## Implications for Neuroautonomic Heart Rate Control in Health and Disease

An expanded and updated version of this material appears in:
Bolis CL, Licinio J, eds. Please cite this publication when referencing this material, and also
include the standard citation for PhysioNet:
Goldberger AL, Amaral LAN, Glass L, Hausdorff JM, Ivanov PCh, Mark RG, Mietus JE, Moody GB, Peng C-K, Stanley HE. PhysioBank, PhysioToolkit, and PhysioNet: Components of a New Research Resource for Complex Physiologic Signals. |

## I. Basic Concepts

Clinicians and basic investigators are increasingly aware of the
remarkable upsurge of interest in nonlinear dynamics, the branch of
the sciences widely referred to as "chaos theory." Those
attempting to evaluate the biomedical relevance of this field confront
a daunting array of terms and concepts, such as *nonlinearity,
fractals, periodic oscillations, bifurcations* and
*complexity*, as well as *chaos* (1-4). Therefore, the
present discussion provides an introduction to some key aspects of
nonlinear dynamics, with a particular emphasis on heart rate control.
A major challenge is in making these concepts accessible, not only to
basic and clinical investigators, but to medical and graduate students
at a formative stage of their training.

### A. Introduction: The Concept of a Time Series

To appreciate the general clinical relevance of dynamics to the heartbeat,
consider the following common problem. What is the best way to compare a
sequence of measurements obtained from two subjects, or from one individual or
experimental procedure under different conditions? Conventionally, clinicians
and investigators rely primarily on a comparison of means using appropriate
statistical tests. However, the limitations of such traditional analyses
become apparent when evaluating the data in Fig. 1, showing
sinus rhythm heart rate plots collected from a healthy subject and one with
congestive heart failure. Recording the instantaneous signal from any system
over a continuous observation period generates a *time series*. What is
noteworthy in this example is that these two time series have nearly identical
means and variances, suggesting no clinically relevant differences. Yet,
visual inspection indicates that the two sequences of data display a markedly
different organization. The healthy heartbeat trace shows a complex,
"noisy" type of variability, whereas the data set from the patient
with heart failure reveals periodic oscillations in heart rate repeating about
1 cycle/minute (~.02 Hz). Time series analysis is concerned with quantifying
the *order* of data points; nonlinear dynamics provides a deeper
understanding of the mechanisms of patterns and differences such as those in
Fig. 1.

Figure 1. Two heart rate time series, one from a healthy subject (

*top*) and the other from a patient with severe congestive heart failure (CHF) (

*middle*) have nearly identical means and variances (

*bottom*), yet very different dynamics. Note that according to classical physiological paradigms based on homeostasis, neuroautonomic control systems should be designed to damp out noise and settle down to a constant equilibrium-like state. However, the healthy heartbeat displays highly complex, apparently unpredictable fluctuations even under steady-state conditions. In contrast, the heart rate pattern from the subject with heart failure shows slow, periodic oscillations that correlate with Cheyne-Stokes breathing.

### B. Linear versus Nonlinear Systems

In linear systems, the magnitude of the output (*y*) is controlled by
that of the input (*x*) according to simple equations in the familiar
form *y=mx+b*. A well-known example of such a relationship is Ohm's
law: V=IR where the voltage (V) in a circuit will vary linearly with
current (I), provided the resistance (R) is held constant. Two
central features of linear systems are *proportionality* and
*superposition*. Proportionality means that the output bears a
straightline relationship to the input. Superposition refers to the
fact that the behavior of linear systems composed of multiple
components can be fully understood and predicted by dissecting out
these components and figuring out their individual input-output
relationships. The overall output will simply be a summation of these
constituent parts. The components of a linear system literally
"add up" - there are no surprises or anomalous
behaviors.

