U.S. patent application number 13/802152 was filed with the patent office on 2013-10-10 for system and method for detecting ventilatory instability.
This patent application is currently assigned to Covidien LP. The applicant listed for this patent is COVIDIEN LP. Invention is credited to Keith Batchelder, James Ochs, Yu-Jung Pinto.
Application Number | 20130267805 13/802152 |
Document ID | / |
Family ID | 41338504 |
Filed Date | 2013-10-10 |
United States Patent
Application |
20130267805 |
Kind Code |
A1 |
Ochs; James ; et
al. |
October 10, 2013 |
SYSTEM AND METHOD FOR DETECTING VENTILATORY INSTABILITY
Abstract
Embodiments described herein may include systems and methods for
detecting events that may be associated with sleep apnea. Some
embodiments are directed to a system and/or method for automated
detection of reduction in airflow events using polysomnograph
signals, wherein the reduction in airflow events may relate to
sleep apnea. The PSG signals may be limited to four signals,
including data from an airflow channel, a blood oxygen saturation
channel, a chest movement channel, and an abdomen movement channel.
Using information from these channels, some embodiments may
automatically identify reduction in airflow events.
Inventors: |
Ochs; James; (Seattle,
WA) ; Batchelder; Keith; (New York, NY) ;
Pinto; Yu-Jung; (Boulder, CO) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
COVIDIEN LP |
Mansfield |
MA |
US |
|
|
Assignee: |
Covidien LP
Mansfield
MA
|
Family ID: |
41338504 |
Appl. No.: |
13/802152 |
Filed: |
March 13, 2013 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
12208087 |
Sep 10, 2008 |
8398555 |
|
|
13802152 |
|
|
|
|
Current U.S.
Class: |
600/324 ;
600/529; 600/538 |
Current CPC
Class: |
A61B 5/14551 20130101;
A61B 5/4818 20130101; A61B 5/0002 20130101; A61B 5/7282 20130101;
A61B 5/1135 20130101; A61B 5/1455 20130101; A61B 5/087 20130101;
A61B 5/7203 20130101; A61B 5/0816 20130101 |
Class at
Publication: |
600/324 ;
600/538; 600/529 |
International
Class: |
A61B 5/08 20060101
A61B005/08; A61B 5/1455 20060101 A61B005/1455; A61B 5/00 20060101
A61B005/00; A61B 5/087 20060101 A61B005/087 |
Claims
1. A method for automated detection of ventilatory instability,
comprising: calculating an output value of breathes per minute for
a data segment based at least in part upon signals of an airflow
channel, a chest channel, and/or an abdomen channel and/or
combinations thereof; determining a baseline value for each of the
airflow channel, the chest channel, and the abdomen channel based
at least in part upon the output value of breathes per minute
and/or a measure of the magnitude of the corresponding signals;
determining whether a reduction in airflow above a minimum
reduction level relative to the baseline value for the airflow
channel has been maintained for a period of time, and identifying a
reduction in airflow event if the reduction in airflow is above the
minimum reduction level; and providing an indication of ventilatory
instability if the reduction in airflow event is identified and
meets a set of criteria.
2. The method of claim 1, further comprising converting the signals
of the airflow channel, the chest channel, and/or the abdomen
channel to a frequency domain to determine a plurality of frequency
spectrums, wherein the plurality of frequency spectrums comprises
an airflow channel frequency spectrum, a chest channel frequency
spectrum, and/or an abdomen channel frequency spectrum.
3. The method of claim 2, wherein calculating the output value of
breathes per minute comprises determining a highest frequency of an
average of the plurality of frequency spectrums.
4. The method of claim 1, wherein calculating the output value of
breathes per minute comprises averaging an estimate of breathes per
minute for each of a plurality of data segments based at least in
part upon the signals of the airflow channel, the chest channel,
and/or the abdomen channel and/or combinations thereof.
5. The method of claim 1, further comprising determining whether a
criterion is met that requires a pulse oximetry channel to change
by a defined percentage within a window of time generally relative
to the reduction in airflow event.
6. The method of claim 5, further comprising determining whether
the pulse oximetry channel has changed by at least 3% relative to
its baseline within 30 second of the reduction in airflow event to
determine if the criterion is met.
7. The method of claim 1, further comprising determining whether a
criterion is met that requires a reduction in the chest channel
and/or the abdomen channel relative to the respective baseline
values for a defined portion of the reduction in airflow event.
8. The method of claim 7, further comprising determining whether
there was at least a 40% reduction relative to the respective
baseline values in the chest channel and/or the abdomen channel for
at least half of the reduction in airflow event to determine if the
criterion is met.
