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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on May 21, 2026 has been entered.
Response to Arguments
The Applicant filed a Request for Continued Examination, Amendments to the Claims, and Remarks on May 21, 2026 in response to the Examiner’s Final Office Action, mailed February 23, 2026.
Amendments to the Claims
At this time, claims 1-3, 5, and 7-18 are pending. Claims 3, 8, and 17 have been amended. The Applicant has not added new claims. The Applicant asserts that no new matter is added. Claims 1 and 8 are in independent form. (Remarks, pg. 6)
Claim Objections
Claims 3 and 12 were previously objected under due improper multiple dependent form. (Remarks, pg. 6) The Applicant has amended claim 3, however, this does not remedy the improper multiple dependent form.
For claim 3 to overcome the improper multiple dependent form objection, the preamble could be amended to recite either “The computer program of claim 1…” or “The computer program of claim 2…”.
For claim 12 to overcome the improper multiple dependent form objection, the preamble could be amended to recite either “The method of claim 10…” or “The method of claim 11…”.
Applicant's arguments have been fully considered but they are not persuasive. The claim objections to claims 3 and 12 are maintained.
Claim Rejections - 35 U.S.C. § 102
Claim 8 was previously rejected under 35 U.S.C. 102(a)(2) as being anticipated by Kremen et al. (US 2020/0337645, previously cited). The Applicant asserts that Kremen's disclosure of a supervised and an unsupervised learning mode is not equivalent to the "two training cycles" as recited in claim 8 as amended. (Remarks, pg. 6-7)
The Examiner respectfully disagrees with the Applicant’s argument that “Kremen presents these as alternative modes of operation rather than a required two-step process where a general training is followed by a patient-specific training, …Kremen teaches two possible modes, not necessarily two sequential cycles applied together.” (Remarks, pg. 7)
Figure 4 of Kremen depicts two sequential cycles applied together via unsupervised version 404 and supervised version 406. The supervised version 406 represents a general training cycle, using a “gold standard” scoring dataset, whereas the unsupervised version 404 represents a patient-specific training, utilizing a higher degree of fine-tuning the output towards an individual patient.
Considering the amendments to claim 8, the Examiner has expanded the previous ground of rejection, in view of Kremen et al. (US 2020/0337645, previously cited), to further clarify the Examiner’s initial stance and to better represent the claimed invention.
Claim Rejections - 35 U.S.C. § 103
Claims 1-3, 5, 7, and 9-18 were previously rejected under 35 U.S.C. 103 as obvious over Kremen in view of Remmert (US 2018/0117308, previously cited) and further in view of Lee et al. (US 2011/0137381, previously cited). The Applicant respectfully disagreed, arguing “that even if the combination of Kremen and Remmert were considered, a person of ordinary skill in the art would not have been motivated to further incorporate the teachings of Lee to arrive at the claimed invention because the combined teachings fail to disclose all the limitations of claim 1”. The Applicant emphasizes steps c21) and c22) of pending, independent claim 1, further stating “Lee does not teach or suggest the specific algorithmic steps of calculating a mean of coordinates (c21) or iteratively counting electrodes within a threshold distance to find an optimal center (c22)”. (Remarks, pg. 7-8; Emphasis added by Applicant)
Applicant’s arguments with respect to the rejection of 1-3, 5, 7, and 9-18 under 35 U.S.C. 103 have been fully considered and are persuasive. Therefore, the rejection has been withdrawn. However, upon further consideration, a new ground of rejection is made in view of the Kremen/Remmert/Lee combination further in view of Guger et. al (US 2019/0082992, hereinafter referred to as Guger).
Claim Objections
Claim 3 is objected to under 37 CFR 1.75(c) as being in improper form because a multiple dependent claim language in claim 3 (for example, but not limited to, "The computer program of claim 1 or 2..."). See MPEP § 608.01(n). For claim 3 to overcome the improper multiple dependent form objection, the preamble could be amended to recite either “The computer program of claim 1…” or “The computer program of claim 2…”. Claim 3 was previously and is currently examined, but the Examiner requests that this be remedied.
Claim 12 is additionally objected to under 37 CFR 1.75(c) as being in improper form because a
multiple dependent claim language in claim 12 (for example, but not limited to, "The method of claim 10 or 11..."). See MPEP § 608.01(n). For claim 12 to overcome the improper multiple dependent form objection, the preamble could be amended to recite either “The method of claim 10…” or “The method of claim 11…”. Claim 12 is currently examined, but the Examiner requests that this be remedied.
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.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim 1 is 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 contains the claim language of “this process is repeated for all EEG electrode positions over the scalp” in pg. 1 of the Claims (in li. 3 of the paragraph beginning with “c22)”). It is uncertain exactly what “process” this language is referring to. Is this in reference to the methodology described in step c22)? Or perhaps in reference to the methodology described in step c21) or c22)? For the purposes of examination, the Examiner has chosen to interpret this as – step c22) is repeated for all EEG electrode positions over the scalp –.
