U.S. patent application number 12/938268 was filed with the patent office on 2011-11-03 for multi-function health monitor.
This patent application is currently assigned to Applied Cardiac Systems, Inc.. Invention is credited to Kenneth L. Burns, Loren A. Manera.
Application Number | 20110270112 12/938268 |
Document ID | / |
Family ID | 44858806 |
Filed Date | 2011-11-03 |
United States Patent
Application |
20110270112 |
Kind Code |
A1 |
Manera; Loren A. ; et
al. |
November 3, 2011 |
Multi-Function Health Monitor
Abstract
A multi-function health monitor is capable of performing a
resting 12-lead ECG test, an ECG stress test, a 24-hour holter
monitor evaluation and or a 30-day MCT monitoring. Using only 3
electrodes, the multifunction health monitor derives 6 channels
(Limb leads & Augmented leads) of data with the noise
cancellation (ground) effect of a virtual dynamic RL electrode. An
electrode resistivity measurement system quantifies and may
compensate for increasing resistance the electrodes and the patient
that results from the length of time the electrodes are installed
on a patient. The multi-function health monitor can provide data
analysis through the gate array as the data is acquired. Data may
also be stored for remote analysis as well as for transmission to
remote stations upon occurrence of one or more threshold events.
Parameters for threshold events may be adjusted remotely to obviate
the need for a patient to travel for system adjustment.
Inventors: |
Manera; Loren A.; (Laguna
Hills, CA) ; Burns; Kenneth L.; (Laguna Hills,
CA) |
Assignee: |
Applied Cardiac Systems,
Inc.
|
Family ID: |
44858806 |
Appl. No.: |
12/938268 |
Filed: |
November 2, 2010 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
61257388 |
Nov 2, 2009 |
|
|
|
Current U.S.
Class: |
600/523 |
Current CPC
Class: |
A61B 2560/0431 20130101;
A61B 5/349 20210101; A61B 5/363 20210101; A61B 5/361 20210101; A61B
5/276 20210101; A61B 5/332 20210101; A61B 5/0006 20130101; A61B
5/333 20210101; A61B 2505/07 20130101; A61B 5/316 20210101; A61B
5/6823 20130101; A61B 5/304 20210101; A61B 5/0245 20130101 |
Class at
Publication: |
600/523 |
International
Class: |
A61B 5/0408 20060101
A61B005/0408 |
Claims
1. A remote ambulatory cardiac monitoring system comprising:
microprocessor for controlling interactions with external devices;
a hardware gate array for controlling acquisition of ECG signals
and other data and analysis of data as it is acquired; a cable
connector operably connected to the microprocessor for engaging a
plurality of different input/output cables, insertion of each of
the plurality of input/output cables causes the microprocessor to
initiate one of a plurality of different modes from the remote
ambulatory cardiac monitoring system; a plurality of electrodes
secured to a patient, each electrode operably connected to at least
one of the plurality of input/output cables for sensing ECG signals
from a patient; and means for quantifying resistance changes
between each of the one or more of the plurality of electrodes and
the patient during acquisition of the ECG signals.
2. The apparatus of claim 1 wherein the mode initiated by the
connection of one of the plurality of different input/output cables
is selected from the group of modes including: 12-lead EKG mode,
Holter recorder mode, MCT mode, and USB data transfer mode.
3. The apparatus of claim 1 wherein the mode initiated by the
connection of one of the plurality of different input/output cables
is MCT mode.
4. The apparatus of claim 1 wherein the mode initiated by the
connection of one of the plurality of different input/output cables
is selected from the group of modes including: 12-lead EKG mode and
Holter recorder mode.
5. The apparatus of claim 1 wherein the means for quantifying
resistance changes is operable to detect when one or more of the
plurality of electrodes is not secured to the patient and upon
detection of one or more electrodes not secured, the means for
quantifying resistance changes initiates an alarm signal.
6. The apparatus of claim 1 further comprising: means for cross
checking alarm triggers before initiating an alarm using one or
more alternate data channels, where the one or more alternate data
channels did not initiate the alarm trigger.
7. A mobile cardiac telemetry system comprising: microprocessor for
controlling interactions with external devices; a hardware gate
array for controlling acquisition and analysis of ECG signals and
other data; a cable connector operably connected to the
microprocessor for engaging a plurality of different input/output
cables, insertion of each of the plurality of input/output cables
causes the microprocessor to initiate one of a plurality of
different modes from the remote ambulatory cardiac monitoring
system; no more than three electrodes secured to a patient, each of
the electrodes operably connected to one of the plurality of
input/output cables for sensing ECG signals from a patient; means
for quantifying resistance changes between one or more of the
electrodes and the patient during data acquisition; and means for
sequentially processing ECG signals from the electrodes to generate
six channels of data and a noise cancellation signal to feed back
to the electrodes.
Description
RELATED APPLICATIONS
[0001] This application claims priority to copending U.S.
Provisional patent application 61/257,388 filed Nov. 2, 2009.
FIELD OF THE INVENTIONS
[0002] The present invention relates generally to the field of
portable health monitors and more specifically to battery operated
multi-function ECG recorder/transmitters.
BACKGROUND OF THE INVENTIONS
[0003] If you or your doctor have concerns about the function of
your heart or other similar organ you may embark on a multi-step
evaluation that may include many different tests for example a
resting 12-lead ECG test, an ECG stress test, a 24-hour holter
monitor evaluation and or a 30-day MCOT monitoring. These tests and
others utilize many different and specialized tools due to some of
the conflicting requirements of the tests or the equipment. For
example, some of the tests must be done in a doctor's office or a
hospital or lab because of the cost and power consumption of the
test tools and it would be unrealistic to conduct a 30-day MCOT
analysis with the patient confined to the hospital, or worse, the
doctors office.
[0004] Some of the evaluation tests are also challenging because of
the limitations enforced by the ability of the medical system to
derive adequate compensation for the work necessary. For example,
an MCOT monitor analysis of a patient may be conducted for up to 30
days. The equipment to perform this monitoring and analysis must be
portable, thus low-power, durable and accurate. The necessity of
low-power and durable equipment forces the equipment designers to
cut corners on real-time data analysis. As a result, many false
positive events generate excessive data that requires expensive
medical professional time to review. Thus the doctor ordering the
test (and the insurance company looking over his shoulder) must be
cognizant that the test, and the analysis of the resulting data
could cost more that anyone will be paid to perform the test.
SUMMARY
[0005] A multi-function health monitor avoids the power hungry
microprocessors that are becoming ubiquitous in favor of a gate
array hardware implementation for data gathering, analysis and data
encryption. The use of a gate array, complex programmable logic
device (CPLD), for data processing minimizes power consumption
while enabling fast processing as well as accommodating many
functions with dissimilar analysis algorithms. The multi-function
health monitor is capable of calculating six channels or leads of
ECG data using only three electrodes connected to the patient. The
lead switching technique calculates six channels of ECG data with
the noise cancellation benefits of a right leg feed system. Use of
only three electrodes makes the multi-function health monitor ideal
for ambulatory data gathering and to enable better long term
compliance by patients.