In contrast, even simple nonlinear systems violate the principles of
proportionality and superposition. An example of a deceptively complex
nonlinear equation is *y = ax (1-x)*,*logistic equation* in population biology (5). The nonlinearity of this
equation, which describes a parabola, arises from the quadratic
(*x ^{2}*) term. Changes in the output as a function of
sequential time steps can be readily plotted by a feedback procedure in which
the current value of the output becomes the next value of the input, and so on.
Iteration of the simple-in-form logistic equation reveals dynamics that are
extraordinarily complex; depending on the value of the single parameter,

*a*, the same equation can generate steady states, regular oscillations, or highly erratic behavior (4, 5). Thus, for nonlinear systems, proportionality does not hold: small changes can have dramatic and unanticipated effects. An added complication is that nonlinear systems composed of multiple subunits cannot be understood by analyzing these components individually. This reductionist strategy fails because the components of a nonlinear network interact, i.e., they are coupled. Examples include the "cross-talk" of pacemaker cells in the heart or neurons in the brain. Their nonlinear coupling generates behaviors that defy explanation using traditional (linear) models (Fig. 2). As a result, they may exhibit behavior that is characteristic of nonlinear systems, such as self-sustained, periodic waves (e.g., ventricular tachycardia) (6, 7); abrupt changes (e.g., sudden onset of a seizure) (8) and, possibly, chaos (see below). Table 1 gives a more complete, but not exhaustive list of nonlinear mechanisms with potential relevance to biology and medicine (1-4, 6-14).

Figure 2. Examples of nonlinear dynamics of the heartbeat. Panels (a-c) are from subjects with obstructive sleep apnea syndrome. Panels (d and e) are from healthy subjects at high altitude (~15,000 ft).

A related and noteworthy property of nonlinear dynamics is referred
to as *universality* (1, 4). Surprisingly, nonlinear systems
that appear to be very different in their specific details may exhibit
certain common patterns of response. For instance, nonlinear systems
may change in a sudden, discontinuous fashion. One important and
universal class of abrupt, nonlinear transitions is called a
*bifurcation* (1, 9). This term describes situations where a
very small increase or decrease in the value of some parameter
controlling the system causes it to change abruptly from one type of
behavior to another. For example, the output of the same system may
suddenly go from being wildly irregular to a highly periodic, or vice
versa. A universal class of bifurcations occurring in a wide variety
of nonlinear systems is the sudden appearance of regular oscillations
that alternate between two values (15). This type of dynamic may
underlie a variety of *alternans* patterns in cardiovascular
dysfunction. A familiar example is the beat-to-beat alternation in
QRS axis and amplitude seen in some cases of cardiac tamponade (16).
This kind of electrical alternans is related to the back and forth
swinging motion of the heart within the pericardial effusion.
Multiple other examples of alternans in perturbed cardiac physiology
have been described, including ST-T alternans which may precede
ventricular fibrillation (17), and pulsus alternans during heart
failure.

### C. Chaos

Although the focus of much recent attention, chaos *per se*
actually comprises only one specific subtype of nonlinear dynamics.
Prior to the work of the renowned French mathematician, Henri
Poincaré, in the early 1900s, science was dominated by the
seemingly inviolable tenet that the behavior of systems for which one
could write out explicit equations (e.g., the solar system) should be,
in principle, fully predictable for all future times (18). What
Poincaré discovered (and what was more recently rediscovered)
is that a complex type of variability can arise from the operation of
even the simplest nonlinear system, such as that governed by the
logistic equation mentioned earlier. Because the equations of motion
which generate such erratic, and apparently unpredictable behavior do
not contain any random terms, this mechanism is now referred to as
*deterministic chaos* (1, 4). The colloquial use of the term
chaos - to describe unfettered randomness, usually with catastrophic
implications - is quite different from this specialized usage.

The extent to which chaos relates to physiological dynamics is a subject of active investigation and some controversy. At first it was widely assumed that chaotic fluctuations were produced by pathological systems such as cardiac electrical activity during atrial or ventricular fibrillation (19). However, this initial presumption has been challenged (20) and the weight of current evidence does not support the view that the irregular ventricular response in atrial fibrillation or that ventricular fibrillation itself represents deterministic cardiac chaos (21). Further, there is no convincing evidence that other arrhythmias sometimes labeled "chaotic," such as multifocal atrial tachycardia, meet the technical criteria for nonlinear chaos. An alternative hypothesis (22) is that the subtle but complex heart rate fluctuations observed during normal sinus rhythm in healthy subjects, even at rest, are due in part to deterministic chaos, and that a variety of pathologies, such as congestive heart failure syndromes, may involve a paradoxical decrease in this type of nonlinear variability (Fig. 1). Because the mathematical algorithms designed for detecting chaos are not reliably applied to nonstationary, relatively short and often noisy data sets obtained from most clinical and physiological studies, the intriguing question of the role, if any, of chaos in physiology or pathology remains unresolved (22-28).