9. The method of claim 1, further comprising discarding invalid
data corresponding to the airflow channel, the pulse oximetry
channel, the chest channel, and the abdomen channel based on
whether certain indicators are present in an associated data
segment.
10. The method of claim 1, wherein the data segment comprises a 10
minute data segment.
11. The method of claim 1, wherein the minimum reduction level
relative to the baseline value for the airflow channel comprises a
40% reduction relative to the baseline value for the airflow
channel.
12. The method of claim 1, comprising determining whether the
reduction in airflow above the minimum reduction level relative to
the baseline value for the airflow channel has been maintained for
10 seconds or more.
13-25. (canceled)
26. The method claim 1, further comprising determining if the
reduction in airflow event is qualified by determining whether a
specified amount of change has occurred in the chest channel or the
abdomen channel during a window of time including the reduction in
airflow event.
27. The method of claim 1, further comprising filtering at least
one of the airflow channel, the chest channel, and the abdomen
channel.
28. A method, comprising: supplying signals of an airflow channel,
a chest channel, an abdomen channel, and a pulse oximetry channel;
calculating a baseline value for each of the airflow channel, the
chest channel, and the abdomen channel based at least in part upon
a value of breaths per minute and a measure of the magnitude of the
corresponding signals; identifying a reduction in airflow event
when a reduction in airflow above a minimum reduction level
relative to the baseline value for the airflow channel has been
maintained for a threshold period of time; determining if the
reduction in airflow event is qualified by determining whether a
specified amount of change has occurred in the chest channel, the
abdomen channel, or the pulse oximetry channel during a window of
time including the reduction in airflow event; identifying
ventilatory instability based at least in part upon a series of
blood oxygen saturation values; and comparing results from the
event detection system and the pulse oximetry pattern recognition
system to facilitate adjustment of the pulse oximetry pattern
recognition system.
29. The method of claim 28, further comprising calculating the
value of breaths per minute for a data segment based on signals of
the airflow channel, the chest channel, and/or the abdomen channel
and/or combinations thereof.
30. The method of claim 28, further comprising calculating the
value of breaths per minute by averaging an estimate of breathes
per minute for each of a plurality of data segments based at least
in part upon the signals of the airflow channel, the chest channel,
and/or the abdomen channel and/or combinations thereof.
31. The method of claim 28, further comprising providing an
indication of ventilatory instability on a display if the reduction
in airflow event is identified and qualified.
32. The method of claim 28, further comprising: converting the
signals of the airflow channel, the chest channel, and/or the
abdomen channel to a frequency domain; and determining a plurality
of frequency spectrums, wherein the plurality of frequency
spectrums comprises an airflow channel frequency spectrum, a chest
channel frequency spectrum, and/or an abdomen channel frequency
spectrum.
33. The method of claim 32, further comprising calculating an
output value of breaths per minute for a data segment based on
determining a highest frequency of an average of the plurality of
frequency spectrums.
Description
[0001] This application is a continuation of U.S. patent
application Ser. No. 12/208,087 filed Sep. 10, 2008, which is
incorporated herein by reference in its entirety.
BACKGROUND
[0002] The present disclosure relates generally to medical devices
and methods and, more particularly, to an automated system and
method for detecting events related to ventilatory instability,
such as sleep apnea.
[0003] This section is intended to introduce the reader to various
aspects of art that may be related to various aspects of the
present disclosure, which are described and/or claimed below. This
discussion is believed to be helpful in providing the reader with
background information to facilitate a better understanding of the
various aspects of the present disclosure. Accordingly, it should
be understood that these statements are to be read in this light,
and not as admissions of prior art.
[0004] Sleep apnea is generally described as a sleep disorder that
is characterized by episodes of paused breathing during sleep.
These episodes of paused breathing may occur repeatedly throughout
sleep, and each episode may last long enough to cause one or more
breaths to be missed. Such episodes may be referred to as apneas. A
typical definition of an apnea may include an interval between
breaths of at least 10 seconds, with a neurological arousal and/or
a blood oxygen desaturation of 3% or greater. The actual duration
and severity of each apnea may substantially vary between multiple
patients. Further, duration and severity of apneas may vary
throughout a period of sleep for a single patient. Indeed, sleep
apnea may have a wide range of severity. For example, sleep apnea
may include mild snoring, which may be related to incomplete and
inconsequential airway obstruction, or severe apneas, which may
result in hypoxemia. Sleep apnea commonly results in excessive
daytime sleepiness. Further, sleep apnea can hinder cognitive
function during the day due to sporadic sleep during the night
resulting from recurrent arousals associated with the sleep
apnea.