[0006] A multi-function health (MFH) monitor capable of performing
a resting 12-lead ECG test, an ECG stress test, a 24-hour holter
monitor evaluation and or a 30-day MCOT monitoring as well as
providing both resting and ambulatory blood pressure monitoring
would provide a tool for generating revenue through insurance
payments for a single device that may be provided to patients
through their doctors and or hospitals.
[0007] The MFH monitor can provide data analysis through the gate
array as the data is acquired. Data may also be stored for remote
analysis as well as for transmission to remote stations upon
occurrence of one or more threshold events. Parameters for
threshold events may be adjusted remotely to obviate the need for a
patient to travel for system adjustment.
[0008] The design of the gate array and the data algorithms permits
one or more sections of the gate array to be powered off when
unused to prevent unnecessary power drain. The advanced data
algorithms implemented in the hardware of the gate array permits a
high level of sophisticated data processing by the remote
ambulatory cardiac monitoring system as data is acquired to
minimize false events and minimize the requirement for remote data
analysis.
[0009] A lead resistivity measurement system quantifies and may
also compensate for increasing resistance between an electrode and
the patient that results from the length of time the electrodes are
installed on a patient. This system can also detect and alert to an
electrode disconnected.
[0010] These and other features and advantages will become further
apparent from the detailed description and accompanying figures
that follow. In the figures and description, numerals indicate the
various features of the disclosure, like numerals referring to like
features throughout both the drawings and the description.
BRIEF DESCRIPTION OF THE DRAWINGS
[0011] FIG. 1 is a block diagram of multifunction health monitor
systems with alternate electrode sets and the system network.
[0012] FIG. 2 is a schematic diagram of an electrode input.
[0013] FIG. 3 is a block diagram of the gate array controller and
connected components of the remote ambulatory cardiac monitoring
system of FIG. 1.
[0014] FIG. 4 is a block diagram of the timing and data paths of a
remote ambulatory cardiac monitoring system.
[0015] FIG. 5 is a block diagram for channel derivation from three
electrodes.
DETAILED DESCRIPTION OF THE INVENTIONS
[0016] In electrocardiography, the word "lead" may refer to either
the electrodes attached to the patient, or, properly, to the
voltage or signal measured between two electrodes. To avoid
confusion, "channel" is used to describe the voltage, signal or
view from a positive electrode to a negative electrode across the
heart.
[0017] Referring now to FIG. 1, remote ambulatory cardiac
monitoring system 10 is a multi-purpose ECG recorder, arrhythmia
detector and alarm, capable of changing its functionality to meet
the normal workflow of a Cardiologist office. For a patient of
concern such as patient 1, a cardiologist might first order a
resting 12-channel ECG study. If the study is negative, then a
24-hour holter would be prescribed next. If the holter study is
still negative then a mobile cardiac telemetry (MCT) study might be
performed for up to 30 days.
[0018] Remote ambulatory cardiac monitoring system 10 can perform
all of these functions as individual instruments have done in the
past and can change the way these tests are usually performed. A
single instrument can save money and improve patient compliance
particularly when going from a holter to a MCT recorder because a
new apparatus is not necessary. Because of the built-in cellular
modem technology, patient demographic and scheduling information
can be "pushed" from an EMR or HIS system to device 10 over a
network such as network 11. Conversely, test results can
automatically be sent back the electronic management system of to a
24 hour monitoring service such as remote monitoring service 12.
Monitoring system 10 is "smart" in that it automatically adapts to
its intended purpose by detecting the ECG electrode set or other
cable inserted into cable socket 14. For example insertion of three
electrode cable 16 will cause monitoring system 10 to initiate MCT
mode. Insertion of ten electrode cable 17 will cause monitoring
system 10 to initiate 12-lead EKG mode. Remote ambulatory cardiac
monitoring system 10 also provides video tutorials and other
information on display screen 18 related to electrode connection,
specific to the modality and can also provide real-time ECG
arrhythmia detection using its built-in hardware DSP engine in any
mode: resting 12-lead; holter; or MCT.
[0019] Monitoring system 10 also integrates various ECG-specific
functions that are well-suited for scalable electrocardiogram (ECG)
applications. The device can also be used in high-performance,
multichannel data acquisition system by powering down the
ECG-specific circuitry in the controller.
[0020] Referring now to FIG. 3, controller 20 has a highly
programmable multiplexer that allows for temperature, supply, input
short, and RLD measurements. Additionally, a multiplexer allows any
of input electrode lines 21 such as electrode line 21A to be
programmed as the patient reference drive. The PGA gain can be
chosen from one of seven settings (1, 2, 3, 4, 6, 8, and 12). The
ADCs in the device offer data rates from 250 SPS to 32 kSPS.
Communication to the device is accomplished using an SPI-compatible
interface. The device provides four GPIO pins for general use.
Multiple devices can be synchronized using the START pin.
[0021] The internal reference can be programmed to either 2.4V or
4V. The internal oscillator generates a 2.048 MHz clock. The
versatile right leg drive (RLD) block allows the user to choose the
average of any combination of electrode lines 21 to generate the
patient drive signal for signal noise reduction.
[0022] Referring now to FIG. 2, lead-off detection and electrode
analysis can be accomplished either by using a pull-up/pull-down
resistor or a current source/sink such as the voltage divider
formed by electrode 23 and reference resistor 24 of electrode input
channel 25. A fixed voltage or current is applied at reference
point 25R and electrode 23 is connected at input point 21B. As
electrode 23 separates from the patient's skin and or the electrode
adhesive decays, the resistance between electrode 23 and the
patient will grow relative to reference resistor 24. Controller 20
can quantify the change of electrode resistance and adjust signal
gain to compensate until the electrode separates from the patient.
Input signal 26 may be analyzed for the contribution of reference
voltage or current at reference point 25R and if signal 26 is
entirely the reference voltage or current, a lead-off alarm is
initiated.
[0023] Monitoring system 10 supports both hardware pace detection
and software pace detection. The Wilson center terminal (WCT) block
can be used to generate the WCT point of the standard 12-lead
ECG.
[0024] Remote ambulatory cardiac monitoring system 10 is a
multipurpose device designed with the ability to perform: a
3-channel (3-electrode) Mobile Cardiac Telemetry complete with
Arrhythmia Detection and Alarm for up to 30 days; a 24-hour or
longer 12-channel or 3-channel (5-electrode) ECG Holter study; and
a resting 12-channel EKG. The device may be used for patient in
home use with remote clinician data analysis as well as use within
the office or home setting by a medical professional.