D. Fractal Anatomy

The term *fractal* is a geometric concept related to, but not
synonymous with chaos (29, 30). Classical geometric forms are smooth and
regular and have integer dimensions (1,2, and 3, for line, surface, and volume
respectively). In contrast, fractals are highly irregular and have
non-integer, or fractional, dimensions. Consider a fractal line. Unlike a
smooth Euclidean line, a fractal line, which has a dimension between 1 and 2,
is wrinkly and irregular. Examination of these wrinkles with the low-power
lens of a microscope, reveals smaller wrinkles on the larger ones. Further
magnification shows yet smaller wrinkles, and so on. A fractal, then, is an
object composed of subunits (and sub-subunits) that resemble the larger scale
structure, a property known as self-similarity (Fig. 3). A
wide variety of natural shapes share this internal look-alike property,
including branching trees and coral formations, wrinkly coastlines, and ragged
mountain ranges. A number of cardiopulmonary structures also have a
fractal-like appearance (2, 3, 30, 31). Examples of self-similar anatomies
include the arterial and venous trees, the branching of certain cardiac muscle
bundles, as well as the ramifying tracheobronchial tree and His-Purkinje
network.

Figure 3. Left, schematic of a tree-like fractal has self-similar branchings such that the small scale (magnified) structure resembles the large scale form. Right, a fractal process such as heart rate regulation generates fluctuations on different time scales (temporal "magnifications") that are statistically self-similar. (Adapted from Goldberger AL. Non-linear dynamics for clinicians: chaos theory, fractals, and complexity at the bedside.

*Lancet*1996;

**347**:1312-1314.)

From a mechanistic viewpoint, these self-similar cardiopulmonary
structures all serve a common physiologic function: rapid and
efficient transport over a complex, spatially distributed system. In
the case of the ventricular electrical conduction system, the quantity
transported is the electrical stimulus regulating the timing of
cardiac contraction (31). For the vasculature, fractal branchings
provide a rich, redundant network for distribution of O_{2}
and nutrients and for the collection of CO_{2} and other
metabolic waste products. The fractal tracheo-bronchial tree provides
an enormous surface area for exchange of gases at the
vascular-alveolar interface, coupling pulmonary and cardiac function
(30). Fractal geometry also underlies other important aspects of
cardiac function. Peskin and McQueen (32) have elegantly shown how
fractal organization of connective tissue in the aortic valve leaflets
relates to the efficient distribution of mechanical forces. A variety
of other organ systems contain fractal structures that serve functions
related to information distribution (nervous system), nutrient
absorption (bowel), as well collection and transport (biliary duct
system, renal calyces) (2, 3, 30).

### E. Scaling in Health and its Breakdown with Disease

An important extension of the fractal concept was the recognition that it applies not just to irregular geometric or anatomic forms that lack a characteristic (single) scale of length, but also to complex processes that lack a single scale of time (29, 33). Fractal processes generate irregular fluctuations on multiple time scales, analogous to fractal objects that have wrinkly structure on different length scales. Furthermore, such temporal variability is statistically self-similar. A crude, qualitative appreciation for the self-similar nature of fractal processes can be obtained by plotting the time series in question at different "magnifications," i.e., different temporal resolutions. For example, Fig. 3 plots the time series of heart rate from a healthy subject on three different scales. All three graphs have an irregular ("wrinkly") appearance, reminiscent of a coastline or mountain range. The irregularity seen on different scales is not visually distinguishable, an observation confirmed by statistical analysis (34, 35). The roughness of these time series, therefore, possesses a self-similar (scale-invariant) property.

Since scale-invariance appears to be is a general mechanism underlying many
physiological structures and functions, one can adapt new quantitative tools
derived from fractal mathematics for measuring healthy variability. Complex
fluctuations with the statistical properties of fractals have not only been
described for heart rate variability, but also for fluctuations in respiration
(36), systemic blood pressure (37), human gait (38) and white blood cell counts
(39), as well as certain ion channel kinetics (3). Furthermore, if
scale-invariance is a central organizing principle of physiological structure
and function, we can make a general, but potentially useful prediction about
what might happen when these systems are severely perturbed. If a functional
system is self-organized in such a way that it does *not* have a
characteristic scale of length or time, a reasonable anticipation would be a
*breakdown* of scale-free structure or dynamics with pathology (35).
How does a system behave after such a pathologic transformation? The
antithesis of a scale-free (fractal) system (i.e., one with multiple scales) is
one that is dominated by a single frequency or scale. A system that has only
one dominant scale becomes especially easy to recognize and characterize
because such a system is by definition *periodic* - it repeats its
behavior in a highly predictable (regular) pattern (Fig. 4). The theory underlying this prediction may account
for a clinical paradox: namely, that a wide range of illnesses are associated
with markedly periodic (regular) behavior even though the disease states
themselves are commonly termed "dis-orders" (39).

