 |
Description  |
|
|
The present invention relates to methods of medical diagnosis and
displaying diagnostic data. More specifically, novel methods of observing
and analyzing the relationship between the ventilatory cycle and the
natural pacing of the heart are disclosed. The information derived from
such observations is a useful clinical tool for diagnosing autonomic
neural dysfunction, natural pacemaker dysfunction, and other physiological
disorders. Specific applications of these methods to isolating and
observing the neural activity governing cardioventilatory interaction are
also disclosed.
BACKGROUND OF THE INVENTION
It is well known that many biological systems involve rhythmic functions
which repeat in an almost regular cyclic fashion. Examples of such
rhythmic functions include, but are not limited to: the motion of one limb
n a gait pattern, the contraction of the heart, the movement of the chest
and diaphragm in respiration, the contractions of segments of the
intestine, the rise and fall of populations of different species of
animals within an ecosystem and the rhythmic twitching symptomatic of
certain neurological disorders. Further, in many cases, it is also known
(and, in other cases, it is believed) that the rates and strengths of
these rhythmic functions are modulated by underlying control functions
which may, in turn, be loosely related to other rhythmic functions.
Examples of such phenomena are the relationships between two different
limbs in a gait pattern, the effect of respiration on heart rate, the
effect of small bowel contractions on rumen contractility in cows, and the
interrelationship between the population cycles of predator and prey in
closed ecosystems.
Perhaps the best known of these interrelated rhythmic functions is that
relationship which exists between respiration and heart rate. It is well
known that for the heart to function efficiently in perfusing lung and
peripheral tissues, its rate and strength of contraction must be
coordinated with several factors including respiration, vascular load and
tissue demand for oxygen.
A cardiovascular system that is responsive to changing physiological states
requires a cooperative interaction between the cardiac, ventilatory, and
vascular systems. While ventilatory activity and the vascular bed are
intrinsically coupled to the heart through mechanical interactions, the
major factor controlling coherence in their function is the autonomic
nervous system (ANS), which communicates with the cardiac pacemaker known
as the sinoatrial (SA) node. Although it is widely accepted that an
altered state of neural interaction with the heart accompanies a variety
of pathological conditions such as congestive heart failure and diabetes,
the details of this interaction remain poorly understood.
At present, the only means of understanding the effect of the autonomic
nervous system is to study the aggregate effect of neural stimulation on
average heart rate, since the origin of any single nerve impulse is
uncertain. Since all physiological variables that contribute to the neural
traffic cannot be accounted for further refinements are difficult to
achieve.
Considered separately, the heart has its own series of pacemakers, the most
prominent of which is the sinoatrial node (SA node), which produces an
electrical depolarization which spreads throughout the heart in a
coordinated way, producing a single contraction of the heart muscle--a
heartbeat. The electrical signal produced by the spreading depolarization
of the heart muscle can be measured at the skin surface of a subject and
visually represented in an electrocardiogram (EKG or ECG). The electrical
signal related to the entire cardiac cycle consists of two distinct
periods: (1) the period of electrical activity when the depolarization
occurs; and (2) a period of electrical quiet in between heartbeats (the
interbeat interval). The measurement of this electrical signal is often
used by scientists veterinarians and physicians as a means of monitoring
certain cardiac and cardiovascular functions.
Much information has been accumulated about the patterns of electrical
discharges through the us of EKG's. By analyzing the waveforms of single
heartbeats, those skilled in the art can interpret such waveforms and make
certain diagnoses based on these single heartbeat patterns. Abnormal
electrical discharges originating in the ventricle of the heart, for
example, are easily detected in the EKG and produce patterns in the EKG
record characteristic of ventricular beats readily distinguishable from
the normal beats originating from the SA node. One catastrophic condition,
ventricular fibrillation, is also easily recognized in the EKG pattern.
