DETAILED ACTION
1. This office action is in response to the Application No. 19546399 filed on 06/18/2026. Claims 1-20 are presented for examination and are currently pending.
Notice of Pre-AIA or AIA Status
2. 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 Arguments
3. The Applicants argument regarding the prior art have been considered and the Examiner is withdrawing the rejections in the previous Office action because Applicant’s amendment necessitated new grounds of rejection presented in this Office Action.
It is noted that arguments regarding independent claims 1 and 11 have been considered but are moot because new references have now been used to remap the independent claims 1 and 11.
The claim amendments of the independent claims 1 and 11 which recites “wherein said convergence comprises energy-dissipating physical settling of the spiking substrate to a stable attractor configuration” of 06/18/2026 has overcome the 101 rejection. As a result, the 101 rejection is withdrawn.
As regards to the prior art rejection, the Applicant argued on page 12 of the remarks that “Langdon Does Not Disclose a Neuromorphic Computing System. Langdon et al. is a review article that surveys and theorizes about neural population-level representations in biological nervous systems. It describes observations and models of how neurons in biological brains collectively encode information on geometric manifolds-it does not disclose, teach, or suggest any neuromorphic computing system, hardware architecture, or method of neuromorphic computing. The Examiner's mapping of Langdon's biological circuit descriptions to the claimed neuromorphic system improperly reads structural and operational hardware requirements out of the claims”.
The Applicant has also argued on page 14 of the remark that “Langdon Describes Biological Observation, Not a Neuromorphic Computing System. The Examiner's anticipation mapping conflates two fundamentally different things: (1) theoretical models and observations concerning how biological neural circuits behave, and (2) engineered neuromorphic computing systems designed to perform geometric cognition through physical substrate dynamics. Langdon et al. falls entirely within the first category. The reference is a neuroscience review article; it neither discloses nor suggests any computing system, neuromorphic or otherwise, that is configured to perform the claimed operations”.
The arguments above are not persuasive because Langdon which has now been used to map the dependent claims is related to neuromorphic computing because neuronal spikes conventionally occur in neuromorphic circuits and Langdon teaches neuron spikes in Figure 2b, page 368. Furthermore, also Langdon teaches neural code can be summarized on the level of individual neurons in the form of tuning curves that describe a neuron’s firing rate, pg. 365, right col., second para. The Examiner notes spikes are fired by neurons and firing rate is also a spike count).
Langdon is also related to hardware because the reference teaches circuits and a substrate (Grid-like responses can emerge in artificial recurrent neural networks (RNNs) trained to perform path integration (Fig. 2d, pg. 368); identify neural circuit elements that carry out specific functions, emphasizing connectivity between neurons as a substrate, abstract).
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.
4. Claims 1-3, 10-13, and 20 are rejected under 35 U.S.C 103 as being unpatentable over Gangopadhyay et al. (“A Spiking Neuron and Population Model Based on the Growth Transform Dynamical System”. Front. Neurosci. 14:425. doi: 10.3389/fnins.2020.00425, published: 12 May 2020) in view of Maslennikov et al. ("Nonlinear dynamics and machine learning of recurrent spiking neural networks." Physics-Uspekhi 192 (2022): 1089-1109.).
Regarding claim 1, Gangopadhyay teaches a neuromorphic computing system (In this paper, we address these limitations by proposing a novel GT spiking neuron and population model, along with a neuromorphic framework (pg. 2, left col., last sentence to right col.); we quantitatively measure the recall performance of the network by computing the mean distance between each pair of original-recall spiking dynamics as they unfold overtime, pg. 14, right col., first para.)
comprising: a dynamical substrate (Individual neuron models are then connected through synapses, bottom-up, to form large-scale spiking neural networks, pg. 2, left col., first para.)
