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 .
Status of Claims
Claims 1 and 9 have been amended.
Claims 1, 5-9 and 11-15 as presented July 22, 2025 are currently pending and considered below.
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, 6-9 and 11-15 are rejected under 35 U.S.C. 103 as being unpatentable over Echauz (US 2002/0103512 A1) in further view of Aguilar (US 2012/0172743 A1), Sirendi (US 2020/0008696 A1) and Grado (US 2018/0365194 A1).
Regarding claim 1, Echauz teaches: A neurological event detection system, comprising:
a detector, configured to obtain a neural oscillation signal; a display; and a processor, coupled to the detector and the display, and configured to: (receiving brain signals from the device connected to a processor to determine and display the probability of having a seizure [0088]-[0090], [0070], [0086], [0060])
obtain, from the detector, a neural oscillation signal and extract a plurality of features from the neural oscillation signal; (receiving brain electrical signals [0060], [0072], Fig, 1 – item 100; feature extraction of the of the data [0072]-[0073], [0063], [0119]-[0123]; Fig. 1 – item 200)
obtain a plurality of classification results of a plurality of time periods according to the plurality of features of the neural oscillation signal by using a classification model, wherein the classification model is constructed based on a plurality of training data in which whether a neurological event occurs is known; (using the training set into the probabilistic neural network to reach a classification and estimating the probability of having a seizure given the input feature vector [0286]-[0292]; a classification tool is applied to the features to perform the analysis for predicting the probability of having a seizure or other neurological event, in order to achieve a classification within a plurality of time frames [0117]-[0123], [0085], [0087], claim 1, Fig. 12; the system provides outputs of one or more probabilities of having a seizure within one or more time frames [0024])
Echauz does not teach:
determine whether the event has occurred in each of the plurality of evaluation time windows, wherein whether the event has occurred in each of the plurality of evaluation time windows is determined […]
the event is determined as having occurred in one of the plurality evaluation time windows when the [parameter] exceeds a preset threshold, the neurological event is determined as having not occurred in the one of the plurality evaluation time windows when the [parameter] does not exceed the preset threshold
However, Aguilar in the analogous art of analysis of brain activity (e.g. see [0013]-[0014]) teaches:
determine whether the event has occurred in each of the plurality of evaluation time windows, wherein whether the event has occurred in each of the plurality of evaluation time windows is determined […] (for each time window, generating a decision as to whether the sliding window included a significant brain response or not, e.g. see [0057], [0039])
the event is determined as having occurred in one of the plurality evaluation time windows when the [parameter] exceeds a preset threshold, the neurological event is determined as having not occurred in the one of the plurality of evaluation time windows when the [parameter] does not exceed the preset threshold (for each time window, generating an output compared to a threshold to indicate 1 as declaring significant brain response and 0 representing a non-significant brain response, e.g. see [0039], [0041])
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Echauz to include determine whether the event has occurred in each of the plurality of evaluation time windows, wherein the event is determined as having occurred in one of the plurality evaluation time windows when the parameter exceeds a preset threshold and the neurological event is determined as having not occurred in the one of the plurality evaluation time windows when it does not exceed the preset threshold as taught by Aguilar, for the purposes enhancing the performance of the classifiers and reducing the false alarm rate (Aguilar [0017]).
Echauz and Aguilar do not teach:
determine whether an event has occurred or has not occurred in each of the plurality of time periods according to determinations of the plurality of classification results by the classification model; assign a respective binary event state, indicating that the event has occurred or has not occurred, to each of the plurality of time periods based on the corresponding classification result obtained for each time period of the plurality of time periods
set a plurality of evaluation time windows, each one of the plurality of evaluation time windows include a subset of the plurality of time periods, wherein each time period in the subset has been associated with the respective binary event state […];
the event is determined by dividing a number of time periods of the subset of the plurality of time periods having the assigned binary event state indicating that the neurological event has occurred with a total quantity of the subset of the plurality of time periods to calculate a ratio for each of the plurality of evaluation time windows; when the ratio exceeds a preset threshold and when the ratio does not exceed the preset threshold
display, on the display, a final determination result for each of the plurality of evaluation time windows based on whether a respective ratio of each of the plurality of evaluation time windows has exceeded the preset threshold.