For the above-cited limitation "this process is repeated for all EEG electrode positions over the scalp" in claim 1, there is additionally insufficient antecedent basis for this limitation in the claim.
Claim Rejections - 35 USC § 102
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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claim 8 is rejected under 35 U.S.C. 102(a)(2) as being anticipated by Kremen et al. (US 2020/0337645, hereinafter referred to as Kremen).
Regarding amended, independent claim 8, Kremen discloses a computer program stored on a non-transitory computer-readable medium (a method 300 in Fig. 3 for classifying the behavioral state of the brain; [0035]: “Note that the method 300 can be implemented using a non-transitory computer readable medium that when executed causes the one or more processors to perform the method.”) for training a neurological condition detection algorithm to be used for neurological condition detection in an implantable neurostimulation device (implanted device 900 in Fig. 9 ; [0030]: “…machine learning and/or supervision can be used to configure, adjust, fine tune or otherwise operate the system in block 212 [in Fig. 2].”; [0066]: “The control algorithm in this case would use behavioral state classifications determined from EEG or other sensors as the input and electrical stimulation is used to modulate and drive the brain to the prescribed state. This algorithm can run on the implanted device…”) having a target electrode arrangement ([0031], [0037]: “…one or more suitable sensor configurations…”), the computer program comprising the following steps:
a) inputting EEG data in a computer ([0028]: “The user interface 104 can be any component or device in which information is transmitted to and/or received from the one or more processors 108. …the user interface 104 can be a connector in which a separate device can be attached, a network interface, a wireless transceiver, a display, a keyboard, a keypad, etc. For example, …computer, can wirelessly connect to and control/monitor the system 102 via the user interface 104.”) which executes the computer program ([0066]: “The control algorithm in this case would use behavioral state classifications determined from EEG or other sensors as the input and electrical stimulation is used to modulate and drive the brain to the prescribed state.”), the EEG data being recorded by at least one EEG from at least one patient (multi-channel iEEG data 402, 418 in Fig. 4) using an electrode system with a plurality of electrode channels ([0062]: “…selects one electrodes from array of available electrodes…”),
b) identifying neurological activity in the EEG data ([0006]; [0029]: “The one or more processors 108 receive a signal from each of a plurality of sensors 118 via the sensor and/or electrode interface 102. The sensors 118 are configured to detect an electrical activity of the brain.”), which corresponds to a neurological condition ([0030]: “…behavioral state classification is performed in block 208 [in Fig. 2]…”; [0032]: “The one or more processors 108 can pre-process the signals by: detecting an abnormal amplitude distortion in the signals; or detecting a seizure or an abnormal electrophysiological condition using the signals; or detecting a high 60 or 50 Hz line interference in the signals; or other desired process.”; [0036]), based upon neurological condition identification tags included in the EEG data and/or input in the computer ([0062]: “The foregoing testing investigated behavioral state classification (wake & N2, and slow wave sleep) using intracranial EEG spectral power features in an unsupervised machine learning method. The method automatically selects one electrodes from array of available electrodes based on unsupervised score of the data and deploys cascade of classifiers using features extracted from selected electrode to classify into AW, N2, and N3 stages.”),
c) selecting a subset of electrode channels out of the available electrode channels in the EEG data ([0032]: “…the one or more processors 108 can select or restrict a number of channels of the sensors 118…”; [0038]; [0052]: “The method 400 is also referred to as the Behavioral State Classifier (BSC). Sleep scoring and multi-channel iEEG data 402 are used in subsequent automated steps for feature extraction or selection (if turned on). The single electrode assesed to yield in the best performance is selected and used for classification and classifier uses features extracted from the selected electrode and then supplied to inputs of hierarchical clustering methods that returns an AW, N2, and N3 classification. The user can select and restrict the number of channels, number of features, and whether automated feature selection and electrode selection is used.”; [0064]: “…FIG. 8 shows a training method 800 and an application of trained classifier 850. In training 800, scalp EEG data 802 and multi-channel iEEG data 804 is provided for manual sleep scoring (e.g., 10 min Awake, 10 min SWS) 806. Automated classification 808 uses the score 810 and iEEG data 812 with channel constriction 814 for feature extraction selection 816.”) depending ci) on the identified neurological activity and/or c2) on characteristic data of the target electrode arrangement ([0062]: “…selects one electrodes from array of available electrodes based on unsupervised score of the data and deploys cascade of classifiers using features extracted from selected electrode to classify into AW, N2, and N3 stages.”),
d) training a neurological condition detection algorithm by using the EEG data only of the selected subset of electrode channels ([0031]: “The one or more processors 108 can also use training signal processing and a machine learning system to identify one or more suitable sensor configurations for an automated or semi-automated classification of the behavioral state.”), and wherein the computer program comprises at least two training cycles of the neurological condition detection algorithm (method 400 in Fig. 4 for classifying the behavioral state of the brain):