Figure 4. Breakdown of a fractal physiological control mechanism can lead ultimately either to a highly periodic output dominated by a single scale or to uncorrelated randomness. The top heart rate time series is from a healthy subject; bottom left is from a subject with heart failure; and bottom right from a subject with atrial fibrillation. (Adapted from Goldberger AL. Non-linear dynamics for clinicians: chaos theory, fractals, and complexity at the bedside.

*Lancet*1996;347:1312-1314.)

## II. Fractal Scaling of the Heartbeat in Health and its Breakdown with Disease

### A. Periodic Disease and the Loss of Fractal Complexity

The appearance of highly periodic dynamics in many disease states
is one of the most compelling examples of the notion of *complexity
loss in disease* (40). Complexity here refers specifically to a
multiscale, fractal-type of variability in structure or function.
Many disease states are marked by less complex dynamics than those
observed under healthy conditions. This *de-complexification* of
systems with disease appears to be a common feature of many
pathologies, as well as of aging (40). When physiologic systems
become less complex, their *information content* is degraded
(41). As a result, they are less adaptable and less able to cope with
the exigencies of a constantly changing environment. To generate
information, a system must be capable of behaving in an unpredictable
fashion (2, 42). In contrast, a highly predictable, regular output
(i.e., a sine wave) is information-poor since it monotonously repeats
its activity. (The most extreme example of complexity loss would be
the total absence of variability - a straightline output.)

Quantitative assessment of periodic oscillations can be obtained by analyzing the time series of interest with a variety of standard mathematical tools. For systems producing a highly periodic output, the most widely used methods are based on spectral analysis. Remarkably, the time series of many severely pathologic systems have a nearly sinusoidal appearance; the spectrum shows a dominant peak at this characteristic frequency. An example is the heart rate variability sometimes observed in patients with severe congestive heart failure (Fig. 1) (43, 44) or with fetal distress syndromes (45). In contrast, systems with a fractal output (such as normal heart rate variability) show a type of broadband spectrum which includes many different frequencies (scales).

Probably the first explicit description of the concept of
*periodic diseases* was provided over 30 years ago by Dr. Hobart
Reimann (46). He called attention to a number of conditions in which
the disease process itself could be shown to *flare* or
*recur* on a regular basis of days to months; ranging from
certain forms of arthritis to some mental illnesses and hereditary
diseases, such as familial Mediterranean fever. In the late 1970s,
Michael Mackey and Leon Glass (47, 48) helped to rekindle interest in
this dormant field when they introduced the term *dynamical
disease* to encompass the types of periodic syndrome Reimann had
catalogued, as well as irregular dynamics thought possibly to
represent deterministic chaos.

Reimann's original list was premised on the assumption that periodic conditions were somewhat unique, and even idiosyncratic, in clinical medicine. However, to the extent that healthy function is often characterized by a multi-scale fractal complexity, we would anticipate that the emergence of single-scale (i.e., non-fractal) states might be considerably more common, if not ubiquitous, in pathophysiology. Indeed, a recent survey of the literature (49) indicates that Reimann, rather than compiling a list of the exceptional, was more likely sampling a widespread, even generic manifestation of the dynamics of disease. From the most general perspective, the practice of bedside diagnosis itself would be impossible without the loss of complexity and the emergence of pathologic periodicities. To a large extent, it is these periodicities and highly-structured patterns - the breakdown of multi-scale fractal complexity under pathologic conditions - that allow clinicians to identify and classify many pathologic features of their patients. Familiar examples include periodic tremors in neurologic conditions, AV Wenckebach patterns, the "sine-wave" ECG pattern in hyperkalemia, manic-depressive alterations, and cyclic breathing patterns in heart failure.