comprising a plurality of interconnected processing elements configured to update their states in response to events (For a network of M neurons with state variables v={vi}∈RM, where the trans-conductance coupling matrix is denoted by Q = {Qij} ∈ RM × RM and the external stimulus vector is denoted by b={bi}∈RM, the time-evolution of the network under bound constraints on the state variables |vi,n|≤vc for all time-indices n, is governed by the following discrete-time updates: ... The composite spike response of the i-th neuron at time-step n is given by ... where the trans-impedance parameter C>0 Ω determines the magnitude of
each spike, pg. 7, left col., first para.),
wherein: the processing elements comprise spiking neurons governed by membrane dynamics (In this paper, we introduce a new spiking neuron and population model where the dynamical and spiking responses of neurons can be derived directly from a network objective or energy functional of continuous-valued neural variables like the membrane potential, abstract) and synaptic conductances (For a single neuron, the state variables are usually its membrane potential and the conductances of ion channels that mediate changes in the membrane potential via flux of ions across the cell membrane, pg. 2, left col., first para.); and
attractor convergence is an energy-dissipating physical process (We then show how gradient discontinuity in the network energy functional can be used to modulate the shape of the action potential while maintaining the local convexity and the location of the steady-state attractor ... The formulation will then allow for affecting network convergence toward the steady-state attractor, pg. 2, right col., third and fourth bullet points) governed by the coupled nonlinear differential equations of said membrane dynamics and network connectivity (We will subsequently extend these dynamics to build coupled networks with interesting properties like memory and global adaptation for energy-efficient neural representation, pg. 8 (right col., third to the last para.); Considering the n-th iteration of the update equation in(9)as the n-th time-step for the neuron i, we can rewrite (9) in terms of the objective function for the neuron model presented in (8),as given below
PNG
media_image1.png
192
504
media_image1.png
Greyscale
Then asymptotically from (1), and as shown in Appendix B,
we have
PNG
media_image2.png
58
474
media_image2.png
Greyscale
pg. 4, right col.);
wherein the system is configured to: transform an input into a representation on a continuous manifold by causing the dynamical substrate to converge to an attractor state (This history-dependent stimulus response could serve as a short-term memory, where residual network energy from a previous external input subserves synaptic interactions among a population of neurons to set specific initial conditions for a future stimulus based on the stimulus history, forcing the network to settle down in a particular attractor state, pg. 13, right col., second to the last para.),
wherein said convergence comprises energy-dissipating physical settling of the spiking substrate to a stable attractor configuration (The proposed work is the first of its kind to treat the spike generation and transmission processes in a spiking network as an energy-minimization problem involving continuous-valued neural state variables like the membrane potential (pg. 16, right col., first para.); The rationale for this approach is that physical processes occurring in nature have a tendency to self-optimize toward a minimum-energy state. This principle has been used to design neuromorphic systems, pg. 2, left col., second para.);
derive geometric properties of the continuous manifold (Corresponding population activities trace different trajectories in the neural subspace, Fig. 7C, pg. 12. The Examiner notes Fig. 7C is illustrates neural trajectories in a continuous neural manifold) from physical characteristics of the dynamical substrate (The system minimizes an appropriate energy functional under realistic physical constraints to produce emergent spiking activity in a population of neurons, pg. 16, right col., first para.),
wherein the geometric properties are derived from spike-timing correlations
(time-to-first spike as a function of the distance d for each neuron in the network, Fig. 6, pg. 11), synaptic weight distributions (synaptic weight matrix for the network, pg. 3, right col., last para.), ..., and population firing rate covariances that are physical quantities realized in hardware (We see from Figures 6C,D that as this value increases, the average firing rate of a neuron (number of spikes in a fixed number of time-steps or iterations) increases, and the time-to-first spike becomes progressively smaller, pg. 12, left col., second to the last para.);
generate trajectories on the continuous manifold through evolution of the dynamical substrate states (The tool enables users to visualize the effects of different modulation functions and other parameters on the neural dynamics, as well as the time-evolution of population trajectories and the network energy function with different inputs and under different initial conditions, pg. 8, right col., third to the last para.),
and modify parameters of the dynamical substrate based on its activity patterns, wherein the modifications alter the geometric properties of the continuous manifold,
wherein said parameter modifications comprise physical changes to at least one of: synaptic weights (Associative memories are neural networks which can store memory patterns in the activity of neurons in a network through a Hebbian modification of their synaptic weights, pg. 13, right col., last para.), conduction delays, or memristive resistance states of the substrate;
Gangopadhyay does not explicitly teach conduction delay patterns; wherein said trajectories emerge from the temporal sequence of network states unfolding according to the physical laws governing the neuromorphic substrate at neuromorphic hardware timescales; wherein the geometric properties emerge from dynamics of the substrate rather than from explicit computation.
Maslennikov teaches conduction delay patterns (Further, during the delay interval, a transition occurs to the neighborhood of stable or saddle points corresponding to two categories, pg. 1027, left col., first full para.);
wherein said trajectories emerge from the temporal sequence of network states unfolding according to the physical laws (the trajectory is in a stable state of equilibrium ...; after the input stimulus is presented, the trajectory leaves its vicinity in the direction of one of the states that correspond toa particular category ... (Fig. 3d), pg. 1027, left col., first full para.) governing the neuromorphic substrate at neuromorphic hardware timescales (we consider examples of several of the most popular large-scale projects in which neuromorphic systems have been implemented in the form of hardware devices, pg. 1032, right col., second to the last para.);
wherein the geometric properties emerge from dynamics of the substrate rather than from explicit computation (It has been shown that a recurrent neural network of the form (11) can be trained to autonomously generate various output patterns in the absence of input signals (y, 0). Fig. 1b, c shows the target space-time activity of the output neurons and the corresponding dynamics of the neurons of the recurrent network trained to reproduce the given target pattern at the output, pg. 1024, right col., last para.).