e) in a first training cycle a general training of the neurological condition detection algorithm (supervised version 406 in Fig. 4) is done using the EEG data of one or more patients to form a pre-trained algorithm comprising multiple layers ([0031]: “…trained on known scalp electrophysiology data in parallel with any simultaneous data (e.g. intracranial, epidural, subscalp, EEG, video recording, EMG, actigraphy, etc.)…”, emphasis added; [0052]: “The bottom part 406 of Figure 4 shows how another user input and/or a supervision and active learning can be implemented if needed or if training data are being available… In the supervised version 406, multi-channel scalp EEG data 418 is processed using gold standard sleep scoring (AW, N1, N2, N3, REM) 420 to define and select the features, and/or automated sleep scoring (AW, N2, N3)N3) 422 from the cascade classification is used to define and select the features.”; The Examiner notes that “gold standard” denotes data that is “pre-trained” into an algorithm by representing the general data from one or more patients.; The Examiner further notes that supervised version 406 is described with multiple layers, as it has an input layer via EEG data 418 and gold standard scoring 420, and an output layer as shown by the arrow going from box 406 to box 404 in Fig. 4.), and
f) in a second training cycle a patient specific training of the neurological condition detection algorithm (unsupervised version 404 in Fig. 4) is done using the EEG data only of the patient to which the neurological condition detection algorithm shall be applied ([0052]: “…multi-channel iEEG data 402 …”) and/or using the EEG data of another patient having similar neurological condition onset pattern as the patient to which the neurological condition detection algorithm shall be applied ([0031]: “…trained on known scalp electrophysiology data in parallel with any simultaneous data (e.g. intracranial, epidural, subscalp, EEG, video recording, EMG, actigraphy, etc.)…”, emphasis added; [0052:]: “The top part 404 of Figure 4 shows its unsupervised method that was trained and tested here in the study and it doesn't require training on gold standard data for each patient. …Here a day and night of multichannel iEEG recording were used as an input of the method for each patient.”), wherein the second training cycle comprises fine-tuning only a subset of the multiple layers of the pre-trained algorithm (The Examiner further notes that supervised version 404 is described with multiple layers, as it has an input layer via EEG data 402 and user inputs and options 414, and output layers as shown by electrode selection 426, cascade classification 428, behavior state classification 444, 446, and as shown by the arrow going from box 404 to box 406 in Fig. 4.) while the remaining layers of the pre-trained algorithm are frozen ([0031] and Fig. 4 show that the two modes/cycles 404, 406 applied together as two sequential cycles, with the arrows depicting adjustment or “fine-tuning” to mode 404 from mode 406 data.; [0053]: “Definition of features and their selection is optional and can be configured in unsupervised version 404 and supervised version 406, and can be automatically adjust when gold standard data are present.”; The Examiner notes that only parts (of the layers) of supervised version 406 are influenced or “fine-tuned” by the unsupervised version 404, as evidenced by cascade classification 428 and behavior state classification 444, 446 acting as inputs to automated sleep scoring 422 in Fig. 4, while the EEG data 418 and gold standard scoring 420 are “frozen”.).
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, 5, 7, and 9-18 are rejected under 35 U.S.C. 103 as being unpatentable over Kremen in view of Remmert (US 2018/0117308) and further in view of Lee et al. (US 2011/0137381, hereinafter referred to as Lee) and Guger et. al (US 2019/0082992, hereinafter referred to as Guger).
Regarding independent claim 1, Kremen discloses tracking human brain activity, and more particularly, to a system and method for classifying and modulating brain behavioral states ([0002]). Kremen further discloses a computer program stored on a non-transitory computer-readable medium (a method 300 in Fig. 3 for classifying the behavioral state of the brain; [0035]: “Note that the method 300 can be implemented using a non-transitory computer readable medium that when executed causes the one or more processors to perform the method.”) for training a neurological condition detection algorithm to be used for neurological condition detection ([0024]; [0030]: “…in FIG. 2, the sensor signals/data are received in block 202, the sensor signals/data are preprocessed in block 204 (optional), one or more sensors are selected in block 206, and behavioral state classification is performed in block 208. A user or other device can be used to configure the system 100 in block 210. In addition, machine learning and/or supervision can be used to configure, adjust, fine tune or otherwise operate the system in block 212.”) in an implantable neurostimulation device (implanted device 900 in Fig. 9 ; [0030]: “…machine learning and/or supervision can be used to configure, adjust, fine tune or otherwise operate the system in block 212 [in Fig. 2].”; [0066]: “The control algorithm in this case would use behavioral state classifications determined from EEG or other sensors as the input and electrical stimulation is used to modulate and drive the brain to the prescribed state. This algorithm can run on the implanted device…”) having a target electrode arrangement ([0031], [0037]: “…one or more suitable sensor configurations…”), the computer program comprising the following steps:
a) inputting EEG data in a computer ([0028]: “The user interface 104 can be any component or device in which information is transmitted to and/or received from the one or more processors 108. …the user interface 104 can be a connector in which a separate device can be attached, a network interface, a wireless transceiver, a display, a keyboard, a keypad, etc. For example, …computer, can wirelessly connect to and control/monitor the system 102 via the user interface 104.”) which executes the computer program ([0066]: “The control algorithm in this case would use behavioral state classifications determined from EEG or other sensors as the input and electrical stimulation is used to modulate and drive the brain to the prescribed state.”), the EEG data being recorded by at least one EEG (multi-channel iEEG data 402, 418 in Fig. 4) from at least one patient,