[0025] Remote cardiac monitor 10 includes:
[0026] 1) a single component ECG monitor with an integrated
cellular modem and
[0027] 2) an interface to five independent cable configurations
through a single connector. Monitor 10 automatically changes
functionality when a specific cable with the same form factor is
inserted with the following configurations: [0028] 3-wire,
ambulatory, snap electrode cable invokes a 1, 2, or 3-channel MCT
mode (Lead I, II, III--no anterior views) [0029] 10-wire,
ambulatory, snap electrode cable invokes the Holter 12-channel mode
[0030] 5-wire, ambulatory, snap electrode cable invokes the Holter
3-channel mode by default (up to 7-leads are available with
anterior views); If programmed to do so, the system can
automatically switch to the MCT mode after 24- or more hours [0031]
10-wire, resting (lengthened for full body), alligator clip
electrode cable invokes the Resting 12-lead EKG mode (8-channels;
Lead III, aVF, aVR, aVL derived) [0032] 4-wire, USB cable invokes
the PC communication service mode. Cable is interchangeable with
ECG lead sets requiring disconnection from the body before
connection to an external device can be made.
[0033] The built-in cellular modem technology pushes and pulls
information to and from the device in a HIPAA compliant fashion
using the cryptographic protocol; Transport Layer Security (TLS).
Additional data integrity is performed by Error Correction Coding
(ECC) and MD5 hash sums.
[0034] Monitor 10 houses a microprocessor for running the algorithm
and an application specific integrated circuit (ASIC) as controller
20, a rechargeable battery, real time ECG arrhythmia detection
using built-in hardware DSP engine in any mode, ECG capture
circuitry provided by the ASIC and the multiple components,
GSM/GPRS/EGPRS/WCDMA/HSPA
[0035] Upon detection of an arrhythmic or patient-activated event
through activation of patient, the ECG signal is transmitted
wirelessly via the cellular network to a remote monitoring center
for additional analysis and intervention by a clinician. When
cellular service is unavailable, the event will be stored until
such time the cellular network becomes available or the patient
transmits the data using a land telephone line.
[0036] If in the resting 12-lead mode, the device can capture,
display, and analyze 12 channels of ECG for up to 1,024 patients in
diagnostic quality. The ECG can be streamed in real-time to a
computer wirelessly via the 802.15.4 network transceiver to be
displayed, printed, and stored. An embedded SQL database is used in
monitor 10 for ECG storage and reporting in all modes--MCT, Holter,
and resting 12-lead as well as storage of other data such as blood
pressure data.
[0037] Conventional 3-lead electrode systems provide a single
channel measurement with no redundancy. Referring now to FIGS. 1
and 4, the MCT mode of remote ambulatory cardiac monitoring system
10 is initiated by connection of cable 16 which uses three
electrodes, electrodes 16A, 16B and 16C, configured as RA (Right
Arm), LA (Left Arm) and LL (Left Leg) to calculate Lead I (channel
I), Lead II (channel II), or Lead III (channel III). Controller 20
dynamically substitutes RL (Right Leg) drive 27 for each electrode
as it makes a measurement across the other two electrodes. This
allows the system to function as if it were a four electrode system
making use of the noise cancellation of the RL output. The RL drive
also has an integrated constant current source 29 which can be
momentarily switched on to make an impedance measurement of each
connected electrode. This allows system to make an immediate
lead-off or high impedance (HiZ) determination and use the
remaining two electrodes to make a single lead measurement.
Controller 20 can measure and calculate six separate data channels
(Lead I, Lead II, Lead III, aVL, aVR, and aVF) with the effect of
full AC noise cancellation via the RL drive provided to each of the
three electrodes are connected. In the case of any electrode
becoming disconnected, the system will still function as a single
channel system.
[0038] RL drive 27 sums the ECG signals from two of the three
electrodes, the LA, RA, and LL electrodes, and applies the combined
signal 180 degrees out of phase and drives the resulting inverted
signal, signal 27, back into the patient for noise cancellation. In
MCT mode, remote ambulatory cardiac monitoring system 10 only uses
three electrodes such as electrodes 16A, 16B and 16C, so the fourth
electrode (RL) has to replace one of the limb lead electrodes
dynamically. This done by electronically switching the LL input
circuitry in ECG instrumentation amplifier 30, shown in FIG. 4,
with the RL output drive circuitry. The LA-RA (channel I)
measurement orientation is show in FIG. 5 as channel 31. LL
Electrode circuitry is replaced by the RL circuitry.
[0039] Controller 20 electronically switches the LA input circuitry
in ECG instrumentation amplifier 30 with the RL output drive
circuitry. The RA-LL (channel II) measurement orientation is shown
in FIG. 5 as channel 32. LA Electrode circuitry is replaced by the
RL circuitry.
[0040] Controller 20 next electronically switches the RA input
circuitry in the ECG instrumentation amplifier with the RL output
drive circuitry. The LA-LL (channel III) measurement orientation is
show in FIG. 5 as channel 33. RA Electrode circuitry is replaced by
the RL circuitry.
[0041] The RL drive circuitry sequentially connects from the LL to
LA to RA so the Noise cancellation system functions without the
fourth RL electrode being necessary. Once the data has been
acquired for Lead I, channel 31, and Lead II, channel 32 then Lead
III, channel 33 can be derived or measured directly. This is
possible because Einthoven's Law states that I+(-II)+III=0. The
equation can also be written I+III=II. It is written this way
(instead of I-II+III=0). Therefore Lead III is Lead II-Lead I. Now
the derived channels, channels 34, 35, and 36 can be
calculated:
channel 34,aVR (Lead I+Lead II)/2
channel 35,aVL Lead I-(Lead II/2)
channel 36,aVF Lead II-(Lead I/2)
Lead III Lead II-Lead I
[0042] Using only 3 electrodes, electrodes 16A, 16B and 16C it is
possible to derive 6 channels (Limb leads & derived leads) of
data with the noise cancellation (ground) effect of a virtual
dynamic RL electrode.
[0043] All ECG signals will be analyzed real-time by the internal
algorithm of controller 20. Upon determination of an arrhythmic or
patient-activated event such as by activation of event button 38,
composite ECG data 39 will be wirelessly transmitted to a remote
Monitoring Center such as monitoring center 12 for additional
analysis and intervention. If cellular transmission is not
possible, the event will be stored for future cellular/TTM/base
station transmission such as through conventional phone 40 or base
station 41.
[0044] Monitoring device 10 will be equipped with a service center
lab changeable flash-based memory 42 in order to record the
receiving signal coming from the sensor and to allow pre and post
processing options through the use of this memory and a mobile
database of ECG data and corresponding event markers. The database
will facilitate query and retrieval of specific ECG based on event
information, time and date, or other data as desired.
[0045] Remote ambulatory cardiac monitoring system 10 will
automatically transmit the recorded ECG, via cellular technology,
to a Remote Management System but will also allow patients to use a
land line such as phone 40, plugged into a base station such as
base station 41 or held to the MCT Device in a standard
Trans-telephonic way (TTM) as illustrated in FIG. 1, to transmit
the data in the case where cellular transmission is not
possible.