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Gangopadhyay to incorporate the teachings of Maslennikov for the benefit of development of hardware neuromorphic systems mimicking the basic principles of brain operations aimed at developing and implementing algorithms that consume much less energy and are easier to scale to massive parallelization of computing architectures (Maslennikov, pg. 1035, last para. to pg. 1036)
Regarding claim 2, Modified Gangopadhyay teaches the system of claim 1, Gangopadhyay teaches wherein the processing elements comprise spiking neurons (The proposed spiking neuron model based on the discrete-time Growth Transform dynamical system is summarized in Table 2, pg. 5, right col., last para.), and
wherein the events comprise discrete spike events (The composite spike signal si,n, shown in Figure 2E, pg. 5, left col., last para.).
Regarding claim 3, Modified Gangopadhyay teaches the system of claim 2, Gangopadhyay teaches wherein the geometric properties comprise metric tensor components derived from at least one of: spike-timing correlations, synaptic weight distributions (time-to-first spike as a function of the distance d for each neuron in the network, Fig. 6, pg. 11; synaptic weight matrix for the network, pg. 3, right col., last para.), conduction delay patterns, or population firing rate covariances.
Regarding claim 10, Modified Gangopadhyay teaches the system of claim 1, Gangopadhyay teaches wherein the dynamical substrate comprises one of: a spiking neural network, (We first remap the synaptic interactions in a standard spiking neural network in a manner that the solution (steady-state attractor) could be encoded as a first-order condition of an optimization problem, pg. 2, right col., first bullet point) a memristive array, a photonic processor, or an analog dynamical system.
Regarding claim 11, claim 11 is similar to claim 1. It is rejected in the same manner and reasoning.
Regarding claim 12, claim 12 is similar to claim 2. It is rejected in the same manner and reasoning.
Regarding claim 13, claim 13 is similar to claim 3. It is rejected in the same manner and reasoning.
Regarding claim 20, claim 20 is similar to claim 10. It is rejected in the same manner and reasoning.
5. Claims 4-9 and 14-19 are rejected under 35 U.S.C 103 as being unpatentable over Gangopadhyay et al. (“A Spiking Neuron and Population Model Based on the Growth Transform Dynamical System”. Front. Neurosci. 14:425. doi: 10.3389/fnins.2020.00425, published: 12 May 2020) in view of Maslennikov et al. ("Nonlinear dynamics and machine learning of recurrent spiking neural networks." Physics-Uspekhi 192 (2022): 1089-1109.) and further in view of Langdon et al. ("A unifying perspective on neural manifolds and circuits for cognition." Nature Reviews Neuroscience 24.6 (2023): 363-377).
Regarding claim 4, Modified Gangopadhyay teaches the system of claim 1, they do not explicitly teach the limitations of claim 4.
Langdon teaches wherein transforming the input comprises inducing competition among groups of processing elements through inhibitory connections, wherein a winning group determines the attractor state (The inhibitory neurons in these networks mediate winner take-all competition between the excitatory populations so that in response to a stimulus, one group elevates its firing rate representing the decision outcome (pg. 370, right col., first para.); For example, recurrent self-excitation in two clusters of excitatory neurons and cross inhibition mediated by a third cluster of inhibitory neurons (see the figure, part a) can generate ‘winner-take-all’ dynamics, in which two discrete attractors support decision-making, pg. 367, left col., first para.).
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Modified Gangopadhyay to incorporate the teachings of Langdon for the benefit of modelling how neural population dynamics unfold along the manifold as the task progresses which provides a dynamical-system description of neural computation (Langdon, pg. 364, right col., last sentence. to right col., first sentence)
Regarding claim 5, Modified Gangopadhyay teaches the system of claim 1, they do not explicitly teach the limitations of claim 5.
Langdon teaches wherein transforming the input comprises propagating activity through feedforward chains of processing elements (Grid firing patterns can also arise in feedforward models in which spatial selectivity is inherited from external inputs, pg. 369, right col., first para.),
wherein convergence of chain activity determines the attractor state (When triggered by sensory input, the network activity converges to one of the attractors, pg. 364, left col., second para.).
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Modified Gangopadhyay to incorporate the teachings of Langdon for the benefit of modelling how neural population dynamics unfold along the manifold as the task progresses which provides a dynamical-system description of neural computation (Langdon, pg. 364, right col., last sentence. to right col., first sentence)
Regarding claim 6, Modified Gangopadhyay teaches the system of claim 1,
they do not explicitly teach the limitations of claim 6.
