DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
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.
Claim(s) 1-2, 4-5, 8-15, and 17-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Zeller et al. (US 2022/0043091 A1, Feb. 10, 2022) (hereinafter “Zeller”) in view of Ankit, Utkarsh (Transformer Neural Networks: A Step-by-Step Breakdown; builtin.com/artificial-intelligence/transformer-neural-network; Jun. 28, 2022) (hereinafter “Ankit”).
Regarding claims 1 and 14: Zeller discloses receiving raw data of the MRI scan of a patient (fig. 2A, [0077] - data sets D1, D2); and determining, by a neural network, a physiological signal of the patient from the received raw data (fig. 2A, [0078]), and wherein an encoding of the raw data comprises a positional encoding that includes information corresponding to a slice position within a slice stack of the MRI scan ([0021], as evidenced by McMahon, Katie L., Gary Cowin, and Graham Galloway. "Magnetic resonance imaging: the underlying principles." journal of orthopaedic & sports physical therapy 41.11 (2011): 806-819).
However, Zeller discloses the use of a recurrent neural network (RNN) which uses LSTM modules ([0039]) and does not disclose that the neural network comprises a transformer architecture.
Ankit discloses that RNNs have several disadvantages, including vanishing gradient and slow training times, and that these disadvantages can be overcome by using a transformer architecture (see whole document but particularly section header Long Short-Term Memory).
It would have been prima facie obvious for one having ordinary skill in the art prior to the effective filing date of the claimed invention to modify the method of Zeller by replacing the LSTM-based RNN with a transformer network in view of the teachings of Ankit that a transformer architecture overcomes several known disadvantages of RNNs.
Regarding claims 2 and 15: Zeller further discloses wherein the physiological signal comprises at least one of a respiration curve, an electrocardiogram curve, or a movement curve ([0079], predicted curve r').
Regarding claims 4 and 17: Zeller further discloses modifying, by the neural network, the received raw data taking into account the determined physiological signal ([0079]-[0081]); and outputting the modified raw data (the modified raw data is “output” to the reconstruction process).
Regarding claims 5 and 18: Zeller further discloses receiving sensor data with regard to the physiological signal of the patient and/or with regard to a movement of the patient, wherein the sensor data was recorded by a sensor during creation of the MRI scan, wherein modifying the received raw data further takes into account the received sensor data ([0076], [0079]-[0081]).
With respect to claims 8-13 and 19-20: While the claims recite details of the implementation of a known type of neural network (transformer) to a known problem (the motion correction based on physiological signals as disclosed by Zeller), there is no evidence that any of the limitations of these claims are anything more than the natural result of adapting a general transformer neural network framework to solving a specific problem.
See, for example, the description of transformer implementation provided by Ankit as well as the evidence provided by Alammar, Jay (Visualizing A Neural Machine Translation Model (Mechanics of Seq2seq Models With Attention); jalammar.github.io/visualizing-neural-machine-translation-mechanics-of-seq2seq-models-with-attention/; May 9, 2018), which provides a detailed breakdown regarding the nature of the input and output to encoder/decoder layers as well as context vectors, and Romano, Nicola (A basic introduction to neural networks – Part 2: Training; www.nicolaromano.net/data-thoughts/training-neural-networks/; retrieved 01/22/2026) which provides a basic overview of supervised training of a neural network including ground-truth training data as well as the application of loss functions for optimization.
In the absence of any evidence to the contrary, the limitations of claims 8-13 and 19-20
are considered to merely be the natural product of implementing the method of Zeller using a transformer architecture as described by Ankit.
Claim(s) 3 and 16 is/are rejected under 35 U.S.C. 103 as being unpatentable over Zeller and Ankit as applied to claim 1 above, and further in view of Nguyen et al. (US 2025/0248617 A1, Aug. 7, 2025) (hereinafter “Nguyen”).
Regarding claims 3 and 16: Zeller as modified by Ankit discloses the method of claim 1 and the neural network of claim 14 but are silent on outputting the determined physiological signal.
Nguyen, in the same field of endeavor, discloses outputting a determined physiological signal ([0045]).
It would have been prima facie obvious for one having ordinary skill in the art prior to the effective filing date of the claimed invention to modify the method and neural network of Zeller and Ankit to output the determined physiological signal as taught by Nguyen in order to allow the user to view the data for verification or confirmation purposes.