[0046] Six elements generally influence physicians' decision on
which provider to use for a particular ECG monitor: [0047]
Compliance: Ensure greatest amount of quality ECG for analysis
[0048] Higher Diagnostic Yield: Provide most clinically-relevant
information [0049] Technology: Reliable transmission, easy to use,
less patient intervention [0050] Reimbursement: highest
reimbursement rates for Partner Physicians [0051] Innovation and
fashion: Best solution offered for arrhythmia detection and
transmission, including continuous storage and monitoring. Engaging
new technology
[0052] Device Overview
[0053] Remote ambulatory cardiac monitoring device 10 is a
battery-powered, portable and ambulatory, multi-purpose ECG device
capable of serving as (1) a resting 12-lead EKG (2) a multi-day
Holter recorder and (3) a Mobile Cardiac Telemetry (MCT) device
with real-time arrhythmia analysis and a cellular modem.
[0054] Cardiac monitoring device 10 is designed to be operated by
both patients and medical professionals and has two user-interfaces
to address each type of user. It can be operated through touch
screen 18 and alternatively using 5-button keyboard 44 with
complete functionality by either method.
[0055] Cardiac monitoring device 10 has four functional modes of
operation: resting 12-lead EKG (medical professionally only); 1-5
day Holter recorder (patient and medical professional interfaces);
30 day MCT operation (patient, medical professional and service
center interfaces); and a USB data transfer mode to download the
data stored in memory to a personal computer for long term storage
and to perform a more extensive analysis (medical professional
only). The built-in cellular modem technology allows patient
demographic and scheduling information to be "pushed" from an EMR
or HIS system to cardiac monitoring device 10. Conversely, the
results can automatically be sent back to same electronic
management system to a 24-hour monitoring service facility.
[0056] Cardiac monitoring device 10 is "smart," automatically
adapting to its intended purpose by detecting the ECG lead set
inserted. It also provides embedded video tutorials on hook-up
specific to the modality and can also provide real-time ECG
arrhythmia detection using its built-in hardware DSP engine in any
mode: resting 12-lead EKG; Holter; or MCOT.
[0057] 12-Lead EKG Mode:
[0058] Cardiac monitoring device 10 can function as a 12-lead EKG
with a special 10-wire, alligator-clip, electrode cable lead set.
It can be set in "Capture Only" mode or "Capture & Stream"
mode.
[0059] "Capture Only" mode allows the operator to acquire data in
monitor or diagnostic quality mode and store up to 1024 patients
EKG on the device. The EKG can later be transmitted wirelessly or
downloaded directly via USB to central cardiovascular system
computer running receiving and analysis software 46.
[0060] The system is intended as a mobile device which can be
carried to the patient in a medical office or home environment.
Each EKG is displayed on display 18 which may be any suitable touch
screen display such as a 3.2'' TFT color display in high-resolution
and all 12 channels are able to be viewed. There is no EKG
interpretive analysis in this mode.
[0061] If in the "Capture & Stream" mode, cardiac monitoring
device 10 will operate as in "Capture Only" but also send the EKG
wirelessly via 802.15.4 transceivers to the central cardiovascular
system PC. Cardiac monitoring device 10 can deliver 3, 6, 12 EKG's,
interpretive (continued) analysis and vector loops where the data
may be viewed, printed, and stored. Additionally, the data is
analyzed in the PC using automated ECG analysis and interpretation
software library 47.
[0062] Holter Mode:
[0063] Cardiac monitoring device 10 can function as a Holter with a
3-wire, 5-wire, or 10-wire, snap, ambulatory, electrode cable lead
set. It can be set to convert to MCT (30 day ambulatory monitoring)
mode automatically if the Holter is negative after 24 or 48 hours.
It can be programmed to record 1, 2, 3, or 12 leads up to five
days. There are two user-interfaces available: one designed for the
medical professional (Tech Mode) which allows recording
configurations to be set and; another for the patient (Patient
Mode) with a simplified interface allowing events and symptoms to
be annotated within the device. There is also a third
user-interface where low-level system settings can be changed
intended for system administrators (Admin Mode). Cardiac monitoring
device 10 also contains hook-up diagrams and instructional videos
to aid in assisting to properly prep and place the electrodes
properly depending on the ECG lead set being used. Additionally,
monitoring system 10 will perform an impedance test, as discussed
above, of the electrode cable currently connected for validation of
the electrode connection. The impedance test quantifies the
resistance of each electrode connection and can compensate for
degrading connections that occur over time and also provides the
ability to detect an electrode off condition. After connection of
the electrodes, all channels of ECG can be reviewed on display
screen 18 by a medical professional.
[0064] The device screen is normally blank during recording but
will show the time and date if a key is pressed along with a
message stating "Recording". If "Event" button 38 is pressed, the
device will present a drop-down menu for both symptoms and
activities. The user may add menu items with the built-in virtual
keyboard through touch screen display 18 if desired. The patient
also has control over the screen brightness, speaker volume, and
the LED indicator.
[0065] Furthermore, the patient is not permitted to turn off the
device in holter mode. If the "Power" button is held in for more
than five seconds, a special "Tech Mode" key prompt (EVENT-SELECT),
only known to the medical technician, will need to be entered to
complete the power down function.
[0066] The holter data is also processed and analyzed software
library 47.
[0067] MCT Mode:
[0068] Cardiac monitoring device 10 can also function as a MCT
device with a 3-wire, snap electrode, ambulatory cable lead set
such as cable 16 of FIG. 1. There are two user-interfaces: one
designed for the medical professional which allows recording
configurations to be set (Tech Mode) and another for the patient
with a simplified interface (Patient Mode) allowing events and
symptoms to be annotated within the device and control over the
screen brightness, speaker volume, and the LED indicator. There is
also a third user-interface where low-level system settings can be
changed intended for system administrators (Admin Mode) including
patient enrollment, downloading all MCT data, and clearing all
memory.
[0069] As in holter mode above, MCT mode also provides hook-up
diagrams and instructional videos as well as performing an
impedance test and. After the electrode connection in MCT mode, all
channels can be reviewed on the screen. If a patient electrode
connection is performed, no ECG is presented; only a good or bad
electrode connection indication message.
[0070] Whenever monitoring device 10 is powered up, the patient
enrollment is verified and must be confirmed before the device may
be used.
[0071] The device screen is normally blank during recording but
will show the time and date if a key is pressed along with a
message stating "Recording". If event button 38 is pressed, device
10 will present a drop-down menu for both symptoms and activities.
The user may add menu items with the built-in virtual keyboard if
desired. The system will use the display screen, and LED indicator,
and audible alarms to notify the user of any messages including a
phone number to make a voice call to the 24-hour monitoring service
facility. The service may also post messages for the patient.
[0072] Cardiac monitoring device 10 is programmed with preset
triggers and alarms for sinus bradycardia [based ventricular rate],
sinus tachycardia [based on ventricular rate], A-Fib, pauses, and
heart-block. If triggers are made then the event is just stored in
memory and downloaded at the end of the study. If an alarm is made,
the event is sent to a 24-hour monitoring service facility via the
cellular modem. The service facility may also adjust the triggers
and alarms remotely.