b) identifying neurological activity in the EEG data ([0006]; [0029]: “The one or more processors 108 receive a signal from each of a plurality of sensors 118 via the sensor and/or electrode interface 102. The sensors 118 are configured to detect an electrical activity of the brain.”), which corresponds to a neurological condition ([0030]: “…behavioral state classification is performed in block 208 [in Fig. 2]…”; [0032]: “The one or more processors 108 can pre-process the signals by: detecting an abnormal amplitude distortion in the signals; or detecting a seizure or an abnormal electrophysiological condition using the signals; or detecting a high 60 or 50 Hz line interference in the signals; or other desired process.”; [0036]), based upon neurological condition identification tags comprising information identifying the time and location of a neurological condition ([0030]: “…automatically map one or more spatial and temporal patterns of the classified behavioral state…”) included in the EEG data and/or input in the computer ([0062]: “The foregoing testing investigated behavioral state classification (wake & N2, and slow wave sleep) using intracranial EEG spectral power features in an unsupervised machine learning method. The method automatically selects one electrodes from array of available electrodes based on unsupervised score of the data and deploys cascade of classifiers using features extracted from selected electrode to classify into AW, N2, and N3 stages.”),
c) selecting a subset of five electrode channels out of the available electrode channels in the EEG data ([0032]: “…the one or more processors 108 can select or restrict a number of channels of the sensors 118…”; [0038]; [0052]: “The method 400 is also referred to as the Behavioral State Classifier (BSC). Sleep scoring and multi-channel iEEG data 402 are used in subsequent automated steps for feature extraction or selection (if turned on). The single electrode assesed to yield in the best performance is selected and used for classification and classifier uses features extracted from the selected electrode and then supplied to inputs of hierarchical clustering methods that returns an AW, N2, and N3 classification. The user can select and restrict the number of channels, number of features, and whether automated feature selection and electrode selection is used.”; [0064]: “…FIG. 8 shows a training method 800 and an application of trained classifier 850. In training 800, scalp EEG data 802 and multi-channel iEEG data 804 is provided for manual sleep scoring (e.g., 10 min Awake, 10 min SWS) 806. Automated classification 808 uses the score 810 and iEEG data 812 with channel constriction 814 for feature extraction selection 816.”) depending
i) on the identified neurological activity ([0062]: “…selects one electrodes from array of available electrodes based on unsupervised score of the data and deploys cascade of classifiers using features extracted from selected electrode to classify into AW, N2, and N3 stages.”) and
ii) on characteristic data of the target electrode arrangement ([0038]: “For example, the one or more selection criteria can be a K-NN clustering algorithm with Euclidean distance measure where inter and intra-cluster distance are used as parameters for selection of only one sensor.”),
wherein electrode selection is performed to obtain an electrode set with the maximum number of electrodes covering a seizure onset zone at inter-electrode distances ([0031]: “The one or more processors 108 can also use training signal processing and a machine learning system to identify one or more suitable sensor configurations for an automated or semi-automated classification of the behavioral state. Moreover, the one or more processors 108 select target brain locations for the sensors from one or more of a cortex, hippocampus, thalamus, brain stem, basal ganglia, subthalamic nucleus, globus pallidus or other movement circuitry structures and muscles via EMG or ENG or actigraphy.”)
d) training a neurological condition detection algorithm by using the EEG data only of the selected subset of electrode channels ([0031]: “The one or more processors 108 can also use training signal processing and a machine learning system to identify one or more suitable sensor configurations for an automated or semi-automated classification of the behavioral state.”).
Kremen is silent to having five electrodes which are arranged in a pseudo-Laplacian pattern having a center electrode and four stimulation electrodes which surround the center electrode; a)… the EEG data being recorded by at least one EEG from at least one patient using a 10-20 or 10-10 or any other high density EEG electrode system; and wherein electrode selection is performed to obtain an electrode set with the maximum number of electrodes covering a seizure onset zone at inter-electrode distances below a threshold mapping to the pseudo-Laplacian pattern design of the implantable system, wherein
c21) the mean of five selected electrode coordinates is calculated and the nearest scalp electrode to this position is found, wherein this electrode is considered as the center electrode, and any of the selected five electrodes that have a distance from the center electrode greater than the threshold distance is excluded from the list, or
c22) one electrode is selected as the central electrode and the number of electrodes from the initial seizure onset zone electrodes whose distance from the central electrode is smaller than the threshold distance is counted;
this process is repeated for all EEG electrode positions over the scalp, and list of the selected initial seizure onset zone electrodes enclosed with each electrode is generated;
the electrode that contained the maximum number of initial seizure onset zone electrodes is chosen as the central electrode;
if required, electrodes with a minimum distance from the central electrode are added to the list to yield exactly five electrodes for seizure detection.