[0005] Although sleep apnea commonly affects obese patients, it may
occur in patients with any body type. Indeed, sleep apnea is fairly
common and causes undesirable symptoms of excessive daytime
sleepiness, morning headache, and decreasing ability to concentrate
during the day. Thus, it is desirable to diagnose and treat sleep
apnea. Traditionally, sleep apnea is diagnosed utilizing an
overnight sleep test referred to as a polysomnogram. This is
generally performed in a sleep lab and involves the continuous and
simultaneous measurement and recording of an encephalogram,
electromyogram, extraoculogram, chest wall plethysmogram,
electrocardiogram, measurements of nasal and oral airflow, and
pulse oximetry. All or some of these and other channels may be
measured simultaneously throughout the night, and complex
recordings of such measurement may then analyzed by a highly
trained clinician to determine the presence or absence of sleep
apnea.
BRIEF DESCRIPTION OF THE DRAWINGS
[0006] Advantages of the disclosure may become apparent upon
reading the following detailed description and upon reference to
the drawings in which:
[0007] FIG. 1 is a block diagram of a medical analysis system in
accordance with some embodiments;
[0008] FIG. 2 is a process flow diagram of a method for detecting
ventilatory instability related events in accordance with some
embodiments;
[0009] FIG. 3 is a block diagram of a device and/or module capable
of receiving and filtering signals in accordance with some
embodiments;
[0010] FIG. 4 is a process flow diagram of a method for estimating
a value for breathes per minute of a patient in accordance with
some embodiments; and
[0011] FIG. 5 is a block diagram of the system of FIG. 1
communicatively coupled with a separate ventilatory instability
detection system to facilitate calibration of the separate system
in accordance with some embodiments.
DETAILED DESCRIPTION
[0012] One or more embodiments of the present disclosure will be
described below. In an effort to provide a concise description of
the embodiments, not all features of an actual implementation are
described in the specification. It should be appreciated that in
the development of any such actual implementation, as in any
engineering or design project, numerous implementation-specific
decisions may be made to achieve the developers' specific goals,
such as compliance with system-related and business-related
constraints, which may vary from one implementation to another.
Moreover, it should be appreciated that such a development effort
might be complex and time consuming, but would nevertheless be a
routine undertaking of design, fabrication, and manufacture for
those of ordinary skill having the benefit of this disclosure.
[0013] Some embodiments are directed to automated systems and
methods for detecting events that may be associated with sleep
apnea. Specifically, some embodiments are directed to a system
and/or method for automated detection of reduction in airflow
events using polysomnograph (PSG) signals, wherein the reduction in
airflow events may relate to sleep apnea. The PSG signals may be
limited to four signals, including data from an airflow channel, a
blood oxygen saturation (SpO.sub.2) channel, a chest movement
channel, and an abdomen movement channel. Using information from
these channels, some embodiments may automatically identify
reduction in airflow events.
[0014] Accordingly, some embodiments may facilitate automated
detection and/or diagnosis of sleep apnea in patients. For example,
some embodiments may be utilized to analyze data that has been
acquired using a separate PSG system to determine whether sleep
apnea related events have occurred. In another example, some
embodiments may be incorporated with a PSG system to automatically
detect and/or diagnose sleep apnea while a patient is being
monitored. Indeed, some embodiments may facilitate detection of
events relating to sleep apnea and/or facilitate diagnosis of sleep
apnea in real time. Further, some embodiments may be utilized to
demonstrate or confirm the accuracy or reliability of other systems
and/or methods for detecting events related to sleep apnea. For
example, some embodiments may be utilized in conjunction with a
device configured to identify ventilatory instability (e.g., sleep
apnea) based on an SpO.sub.2 pattern recognition algorithm.
[0015] In some embodiments, a distinction may be made between
whether identified sleep apnea events correspond to central sleep
apnea or obstructive sleep apnea. Central and obstructive sleep
apnea may be distinguished based on the nature of their occurrence.
For example, a lack of effort in breathing is generally the cause
of interrupted breathing associated with central sleep apnea, while
a physical block in airflow despite effort is generally the cause
of interrupted breathing associated with obstructive sleep apnea.
If a patient is or is not making an effort to breathe, the
patient's chest and abdomen activity may be indicative.
Accordingly, some embodiments may distinguish between the two types
of apnea by including devices or modules that are capable of
quantifying phase differences between chest and abdomen signals.
For example, some embodiments may include an algorithm stored on a
memory that receives chest and abdomen signals and determines that
certain events correspond to obstructive sleep apnea when the chest
and abdomen signals are out of phase, or that the events correspond
to central sleep apnea when there is no chest and/or abdomen
movement, or there is a decrease in chest and abdomen movement but
signals are in phase.