Langdon teaches the system is further configured to estimate curvature of the manifold by introducing perturbations to the dynamical substrate (Interpretable mechanisms enable us to predict the behavioural effects of specific circuit perturbations, making it possible to experimentally test circuit mechanisms, pg. 367, left col., second para) and measuring divergence of trajectories (The population activity varies over time in an orthogonal non-coding subspace (right; coloured trajectories) ... During the sample period, trajectories diverge (red and blue circles) and then approach one of two distinct states representing the memory of the sample category during the delay (red and blue triangles). During the test period, trajectories again diverge towards two other states representing match or non-match decisions (green and black crosses), pg. 371, Fig.3).
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Modified Gangopadhyay to incorporate the teachings of Langdon for the benefit of modelling how neural population dynamics unfold along the manifold as the task progresses which provides a dynamical-system description of neural computation (Langdon, pg. 364, right col., last sentence. to right col., first sentence)
Regarding claim 7, Modified Gangopadhyay teaches the system of claim 6, Langdon teaches wherein the perturbations comprise at least one of: injection of additional events, phase shifts in periodic activity, or transient bias signals (In a similar manner, circuit models have been used to relate connectivity structure to the dynamical-system description of neural computation across many cognitive tasks. Such links are powerful because they enable us to predict the behavioural effects of circuit perturbations (such as changes in the excitation–inhibition balance, pg. 364, left col., second para. The Examiner notes that changes in the excitation-inhibition balance are due to injection of additional events).
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Modified Gangopadhyay to incorporate the teachings of Langdon for the benefit of modelling how neural population dynamics unfold along the manifold as the task progresses which provides a dynamical-system description of neural computation (Langdon, pg. 364, right col., last sentence. to right col., first sentence)
Regarding claim 8, Modified Gangopadhyay teaches the system of claim 1, they do not explicitly teach the limitations of claim 8.
Langdon teaches wherein the trajectories follow geodesic paths determined by the geometric properties (In navigational systems, the regularity of single-neuron responses and simple manifold geometry naturally suggest the underlying connectivity structure and form the basis for theoretical circuit models, pg. 374, right col., second para.), and
wherein the paths emerge from winner-take-all competition for propagation direction (The inhibitory neurons in these networks mediate winner take-all competition between the excitatory populations so that in response to a stimulus, one group elevates its firing rate representing the decision outcome (pg. 370, right col., first para.)).
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Modified Gangopadhyay to incorporate the teachings of Langdon for the benefit of modelling how neural population dynamics unfold along the manifold as the task progresses which provides a dynamical-system description of neural computation (Langdon, pg. 364, right col., last sentence. to right col., first sentence)
Regarding claim 9, Modified Gangopadhyay teaches the system of claim 1, they do not explicitly teach the limitations of claim 9.
Langdon teaches wherein modifying parameters comprises adjusting connection strengths between processing elements based on relative timing of their activity (In a two-dimensional continuous attractor network model of grid cells, the cells are spatially arranged on a two-dimensional torus, with the strength of the recurrent excitatory connections between the cells decreasing in proportion to the distance that separates them, pg. 369, right col., first para.),
implementing plasticity rules that reshape the manifold geometry through experience (The ring manifold is nonlinearly embedded in the neural population state space. The shape of this embedding is determined by the heterogeneous and nonlinear tuning curves of individual neurons, pg. 366, right col., Fig. 1).
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Modified Gangopadhyay to incorporate the teachings of Langdon for the benefit of modelling how neural population dynamics unfold along the manifold as the task progresses which provides a dynamical-system description of neural computation (Langdon, pg. 364, right col., last sentence. to right col., first sentence)
Regarding claim 14, claim 14 is similar to claim 4. It is rejected in the same manner and reasoning.
Regarding claim 15, claim 15 is similar to claim 5. It is rejected in the same manner and reasoning.
Regarding claim 16, claim 16 is similar to claim 6. It is rejected in the same manner and reasoning.
Regarding claim 17, claim 16 is similar to claim 7. It is rejected in the same manner and reasoning.
Regarding claim 18, claim 18 is similar to claim 8. It is rejected in the same manner and reasoning.
Regarding claim 19, claim 19 is similar to claim 9. It is rejected in the same manner and reasoning.
Conclusion
Applicant's amendment necessitated the new ground(s) 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 extension fee 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 date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MORIAM MOSUNMOLA GODO whose telephone number is (571)272-8670. The examiner can normally be reached Monday-Friday 8am-5pm EST.
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, Michelle T Bechtold can be reached on (571) 431-0762. 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.
/M.G./Examiner, Art Unit 2148
/MICHELLE T BECHTOLD/Supervisory Patent Examiner, Art Unit 2148
Prosecution Insights