However, Remmert teaches an electrode for the electrical stimulation of brain tissue or other tissue of a patient is configured for location between skull and scalp of the patient. Remmert further teaches having five electrodes which are arranged in a pseudo-Laplacian pattern having a center electrode and four stimulation electrodes which surround the center electrode ([0053]: “FIG. 2 shows a preferred pseudo-Laplacian arrangement of the electrodes of the neurostimulation system. An electrode pad 2 comprises a stimulation electrode 20 and for secondary electrodes 21, 22, 23, 24.”).
Remmert teaches a similar pursuit to that of the instant application and Kremen in teaching an implantable neurostimulator adapted for a particular treatment, such as epilepsy. It would have been obvious to one having ordinary skill in the art at the effective filing date of the invention to modify the invention of Kremen to further include electrodes configured in a pseudo-Laplacian pattern in order to focalize stimulation and provide optimal treatment to a patient experiencing the onset of a seizure.
The Kremen/Remmert combination is silent to:
a)… the EEG data being recorded by at least one EEG at least one patient using a 10-20 or 10-10 or any other high density EEG electrode system; and
c)… wherein electrode selection is performed to obtain an electrode set with the maximum number of electrodes covering a seizure onset zone at inter-electrode distances below a threshold mapping to the pseudo-Laplacian pattern design of the implantable system, wherein
c21) the mean of five selected electrode coordinates is calculated and the nearest scalp electrode to this position is found, wherein this electrode is considered as the center electrode, and any of the selected five electrodes that have a distance from the center electrode greater than the threshold distance is excluded from the list, or
c22) one electrode is selected as the central electrode and the number of electrodes from the initial seizure onset zone electrodes whose distance from the central electrode is smaller than the threshold distance is counted;
this process is repeated for all EEG electrode positions over the scalp, and list of the selected initial seizure onset zone electrodes enclosed with each electrode is generated;
the electrode that contained the maximum number of initial seizure onset zone electrodes is chosen as the central electrode;
if required, electrodes with a minimum distance from the central electrode are added to the list to yield exactly five electrodes for seizure detection.
However, Lee teaches the prevention and/or treatment of neurological disorders via electrical stimulation. Lee further teaches:
a)… the EEG data being recorded by at least one EEG at least one patient using a 10-20 or 10-10 or any other high density EEG electrode system ([0135]-[0141]; [0136]-[0137]: “As an example, shown in FIG. 10A, is an array 11 of 16.times.14 equally spaced disc electrodes… With a switching network, any combination of electrodes in the n.times.m array can be connected together.”); and
c) …wherein electrode selection is performed to obtain an electrode set with the maximum number of electrodes covering a seizure onset zone at inter-electrode distances below a threshold mapping to the pseudo-Laplacian pattern design of the implantable system ([0106]: “If the position of each electrode is known, the data can be processed to yield a 3-dimensional map of brain electrical activity. Using this map, the appropriate electrodes can be energized and the areas of the brain to be treated can be limited to only those areas in which abnormal electrical activity is present.”).
Lee teaches a similar pursuit to that of the instant application and the Kremen/Remmert in teaching an implantable neurostimulator adapted for a particular treatment, such as epilepsy. It would have been obvious to one having ordinary skill in the art at the effective filing date of the invention to combine the neurological condition detection algorithm of Kremen and pseudo-Laplacian electrode configuration of Remmert with Lee’s high density EEG electrode system and “Dynamic Adaptation of Targeting Disc-Array Electrodes” ([0142]-[0146] and Fig. 11 of Lee) to establish the optimum stimulation locations of a patient in order to provide therapy to prevent the onset or escalation of seizures.
The Kremen/Remmert/Lee combination is silent to the specific algorithmic steps of wherein
c21) the mean of five selected electrode coordinates is calculated and the nearest scalp electrode to this position is found, wherein this electrode is considered as the center electrode, and any of the selected five electrodes that have a distance from the center electrode greater than the threshold distance is excluded from the list, or
c22) one electrode is selected as the central electrode and the number of electrodes from the initial seizure onset zone electrodes whose distance from the central electrode is smaller than the threshold distance is counted;
this process is repeated for all EEG electrode positions over the scalp, and list of the selected initial seizure onset zone electrodes enclosed with each electrode is generated;
the electrode that contained the maximum number of initial seizure onset zone electrodes is chosen as the central electrode;
if required, electrodes with a minimum distance from the central electrode are added to the list to yield exactly five electrodes for seizure detection.