[0016] FIG. 1 is a block diagram of a medical analysis system in
accordance with some embodiments. The medical analysis system is
generally indicated by reference numeral 10. The system 10 includes
a PSG system 12 and an event detection (ED) system 14. The PSG
system 12 and the ED system 14 may be separate or integrated in
accordance with some embodiments. The PSG system 12 may include,
for example, hardware and/or software products from Nellcor Puritan
Bennett's Sandman.RTM. sleep diagnostics line of products.
[0017] According to an embodiment, the PSG system 12 may include a
memory 16 and a processor 18, and the ED system 14 may include a
memory 20 and a processor 22. Programming stored on the respective
memories 16 and 20 for the PSG system 12 and the ED system 14 may
be utilized in conjunction with each system's respective processor
18 and 22 to facilitate performance of certain functions by the PSG
system 12 and the ED system 14. For example, the memories 16 and 20
may include coded instructions that may be utilized with the
respective processors 18 and 22 to perform automated methods in
accordance with some embodiments. It should be noted that, in some
embodiments, the memories 16 and 20 may be integral with the
respective processors 18 and 22. Further, if the PSG system 12 and
ED system 14 are integral components of the system 10, the
integrated PSG system 12 and ED system 14 may share a single memory
and/or a single processor.
[0018] According to an embodiment, the PSG system 12 may be capable
of receiving signals from various sensors 24 that are capable of
measuring certain physiologic parameters or patient activities.
Specifically, the sensors 24 may include an airflow sensor 26, a
chest sensor 28, an abdomen sensor 30, and a pulse oximeter sensor
32. Each of the sensors 24 may be attached to a patient 34 to
facilitate measuring physiologic parameters and/or physical
activity (e.g., movement) of the patient 34. These four measured
values from the sensors 24 may be transmitted as signals to the PSG
system 12 for processing. Thus, the sensors 24 may cooperate with
the PSG system 12 to provide a polysomnogram. A polysomnogram may
be described as a recording of an individual's sleep
characteristics (e.g., activities and physiological events
occurring during sleep), which may include an output of various
measurements obtained by the sensors 24. Such an output may be
recorded on a memory (e.g., a flash memory or hard drive),
presented on other tangible medium (e.g., printed on paper), or
visually displayed (e.g., displayed as video). For example, the
polysomnogram, along with other data, may be displayed on a video
screen 36 of the medical analysis system 10.
[0019] According to an embodiment, the ED system 14 may be capable
of using data acquired by the PSG system 12, such as the
polysomnogram, to automatically detect events that may be
associated with sleep apnea. For example, the ED system 14 may
automatically detect reduction in airflow events, which relate to
sleep apnea, using four PSG signals from the PSG system 12. The PSG
signals may be transferred to the ED system 14 from the PSG system
14 in a number of ways. For example, data corresponding to signals
from the PSG system 12 may be manually entered into the ED system
14, transferred via a memory device (e.g., a flash memory), or
directly transmitted to the ED system 14 from the PSG system 12 for
analysis. The PSG system 12 and the ED system 14 may receive and/or
transmit signals corresponding to particular signal types based on
data in a memory (e.g., a flash memory or a memory of the PSG
system 12) and/or directly from each of the sensors 24. The PSG
signals and/or signal data may be referred to as channels
corresponding to each signal, such as airflow, chest, abdomen, and
pulse oximetry channels.
[0020] According to an embodiment, once the ED system 14 receives
the PSG signals, which may be in the form of signal data, the ED
system 14 may perform an automated process on the PSG signals to
identify events related to sleep apnea, such as reduction in
airflow events. If certain events are identified in accordance with
specified rules or criteria, the ED system 14 may provide an
indication of the presence of ventilatory instability. For example,
the ED system 14 may include hardware and/or software components
that filter one or more of the PSG signals and identify
characteristics of the signals that combine to suggest the presence
of events related to sleep apnea. Specifically, for example, after
checking for invalid data, filtered and unfiltered PSG signals may
be utilized by the ED system 14 to estimate a value of breathes per
minute (BPM) of the patient 34. The estimated BPM may then be
utilized by the ED system 14 to calculate a baseline for use in
determining whether a threshold level of airflow reduction is
present. Further, the ED system 14 may determine whether certain
other events indicative of respiratory instability are present
based on the PSG signals, as will be discussed in further detail
below. The ED system 14 may output values indicative of whether
respiratory instability is present based on whether certain events
were identified. For example, the ED system 14 may indicate the
presence of sleep apnea and/or indicate a type of sleep apnea
(e.g., central or obstructive) based on results obtained through
analysis of a neural network of the ED system 14. It should be
noted that the identification of events by the ED system 14 may be
facilitated by an algorithm stored in the memory 20, which may
cooperate with the processor 22 to implement methods in accordance
with some embodiments. It should further be noted that the
algorithm may be tuned by adjusting constant values in the
algorithm to correspond to particular situations or patients (e.g.,
a patient with a heart condition).