[0073] Controller 20 is also configured to interact with software
algorithm such as algorithm 55 to cross check each alarm trigger
before initiating an alarm. Using the sequential connection to data
channels discussed above, an alarm trigger from a data channel will
be checked using one or more other data channels to verify the
validity of the alarm trigger. The verification cross-check
prevents false positive alarms.
[0074] In the case where there is no extended cellular coverage,
there is a TTM method provided to send events by a conventional
(wired) telephone which requires no disconnection of the ECG lead
set.
[0075] Furthermore, the patient is not permitted to turn off the
device unless the "Power" button is held in for more than five
seconds; a software confirmation is made by the user to complete
the power down function. This would normally be done to perform
battery charging, to bathe or change electrodes.
[0076] USB Connection Mode:
[0077] The USB connection mode is performed by medical
professionals to download data from the device or to program the
internal settings including patient enrollment. Patients are not
permitted to enter this mode.
[0078] Hardware Description
[0079] Referring now to FIG. 3, the design of cardiac monitoring
device 10 is based on microprocessor 48 such as an Atmel
AT92SAM9G20 ARM 9 providing very low power consumption, (50 mA with
all internal peripherals active) and a Field Programmable Gate
Array (FPGA) controller 20. Any suitable microprocessor and gate
array may also be used.
[0080] Controller 20 controls the functionality of remote
ambulatory cardiac monitoring device 10 at the lowest level. There
are several modules within the system controller 20 and external to
the controller. External components include: the ECG Amplifier, the
Analog to Digital Converter, the Digital Pots for programming
offset and gains to the ECG amplifier, 3.2'' TFT color Display, a
keyboard and touch screen interface for the display, a Real Time
Clock, a Zigbee 802.15.4 interface, a Bluetooth interface, and
Accelerometer interface, a Cell Phone interface, Cellular SIM card,
SDHC flash memory, Power control circuit, Speaker, Microphone, RGB
LED indicator, Event and Power Buttons, and microprocessor 48.
[0081] Inside system controller 20 there are also a number of
modules. They are the Data Acquisition Module, the Processor Decode
module, the SDRAM controller module, the Digital Pot control
module, the Hardware Watchdog module, the Video Control module, the
Zigbee Control module, the Bluetooth Control module, the
Accelerometer Control module and the Cell Phone Control module.
Each of these modules has the appropriate Register sets to allow
total control via microprocessor 48. [0082] Zigbee transceiver
(802.15.4) [0083] Bluetooth transceiver [0084] Accelerometer [0085]
Cell phone/GPS [0086] Video Controller [0087] Touch panel
interface. [0088] ECG Data Acquisition [0089] Battery monitor
[0090] Digital POTS interface [0091] GPIO interface. [0092] DSP
core interface [0093] 8051 secondary Microprocessor core [0094]
Speaker/Microphone interface [0095] SDRAM [0096] RGB LED registers
& control [0097] Keyboard button controller [0098] System
decoder [0099] Resource State Machine & 8051 Arbiter [0100]
8051 Local decoder & Timing Generator [0101] System control
& Status Registers [0102] I2C Serial modules [0103] DCM Clock
Synthesis [0104] Real-Time Clock controller [0105] Byte swapper
[0106] Watch-Dog Timer Control [0107] Refresh Controller [0108]
SRAM Controller
[0109] The selection of a suitable processor as microprocessor 48
is based on the integration of fast ROM and RAM memories and a wide
range of peripherals. A suitable processor will embed a USB Device
Port and a USB Host controller. It may also integrate several
standard peripherals, such as the USART, SPI, TWI, Timer Counters,
Synchronous Serial Controller, ADC and Multimedia Card Interface.
Microprocessor 48 should include a maximum of six concurrent high
bandwidth 32-bit busses, and it should also feature an external bus
interface capable of interfacing with a wide range of memory
devices.
[0110] System controller and hardware accelerator block is directly
connected to the External Bus Interface of microprocessor 48.
Controller 20 is the interface controls all aspects of hardware not
integrated into the processor. The following blocks are implemented
internal to the controller, and all control and data transfers are
via memory mapped I/O for setup, control, and processor data
transfers. [0111] Acquire Module. This module controls all aspects
of the data acquisition functionality of the design. This module is
responsible for the acquisition Channel List Control, Frequency of
Scan, Transfer Interrupts, Interrupt Threshold Control, Analog To
Digital SPI interface control, Conversion Control, and houses the
buffering RAM that feeds the SDRAM Interface block. [0112] Local
Decoder. This module is responsible for the decoding of processor
accesses to all Read/Write registers accessible to the local
processor. It generates all timing and direction control to all
modules within the CORE system controller. [0113] Memory
Controller. This module directly controls the local SDRAM in all
aspects. The local refresh timer, SDRAM power down and power saving
control, local DMA transfers between the Acquire block and the
SDRAM, Processor access to the SDRAM, (both in random access and
FIFO mode access), arbitration control between the DMA requests,
Video data requests, and the local processor accesses, and all
addresses generated to the SDRAM. [0114] Local Watchdog Timer. This
module is programmable for general purpose use in timing for
whatever type of wakeup might be needed based on time. [0115]
Digital Pot Controller. This module controls all SPI accesses,
Digital Pot selection, timing control, and registering of all
processor accesses to and from the Digital pots. [0116] Bluetooth
Controller. This module interfaces to the Bluetooth hardware.
[0117] ZigBee Controller. This module interfaces to the ZigBee
hardware. [0118] Video Controller. This module controls the LCD
accesses by the processor, the raster scan, video data access. All
timing generation and sequencing is controlled from this module to
present whatever graphics are required. [0119] Touch Screen
Interface. This module controls the capacitive virtual button
interface that allows the user access to setup and control of the
hardware. [0120] Accelerometer Interface. This module is a direct
processor access to the 3 axis accelerometer so the display screen
orientation flips on position and detects patient position.
[0121] Operations
[0122] Remote ambulatory cardiac monitoring system 10 is a
multi-purpose ECG recorder based on a modular approach. Each
requirement of functionality of the design is isolated in its own
module, and has a standard intercommunication architecture with
other modules of functionality, and the local processor. At the
highest level of description, there are several modules within gate
array controller 20, and external to the gate array. There are the
ECG Amplifier, the Analog to Digital Converter, the Digital Pots
for programming offset and gains to the ECG amplifier, a capacitive
touch sense interface for a from panel, the front panel itself, a
Real Time Clock, a Zigbee interface, a Bluetooth interface, and
Accelerometer interface, a Cell Phone interface, and a processor
interface are all external to the gate array. Gate array controller
20 controls all of these above mentioned modules that are external
to the gate array.
