising: a sensing device for capturing the brain signal from the subject;].”
2. a machine for performing the movement action;
(Francis, ¶0106)
“Motor signals from a motor cortex of the subject's brain can be received to determine and intended action to be taken by an external device [i.e. a machine for performing the movement action;] Command signals can be provided to the external device based on (1) the motor signals and (2) a policy of the reinforcement learning agent.”
3. and a computer configured to execute a computing process of processing the brain signal to determine the movement action and controlling the machine to perform the movement action according to the method of claim 2.
(Francis, ¶0110)
“The environment 100 includes an actor 102 (in some embodiments also referred to as an RL-BMI agent or a BMI agent), a critic 104, a brain 106, a target 108, a user 110, and an end effector 112. The components included in the environment 100 can be further broken down into more than one component and/or combined together in any suitable arrangement. Further, one or more components can be rearranged, changed, added, and/or removed. The actor 102 receives motor signal from the brain 106, and the critic 104 receives reward signal from the brain 106 [i.e. and a computer configured to execute a computing process of processing the brain signal to determine the movement action]. The RL-BMI agent (actor) 102 can be viewed as trying to map the neural activity in the brain 106 pertaining to the intended action to an appropriate action of the external actuator (end effector) 112 [i.e. and controlling the machine to perform the movement action according to the method of claim 2].”
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to modify Francis with Wu. The motivation is the same as claim 1.
Regarding claim 13:
Francis and Wu teach the method of claim 1.
Francis teaches:
1. A system for capturing a brain signal of a subject and performing a movement action determined by the brain signal, the system comprising: a sensing device for capturing the brain signal from the subject;
(Francis, ¶0120)
“96 channel platinum microelectrode arrays (for example, from Blackrock Microsystems, Salt Lake City, Utah) were implanted in the contralateral and ipsilateral primary motor cortex (M1) of Monkey A and Monkey Z respectively [i.e. A system for capturing a brain signal of a subject and performing a movement action determined by the brain signal, the system comprising: a sensing device for capturing the brain signal from the subject;].”
2. a machine for performing the movement action;
(Francis, ¶0106)
“Motor signals from a motor cortex of the subject's brain can be received to determine and intended action to be taken by an external device [i.e. a machine for performing the movement action;] Command signals can be provided to the external device based on (1) the motor signals and (2) a policy of the reinforcement learning agent.”
3. and a computer configured to execute a computing process of processing the brain signal to determine the movement action and controlling the machine to perform the movement action according to the method of claim 3.
(Francis, ¶0110)
“The environment 100 includes an actor 102 (in some embodiments also referred to as an RL-BMI agent or a BMI agent), a critic 104, a brain 106, a target 108, a user 110, and an end effector 112. The components included in the environment 100 can be further broken down into more than one component and/or combined together in any suitable arrangement. Further, one or more components can be rearranged, changed, added, and/or removed. The actor 102 receives motor signal from the brain 106, and the critic 104 receives reward signal from the brain 106 [i.e. and a computer configured to execute a computing process of processing the brain signal to determine the movement action]. The RL-BMI agent (actor) 102 can be viewed as trying to map the neural activity in the brain 106 pertaining to the intended action to an appropriate action of the external actuator (end effector) 112 [i.e. and controlling the machine to perform the movement action according to the method of claim 3].”
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to modify Francis with Wu. The motivation is the same as claim 1.
Regarding claim 18:
Francis and Wu teach the method of claim 1.
Francis teaches:
1. A brain-machine interface (BMI) apparatus for capturing a brain signal of a subject and controlling a machine to perform a movement action determined by the brain signal, the BMI apparatus comprising: a sensing device for capturing the brain signal from the subject;
(Francis, ¶0111)
“FIG. 1B illustrates an exemplary architecture of a supervised actor-critic reinforcement learning brain machine interface (SAC-BMI) environment 200 in accordance with certain embodiments of the disclosed subject matter [i.e. A brain-machine interface (BMI) apparatus for capturing a brain signal of a subject and controlling a machine to perform a movement action determined by the brain signal, the BMI apparatus comprising:].”