[0021] FIG. 2 is a process flow diagram illustrating a method in
accordance with some embodiments. The method is generally indicated
by reference number 100 and includes various steps or actions
represented by blocks. It should be noted that the method 100 may
be performed as an automated procedure by a system, such as the
analysis system 10. Further, certain steps or portions of the
method may be performed by separate devices. For example, a first
portion of the method 100 may be performed by the PSG system 12 and
a second portion of the method 100 may be performed by the ED
system 14. In this embodiment, the method includes receiving and
filtering signals (block 102), checking for invalid data (block
104), estimating BPM (block 106), determining a signal baseline for
certain signals (block 108), determining whether certain criteria
are met by acquired data (block 110), and outputting a result
(block 112).
[0022] According to an embodiment, the method 100 begins with
receiving and filtering signals, which may include signal data, as
represented by block 102. Block 102 may include receiving and/or
filtering signal data from several sensors, which may include the
airflow sensor 26, the chest sensor 28, the abdomen sensor 30,
and/or the pulse oximeter sensor 32. Specifically, block 102 may
represent receiving data or signals from airflow, chest, abdomen,
and pulse oximetry channels and filtering signals on a subset of
the channels. Signals on the chest, airflow, and abdomen channels
may be filtered, while signals on the pulse oximetry channel, which
may relate to SpO.sub.2 data, may remain unfiltered. For example,
FIG. 3 illustrates a component 300 capable of filtering the chest,
airflow, and abdomen channels in accordance with some embodiments.
In some embodiments, the component 300 may be a feature of the
processor, or it may be implemented using electronic circuits which
preprocess and filter some or all of the PSG signals, 22.
Specifically, as illustrated in FIG. 3, a raw signal 302 from the
PSG system 12, such as a signal on the chest, airflow or abdomen
channel, may be received into the component 300 of the ED system
114. The raw signal 302 may pass through the component 300, which
may include passing the raw signal through a band-pass filter 304
and a low-pass filter 306 of the component 300.
[0023] According to an embodiment, the band-pass filter 304 may
pass frequencies within a selected range and attenuate frequencies
that are outside of the selected range. Specifically, for example,
the band-pass filter 30 may include a 0.0833-1 Hz band-pass filter
that operates to pass frequencies corresponding to 5-60 BPM and
filter out frequencies above and below that range. An initial or
default 5-60 BPM pass band may be used since most human breathing
rates will be within this range under typical sleep-lab conditions.
The pass band may be tuned (moved, narrowed, and/or widened) by the
operator based on known patient conditions to increase the
specificity and/or sensitivity of the ED system. Once the raw
signal 302 has passed through the band-pass filter 304, it becomes
a band-passed signal 308. The band-passed signal 308 may be
utilized in conjunction with other input in the analysis and
identification of ventilatory instability, as will be discussed in
further detail below.
[0024] According to an embodiment, the low-pass filter 306 may pass
frequencies below a cutoff level and attenuate frequencies higher
than the cutoff level. Specifically, for example, the low-pass
filter 306 may attenuate frequencies lower than 0.04166 Hz, which
corresponds to 2.5 BPM. 2.5 BPM may be selected as the initial
cutoff level since it is unlikely that any patient will breathe
slower than this rate, and therefore frequencies below this level
can be considered noise, or more specifically these frequencies can
be considered the DC offset of the PSG signal. The operator may
move the low pass filter level based on known patient conditions in
an attempt to increase the specificity and sensitivity of the ED
system. Once the raw signal 302 has passed through the low-pass
filter 306, it becomes a low-passed signal 310. The low-passed
signal 310 may be used to calculate a noise signal 312 by passing
it through a subtraction block 314 with the raw signal 302 to get a
difference between the raw signal 302 and the low-passed signal
310. This and other features of noise calculation will be discussed
in further detail below. The calculated noise level may be utilized
to clarify signal features by reducing noise content. In some
embodiments, features may be included that facilitate reading data
through the noise, rather than merely removing the noise.
[0025] According to an embodiment, after the signals have been
received and/or filtered in block 102, the method 100 may proceed
to block 104, which includes checking for invalid data.