### B. Irregular Dynamics and the Breakdown of Fractal Mechanisms

While fractals are irregular, not all irregular structures or
erratic time series are fractal. A key feature of the class of
fractals seen in biology is a distinctive type of long-range order.
This property generates *correlations* that extend over many
scales of space or time. For complex processes, fractal long-range
correlations are the mechanism underlying a "memory" effect;
the value of some variable, e.g., heart rate at a particular time, is
related not just to immediately preceding values, but to fluctuations
in the remote past. Certain pathologies are marked by a breakdown of
this long-range organization property, producing an uncorrelated
randomness similar to "white noise." An example is the
erratic ventricular response in atrial fibrillation over relatively
short time scales. Peng et al. (50) have recently described a simple
algorithm for quantifying the breakdown of long-range (fractal)
correlations in physiological time series.

### C. Future Applications

Practical applications of nonlinear dynamics are likely within the next few years. Probably the first bedside implementations will be in physiological monitoring. A number of indices derived from chaos theory have shown promise in forecasting subjects at high risk of electrophysiologic or hemodynamic instability, including

- automated detection of the onset and end of pathologic low frequency (<.10 Hz) heart rate oscillations (Figs 1, 2 and 4) (43, 52-59);
- detection of subtle ST-T alternans (17, 51);
- detection of a breakdown in fractal scaling with disease and aging (43); and
- quantification of differences or changes in the nonlinear complexity or dimension of a physiological time series (59, 60).

In addition to these diagnostic applications, perhaps the most
exciting prospects are related to novel therapeutic interventions. An
important recent finding is that certain mathematical or physical
systems with complex dynamics can be controlled by properly timed
external stimuli: chaotic dynamics can be made more regular (*chaos
control*) and periodic ones can be made chaotic (*chaos
anti-control*) (61-63). One proposal, based on the earlier notion
that certain arrhythmias, particularly ventricular fibrillation,
represent cardiac chaos, is to develop chaos control algorithms to
electrically pace the heart beat back to sinus rhythm (63). A more
recent proposal is to use chaos anti-control protocols to treat or to
prevent cardiac arrhythmias or epilepsy based on the hypothesis that
restoration of a kind chaotic-like variability may actually be
advantageous (62).

Chaos theory also holds promise for illuminating a number of major problems
in contemporary physiology and molecular biology. Nonlinear wave mechanisms
may underlie certain types of reentrant ventricular tachyarrhythmias (6, 7).
Appreciation for the rich nonlinearity of physiological systems may have
relevance for modeling enormously complicated signal-transduction cascades
involved, for example, in neuroautonomic dynamics in which interactions and
"cross-talk" occur over a wide range of temporal and spatial scales,
as well as for understanding complex pharmacologic effects. Fractal analysis
of long DNA sequences has recently revealed that non-coding, but *not*
coding sequences possess long-range correlations among nucleotides (64). This
finding has implications for possible functions of introns as well as for
understanding molecular evolution (65) and developing new methods for
distinguishing coding from non-coding sections of long DNA sequences (66).
Findings from nonlinear dynamics have also challenged conventional mechanisms
of physiological control based on classical homeostasis, which presumes that
healthy systems seek to attain a constant steady state. *In contrast,
nonlinear systems with fractal dynamics, such as the neuroautonomic mechanisms
regulating heart rate variability, behave as if they were driven far from
equilibrium under basal conditions. This kind of complex variability, rather
than a regular homeostatic steady state, appears to define the free-running
function of many biological systems* (Fig. 1) (2, 39).
Finally, a fundamental methodologic principle underlying these new applications
and interpretations is the importance of analyzing continuously sampled
variations in physiological output, such as heart rate, and not simply relying
on averaged values or measures of variance. Dynamical analysis demonstrates
that there is often hidden information in physiological time series and that
certain fluctuations previously considered "noise" actually represent
important signals (67-70).

**Table 1**

- Abrupt Changes (10)
Bifurcations (1, 9)

Intermittency / Bursting (9, 10)

Bistability; Multistability (11)

Phase transitions - Hysteresis (12)
- Nonlinear Oscillations (1, 2)
Limit cycles

Phase resetting

Entrainment

Pacemaker annihilation - Fractals (2, 3)
Scale-invariance

Long-range correlations

Self-organized criticality

Diffusion limited aggregation - Alternans (9)
- Nonlinear waves: spiral; scroll (6, 7)
- Complex periodic cycles; quasiperiodicities (13)
- Stochastic resonance (14)
- Time irreversibility
- Chaos