[0123] Inside gate array, controller 20, there are also a number of
modules. They are the Data Acquisition Module, the Processor Decode
module, the SDRAM controller module, the Digital Pot control
module, the Hardware Watchdog module, the Video Control module, the
Zigbee Control module, the Bluetooth Control module, the
Accelerometer Control module and the Cell Phone Control module. Any
other suitable control module may also be added for additional
functionality such as blood pressure module 49 which captures blood
pressure data using ambulatory blood pressure cable 50. Each of
these modules has the appropriate Register sets to allow total
control via the local microprocessor.
[0124] The design of controller 20 is totally synchronous and
supports a single software controllable clock rate. There are many
different VHDL modules in this design.
[0125] Top Level Module (holter_top.vhd)
[0126] This is the top level VHDL module. It is a wrapper for all
of the other modules listed above. This module is responsible for
all port mapping of sub-modules. It also contains the system
control register, the potentiometer select register, write/read
strobe generation for I/O, The refresh timer, processor transaction
flow control, and reset synchronization. All other modules to
follow, are instantiated within this module.
[0127] Local Decode Module: (lcl_dec.vhd)
[0128] This module is responsible for the decoding local processor
accesses to the FPGA. It is capable of up to 256 word only
locations, (16 bits). The read/write strobe generation is done in
the top level module to be used in conjunction with the output of
this module. This decode is based on the activation of io_acc,
signal by hardware, and the processor request for read or write.
This module also decodes the address space for the SDRAM when CS7
is active. An SDRAM access is only decoded here, and not acted
upon. The action is when the SDRAM module sees the decode of SDRAM
space. The decodes are based on a simple 3 state machine.
[0129] 1. Idle_state: This is where the state machine spends most
of its time until an access to the decode module is sensed. When in
this state, the hardware is looking for an I/O access. This is done
with a comparator. Once detected, the io_acc signal goes active to
the state machine, and forces the state machine to put out the hold
signal to the processor during the transition to the next state,
decode1. It also latches the chip select lines, and the lower 9
down to 1 bits of address for decode1.
[0130] 2. Decode1: This state is entered one clock after the
detection of the leading edge of an I/O access. Bits 9 down to 1,
(8 bits) are then fed into a look-up table for decode, the proper
decode enables are activated to the rest of the FPGA, and is
unconditionally sent to the next state, decode2.
[0131] 3. Decode2: This state does an immediate release of the
io_hold generated on transition from the idle state to decode1. The
io_hold signal is also used by the top level module to form the
trailing edge of a write strobe if it is a write transaction.
During a read transaction, the read strobe is generated for the
duration of decode. In either case, decode2 waits for the io_acc to
clear as a result of the release of the io_hold signal. It then
advances back to the starting point, idle state, as soon as the
processor terminates the transaction. Transaction termination is
based on the negation of all Chip Selects.
[0132] Data Acquisition Module: (acquire.vhd)
[0133] This module controls, (after programming), all aspects of
the data acquisition sequence.
[0134] The major blocks within this module are:
[0135] 1. Scan Sequencer
[0136] 2. Programmable clock divider to program the
acquisitions/sec
[0137] 3. Processor interface
[0138] 4. A2D SPI interface
[0139] 5. Pot Selection
[0140] 6. Wake Up Count Register
[0141] 7. Memory Threshold Register
[0142] 8. DMA Count Register.
[0143] 9. A2D Mux Address
[0144] 10. Calibration Voltage Select
[0145] 11. Input Select
[0146] 12. ADC Pot Select
[0147] 13. Wake Up Interrupt Generation
[0148] 14. A2D Converter control
[0149] 15. A2D SPI Data accesses
[0150] 16. Conversion Frequency
[0151] 17. DMA transfers to SDRAM requests
[0152] Scan Sequencer
[0153] The Scan Sequencer is made up of a 16.times.5 bit RAM that
holds the following control:
[0154] 1. Channel to be converted, (bits 3:0)
[0155] 2. End Sequence Flag, (bit 4)
[0156] 3. Processor Multiplexer Address
[0157] 4. Data acquisition sequence address
[0158] After the processor loads the desired address into the
Processor Multiplexer Address register, the required processor
access location is setup. When a decode of the scan ram sequencer
is decoded on a processor access, the value of the Processor
Multiplexer Address is sent to the address inputs of the scan ram.
The Data transaction (read/write), will take place to the channel
location set up previously in the Processor Mux Address register.
This location is one of 16 channels feeding the Analog to Digital
Converter addressed by the Scan Ram.
[0159] The Scan Ram Sequencer is simply a four bit counter that
increments after every scan. The address, (count), comes from the
Scan Sequencer logic when running. If location 0 of the scan ram
wanted to scan channel 5, then the processor would set up the
Processor Mux Address register to 0, and write a 5 into that
location. The scan list can be programmed to acquire any channel in
any order to provide for maximum flexibility
[0160] This RAM is accessed in two different manners. Either the
processor addresses it to read/write the RAM, (Holter_Run=`0` and
Mux_Ctl_Mode=`1`), or the acquisition control state machine scan
sequence register, (Holter_run=`1`).
[0161] When in the processor mode of access, the processor accesses
this RAM as a Memory Mapped resource. When the local decoder
decodes processor access to this RAM, the hardware passes the
address in the processor multiplexer address register to the
address inputs of the scan ram, along with the lower five bits of
the system data bus, (sys_dat) to the data inputs of the RAM, on a
write. During a read, the internal sys_data is tri-stated, and the
contents of the RAM are put on bits (4:0) for a read access. Based
on the read strobe, (rstrb), and write strobe, (wstrb), the data is
either read from the RAM to bits (4:0) of the sys_data bus from the
address specified in the processor multiplexer address, or a write
to the same location from sys_data (4:0).
[0162] When running, (Holter_Run=`1`), the address presented to the
Scan RAM is from the data acquisition state machine, (scan_seq).
These accesses are always reads. At the beginning of a convert
cycle, just before the conversion command is given to the A2D
converter, the address of the scan sequence is presented to the
Scan RAM, and its contents are sent to the external analog
multiplexer that feeds the input to the A2D converter. After this,
on the next clock, the convert command is given. After the
conversion, the scan sequence address is incremented with one
exception. If bit 4 of the Scan RAM contents is `1` for the address
given, after the conversion the scan sequencer is reset to "0000".
This restarts the scan list. The Scan list length is programmable.
To make the scan list 1 deep, (then it repeats itself), simply set
BIT 4 of the data, (channel included in the data). BIT 4 of the
output of the Scan Ram directly resets the counter addressing the
scan ram. Therefore, bit 4 being set at location zero, will lock
the counter at location 0 when running. For a scan list of 2, write
BIT 4 of location 1 to a 1. This causes a counter reset when
location 1 is accessed allowing the counter to cycle between
locations 0 and 1. Writing a 1 to BIT 4 of location 0xF makes the
Scan list 16 long. The scan list can be programmed to acquire any
channel in any order to provide for maximum flexibility. For a scan
list of 16, writing the same channel address in all locations would
be logically the same as a single length list. All 16 locations
would be accessed, but they would all have the same channel #. This
is true for any sequence. This allows the setup to scan one channel
more times than another during a scan run allowing individual
channels to be acquired at different sampling rates.