Specifically, block 104 may represent receiving and processing a
stream of data to determine whether any portions of the data meet
criteria indicating that the data should be discarded. For example,
block 104 may represent receiving a 10 minute segment of data that
is analyzed to determine if the following criteria are present: (1)
the SpO.sub.2 signal is less than a designated value (e.g., less
than a value of 20% SpO2, which may indicate that the sensor is off
or disconnected), (2) the signal to noise ratio for a signal on the
airflow channel is less than a designated value (e.g., 5 dB), (3)
the signal to noise ratio for a signal on the chest channel is less
than a designated value (e.g., 0 dB), and/or (4) a signal to noise
ratio (SNR) for a signal on the abdomen channel is less than a
designated value (e.g., 0 dB). If it is determined in block 104
that any of the designated criteria, such as the criteria set forth
above, are present for a designated period, e.g., 2 minutes, within
the minute segment of data, all or part of the 10 minute segment of
data may be discarded as being invalid. In some embodiments,
features may be included that facilitate reading through data with
a SNR below a threshold, rather than merely invalidating data
because of low SNR values.
[0026] It should be noted that the signal to noise ratios
referenced above may be automatically calculated in block 104 along
with other determinations relating to identifying invalid data. For
example, block 104 may represent calculating the signal to noise
ratio for a particular signal by first removing DC from the raw
signal 302. That is, the low-passed signal 310 (LPFilteredSignal)
may be subtracted from the raw signal 302 (RawSignal) to obtain a
raw signal without DC (RawSignalNoDC). This procedure may be
represented by the following equation:
RawSignalNoDc=RawSignal-LPFilteredSignal.
[0027] Noise may then be calculated by subtracting the band-passed
signal 308 (BPSignal) from the raw signal without DC. This
procedure may be represented by the following equation:
Noise=RawSignalNoDC-BPSignal.
[0028] A statistical measure of magnitude, such as a root mean
square, may then be obtained for both the noise and the band-passed
signal, and the signal to noise ratio (SNR) may be calculated by
dividing the root mean square of the band-passed signal
(RMS(Signal)) by the root mean square of the noise (RMS(Noise)), as
represented by the following equation:
SNR=RMS(Signal)/RMS(Noise).
As will be appreciated, in some embodiments, different calculations
or procedures may be utilized to obtain the signal to noise
ratio.
[0029] According to an embodiment, after invalid data has been
removed in block 104, the method 100 may proceed to determine an
estimated BPM based on the received data, as represented by block
106. As illustrated in FIG. 4, the BPM estimate may be calculated
using the chest, abdomen, and airflow signals. FIG. 4 is a process
flow diagram of a method of estimating BPM in accordance with some
embodiments. The method of estimating BPM is generally indicated by
reference number 400 and includes various steps or actions
represented by blocks. As with the other steps of the method 100,
the method 400 may be performed as an automated procedure by the
system 100 in accordance with some embodiments.
[0030] According to an embodiment, the method 400 begins by
receiving band-passed airflow, chest, and abdomen signals, and
converting the signals to the frequency domain by performing a fast
Fourier transform (FFT) on all three channels, as represented by
block 402. Specifically, block 402 may represent determining an FFT
for the three channels a certain number of times per time period.
For example, block 402 may represent calculating an FFT for the
band-passed airflow, chest, and abdomen signals once every 2
minutes.
[0031] According to an embodiment, after performing the frequency
conversion in block 402, the method 400 may proceed to block 404,
which represents normalizing the frequency spectrums obtained in
block 402. The procedure represented by block 404 may include
various different types of normalization, which may result in
ranging the frequency spectrums from 0.0 to 1.0. For example, peak
normalization may be performed by dividing the amplitude at each
point in the spectrum of each signal by the maximum amplitude of
that particular spectrum. Thus, the normalized spectrum may include
intensities that range from as low as 0.0 to as high as 1.0. By
normalizing the spectrums, certain discrepancies between the
spectrums may be removed to facilitate proper combination or
comparison of the different spectrums. It should be noted that each
spectrum may be based on a segment of data, such as a 10 minute
segment of recorded sensor data.
[0032] According to an embodiment, after the spectrums have been
normalized in block 404, all three frequency spectrums may be
averaged, as represented by block 406, and an estimate of BPM may
be calculated based on the average, as represented by block 408.
Specifically, blocks 406 and 408 may include averaging all of the
histograms of the frequency spectrums and selecting a frequency
from the average that has the highest amount of energy to determine
the estimate BPM for the associated segment of data. This procedure
may be represented by the following equation:
BPM Estimate=max(Average Frequency Spectrum).
[0033] According to an embodiment, the process of providing BPM
estimates for a segment of data may be repeated for a series of
data segments, as represented by arrow 410. The most recent BPM
estimate obtained for a data segment may be referred to as the
current estimate and the penultimate BPM estimate obtained for a
previous data segment may be referred to as the previous estimate.