### References

1. Glass L, Mackey MC. *From Clocks to Chaos: the Rhythms of Life*.
Princeton: University Press, 1988.

2. Goldberger AL, Rigney DR, West BJ. Chaos and fractals in human
physiology. *Sci Am* 1990;**262**:42-9.

3. Bassingthwaighte JB, Liebovitch LS, West BJ. *Fractal Physiology.*
New York: Oxford University Press, 1994.

4. Kaplan DT, Glass L. *Understanding Nonlinear Dynamics*. New York:
Springer-Verlag, 1995.

5. May RM. Simple mathematical models with very complicated dynamical
behavior. *Nature* 1976;**261**:459-67.

6. Winfree AT. *When Time Breaks Down: The Three-Dimensional Dynamics of
Electrochemical Waves and Cardiac Arrhythmias*. Princeton, NJ: Princeton
University Press, 1987.

7. Davidenko JM, Pertsov AV, Salomonosz R, Baxter W, Jalife J. Stationary
and drifting waves of excitation in isolated cardiac muscle. *Nature*
1992;**355**:349-51.

8. Babloyantz A, Destexhe A. Low dimensional chaos in an instance of
epilepsy. *Proc Natl Acad Sci USA* 1986;**83**:3513-7.

9. Goldberger AL, Rigney DR. Nonlinear dynamics at the bedside. In: Glass
L, Hunter P, McCulloch A. eds. *Theory of Heart*. New York: Springer,
1991; 583-605.

10. Chay TR, Lee YS. Bursting, beating, and chaos by two functionally
distinct inward current inactivations in excitable cells. *Ann NY Acad
Sci* 1990;**591**:328-50.

11. Longtin A, Bulsara A, Pierson D, Moss F. Bistability and the dynamics
of periodically forced sensory neurons. *Biol Cybern*
1994;**70**:569-78.

12. Delmar M, Ibarra J, Davidenko J, Lorente P, Jalife J. Dynamics of the
background outward current of single guinea pig ventricular myocytes: ionic
mechanisms of hysteresis in cardiac cells. *Circ Res*
1991;**69**:1316-26.

13. Courtemanche M, Glass L. Rosengarten MD, Goldberger AL. Beyond pure
parasystole: promises and problems in modelling complex arrhythmias. *Am J
Physiol* 1989; **257** (*Heart Circ Physiol*
**26**):H693-706.

14. Wiesenfeld K, Moss F. Stochastic resonance and the benefits of noise:
from ice ages to crayfish and SQUIDS. *Nature* 1995;**373**:33-6.

15. Glass L, Guevara MR, Shrier A. Universal bifurcations and the
classification of cardiac arrhythmias. *Ann NY Acad Sci*
1987;**504**:168-78.

16. Rigney DR, Goldberger AL. Nonlinear mechanics of the heart's swinging
during pericardial effusion. *Am J Physiol* 1989;257(*Heart Circ
Physiol* **20**):H:1292-1305.

17. Rosenbaum DS, Jackson LE, Smith JM, Garan H, Ruskin JN, Cohen RJ.
Electrical alternans and vulnerability to ventricular arrhythmia. *N Engl J
Med* 1994;**330**:235-41.

18. Peterson I. *Newton's Clock: Chaos in the Solar System*. New
York: WH Freeman, 1993.

19. Smith JM, Cohen RJ. Simple finite-element model accounts for wide range
of cardiac dysrhythmias. *Proc Natl Acad Sci USA*
1984;**81**:233-7.

20. Goldberger AL, Bhargava V, West BJ, Mandell AJ. Some observations on
the question: Is ventricular fibrillation "chaos?" *Physica D*
1986;**19**:282-9.

21. Kaplan DT, Cohen RJ. Is fibrillation chaos? *Circ Res*
1990;**67**:886-92.

22. Goldberger AL. Is the normal heartbeat chaotic or homeostatic? *News
in Physiological Science* 1991;**6**:87-91.

23. Prank K, Harms H, Dammig M, Brabant G, Mitschke F, Hesch RD. Is there
low-dimensional chaos in pulsatile secretion of parathyroid hormone in normal
human subjects? *Am J Physiol* 1994;**266**:E653-8.

24. Kanters JK, Holstein-Rathlou NH, Anger E. Lack of evidence for
low-dimensional chaos in heart rate variability. *J Cardiovasc
Electrophysiol* 1994;**5**:591-601.