However, Sirendi in the analogous art of analyzing patient health data (e.g. see [0004]) teaches:
determine whether an event has occurred or has not occurred in each of the plurality of time periods according to determinations of the plurality of classification results by the classification model; assign a respective binary event state, indicating that the event has occurred or has not occurred, to each of the plurality of time periods based on the corresponding classification result obtained for each time period of the plurality of time periods, (assigning a probability to each heartbeat that reflects the likelihood of the heartbeat to lead to an arrhythmic episode, where around 3600 heartbeats is roughly an hour or 350 beats is roughly 5-6 minutes (each heartbeat represents a time period of 1 second), e.g. see [0119], [0014]; a classifier separates arrhythmic and normal beat sequences and determines “'abnormal' heartbeats (e.g. which cross a threshold probability)” (a binary event state of “abnormal” or “normal” is assigned to each time period (heartbeat)), e.g. see [0095], [0120])
set a plurality of evaluation time windows, each one of the plurality of evaluation time windows include a subset of the plurality of time periods, wherein each time period in the subset has been associated with the respective binary event state […]; (“In order to arrive at a robust decision, the number of 'abnormal' heartbeats (e.g. which cross a threshold probability) are counted, and the fraction of said 'abnormal' heartbeats occurring in a given time window (for example, five minutes) is computed.”, e.g. see [0120]; “This probability is preferably updated periodically, where the time intervals for the updates may depend on the situation…the time intervals may be small enough to be effectively continuous” (Periodic or continuous updates imply the probability analysis is repeated over multiple sequential periods or windows of time), e.g. see [0025]; “The probability preferably comprises an indication of a corresponding time, where this may be an amount of time…or one or more time windows (e.g. a probability P1 of an event between x and y minutes, and a probability P 2 of an event between y and z minutes).”, e.g. see [0027])
the event is determined by dividing a number of time periods of the subset of the plurality of time periods having the assigned binary event state indicating that the neurological event has occurred with a total quantity of the subset of the plurality of time periods to calculate a ratio for each of the plurality of evaluation time windows; when the ratio exceeds a preset threshold and when the ratio does not exceed the preset threshold (the number of abnormal heartbeats are counted, and the fraction (i.e. ratio) of abnormal heartbeats in a given time window, e.g. five minutes, is computed to lead to an abnormality fraction F; a yes/no decision is made based on comparing this fraction to an abnormality threshold; a positive/yes decision on this fraction exceeding an abnormality threshold indicating a cardiac event is predicted, e.g. see [0120], [0018])
display, on the display, a final determination result for each of the plurality of evaluation time windows based on whether a respective ratio of each of the plurality of evaluation time windows has exceeded the preset threshold. (for a given time window, comparing the fraction against a predetermined threshold to indicate a probability of experiencing a cardiac event (i.e. a final determination result); displaying an output of one or more probabilities of the cardiac event, e.g. see [0167]-[0173])
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Echauz and Aguilar to include determine whether an event has occurred or has not occurred in each of the plurality of time periods according to determinations of the plurality of classification results by the classification model, assign a respective binary event state, indicating that the event has occurred or has not occurred, to each of the plurality of time periods based on the corresponding classification result obtained for each time period of the plurality of time periods, set a plurality of evaluation time windows, each one of the plurality of evaluation time windows include a subset of the plurality of time periods, wherein each time period in the subset has been associated with the respective binary event state, the event is determined by dividing a number of time periods of the subset of the plurality of time periods in which the event has occurred with a total quantity of the subset of the plurality of time periods to calculate a ratio for each of the plurality of evaluation time windows, when the ratio exceeds a preset threshold and when the ratio does not exceed the preset threshold and display a final determination result for each of the plurality of evaluation time windows based on whether a respective ratio of each of the plurality of evaluation time windows has exceeded the preset threshold as taught by Sirendi, for the purposes of enabling timely preventative action or determining periods where increased monitoring may be beneficial (Sirendi [0001]).
Echauz and Aguilar teach each evaluation time window being displaced relative to either a previous evaluation time window of the plurality of evaluation time windows or a subsequent evaluation time window of the plurality of evaluation time windows (e.g. see Echauz [0085] and Aguilar [0056]). Sirendi teaches time periods as described above.