However, Guger teaches a method and an apparatus provide electro stimulation to a test subject. Guger further teaches wherein
c21) the mean of five selected electrode coordinates is calculated and the nearest scalp electrode to this position is found ([0058]: “The mean value can be ascertained by averaging the electrodes 21u directly adjoining the respective electrode 21z (FIG. 3). In this case, the filter value is computed by subtracting the total of the measured signal values of the adjacent electrodes divided by 4 from the measured signal value. The value thus ascertained substantially corresponds to the discretely ascertained Laplace operator or a multiple of the discretely ascertained Laplace operator.”), wherein this electrode is considered as the center electrode (Figs. 3-5 show a center electrode 21z in proximity to individual adjacent electrodes 21u.; [0057]: “…can be utilized in order to suppress effects in the surroundings around an electrode 21z.”), and any of the selected five electrodes that have a distance from the center electrode greater than the threshold distance is excluded from the list ([0052]: “…individual electrodes 21, at which special properties have been established in the signals on the basis of the analysis, are selected for the emission of a stimulus S.”), or
c22) one electrode is selected as the central electrode and the number of electrodes from the initial seizure onset zone electrodes whose distance from the central electrode is smaller than the threshold distance is counted (Figs. 3-5; [0056]-[0060]);
this process is repeated for all EEG electrode positions over the scalp ([0062]: “The analysis unit 13 is designed for the purpose of processing continuously derived and at best filtered measured values for each individual electrode 21, wherein measured values produced within a predefined period of time are combined into windows (FIG. 6).”), and list of the selected initial seizure onset zone electrodes enclosed with each electrode is generated (such as via display unit 141 in Fig. 8; [0011]: “A particularly simple overview of the analysis results is achieved by the selection and actuating unit having a display unit, which represents the electrodes and also the analysis results ascertained by the analysis unit, in particular the preselection results, on the basis of the analysis for the individual electrodes at positions of the display unit graphic visualizations of the analysis results.”; [0014]: “…the selection or actuating elements each being associated with one electrode and being arranged on the display unit in the region of the position at which the graphic visualizations for the relevant electrode are represented…”); the electrode that contained the maximum number of initial seizure onset zone electrodes is chosen as the central electrode ([0059]: “…the possibility also exists of using the eight adjacent electrodes 21u′ surrounding an electrode 21z for ascertaining the mean value in a square electrode grid (FIG. 4). In this case, those adjacent electrodes 21u′ which are located diagonally in relation to the central electrode 21z can be weighted with a lesser weighting factor. In particular, this weighting factor can be dependent on the spacing of the adjacent electrodes, and therefore diagonally located adjacent electrodes 21u′ are weighted more weakly by a factor of 1 divided by √2 than directly adjoining adjacent electrodes 21u′”); if required, electrodes with a minimum distance from the central electrode are added to the list to yield exactly five electrodes for seizure detection ([0060]: “… instead of the four electrodes 21u directly adjoining the electrode 21z, those four electrodes 21u″ within a square grid for the determination of the mean value, the one coordinate position of which deviates from the relevant coordinate position of the electrode 21z by two, and the other coordinate positions of which corresponds to the relevant coordinate position of the middle electrode (FIG. 5).”; [0075]: “… a selection and actuating unit 14, using which one or more electrodes 21 among the electrodes preselected by the analysis unit can be selected to emit a predefined electrical stimulus S. …the selection and actuating unit 14 has a display unit 141, which represents the electrodes 21 and the analysis results ascertained by the analysis unit 13, in particular in the present case the preselection results ascertained on the basis of the correlation coefficient k, preferably by threshold value comparison, on the basis of the analysis for the individual electrodes 21 at positions 142 of the display unit in the form of graphic visualizations 143a, 143b. The illustrated selection and actuating unit 14 has individual selection and actuating elements 144 for each individual electrode 21 in the region of the display unit 141. The selection and actuating elements 144 are each associated with one electrode 21 and are arranged on the display unit 141 in the region of the position 142, at which the graphic visualizations 143a, 143b for the relevant electrode 21 are also shown. … A particularly preferred selection of electrodes 21 by the selection and actuating unit 144 is carried out by also selecting an adjacent electrode 21 upon selection of a respective electrode 21. …If one electrode 21 is preselected by the selection and actuating unit 14 as a result of the analysis, the selection and actuating unit 14 thus proposes an electrode 21 adjacent to this selected electrode 21 for selection or selects it itself. …It is also advantageous that multiple electrodes 21 can be stimulated simultaneously or in rapid succession, in order to amplify the effect and make the mapping faster.”; see also Fig. 8).
Guger is of a similar pursuit to the instant application in teaching selection of the electrode configuration of an implantable neurostimulator. Therefore, it would have been obvious to one having ordinary skill in the art at the effective filing date of the invention to combine the algorithmic electrode selection steps of Guger with the Kremen/Remmert/Lee combination in order to optimize the treatment to a patient suffering from seizures.
Regarding claim 2, in view of the combination set forth in claim 1, Kremen discloses that such electrode channels are selected out of the available electrode channels ([0062]: “…selects one electrodes from array of available electrodes based on unsupervised score of the data and deploys cascade of classifiers using features extracted from selected electrode to classify into AW, N2, and N3 stages.”) which are in closest proximity to the location of the identified neurological activity which corresponds to the neurological condition ([0031]: “The one or more processors 108 can also use training signal processing and a machine learning system to identify one or more suitable sensor configurations for an automated or semi-automated classification of the behavioral state. Moreover, the one or more processors 108 select target brain locations for the sensors from one or more of a cortex, hippocampus, thalamus, brain stem, basal ganglia, subthalamic nucleus, globus pallidus…”).