[0163] Programmable Clock Divider to Program the
Acquisitions/Sec
[0164] This is a 16 Bit register that counts 100 ns intervals.
Loading this register with a 1 would cause an acquisition clock
every 200 ns. This is an illegal number and this register should
never be loaded with a value of less than the maximum speed of the
A2D converter. Resolution is 100 ns. The slowest clock rate is
65535.times.100 nS, or 1 channel every 6.5546 mS per channel. A
scan rate of 1 MHz, (1 channel every microsecond), would load a
value of 9 to this register, (0x0009). Clock rate=N+1 (100 ns),
where N=value of this register. A value of zero is illegal and will
produce undetermined results.
[0165] Processor Interface
[0166] The Processor interface accepts decodes from the local
decoder pertinent to this module of the design. This interface is
based on the output of the local decoder module. When the local
processor is making an access, the local decoder sends the
appropriate decode signals to this interface. The following signals
are pertinent to the update/access to the following blocks of this
module:
[0167] 1. Clock Divider Register
[0168] 2. Acquisition Control Register, (described earlier).
[0169] 3. Scan RAM
[0170] 4. Processor MUX Address.
[0171] 5. Wake Up Count Register
[0172] 6. Memory Threshold Register
[0173] 7. DMA Count Register.
[0174] 8. A2D Data Path
[0175] 9. A2D FIFO Path
[0176] 10. SPI Data Path/Control
[0177] A2D SPI interface
[0178] The A2D SPI interface is for the purpose of all local
processor to A2D protocol conversion and control. The A2D is an SPI
based device that has a 16 bit transfer in either direction. There
is a state machine to handle both sides of this interface. On one
side of the state machine is the local processor. The local
processor sets appropriate bits to the state machine for high level
control, and for data writes and data reads. On the other side of
the state machine is the A2D Converter SPI Interface. This part of
the interface, (acquisition control and translation), is also fed
by the Base Time Acquisition clock, (adc_clk), that is a direct
derivative of the Programmable Clock Divider described above. Once
the local processor asserts the holter_run bit, every adc_clk
starts another conversion from the channel addressed from the Scan
RAM. This can be confusing due to the fact that the Scan RAM both
produces addresses and accepts them for look-up. The address in
this discussion is the output of the Scan RAM. This is done with a
multiple, (8 state), state machine.
[0179] IDLE: This is the state entered after a power up, a reset,
or a completion of a conversion. While in this state, the state
machine is looking at the holter_run bit. This bit is set on the
next leading edge of a conversion clock, (adc_clk). At this time
the SPI data out pin of the FPGA is driven high to do the CS mode
of operation, this forces the busy bit active mode. The state
machine then transitions to START_CNV.
[0180] START_CNV: During this state the adc_cnv clock is asserted.
This instructs the A2D converter to convert the analog value
presented at its input. There are 16 data bits in conversion
transfer. The lower 16 bits are converted data and the 17th bit,
(busy) is the MSB of the serial chain. Then the next state BUSY_W8
is entered.
[0181] BUSY_W8: Upon entering this state, the state machine waits
for the data out of the SPI, (SD0 from the ADC) to go to a zero
state. This signifies the negation of the busy signal from the
previous convert pulse, (ADC_CNV). Once this is detected, the state
machine negates A2D_SDI, and asserts the SPI_CLK signal, shifting
out the busy bit. The state machine then transitions to the
BUSY_SHIFT state.
[0182] BUSY_SHIFT: Upon entering this state, the state machine
negates the SPI_CLK, and forms the trailing edge of the first
SPI_CLK. There is one system clock skipped to allow for minimum
pulse width of the ADC SPI interface. After this delay, the state
machine asserts the leading edge of the next pulse. We then
transition the BUSY_SHIFT1 state.
[0183] BUSY_SHIFT1: This state, when entered skips one machine
clock to not violate the minimum pulse width of the SPI_CLK. After
this delay, SPI_CLK is negated and the state machine transitions to
GET_DATA1.
[0184] GET_DATA1: This state, when entered again skips one machine
clock to allow for setup time from the SPI_DATA pin, but at the
same time negates the SPI_CLK. After this delay the internal shift
count is incremented, and the data from the SPI is shifted into the
LSB of the serial to parallel register, and at the same time, the
MSB is shifted out to the SPI interface. The information shifted
into the serial to parallel register is a function of the command.
If the command is a read command, the register specified in the
command is accessed. If it is a write command, the command word is
written right back into the serial to parallel register as a
function of the SPI interface of the A2D converter. On this same
clock edge, the SPI_CLK is again asserted, and the transition to
GET_DATA2 takes place.
[0185] GET_DATA2: Upon entering this state, the SPI_CLK is negated.
On the same clock edge, the shift count is checked for 16. If the
result is false, the state machine transitions back to GET_DATA1.
If the count is 16, the load a2d fifo signal is activated,
(LD_A2D_FIFO). This causes a shift of the parallel word just
captured, and the state machine transitions to TERMINATE1.
[0186] TERMINATE1: This state negates the SPI_CLK signal and jumps
to the IDLE state.
[0187] Algorithm Specifications
[0188] This section shall define the functional requirements of the
system and sub-system of controller algorithm 20A.
[0189] Sinus Pause:
[0190] A sinus pause will be documented whenever a predetermined
period of time elapses without a QRS complex of any morphology.
This algorithm will not discriminate between a Sinus Pause and
Sinus Block as this interval timer will be based on R-Wave to
R-Wave intervals. The programmable time period for a Sinus Pause is
1.5, 2.0, 2.5, 3.0, 3.5 or 4.0 seconds, with 2.0 seconds being the
factory default setting. The detection of Sinus Pause can be turned
on or off without affecting any of the other arrhythmia
settings.
[0191] Bradycardia:
[0192] Bradycardia occurs when the patient's heart rate goes below
a preset limit. The heart rate for calculating Bradycardia is based
on a 6-beat rolling average, a 10-beat rolling average or a 16-beat
rolling average that can be selected in the configuration software.
The factory default setting is a 10-beat rolling average. The
programmable heart rate settings for Bradycardia are 30, 35, 40,
45, 50, 55, 60, 65 and 70 beats per minute (BPM) with 40 BPM being
the factory default setting. The detection of Bradycardia can be
turned on and off without affecting any of the other arrhythmia
configuration settings.
[0193] Bradycardia Time Offset:
[0194] Bradycardia within a patient can often go below the preset
limit when sleeping. In order to accommodate for this phenomena,
the heart rate for calculating Bradycardia can be lowered by 5, 10,
15, 20 beats per minute (BPM) with 10 BPM being the factory
default. The selection for window set in Bradycardia will be
utilized for this setting. The Bradycardia Time Offset can be
turned on and off without affecting any of the other arrhythmia
configuration settings and is OFF by default.