This effect on the heart rate induced by ventilatory activity is known as
respiratory sinus arrythmia. In the simplest terms, the heart rate
increases on inspiration and decreases upon expiration. Research has shown
that this modulation of the heartbeat is controlled through the interplay
of two branches of the autonomic nervous system, which involuntarily
transmits impulses to internal organs. See, A.D. Jose and R.R. Taylor,
"Autonomic blockade by propanol and atropine to study intrinsic myocardial
function in man", J.Clin.Inves. 48, 2019-31 (1969); J.A. Hirsch and B.
Bishop, "Respiratory sinus arrythmia in humans: how breathing pattern
modulates heart rate", Am.J.Physiol 241, H620-29 (1981), both of which are
incorporated by reference as if fully reproduced herein. Of the two neural
branches, the parasympathetic branch, which is the craniosacral portion of
the autonomic nervous system, is of particular interest. It has been found
that a decrease in parasympathetic activity during the inspiratory phase
accounts for much of the observed increased heart rate. See, T.A. Bruce,
et al., "The role of autonomic and myocardial factors in cardiac control",
J.Clin.Inves. 42, no.5, 721-26 (1963); P.G. Katona, et al., "Cardiac
vagal efferent activity and heart period in the carotid sinus reflex",
Am.J.Physiol. 218, no. 4, 1030-37 (1970), both of which are incorporated
by reference as if fully reproduced herein.
It has also long been known that the heart rate, as measured either by EKG
or by pulse counting, is not constant and varies with a number of
parameters. Prominent among the parameters that affect heart rate in
resting subjects is the respiratory phase. Respiration itself is a
variable rhythmic event under control of the central nervous system (CNS)
in all animals that occurs with much slower frequencies than the heart
rate. As explained above, it is known that during the relatively long
inspiratory phase of respiration in normal individuals and animals, the
heart rate increases and, conversely, during expiration, the heart rate
decreases. However, it is further known that this alteration in heart rate
occurs as a result of neural input to the SA node, principally from the
parasympathetic portion of the autonomic nervous system coursing in the
right vagus nerve. In persons with certain conditions, such as diabetes,
heart transplants, and some forms of congenital anomalies, this increase
and decrease in heart rate in loose synchrony with the inspiration and
expiration is absent or minimal in magnitude. This absence of synchrony,
and a belief that quantitating the effects of neural input to the heart
would lead to a better understanding of cardiac function and dysfunction
have resulted in a long-felt, yet unsolved need for a method of
quantitating neural effects on cardiac rhythm. Those of ordinary skill
recognize that more specific diagnostic and prognostic information about
human and veterinary patients suffering from cardiac and other diseases
can be obtained via such quantiative methods which, prior to the present
invention, was unobtainable in reliable form.
The neural conduction system of the heart originates at the sinoatrial
(sinus or SA) node which is located at the junction of the superior vena
cava (SVC) and the right atrium. This node is the connection point for the
right vagus nerve, which communicates parasympathetic neural information.
At least two distinct neural components are expected to be related to the
ventilatory phase. The first of these is initiated by signals transmitted
to the brain by the lung and thoracic stretch receptors. These receptors
generate afferent neural impulses in response to air intake during
ventilation, which communicate with the sinus node via the brain stem. A
second neural component originates at the carotid and atrial
baroreceptors, the sensory receptors located in the arteries and within
the heart which respond to pressure variations and relay signals
representative of this information to the brain. The brain then transmits
this information to the sinus node via the parasympathetic nervous system.
This neural control of the natural pacemaker activity of a healthy heart
adds great complexity to any detailed understanding of the coupling
between the cardiovascular and ventilatory systems.
A crude method of determining the effects of parasympathetic nerve
stimulation on heart rate, known to those skilled in the art, is to
measure heart rate by counting beats under normal, at rest conditions and
then comparing this rate with the observed heart rate while applying
pressure to one eyeball, which is believed to induce a parasympathetic
neural decrease in heart rate. It is also known, for example, that direct
stimulation of the right vagus nerve will dramatically slow the heart rate
in individuals with functional neural input to the SA node.
Another method, more quantitative than either of the above and applicable
to human medicine, has been used by certain cardiologists and
physiologists, but has met with minimal success. This method relies on
frequent analysis of the electrocardiogram and of the respiratory cycle.