Regarding amended claim 3, in view of the combination set forth in claim 1, Kremen discloses that such electrode channels are selected out of the available electrode channels ([0062]: “…selects one electrodes from array of available electrodes…”) which have the closest geometrical match with the electrodes of the target electrode arrangement ([0038]: “For example, the one or more selection criteria can be a K-NN clustering algorithm with Euclidean distance measure where inter and intra-cluster distance are used as parameters for selection of only one sensor.”).
Regarding claim 5, in view of the combination set forth in claim 1, Kremen discloses that step d) ([0031]: “The one or more processors 108 can also use training signal processing and a machine learning system to identify one or more suitable sensor configurations for an automated or semi-automated classification of the behavioral state.”) comprises the steps:
d2) training the neurological condition detection algorithm ([0030]: “…machine learning and/or supervision can be used to configure, adjust, fine tune or otherwise operate the system in block 212 [in Fig. 2].”; [0066]: “The control algorithm in this case would use behavioral state classifications determined from EEG or other sensors as the input and electrical stimulation is used to modulate and drive the brain to the prescribed state. This algorithm can run on the implanted device…”) by using calculated linear combinations of the EEG data.
The Kremen/Remmert combination is silent to d1) calculating linear combinations of the EEG data of the selected subset of electrode channels, e.g. calculating linear combinations representing bipolar or quadrupolar electrode channels.
However, Lee teaches d1) calculating linear combinations of the EEG data of the selected subset of electrode channels ([0103]; Fig. 11), e.g. calculating linear combinations representing bipolar or quadrupolar electrode channels ([0078]: “FIG. 2K illustrates an elliptical tripolar ring electrode or tripolar Laplace electrode.”; [0081]: “…the ring electrodes may not function as Laplace but rather function as multi bipolar electrodes.”).
It would have been obvious to one having ordinary skill in the art at the effective filing date of the invention to modify the Kremen/Remmert combination with calculating linear combinations of the EEG data of the selected subset of electrode channels and additionally using such calculations in step d2) in order to refine the spatial signal resolution of the electrode pattern, without needing further hardware electrode channels.
Regarding claim 7, in view of the combination set forth in claim 1, Kremen discloses that the neurological condition detection algorithm ([0030]: “…machine learning and/or supervision can be used to configure, adjust, fine tune or otherwise operate the system in block 212 [in Fig. 2].”; [0066]: “The control algorithm in this case would use behavioral state classifications determined from EEG or other sensors as the input and electrical stimulation is used to modulate and drive the brain to the prescribed state. This algorithm can run on the implanted device…”) is an artificial intelligence algorithm (Support Vector Machine 858 in Fig. 8), e.g. Random Forest, Support Vector Machine ([0064]: “A single lead is selected as an input in block 854, feature vectors are calculated in block 856 and a support vector machine 858 is used in block 858 to classify the behavior state as Awake or SWS 860.”), Multi-layer Perceptron, Convolutional Neural Network, Long Short-Term Memory Network.
Regarding claim 9, in view of the combination set forth in claim 1, Kremen discloses that the computer program is arranged for evaluating data tags which are assigned to the EEG data which are input in the computer which executes the computer program (a method 300 in Fig. 3 for classifying the behavioral state of the brain; [0035]: “Note that the method 300 can be implemented using a non-transitory computer readable medium that when executed causes the one or more processors to perform the method.”), wherein the data tags are used for selecting the subset of electrode channels out of the available electrode channels ([0062]: “The foregoing testing investigated behavioral state classification (wake & N2, and slow wave sleep) using intracranial EEG spectral power features in an unsupervised machine learning method. The method automatically selects one electrode from array of available electrodes based on unsupervised score of the data and deploys cascade of classifiers using features extracted from selected electrode to classify into AW, N2, and N3 stages.”).
Regarding claim 10, in view of the combination set forth in claim 1, Kremen discloses a method of programming an implantable neurostimulation device (implanted device 900 in Fig. 9), comprising the following steps:
a) running a computer program of claim 1 on a computer (a method 300 in Fig. 3 for classifying the behavioral state of the brain; [0035]: “Note that the method 300 can be implemented using a non-transitory computer readable medium that when executed causes the one or more processors to perform the method.”),
b) programming the neurological condition detection algorithm trained by the computer program into the implantable neurostimulation device (implanted device 900 in Fig. 9; [0030]: “…machine learning and/or supervision can be used to configure, adjust, fine tune or otherwise operate the system in block 212 [in Fig. 2].”; [0066]: “The control algorithm in this case would use behavioral state classifications determined from EEG or other sensors as the input and electrical stimulation is used to modulate and drive the brain to the prescribed state. This algorithm can run on the implanted device…”).