(Francis, ¶0111)
“The multisensory feedback to the brain 106 with respect to the action performed results in a critic signal, which is labeled as rewarding or non-rewarding by a classifier [i.e. a sensing device for capturing the brain signal from the subject;].”
2. and a computer configured to execute a computing process of processing the brain signal to determine the movement action and controlling the machine to perform the movement action according to the method of claim 1.
(Francis, ¶0110)
“The end effector 112 can include, without limitations, computer cursor, robotic arm, virtual arm, prosthetic limb, or speech generation device [i.e. and a computer configured to execute a computing process of processing the brain signal to determine the movement action and controlling the machine to perform the movement action according to the method of claim 1].”
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to modify Francis with Wu. The motivation is the same as claim 1.
Regarding claim 20:
Francis and Wu teach the method of claim 1.
Francis teaches:
1. A brain-machine interface (BMI) apparatus for capturing a brain signal of a subject and controlling a machine to perform a movement action determined by the brain signal, the BMI apparatus comprising: a sensing device for capturing the brain signal from the subject;
(Francis, ¶0111)
“FIG. 1B illustrates an exemplary architecture of a supervised actor-critic reinforcement learning brain machine interface (SAC-BMI) environment 200 in accordance with certain embodiments of the disclosed subject matter [i.e. A brain-machine interface (BMI) apparatus for capturing a brain signal of a subject and controlling a machine to perform a movement action determined by the brain signal, the BMI apparatus comprising:].”
(Francis, ¶0111)
“The multisensory feedback to the brain 106 with respect to the action performed results in a critic signal, which is labeled as rewarding or non-rewarding by a classifier [i.e. a sensing device for capturing the brain signal from the subject;].”
2. and a computer configured to execute a computing process of processing the brain signal to determine the movement action and controlling the machine to perform the movement action according to the method of claim 2.
(Francis, ¶0110)
“The end effector 112 can include, without limitations, computer cursor, robotic arm, virtual arm, prosthetic limb, or speech generation device [i.e. and a computer configured to execute a computing process of processing the brain signal to determine the movement action and controlling the machine to perform the movement action according to the method of claim 2].”
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to modify Francis with Wu. The motivation is the same as claim 1.
Claims 9 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over US Pre-Grant 2019/0025917 (Francis et al; Francis) in view of “Bayesian Population Decoding of Motor Cortical Activity Using a Kalman Filter,” Neural Computation, January 2006, 18(1):80-118, Wu et al; Wu, further in view of “Non-invasive electroencephalographical (EEG) recording system in awake monkeys,” Heliyon 6 (2020), Nakamura et al; Nakamura.
Regarding claim 9 and analogous claim 19:
Francis and Wu teach the method of claim 1.
Nakamura teaches:
1. wherein the sensing device is an electroencephalogram (EEG) sensing device.
(Nakamura, pg. 2, col. 2, Sect. 2.4, ¶1)
“EEG electrodes were placed on the subject's head according to the International 10–20 system (Figure 2, left panel). A total of eleven active electrodes (F3, FZ, F4, C3, CZ, C4, P3, PZ, P4, A1, A2) and a passive electrode for ground (G) were placed on the scalp (Figure 2) [i.e. wherein the sensing device is an electroencephalogram (EEG) sensing device].”
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to modify Francis and Wu with Nakamura. The motivation is to use a well-known method of sensing brain signals in an ethical manner, as “However, most previous studies in monkeys recorded EEGs invasively using implanted electrodes… it is necessary to fix the subject's head to prevent movement related-artifacts (Nakamura, Introduction, pg. 1, ¶1).”
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to PAUL JUSTIN BREENE whose telephone number is (571)272-6320. 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, Michael J Huntley can be reached on 303-297-4307. 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.
/P.J.B./ Examiner, Art Unit 2129
/MICHAEL J HUNTLEY/Supervisory Patent Examiner, Art Unit 2129