Claim(s) 6 is/are rejected under 35 U.S.C. 103 as being unpatentable over Zeller and Ankit as applied to claim 1 above, and further in view of Zeller (EP 3748387 A1, Dec. 09, 2020) (hereinafter “Zeller ‘387”).
Regarding claim 6: Zeller and Ankit disclose the method of claim 1, but are silent on wherein the raw data comprises temporally sorted k-space lines or Fourier-transformed k-space lines.
Zeller ‘387, in the same field of endeavor, discloses temporal binning (“sorting”) of the k-space data for determining a physiological signal ([0046]-[0048]). Zeller ‘387 further discloses that this binning improves the accuracy of the motion correction ([0047]).
It would have been prima facie obvious for one having ordinary skill in the art prior to the effective filing date of the claimed invention to modify the method of Zeller and Ankit to include temporal binning (“sorting”) of the k-space data as taught by Zeller ‘387 in order to improve the accuracy of the motion correction.
Response to Arguments
Rejection of claims 14-20 under 35 U.S.C. §101 is withdrawn in light of the amendments to the claims.
Rejection of claim 7 under 35 U.S.C. §112(b) is withdrawn in light of the cancelation of the claim.
Applicant’s arguments with respect to prior art rejection of all pending claims, filed 05/05/2026, have been fully considered but are not persuasive.
Applicant argues that Zeller fails to disclose “wherein an encoding of the raw data comprises a positional encoding that includes information corresponding to a slice position within a slice stack of the MRI scan.” Applicant further argues that Zeller does not disclose creating a “positional encoding” as a specific input feature for a neural network and does not teach the inventive step of actively encoding that slice information into a vector to be used as an input to a transformer.
Examiner respectfully notes that these arguments are directed to elements not included in the claims. The claims say nothing about creating a “positional encoding” as a specific input feature for a neural network or the inventive step of actively encoding that slice information into a vector to be used as an input to a transformer. Applicant is reminded that, although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993). The claims as written merely require that “an encoding” of the raw data comprises a positional encoding that includes information corresponding to a slice position within a slice stack of the MRI scan and does not make any connection between this encoding and any specific, inventive input to the neural network. There is nothing in the claims to differentiate “an encoding” as a specific input to a neural network as opposed to the positional encoding that is inherent to the generation of MRI data. See McMahon, Katie L., Gary Cowin, and Graham Galloway. "Magnetic resonance imaging: the underlying principles." journal of orthopaedic & sports physical therapy 41.11 (2011): 806-819. Specifically, in the section titled “Spatial Encoding,” McMahon describes the conventional process of MRI data generation with positional encoding which includes frequency, phase and slice encoding corresponding to the Gx, Gy, and Gz gradients. Nothing in the claims differentiates between “an encoding” as traditionally understood as a conventional and necessary part of MRI data generation and Applicant’s alleged intended meaning of a specific, inventive input to a neural network. Any conventionally generated raw MRI data would have “an encoding” comprising “a positional encoding that includes information corresponding to a slice position within a slice stack of the MRI scan.”
Applicant further argues that “positional encoding” is a term of art from transformer architectures and is modified to include slice position information.
Examiner respectfully notes that “positional encoding” or “spatial encoding” is also a term of art for general magnetic resonance imaging and includes slice position information. Applicant has made no distinction between any alleged inventive encoding and the ordinary encoding that is part of any conventionally generated MRI data.
If Applicant intends a specific and novel meaning to “an encoding” as recited in the independent claims, Applicant should clearly distinguish between such meaning and the conventional meaning of encoding as it is known in the art of magnetic resonance imaging.
The rejections are updated and maintained.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Tsai, Tsung-Han, Yz-Heng Lin, and Tsung-Hsien Lin. "Motion artifact correction in mri using gan-based channel attention transformer." 2023 IEEE Biomedical Circuits and Systems Conference (BioCAS). IEEE, 2023.
Katharopoulos, Angelos, et al. "Transformers are rnns: Fast autoregressive transformers with linear attention." International conference on machine learning. PMLR, 2020.
THIS ACTION IS MADE FINAL. 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 CAROLYN A PEHLKE whose telephone number is (571)270-3484. The examiner can normally be reached 9:00am - 5:00pm (Central Time), Monday - Friday.
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, Chris Koharski can be reached at (571) 272-7230. 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.
/CAROLYN A PEHLKE/Primary Examiner, Art Unit 3799
Prosecution Insights