Echauz, Aguilar and Sirendi do not teach:
each evaluation time window is displaced by one [unit] of the plurality of [units]
However, Grado in the analogous art of processing physiological signals including neural oscillations (e.g. see [0027], [0032]) teaches:
each evaluation time window is displaced by one [unit] of the plurality of [units] (“The SWIFT algorithms may be implemented on an EEG device or another sensor that is operable to measure neural oscillations…receiving measured neural oscillations as input data, applying the input data to one or more SWIFT algorithms”, e.g. see [0032]; “compute a discrete-time Fourier transform ("DTFT") of an input signal over an infinite-length temporal window that is slid from one sample in the input signal to the next”, e.g. see [0022]; “the SWIFT's window advances one sample at a time” (sliding a window by one base unit (“one sample”)), e.g. see [0051])
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Echauz, Aguilar and Sirendi to include each evaluation time window is displaced by one [unit] of the plurality of [units] as taught by Grado, for the purposes of “improved computational efficiency, improved frequency resolution, improved sampling, reduced memory requirements, and reduced spectral leakage” (Grado [0022]).
Regarding claim 6, Echauz, Aguilar, Sirendi and Grado teach the system of claim 1 as described above.
Echauz further teaches:
wherein the classification model comprises one of or a combination of an artificial neural network (ANN) model, a support vector machine (SVM) and linear classification model, a fuzzy logic model and an auto-learn system (classification tool for predicting includes neural network or a fuzzy logic algorithm [0087], [0117], claim 30).
Regarding claim 7, Echauz, Aguilar, Sirendi and Grado teach the system of claim 1 as described above.
Echauz further teaches:
wherein the plurality of features comprise one of or a combination of a temporal feature and a spatial feature (features consist of spatial features [0121], claim 36).
Regarding claim 8, Echauz, Aguilar, Sirendi and Grado teach the system of claim 1 as described above.
Echauz further teaches:
wherein the neural oscillation signal is an electroencephalography (EEG) signal (EEG data, [0282], Fig. 12, claim 177).
Regarding claim 9, Echauz, Aguilar, Sirendi and Grado teach the system of claim 1 as described above.
Echauz further teaches:
wherein the neurological event is an occurrence of seizure (detection of a seizure [0024]).
Regarding claim 11, Echauz, Aguilar, Sirendi and Grado teach the system of claim 11 as described above.
Echauz does not teach:
wherein the preset threshold is ½
However, Aguilar in the analogous art teaches:
wherein the preset threshold is ½ (the threshold n having a value in [0,1])
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Echauz to include the preset threshold is ½ as taught by Aguilar, for the purposes of determining the measurement belongs to the target class (Aguilar [0041]).
Regarding claim 12, Echauz, Aguilar, Sirendi and Grado teach the system of claim 11 as described above.
Echauz does not teach:
wherein there are at least 5 evaluation time windows in the plurality of evaluation time windows
However, Aguilar in the analogous art teaches:
wherein there are at least 5 evaluation time windows in the plurality of evaluation time windows (the EEG data may be captured with a plurality of sliding windows that overlap, e.g. for a duration of 150 ms, there may be 5 sliding windows of 50 ms, e.g. see [0056], [0014]).
Regarding claim 13, Echauz, Aguilar, Sirendi and Grado teach the system of claim 12 as described above.
Echauz teaches the event of a seizure as described above.
Echauz does not teach:
wherein in response to the classification model having determined that seizure has occurred in at least one of the plurality of time periods within the plurality of evaluation time windows and the processor has determined that the ratio does not exceed the preset threshold, all of the plurality of evaluation time windows shows the event of the seizure having not been occurred
However, Sirendi in the analogous art teaches:
wherein in response to the classification model having determined that seizure has occurred in at least one of the plurality of time periods within the plurality of evaluation time windows and the processor has determined that the ratio does not exceed the preset threshold, all of the plurality of evaluation time windows shows the event having not been occurred (the number of abnormal heartbeats are counted, and the fraction (i.e. ratio) of abnormal heartbeats in a given time window, e.g. five minutes, is computed to lead to an abnormality fraction F; a yes/no decision is made based on comparing this fraction to an abnormality threshold, e.g. see [0120], [0018]; for a given time window, comparing the fraction against a predetermined threshold to indicate a probability of experiencing a cardiac event; displaying an output of one or more probabilities of the cardiac event (it is understood that the probability would indicate the patient is not at risk for an event occurrence if the comparison of the fraction to the threshold was below the threshold), e.g. see [0167]-[0173]).