In other words, once a new segment of data has been received and
processed, the BPM estimate that was the current estimate may
become the previous segment and the BPM estimate for the most
recent segment of data may become the current estimate. As
represented by block 412, the current and previous estimates may be
averaged to provide a final BPM estimate, as represented by the
following equation:
Final BPM Estimate=0.5.times.(Current Estimate+Previous
Estimate).
It should be noted that in some embodiments, more than two
estimates may be averaged to determine the final BPM estimate. For
example, three or more estimates for a series of data segments may
be stored and averaged to determine a final BPM estimate in
accordance with some embodiments.
[0034] According to an embodiment, after estimating BPM in block
106, the method 100 may proceed to determining a signal baseline
for each signal, as represented by block 108. Specifically, for
example, block 108 may include calculating a signal baseline for
the airflow, chest, and abdomen signals. In performing this
calculation, a root mean square value for the signal (e.g., the raw
signal 302) may first be obtained using the final BPM estimate and
a running root mean square, as indicated by the following
equation:
Xrms=0.5 BPM running RMS signal X.
This calculation may take into account every point of a signal such
that the results are indicative of the power of the signal at each
point. The baseline value (X-Baseline) for each particular signal
may then be determined based on this root mean square value (Xrms).
Specifically, the baseline may be determined as the top 10% or
ninetieth percentile of the root mean square value, as represented
by the following equation:
X-Baseline=Top 10% (90th percentile) of Xrms.
[0035] According to an embodiment, once a baseline has been
determined for the airflow, chest, and abdomen signals, as
represented by block 108, the method 100 may proceed with
determining whether certain criteria are met, as represented by
block 110. The presence of characteristics that meet certain
criteria may indicate and confirm that a reduction in airflow (RAF)
event occurred relative to a normal airflow. For example, a
determination may be made as to whether a certain level of
reduction (e.g., a 40% reduction) in airflow has occurred for at
least a threshold amount of time (e.g., 10 seconds). This may be
referred to as an RAF event. Specifically, in one embodiment, an
RAF event may be described as an interval where the amplitude
envelope of a signal on the airflow channel from the PSG system 12
is reduced at least 40% relative to the baseline for at least 10
seconds consecutively. It should be noted that the amplitude
envelope may refer to the value of a function describing how the
maximum amplitude of the airflow signal changes over time.
[0036] According to an embodiment, once an RAF event has been
identified, it may be qualified for consideration based on certain
scoring rules. For example, an RAF event may be disqualified if
there is not a specified amount of change in one or more other
signal measurements within a certain window of time with respect to
the time the RAF event occurred. For example, if the measured
SpO.sub.2 value does not change at least a certain amount, e.g.,
3%, during the RAF event or within a certain time, e.g., 30
seconds, after the RAF event, the RAF event may be disqualified.
Similarly, for example, if there is not at least a certain change,
e.g., a 40% reduction, in the chest or abdomen signal from the PSG
system 12 for at least part, e.g., half, of the interval of the RAF
event, the RAF event may be disqualified. Alternatively, if the
available data meets these criteria, the RAF event may be qualified
and used to determine whether the data is indicative of ventilatory
instability in the patient that was or is being monitored.
[0037] According to an embodiment, block 110 may also represent
determining whether a segment of data includes an indication of
ventilatory instability based on a quantity of qualified RAF events
that occur within the data segment. For example, block 110 may
determine that a particular data segment is indicative of
ventilatory instability if a certain number, e.g., at least 5, of
consecutive RAF events are qualified within a given portion, e.g.,
a 10 minute period, of the segment of data. Thus, segments of data
may be divided into intervals, e.g., 10 minute intervals, in
accordance with some embodiments. It should be noted that in order
for the RAF events to be qualified as being consecutive, certain
relative timing criteria may have to be met. For example, in one
embodiment, for a pair of qualified RAF events to be consecutive,
the RAF events must occur within a certain time, e.g., 120 seconds,
of one another.
[0038] According to an embodiment, block 110 may also represent
determining a type of respiratory event, such as determining the
presence of central or obstructive sleep apnea based on
correlations between different signals, such as the chest and
abdomen signals being in or out of phase. As discussed above,
central and obstructive sleep apnea may be distinguished based on
the nature of their occurrence. For example, a lack of effort in
breathing is generally associated with central sleep apnea, while a
physical block in airflow despite effort is generally associated
with obstructive sleep apnea. A patient's chest and abdomen
activity may be utilized to distinguish between the two types of
apnea. Accordingly, some embodiments may include features that are
capable of quantifying phase differences between chest and abdomen
signals. For example, some embodiments may include features that
determine that certain events correspond to obstructive sleep apnea
when the chest and abdomen signals are out of phase, or that the
events correspond to central sleep apnea when there is no chest
and/or abdomen movement, or there is a decrease in chest and
abdomen movement but signals are in phase.