Regarding claim 11, in view of the combination set forth in claim 1, Kremen discloses that the implantable neurostimulation device (implanted device 900 in Fig. 9 ) is a closed-loop neurostimulator which is arranged for recording EEG signals, for calculating stimulation signals based upon the recorded EEG signals and for outputting the stimulation signals ([0065]: “The technology for brain state (behavioral state) determination and tracking described above can be used to dynamically follow and modulate brain state using electrical stimulation. As illustrated in FIGS. 9 and 10, the closed-loop system can modulate sleep and wake states using electrical stimulation, and fine-tune overall sleep-wake dynamics to meet any desired, pre-determined behavioral state patterns by tracking behavioral state and modulating via a control algorithm.”).
Regarding claim 12, in view of the combination set forth in claim 1, Kremen discloses that in step g) the computer program is run on an external computer ([0034]: “The system may also include a remote device communicably coupled to the one or more processors 108, in which the one or more processors 108 transmit the classified behavioral state to the remote device, and receive one or more control signals for the electrical stimulation from the remote device. The remote device can be a handheld device, a cloud computing resource, a computer or any other type of control or processing device.”) which is not part of the implantable neurostimulation device (implanted device 900 in Fig. 9).
Regarding claim 13, in view of the combination set forth in claim 1, Kremen discloses in which a neurological condition detection algorithm or classifier for detecting neurological conditions from EEG data has been trained and/or is being trained by a computer program (a method 300 in Fig. 3 for classifying the behavioral state of the brain; [0035]: “Note that the method 300 can be implemented using a non-transitory computer readable medium that when executed causes the one or more processors to perform the method.”; [0064]: “…FIG. 8 shows a training method 800 and an application of trained classifier 850. In training 800, scalp EEG data 802 and multi-channel iEEG data 804 is provided for manual sleep scoring (e.g., 10 min Awake, 10 min SWS) 806. Automated classification 808 uses the score 810 and iEEG data 812 with channel constriction 814 for feature extraction selection 816.”).
Regarding claim 14, in view of the combination set forth in claim 1, Kremen discloses that the computer program is configured for implementation on a microcontroller (“one or more processors” further described in [0008]).
Regarding claim 15, in view of the combination set forth in claim 1, the Kremen/Remmert combination is silent to that the computer program is optimized for lowest power consumption.
However, Lee in teaching “All other electrodes in the array are not used and left uncharged or neutral.” ([0137]) would inherently optimize the power consumption of the system, as minimizing the number of electrodes charged in an array reduces power usage. Therefore, it would have been obvious to one having ordinary skill in the art at the effective filing date of the invention that the Kremen/Remmert/Lee combination teaches that the computer program is optimized for lowest power consumption.
Regarding claim 16, in view of the combination set forth in claim 1, Kremen discloses that the neurological condition detection algorithm or classifier for detecting neurological conditions ([0030]: “…machine learning and/or supervision can be used to configure, adjust, fine tune or otherwise operate the system in block 212 [in Fig. 2].”; [0066]: “The control algorithm in this case would use behavioral state classifications determined from EEG or other sensors as the input and electrical stimulation is used to modulate and drive the brain to the prescribed state. This algorithm can run on the implanted device…”) is an artificial intelligence algorithm, e.g. Random Forest, Support Vector Machine ([0064]: “A single lead is selected as an input in block 854, feature vectors are calculated in block 856 and a support vector machine 858 is used in block 858 to classify the behavior state as Awake or SWS 860.”), Multi-layer Perceptron, Convolutional Neural Network, Long Short-Term Memory Network.
Regarding amended claim 17, in view of the combination set forth in claim 1, Kremen discloses that the computer program (a method 300 in Fig. 3 for classifying the behavioral state of the brain; [0035]: “Note that the method 300 can be implemented using a non-transitory computer readable medium that when executed causes the one or more processors to perform the method.”) is configured to run on an implantable neurostimulation device (implanted device 900 in Fig. 9).
Regarding claim 18, in view of the combination set forth in claim 1, Kremen discloses a method of treatment of a neurological condition in a subject (a method 300 in Fig. 3 for classifying the behavioral state of the brain; [0035]: “Note that the method 300 can be implemented using a non-transitory computer readable medium that when executed causes the one or more processors to perform the method.”), comprising implanting the implantable neurostimulation device according to claim 17 into the subject (implanted device 900 in Fig. 9; [0067]: “FIG. 9 is an embodiment and application of proposed system that integrates an implanted device 900 with brain electrodes and peripheral nerve electrodes 902 that provides both sensing and electrical stimulation and couples this capability with a bi-directional connectivity with a handheld device 904 and cloud computing environment 906.”).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MARY G SCHLUETER whose telephone number is (703)756-4601. The examiner can normally be reached M-F 9:00am-5:30pm 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, Unsu Jung can be reached at (571) 272-8506. 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.S./Examiner, Art Unit 3796
/LYNSEY C Eiseman/Primary Examiner, Art Unit 3796
Prosecution Insights