Regarding claim 14, Echauz, Aguilar, Sirendi and Grado teach the system of claim 13 as described above.
Echauz further teaches:
wherein in response to the processor having determined that the ratio has exceeded the preset threshold, all of the plurality of evaluation time windows shows the event of the occurrence of seizure (the number of abnormal heartbeats are counted, and the fraction (i.e. ratio) of abnormal heartbeats in a given time window, e.g. five minutes, is computed to lead to an abnormality fraction F; a yes/no decision is made based on comparing this fraction to an abnormality threshold; a positive/yes decision on this fraction exceeding an abnormality threshold indicating a cardiac event is predicted, e.g. see [0120], [0018]; for a given time window, comparing the fraction against a predetermined threshold to indicate a probability of experiencing a cardiac event; displaying an output of one or more probabilities of the cardiac event (it is understood that the probability would indicate the patient is at risk for an event occurrence if the comparison of the fraction to the threshold was above the threshold), e.g. see [0167]-[0173]).
Regarding claim 15, Echauz, Aguilar, Sirendi and Grado teach the system of claim 12 as described above.
Echauz, Aguilar, Sirendi and Grado teach analyzing data using an evaluation time window containing a subset of time periods, where the time window is displaced by one time period as described above.
Echauz does not teach:
wherein for every 5 evaluation time windows, the plurality of time periods is 9 as each subset of the plurality of time periods is 5
However, Aguilar in the analogous art teaches:
wherein for every 5 evaluation time windows, the plurality of time periods is 9 as each subset of the plurality of time periods is 5 (the EEG data may be captured with a plurality of sliding windows that overlap, e.g. for a duration of 150 ms, there may be 5 sliding windows of 50 ms (the time windows are displaced by a time period of 25 ms, therefore, there are 6 time periods as each subset of time periods is 2), e.g. see [0056], [0014])
The claim limitation of “the plurality of time periods is 9 as each subset of the plurality of time periods is 5” is nonfunctional descriptive language. The fact that the claim recites the plurality of time periods is 9 as each subset of the plurality of time periods is 5 and the art recites the plurality of time periods is 6 as each subset of the plurality of time periods is 2 is immaterial, as the invention would perform the same regardless of whether, for every 5 evaluation time windows, the plurality of time periods is 9 or 6 and the plurality of time periods is 5 or 2.
Claims 5 is rejected under 35 U.S.C. 103 as being unpatentable over Echauz, Aguilar, Sirendi and Grado in further view of Conradsen (US 2014/0163413 A1).
Regarding claim 5, Echauz, Aguilar, Sirendi and Grado teach the system of claim 1 as described above.
Echauz further teaches:
further comprising: receiving a setting signal and adjusting […] (the setting adjustment unit may automatically perform parameter tuning [0076]-[0077]; the features or parameters can be tuned according to the patient [0070])
Echauz, Aguilar, Sirendi and Grado do not teach:
adjusting at least one of the preset time window width and the preset threshold according to the setting signal
However, Conradsen in the analogous art of analyzing patient data using classification methods for detection seizures ([0002]) teaches:
adjusting at least one of the preset time window width and the preset threshold according to the setting signal (the adaptive update module adjusts seizure threshold values and time windows [0040])
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Echauz, Aguilar, Sirendi and Grado to include adjustment of the preset time window width or the preset threshold as taught by Conradsen. This increases the accuracy of detecting seizures and reduces the number of false positives (Conradsen [0040], [0013]).
Response to Arguments
Regarding the rejection under 35 U.S.C. § 103 of Claims 1, 5-9 and 11-15, the Examiner has considered the Applicant’s arguments; however the arguments are not persuasive.
Applicant argues Sirendi fails to teach the two-stage process of the claims of first assigning a binary event state to each time period and then calculating a ratio based on a count of those binary state.