[0039] Based on the criteria or rules discussed above, the method
100 may output an epoch score for each data segment, as represented
by block 112. For example, an epoch score may be repeated once
every time a segment of data has been analyzed (e.g., once every 10
minutes). In some embodiments, a system may utilize such a score in
an algorithm to provide an indication of certain conditions
relating to ventilatory instability. For example, the epoch score
may include a value of 1 for an indication that sleep apnea has
been detected, 0 for an indication that sleep apnea has not been
detected, or -1 for an indication of invalid data. The indicator
provided by the system based on the score may include a textual
indication of the detected condition, such as "sleep apnea
detected," "sleep apnea not detected", or "unknown or invalid
data." In some embodiments, a series of epoch scores may be
combined for all or part of a sleep study to generate an aggregate
score. For example, an average of all of the scores for each 10
minute segment of a sleep cycle may be used to determine a summary
score for a particular patient.
[0040] Further, in some embodiments, the severity of the
ventilatory instability or apnea events may be quantified instead
of merely providing binary outputs. For example, the depth of the
airflow reduction may be used to quantify a severity of the
ventilatory instability. In some embodiments different aspects
associated with qualified RAF events may be used to determine
levels of severity. For example, a series of continually larger
drops in airflow and/or continual failures to return to a baseline
level of airflow may be correlated to a higher level of severity.
Further, numerous clusters of RAF events or data segments including
RAF events may be considered in determining a severity level. For
example, a large segment of data, e.g., 4 hours of data, may
include various sub-segments having different levels of apnea. Each
sub-segment may be analyzed separately based on certain criteria,
such as a number of RAF events per time period, an amount of
reduced airflow, or other characteristics, and the values
associated with the sub-segments may be combined to provide an
overall ventilatory instability level for the large segment of
data.
[0041] The ED system 14 may be utilized to demonstrate or confirm
that other systems involving detection of ventilatory instability
related events are properly calibrated and/or providing results
that correlate to the results obtained by the ED system 14. For
example, FIG. 5 is a block diagram of the ED system 14
communicatively coupled with a separate ventilation analysis system
500 to facilitate calibration of the system 500 or to confirm its
proper operation. The separate ventilation analysis system 500 may
include a system such as that described in U.S. Pat. No. 6,223,064.
For example, the system 500 may include an SpO.sub.2 pattern
recognition system that is utilized to identify ventilatory
instability based on a series of SpO.sub.2 values. The system 500
may include a memory 502 and a processor 504 that are capable of
analyzing input received or previously acquired from a single
SpO.sub.2 sensor. Specifically, the system 500 may be capable of
identifying certain patterns or clusters of measured SpO.sub.2
values to identify events related to ventilatory instability.
[0042] According to an embodiment, the ED system 14 may facilitate
adjustment and/or calibration of the system 500 to correlate with
BPM estimates determined by the ED system 14. For example, both
systems 14 and 500 may be provided with data from a storage device
506, wherein the data corresponds to certain ventilatory
instability events that have been observed. During a calibration
period, the system 500 may be adjusted to correspond to the ED
system 14. In other words, the system 500 may be adjusted such that
it detects a high percentage of the ventilatory instability events
detected by the ED system 14. For example, certain coefficients of
an algorithm stored on the memory of the system 500 may be adjusted
to improve correlations between the results for the system 500 and
the results for the ED system 14.
[0043] In a specific example, the ED system 14 may utilize
automated analysis of the data obtained via the various sensors 24
to confirm that the system 500 is properly tuned and/or providing
corresponding output. Specifically, the ED system 14 may receive
data from the storage device 506 corresponding to airflow, chest
impedance, abdomen impedance, and blood oxygen saturation, and use
this data to provide results that include a BPM estimate. These
results may be compared with similar results obtained by the system
500 using SpO.sub.2 pattern recognition. Based on the comparison,
certain features of the system 500 (e.g., features of a pattern
recognition algorithm stored on the memory 502) may be adjusted to
achieve a desired correspondence. The automated analysis provided
by the ED system 14 may facilitate rapid adjustment and/or testing
of systems such as the system 500.
[0044] Specific embodiments have been shown by way of example in
the drawings and have been described in detail herein. However, it
should be understood that the claims are not intended to be limited
to the particular forms disclosed. Rather, the claims are to cover
all modifications, equivalents, and alternatives falling within
their spirit and scope.
* * * * *