[0195] Tachycardia:
[0196] Tachycardia occurs when the patient's heart rate goes above
a preset limit or in the event of 3 or more sequential Ventricular
beats. The heart rate for calculating Tachycardia is based on a
6-beat rolling average, a 10-beat rolling average or a 16-beat
rolling average that can be selected in the configuration software.
The factory default setting is a 10-beat rolling average. The
algorithm will calculate the rate for both Normal and Ventricular
beat such that it will detect Sinus Tachycardia, Supraventricular
Tachycardia and Ventricular Tachycardia. The programmable heart
rate settings for Tachycardia are 70, 75, 80, 85, 90, 95, 100, 105,
110, 115, 120, 125, 130, 135, 140, 145, 150, 155, 160, 165, 170,
175, 180, 185, 190, 195, and 200 beats per minute (BPM) with 100
BPM being the factory default setting. The detection of Tachycardia
can be turned on and off without affecting any of the other
arrhythmia configuration settings.
[0197] Atrial Fibrillation:
[0198] The AFib detection algorithm uses a probability density
function (PDF) method based on a reconstructed attractor of R-R
intervals. By constructing the PDF of distance between two points
in the reconstructed phase space of R-R intervals of normal sinus
rhythm (NSR) and atrial fibrillation (AF), the distributions of PDF
of NSR and AF R-R intervals have significant differences. By taking
advantage of their differences, a characteristic quantity parameter
kn, which represents the sum of n points slope in filtered PDF
curve, is used to detect both NSR and AF R-R intervals. AFib
detection is based on dynamically created PDFs and is not dependent
on reference PDFs from known published databases from MIT-BIH, AHA,
etc. Those databases were, however, used in preliminary algorithm
development to determine the optimal embedding dimension m, the
characteristic quantity Kn, and the slope number used by the
algorithm for optimal performance as described below.
[0199] Stage 10 consists of Phase Space Reconstruction. The R-R
interval is just one observable variable from the multi-variable
cardiac system. According to Takens' embedding theorem, the state
of a dynamic system can be represented in a reconstructed phase
space by time delay embedding. Taken's theorem guarantees that the
dynamic characteristics of the real physiological) and the
reconstructed system are the same. Therefore, it is possible to
reconstruct the phase space of the original cardiac system from the
R-R intervals while preserving the system dynamics. If x(i) is an
element of the R-R intervals vector X, then a reconstructed vector
Yi is introduced as
Yi=[x(i),x(i+.tau.), . . . ,x(i+(m-1).tau.)] (Equation 1)
where .tau. is the time delay and m is the embedding dimension. The
reconstructed vector Yi represents a point in the m-dimensional
phase space. By reconstructing the R-R intervals and investigating
the structure of the reconstructed attractor, it is possible to
provide more information and new diagnostic potential of the
analyzed cardiac system. The phase space reconstruction consists of
5 dimensions (embedding dimension m=5) based on a 40 beat window
([40,5] phase-space array) to a 100 beat window ([100,5] phase
space array). A 40-beat windows gives speed the priority (real-time
application) where moving towards a 100-beat window gives accuracy
the priority (real-time application with DSP assist or PC-based
system).
[0200] Stage 11 constructs the Probability density function. The
basic idea behind the PDF method is to construct the correlation
function C(r, m, .tau.). C(r, m, .tau.) describes the probability
that the distance between arbitrary two points Yi and Yj in the
phase space is shorter than distance r. C(r, m, .tau.) can be
written as
C(r,m,.tau.)=2/(N(N-1)).SIGMA._(i=1) (N-1).SIGMA._(j=i+1) N
.theta.(r-.parallel.Y.sub.--i-Y.sub.--j.parallel. (Equation 2)
where N is the number of phase space points, symbol
.parallel..cndot..parallel. represents the Euclidean distance, and
.theta.(x) is the Heaviside unit-step function which is defined
by
.theta.(z)={1Z.gtoreq.0}{0Z<0} (Equation 3)
Equations (1)-(3) are essential steps of the Grassberger and
Procaccia (GP) method to calculate the correlation dimension. The
determination of the correlation dimension from R-R intervals are
commonly used for gaining information about the nature of the
underlying cardiac dynamics. Since the correlation function C(r, m,
.tau.) has a maximum of 1, minimum of 0 and is a continuous
distribution function, we define the probability density function
as
p(r,m,.tau.)=dC(r,m,.tau.)/dr
[0201] The PDF is now used for the analysis of the reconstructed
attractor structure. A low pass filter (IIR Butterworth type, eight
orders with cutoff frequency 1.5 Hz) is used to filter the high
frequency fluctuating components of the PDF curves. After low pass
filtering, PDF curves of short-time R-R intervals have a near
Gaussian distribution.
[0202] Stage 12 calculates the Characteristic Quantity kn from the
Receiver Operating Curve (ROC). rmin, rtop and rmax are defined as
the abscissas of the starting point, the peak and the ending point
of the filtered PDF curves, respectively. The values of rmin, rtop
and rmax of NSR R-R intervals are smaller than those of AF R-R
intervals. This is due to the fact that the neighboring NSR R-R
intervals are more correlated, so points in the reconstructed phase
space are more concentrated; whereas the neighboring AF R-R
intervals are less correlated, so points in the reconstructed phase
space are more dispersed. However, rmin, rtop and rmax are not good
indexes for differentiating AF from NSR. Therefore, the slope
information of the filtered PDF curves is used to improve the
detection precision
k.sub.--n=[PDF]_filtered/r_max(n).times.100
rrmax is divided evenly into 100 segments and d is the length of
each segment. PDFfiltered (n) is the value of the nth point in the
filtered PDF curve. Naturally, the 0th point in the filtered PDF
curve is 0. Therefore, kn denotes the sum of the slope of n points
in the filtered PDF curve. The characteristic quantity kn fully
takes advantage of the following characteristics of filtered PDF
curves: (1) most of the values of rmax of AF are bigger than those
of NSR; (2) most of the values of rtop of AF are bigger than those
of NSR, which means filtered PDF curves of NSR rise much faster
than those of AF; (3) most of the values of rmin of AF are bigger
than those of NSR, i.e., among the n points slope, AF has more
small slope than NSR. All these characteristics make the value of
kn of NSR bigger than that of AF as long as n is carefully
selected. Stage 13 uses Kn to determine whether the PDF represents
values of NSR or AFib. If a 40 beat window is used then the number
of points used to determine the slope is 20 and a threshold of 210
for Kn is used to determine whether the PDF value is characteristic
of AFib of NSR. If a 100 beat window is used, the number of the
slope points n is 10 and 78 is used for the threshold of Kn. The
same concept applies to 60 and 80 beat windows.
[0203] Thus, while the preferred embodiments of the devices and
methods have been described in reference to the environment in
which they were developed, they are merely illustrative of the
principles of the inventions. Other embodiments and configurations
may be devised without departing from the spirit of the inventions
and the scope of the appended claims.
* * * * *