Ex Parte Kidmose et al

Patent Trials and Appeals Board

Mar 15, 2019

13335901 - (D) (P.T.A.B. Mar. 15, 2019)

UNITED STA TES p A TENT AND TRADEMARK OFFICE
APPLICATION NO. FILING DATE
13/335,901 12/22/2011
23373 7590 03/19/2019
SUGHRUE MION, PLLC
2100 PENNSYLVANIA A VENUE, N.W.
SUITE 800
WASHINGTON, DC 20037
FIRST NAMED INVENTOR
Preben Kidmose
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O. Box 1450
Alexandria, Virginia 22313-1450
www .uspto.gov
ATTORNEY DOCKET NO. CONFIRMATION NO.
Ql28176 4746
EXAMINER
BERHANU, ETSUB D
ART UNIT PAPER NUMBER
3791
NOTIFICATION DATE DELIVERY MODE
03/19/2019 ELECTRONIC
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
Notice of the Office communication was sent electronically on above-indicated "Notification Date" to the
following e-mail address(es):
PPROCESSING@SUGHRUE.COM
sughrue@sughrue.com
USPTO@sughrue.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte PREBEN KIDMOSE and
SOREN ERIK WESTERMANN
Appeal2017-005655
Application 13/335,901
Technology Center 3700
Before BENJAMIN D. M. WOOD, ANNETTE R. REIMERS, and
SEAN P. O'HANLON, Administrative Patent Judges.
WOOD, Administrative Patent Judge.
DECISION ON APPEAL
STATEMENT OF THE CASE
Appellants appeal under 35 U.S.C. § 134(a) from a rejection of claims
1-16 and 18-21. An oral hearing in accordance with 37 C.F.R. § 41.47 was
held on March 7, 2019. We have jurisdiction under 35 U.S.C. § 6(b).
We reverse.
THE INVENTION
The claims are directed to a wearable system for monitoring a diabetic
patient's brain activity (i.e., an EEG monitoring system) that detects
imminent hypoglycemic attacks. Spec. 1-2. Claim 1, reproduced below
Appeal2017-005655
Application 13/335,901
(and reformatted slightly for clarity), is illustrative of the claimed subject
matter:
Drew
1. A portable EEG monitoring system, said system
compnsmg:
electrodes for measuring at least one EEG
signal from a person carrying the EEG monitoring
system,
a signal processor adapted to receive, process and
analyze at least a part of said EEG signal, wherein
said signal processor comprises
a feature extractor for extracting a feature
vector from said EEG signal,
a classifier adapted for monitoring said feature
vector for identifying an event and outputting an
event signal to
an event integrator adapted to integrate the
event signals over time, in order to produce an
event level signal,
a data logger adapted to log data relating to
said EEG signal, including at least one feature
vector extracted from said EEG signal, and
a memory for storing said data relating to said
EEG signal;
wherein
said data logger includes a circular buffer
for storing data over no more than a
predetermined period of time, and
said memory stores data logged in said
data logger, including said feature vector,
when said event level signal exceeds an event
detection threshold.
Beck-Nielsen
REFERENCES
US 2006/0094972 Al
WO 2007 /144307 A2
May 4, 2006
Dec. 21, 2007
2
Appeal2017-005655
Application 13/335,901
REJECTION
Claims 1-16 and 18-21 are rejected under pre-AIA 35 U.S.C.
§ 103(a) as unpatentable over Beck-Nielsen and Drew.
ANALYSIS
Independent claims 1 and 7 recite an EEG monitoring system
comprising, inter alia: (I) a data logger adapted to log data relating to a
received EEG signal, including at least one feature vector; and (2) a memory
that stores data logged in said data logger, including the feature vector,
"when said event level signal exceeds an event detection threshold." We
interpret the claims to require the data logger and memory to be separate
structures. See Engel Indus., Inc. v. Locliformer Co., 96 F.3d 1398, 1404---05
(Fed. Cir. 1996) ( construing separate claim terms as reciting separate
structures). Thus, claims 1 and 7 require the EEG system to be adapted to
store a feature vector twice: once in the data logger, and again in a separate
memory "when said event level signal exceeds an event detection
threshold."
The Examiner finds that Beck-Nielsen teaches both of these
limitations. Final Act. 2 ( citing, e.g., Beck-Nielsen, 18:6-22). According to
the Examiner, Beck-Nielsen inherently stores feature vectors because "the
feature vectors would need to be logged/stored in a memory in order to
perform classification of the EEG signals." Id. at 5. The Examiner further
asserts that "event detections must be stored ... as an integration of the
event detections is performed over a lengthy period of time (300/600/900)
seconds." Id. at 6. The Examiner thus determines that "[a] prior art
reference which teaches continuously storing a feature vector reads on the
claims because a feature vector would inherently be stored (by virtue of
3
Appeal2017-005655
Application 13/335,901
always being stored) when an event level signal exceeds an event detection
threshold." Id. at 5.
Appellants dispute that Beck-Nielsen teaches storing a feature vector
when an event level signal exceeds an event detection threshold. Appellants
argue that
[ w ]hile it may well be true that a device/system that stores data
constantly, including after an event level is exceeded, could be
covered by [ claims 1 and 7], this would only be the case if, in
addition to the continuous storage, there is some storage of a
feature vector that is actually in response to the event threshold
being exceeded.
App. Br. 9. Appellants also dispute that Beck-Nielsen's system
continuously stores feature vectors in a memory that is separate from a data
logger. Id. at 10.
The Examiner premises the rejection on the finding that Beck-
Nielsen's system continuously stores feature vectors in a memory. We agree
with Appellants, however, that this finding lacks support in the record.
Beck-Nielsen's system captures raw EEG signals and extracts power
averages in five frequency bands over a given time segment, e.g., one
second, to compile a feature vector for that time segment. Beck-Nielsen,
4:21-25, 7:7-11, 13:7-9. A classifier analyzes each feature vector to
determine if a pattern indicating hypoglycemia is present; if so, it records an
"event." Id. at 8:6-10, 18:6-22. The system integrates the events during a
selected time period, e.g., 300 seconds, and if the total number of events
exceeds a threshold value, notifies the user that a hypoglycemic condition is
present. Id. at 8: 11-19, 20:8-22. Although Beck-Nielsen's system may
inherently log feature vectors in a data logger so that they may be processed
4
Appeal2017-005655
Application 13/335,901
by the classifier, we are not persuaded that any feature vector is
subsequently saved in a separate memory, continuously or otherwise, after
such processing. While the Examiner is correct that "[t]here is no teaching
in Beck-Nielsen ... that the feature vectors are erased from memory once an
event is detected, or erased from memory once an event level threshold is
reached," Final Act. at 5-6, we do not consider the absence of a teaching to
erase feature vectors as constituting a teaching to save feature vectors to
memory.
Further, even if the Examiner is correct that each separate "event
detection[]" is saved as they are integrated over the selected time period,
"event detections" are not the same as feature vectors. Rather, they are what
the classifier outputs after analyzing the feature vectors. Beck-Nielsen, 8:6-
10, 18:6-22. Because we are not persuaded that Beck-Nielsen teaches the
claimed memory, we do not sustain the Examiner's rejection of claims 1 and
7, and dependent claims 2---6, 8-16, and 18-21, as unpatentable over Beck-
Nielsen and Drew.
DECISION
For the above reasons, the Examiner's rejection of claims 1-16 and
18-21 is reversed.
REVERSED
5