Utilizing the principle of Fourier Analysis, the EKG is broken down into
various imaginary constant components of differing frequencies and
amplitudes. On of ordinary skill will readily appreciate that since the
actual depolarization of the cardiac mass occurs in a regular pattern of
much shorter duration than the overall cardiac cycle, the frequency
components of the electrically active period in the EKG are of higher
frequency than the overall cardiac cycle. Thus, changes in the heart rate
are most easily observed as changes in the heartbeat interval (hbi).
Consequently, in a power spectrum analysis of many sequential heartbeats,
the power in the higher frequencies will be due principally to the
electrical signal produced during depolarization, while the power in the
lower frequencies will be more related to the interbeat interval. If the
heart rate is varying considerably, the power in the lower frequencies
will be spread out, while if there is no variation in heart rate (and,
consequently, no variation in interbeat interval) the power in the lower
frequencies will be more concentrated.
There are, however, severe deficiencies in the results obtained from this
method. First, the heartbeat is not exactly periodic; since the
calculations that must be performed to estimate frequency information can
only be done practically through digital Fourier Transforms--which assume
perfectly periodic signals--several approximations must be made in data
interpretation. These approximations, however, may mask the underlying
behavior of the system. Second, by using power spectral information, all
phase information is lost. Since phase may be an important consideration
in obtaining meaningful results, any result obtained which is unrelated to
this parameter is at best incomplete. Third, because the power spectrum is
dependent upon the amplitude of the electrical signal, this method is
extremely sensitive to such factors as electrode placement, patient
position, and disease conditions such as fluid in the chest or
pericardium. Finally, these difficulties are exacerbated exponentially
when similar approximations to the power spectrum are used to correlate
the frequencies of the respiratory cycle with those of the cardiac cycle
through ratio calculations. Thus, this method is also inadequate to fully
study the effects of parasympathetic nerve stimulation on the heart.
Studies of cardioventilatory interaction typically utilize data taken on
mechanically ventilated subjects. In order to provide reliable results,
the extent to which the neural activity associated with free breathing has
been reproduced must be determined. During both free breathing and
inspiration imposed by a mechanical positive pressure ventilator,
pulmonary and thoracic stretch receptors which initiate the transmission
of afferent impulses to the respiratory center in the brain stem via the
right and left vagus nerves are activated. From the brain stem, efferent
nerve discharges are delivered through the phrenic nerve to the diaphragm
and through the right vagus nerve to the SA node. This feedback mechanism
results in synchronization between ventilation and the neural control of
the heartbeat, whether the ventilation is naturally o externally
controlled. In an artificially ventilated subject, neural activity which
is normally associated with "free" breathing is essentially reproduced
when the respiration is externally and consistently imposed, but this
neural activity is now synchronized with the externally imposed rhythm.
The neural discharges which govern cardioventilatory interaction ar
mediated through the parasympathetic branch of the autonomic nervous
system. These neural discharges result in the deposition of acetylcholine
at the SA node, which induces perturbations in the ion flows across the
cell membranes, thus altering the excitation interval of the cardiac
pacemaker. In general, the effect of the neural impulse is dependent upon
the phase within the heartbeat interval at which the acetylcholine is
delivered, as well as upon the heart rate and the sympathetic tone, but
the result is a discrete change in the heartbeat interval which spans the
neural discharge.
Therefore, although comparing neural data from mechanically respirated and
naturally respirated subjects is conceptually valid, there is still a long
felt but unsolved need for methods which allow neural data to be analyzed
and processed, and to identify discrete neural impulses associated with
specific pathologies.