The Examiner respectfully disagrees. Sirendi teaches “the number of 'abnormal' heartbeats (e.g. which cross a threshold probability) are counted, and the fraction of said 'abnormal' heartbeats occurring in a given time window…is computed” ([0120]). The binary determination (abnormal/normal) is made for each base unit (heartbeat) when the threshold probability is crossed. Next, Sirendi states the “number of” these binary-state heartbeats are “counted” and a fraction (ratio) is computed from this count. Applicant’s argument that “Sirendi prefers to directly perform mathematical operations” is refuted by the reference’s plain language that specifies the heartbeats are counted after they have been classified by “cross[ing] a threshold”. Thus, Sirendi explicitly teaches making the binary judgement for each heartbeat before calculating the fraction (ratio).
Applicant argues the claimed limitation “each evaluation time window is displaced by one time period of the plurality of time periods”.
The argument is moot given the new grounds of rejection as necessitated by amendment.
Applicant argues claim 13.
The typographical error in the previous Office Action is acknowledged and has been corrected. Applicant’s arguments that Sirendi doesn’t match Figs. 8-10 of the claimed invention is not persuasive. The claims and not the drawings or the specification define the scope of the invention. Specific examples or figures from the specification cannot be relied upon to read limitations into the claim that are not explicitly recited therein. MPEP 2111.01 states, “Though understanding the claim language may be aided by explanations contained in the written description, it is important not to import into a claim limitations that are not part of the claim. For example, a particular embodiment appearing in the written description may not be read into a claim when the claim language is broader than the embodiment.”. Claim 13 describes a result of the method recited in claim 1, where when the count of abnormal periods is greater than zero but insufficient to cause the ratio to “exceed the preset threshold”, the final decision is “no”. Sirendi explicitly teaches this logic. Sirendi teaches “an alert may be issued (or another action taken) for positive decision”, where the “positive decision” is indicative of the threshold being exceeded and abnormal (e.g., arrhythmic) ([0120]; also see [0167]). This is further supported by Sirendi in that “Alerts are only issued after the fraction has been evaluated in an appropriate time window and found to exceed a value that optimally separates normal and arrhythmic patient groups.” ([0048]). Therefore, Sirendi teaches that if the fraction does not exceed the threshold, no alert is issued and the decision is not positive (i.e. the event is determined not to have occurred). Notably, Applicant admits this very point in their response, stating that the “core logic” of Sirendi is “that if the ratio does not reach the threshold, it is judged as no event” (see Arguments pg. 9 para. 2).
Applicant argues the numerical values of claim 15 are specific and are not taught by Aguilar.
The Examiner respectfully disagrees. Applicant correctly notes the broadest reasonable interpretation (BRI) applies to claims and not the prior art. However, the Examiner’s mapping of Aguilar is not an application of BRI, but rather a determination of what Aguilar explicitly teaches a person of ordinary skill in the art. Aguilar teaches a specific, fixed numerical relationship. Aguilar teaches “5 sliding windows” and defines the size of the windows as “50 ms” and the displacement as 25 msec ([0056]). The claim specifies the time window as a “subset” and the base unit as a “time period”. Therefore, Aguilar explicitly teaches a structure where each subset (time window) comprises 2 time periods (displacement units). This sequence of 5 time windows necessarily covers a total of 6 unique time periods. Applicant further argues that the claimed numbers (9 and 5) are a “significant difference” from the taught numbers (6 and 2). This argument is not persuasive. Aguilar teaches the concept of a sliding window analysis using a fixed, defined relationship between the window size (subset) and displacement unit (time period). The claimed invention merely selects different numbers for this known relationship. The change from a 2 time period subset to a 5 time period subset is a mere matter of design choice. A person of ordinary skill in the art would readily recognize that the window size can be varied and selecting a different number for the window size is a routine, predictable variation of the structure taught by Aguilar. In addition, the recitation of “9” is immaterial as it is the necessary result of applying the independent claim’s slide-by-one mechanism to 5 time windows of 5 time periods each. Therefore, the rejection is proper and the distinction in the specific numbers is immaterial and “nonfunctional descriptive language” because the underlying structural and functional concept of using a fixed-parameter sliding window analysis is taught by the prior art.
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 mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Aaisha Abdullah whose telephone number is (571)272-5668. The examiner can normally be reached on Monday through Friday 8:00 am - 5:00 pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Peter H Choi can be reached on (469) 295-9171. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
/A.A./Examiner, Art Unit 3686
/PETER H CHOI/Supervisory Patent Examiner, Art Unit 3681