Thus, it is known that the normal functioning of the cardiovascular system
requires a cooperative interaction between the heart and the respiratory
system. It is further known that the respiratory activity may couple
directly to the heart through mechanical interactions, a condition known
as phase locking. For example, there can be effects due to the local
physical environment of the heart changing as the chest cavity expands
during breathing. As explained at the outset, phase locking is a very
general phenomenon in any dynamical system, whether physical or biological
in origin. See Levy, et al., "Paradoxical effects of vagus nerve
simulation of heart rate in dogs," Circ. Res., vol. 25, pp. 303-14 (1969);
Jalife et al., "Dynamic vagal control of pacemaker activity in the
mammalian sinoatrial node," Circ. Res., vol. 52, pp. 642-56 (1983); and
Glass et al., "Global bifurcations of a periodically forced biological
oscillator," Phys. RevA, vol. 29, p. 1348 (1984), all of which are
incorporated by reference as if fully reproduced herein. In addition to
these direct mechanical couplings, the natural pacemaker of the heart is
also affected via the nervous system. The basic physiological pathways
involved in this feedback loop are known; in a healthy heart there are
direct repetitive neural impulses emanating from the brain stem which are
synchronous with ventilation, these neural couplings are mediated through
the carotid and atrial barroreceptors which have direct feedback to the SA
node. The simultaneous interaction of all these influences results in a
highly complex dynamical system.
OBJECTS OF THE INVENTION
It is thus an object of the present invention to provide means of
quantifying and graphically representing the effect of one modulatory
input system which is loosely correlated with a recordable rhythmic event
upon the rate of another recordable rhythmic event in a biological system.
It is a further object of the present invention to provide a means for
quantifying and graphically representing the effect of a cyclic but
continuously varying modulatory neural input, derived indirectly from the
respiratory cycle, on the underlying beat of the heart produced by its own
irregulatory cyclic pacemaker, the SA node.
It is also an object of this invention to provide methods for identifying
neural impulses associated with cardiac activity and a physiological
cycle, such as the ventilatory cycle.
It is another object of this invention to provide methods for obtaining
useful data from the cardio-respiratory system and to provide methods for
processing and converting such data into a format which permits analysis
of cardiac behavior resulting from neurological input.
It is a further object of the present invention to provide methods of
displaying cardio-respiratory data in a format which reveals the presence
or absence of cardioventilatory phase locking.
It is a still further object of the present invention to provide methods
for presenting information useful for diagnosing and treating heart
disease.
It is another object of this invention to provide methods for analyzing the
effects of drugs upon the cardiac neural function.
It is also an object of this invention to provide methods for determining
cardioventilatory phase locking which are incorporated into the control
systems of monitoring apparatus.
SUMMARY OF THE INVENTION
The present invention presents methods and apparatus for analyzing data
collected from rhythmic systems. In accordance with the present invention,
data representative of a first rhythmic activity and data representative
of a second substantially non-periodic rhythmic activity are concurrently
collected. The first and second data collected are then processed to
relate the first activity to the phase of said second activity, phase
being the measure of a point in the cycle of rhythmic activity relative to
its beginning and end points. Finally, a determination is made as to the
presence or absence of ordered data structures indicative of the
interrelationship which exists between the first and second activities.
In a preferred embodiment, cardiac and ventilatory activity of a subject ar
the activities from which data is taken. In a first application, a
determination is made as to whether the ordered data structure comprise
substantially repetitive data points representing cardiac activity
occurring at a plurality of distinct locations within the phase of the
ventilatory activity, such data structures being indicative of
cardioventilatory phase locking. This application of the methods of the
present invention permits the examination of the effects of respiratory
activity on the time interval between heart beats. Means for interpreting
the processed data, such as phase maps of the difference in heart beat
intervals versus the ventilator phase, are generated to allow a comparison
of successive heart beat intervals and their relation to respiratory
phase. These phase maps can be used to determine the presence of
cardioventilatory phase locking, which has been shown to be diagnostic of
cardiac neural feedback mechanisms which have been obstructed or removed.
Cardioventilatory phase locking is evidenced by the grouping of values of
heart beat interval difference at several locations within the ventilator
phase. The existence of this type of data structure over a period of time
is indicative of cardioventillatory phase locking, which may be brought
about by a decrease in the neural activity which mediated the
interrelationship between cardiac and ventilator activity.
In another application, a determination is made as to whether data points
within a region of the phase of the ventilatory activity exhibit a
substantially greater variation in the maximum and minimum data points
representing cardiac activity, as compared to all other regions of the
ventilatory phase; large variations within a well-defined region being
indicative of neural activity.
In this application of the methods of the present invention by rigorously
controlling a single variable within a rhythmic system (e.g., ventilation)
and appropriately processing the data collected, the effect of neural
impulses associated with particular activities and conditions can be
observed in the data. In accordance with a preferred embodiment,
variations of the heartbeat interval are processed by averaging these
values over the ventilator phase. As long as the heart rate is not
synchronized with ventilation, a plot of the differences in heartbeat
interval versus respiratory phase will show sharp variation near that
region of the ventilator phase where nerve impulses which are in synchrony
with the ventilatory cycle arrive. The waveforms which emerge from such a
plot are qualitatively similar to each other over time, independent of the
precise mechanism by which the neural impulses are interpreted by the SA
node. Using the methods of the present invention neural impulses which
occur regularly, but which are not periodic in time can be identified.
Thus, it is now possible to determine the presence or absence of
particular neural impulses non-invasively, and in real time.
The methods and apparatus of the present invention therefore allow the
diagnosis of these and other disorders associated with natural cardiac
pacemaker and neural dysfunction, and are further applicable to the study
of the effects of drugs which modify natural pacemaker activities and to
other related areas of biomedical research or other clinical applications.
BRIEF DESCRIPTION OF THE DRAWINGS
FIGS. 1A-1F are six representative phase maps, created using the methods of
the present invention. FIGS. 1A-1C display the behavior of a heart prior
to transplantation, while FIGS. 1D-1F reveal phase-locking present in the
same heart, denervated by transplantation.
FIGS. 2A-2C illustrate the behavior of a heart using phase maps generated
at successive time intervals after the injection of a neural blocking
drug. FIG. 2C reveals the presence of phase locking after the introduction
of a neural blockade.
FIGS. 3A-3C illustrate data from which ventilator phase maps are generated.
In FIG. 3A, the raw data representing heart beat interval and ventilator
phase are shown. FIG. 3B represents a modification of the data shown in
FIG. 3A.
FIG. 4 illustrates three plots, successive in time. These data were
overlaid in FIG. 3B. The dashed lines highlight the pronounced features
which persist regularly at substantially the same ventilator phase.
FIG. 5A and 5B illustrate a comparison of phase maps taken from innervated
(FIG. 5A, pre-transplant) and denervated (FIG. 5B, post-transplant)
hearts.
FIG. 6 is a simplified flowchart of a method of the present invention.
FIG. 7 illustrates a simplified block diagram of preferred for carrying out
the present invention.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
The novel methods of the present invention utilize the concepts of
quasiperiodic systems theory to achieve an understanding of the dynamics
of cardiac rhythms. The theory of quasiperiodic systems has been applied
to many dynamic systems during the last few years and important lessons
have been learned about how to uncover underlying system behavior from
time series data. See, D. Rand, et al., Phys.Rev.Lett. 49, 405 (1982); D.
Rand, et al., Physica 5D, both of which are incorporated by reference as
if fully reproduced herein.
A quasiperiodic time series contains a component called the hull function
which can be used to determine the essential dynamics of the system. One
preferred embodiment of the present invention discloses novel methods of
uncovering a hull function describing cardiac activity from time series
data. Further explanation of the application of the quasiperiodic systems
theory as applied to cardiac neural control can be found in Appendix A,
Appendix B, and Appendix C which are attached hereto. Preferably, the data
representing the time series consists of a series of successive
differences of heart-beat intervals (/ hbi), measured to about 10
microsecond accuracy.
As previously described, the normal functioning of the cardiovascular
system requires a cooperative interaction between the heart and
respiratory system. It is clear that respiratory activity must couple
directly to the heart beat through mechanical interactions. It is also
known that the natural pacemaker of the heart is affected via the
parasympathetic branch of the autonomic nervous system. Therefore, the
methods of the present invention are applied to show that respiratory
activity must contribute a quasiperiodic component to the time series.
The hull function governing the cardioventillatory relationship is composed
of several separable functions. For a given hull function,
V(x)=S(x)+N(x)+R(x). The first of these, S(x), is described as the
"smooth" component due to the direct interaction of the structure and form
of the chest cavity reflected on the ECG trace, which is used to determine
the interbeat interval measurement. If discrete neural activity in phase
with the breathing cycle exists, there is a discontinuous or
non-differentiable component, N(x). Finally, there will be an apparently
random component, R(x), that contains activity which is either truly
random or apparently random since it is decoupled from respiratory
activity.
The methods for the present invention make the observation of autonomic
neural activity possible by detailed measurements of its effect through
the function N(x described above. Quasiperiodic methods of processing time
series data possess the ability to uncover information which is not
necessarily periodic. This characteristic is particularly useful t the
study the dynamics of physiological systems. In principle, the behavior
uncovered by quasiperiodic techniques is impossible to uncover using
Fourier Transform techniques, such as those previously described, due to
the inherent non-periodicity of the data.
One of ordinary skill will appreciate that there are numerous systems to
which the methods of the present invention are applicable. For example,
neurological information derived from electroencephalographs and
physiological data extracted from electrocardiograms can now be analyzed
to provide diagnostic information previously unavailable. The methods
presented are also particularly useful for examining a variety of rhythmic
interrelations, such as those previously described. The methods of the
present invention place in the hands of the physician or researcher a tool
which "freezes" a particular non-periodic pattern in a manner analogous to
the apparent standstill produced by a strobe light and pendulum
synchronized to the same frequency. Using this new tool, the discovery and
analysis of an entire class of previously hidden patterns within
physiological data may now be examined, creating insights into the
underlying pathologies which cause them.
As applied to the cardioventillatory system, the present invention
comprises methods whereby the discontinuous component, N(x), is uncovered
from biological data in order to obtain a more precise picture of the
neural activity associated with respiration. Previously, the overall
patterns created by the three components created an indecipherable
picture, of limited use as either a clinical or diagnostic tool. It is now
possible to collect data and ascertain the effect of neural activity o the
cardioventillatory system, by isolating the N(x) component of the hull
function. Until now, the isolation and measurement of this neural traffic
was only performed to a limited extent on laboratory animals and was not
obtainable from human subjects.
In accordance with a preferred embodiment of the data evaluation methods of
the present invention, the neural feedback mechanisms regulating the
natural pacemaker are removed, either surgically or pharmacologically, and
data reflecting the change in heart beat interval and ventilatory phase
are collected. By correlating these data in accordance with the novel
method of the present invention, it is now possible to examine the
relationship between ventilation and cardiac activity. These correlated
data may be analyzed for the presence or absence of cardioventilatory
phase locking or other ordered data structures, such information being
useful in the clinical study and diagnosis of autonomic neural
dysfunction, natural pacemaker disorders, as well as other conditions.
Preferably, the data collected consist of two time series. The first,
t.sub.n, consists of the time of the n'the QRS complex, as measured by an
electrocardiograph, preferably measured by 750,000 Hz clock to ensure
sufficient data accuracy. One of ordinary skill in the art will appreciate
that the QRS complex, that portion of an electrocardiogram representing
the period of the depolarization of the ventricles, represents one way of
obtaining these data. Other methods or apparatus which collect the same
data or provide data from which this time series can be derived are
equally applicable to the analysis and method of the present invention. By
using the differences between successive t.sub.n, the heartbeat interval,
hbi, may be obtained from the equation hbi.sub.n =t.sub.n+1 -t.sub.n. The
other data series consists of chest expansion measurements taken about
every 0.25 seconds using the digitized output of an impedance
plethysmograph. An impedance plethysmograph measures volume changes in
terms of the change in bulk impedance between electrodes placed at two or
more points on the skin surface. The chest expansion measurements are
thereby used to obtain the ventilator frequency converted to the units of
t.sub.n. Using the frequency thus obtained, it is now possible to compute
the | | |