25. Elbert T, Ray WJ, Kowalik ZJ, Skinner JE, Graf KE, Birbaumer N. Chaos
and physiology: deterministic chaos in excitable cell assemblies. *Physiol
Rev* 1994;**74**:1-47.

26. Griffith TM, Edwards DH. Fractal analysis of role of smooth muscle
Ca2^{+} fluxes in genesis of chaotic arterial pressure oscillations.
*Am J Physiol* 1994;**266**:H1801-11.

27. Wagner CD, Persson PB. Nonlinear chaotic dynamics of arterial blood
pressure and renal blood flow. *Am J Physiol* 1995;**268**:H621-7.

28. Sugihara G. Allan W, Sobel D, Alan KD. Nonlinear control of heart rate
variability in human infants. *Proc Natl Acad Sci USA*
1996;**93**:2608-13.

29. Mandelbrot BB. *The Fractal Geometry of Nature.* New York: WH
Freeman, 1982.

30. Weibel ER. Fractal geometry: a design principle for living organisms.
*Am J Physiol* 1991;**261**(*Lung Cell Mol Physiol*
**5**):L361-369.

31. Abboud S, Berenfeld O, Sadeh D. Simulation of high-resolution QRS
complex using a ventricular model with a fractal conduction system. Effects of
ischemia on high-frequency QRS potentials. *Circ Res*
1991;**68**:1751-60.

32. Peskin CS, McQueen DM. Mechanical equilibrium determines the fractal
fiber architecture of aortic heart valve leaflets. *Am J Physiol*
1994;**266**(*Heart Circ. Physiol.* **35**):H319-28.

33. Shlesinger MF. Fractal time and *1/f* noise in complex systems.
*Ann NY Acad Sci* 1987;**504**:214-28.

34. Kobayashi M, Musha T. *1/f* fluctuation of heartbeat period.
*IEEE Trans Biomed Eng* 1982;**29**:456-7.

35. Peng C-K, Mietus J, Hausdorff JM, Havlin S, Stanley HE, Goldberger AL.
Long-range anti-correlations and non-Gaussian behavior of the heartbeat.
*Phys Rev Lett* 1993;**70**:1343-6.

36. Szeto H, Chen PY, Decena JA, Cheng YI, Wu Dun-L, Dwyer G. Fractal
properties of fetal breathing dynamics. Am J Physiol
1992;**263**(*Regulatory Interactive Comp Physiol.*
**32**):R141-7.

37. Marsh DJ, Osborn JL, Cowley AW. *1/f* fluctuations in arterial
pressure and regulation of renal blood flow in dogs. *Am J Physiol*
1990;**258**:F1394-1400.

38. Hausdorff JM, Peng C-K, Ladin Z, Wei JY, Goldberger. Is walking a
random walk? Evidence for long-range correlations in the stride interval of
human gait. *J Appl Physiol* 1995;**78**:349-58.

38. Goldberger AL, Kobalter K, Bhargava V. *1/f*-like scaling in
normal neutrophil dynamics: Implications for hematologic monitoring. *IEEE
Trans Biomed Eng 1986*;**33**:874-6.

39. Goldberger AL. Fractal variability versus pathologic periodicity:
complexity loss and stereotypy in disease. *Perspect Biol Med*
1997;**40**:543-61.

40. Lipsitz LA, Goldberger AL. Loss of "complexity" and aging:
potential applications of fractals and chaos theory to senescence. *JAMA*
1992;**267**:1806-9.

41. Goldberger AL, Findley LJ, Blackburn MR, Mandell AJ. Nonlinear dynamics
in heart failure: implications of long-wavelength cardiopulmonary oscillations.
*Am Heart J* 1984;**107**:612-5.

42. Freeman WJ. Role of chaotic dynamics in neural plasticity. *Prog
Brain Res* 1994;**102**:319-33.

43. Goldberger AL, Rigney DR, Mietus J, Antman EM, Greenwald S. Nonlinear
dynamics in sudden cardiac death syndrome: heartrate oscillations and
bifurcations. *Experientia* 1988;**44**:983-7.

44. Saul JP, Arai Y, Berger RD, Lilly LS, Colucci WS, Cohen RJ. Assessment
of autonomic regulation in chronic congestive heart failure by heart rate
spectral analysis. *Am J Cardiol* 1988;**61**:1292-9.

45. Katz M, Meizner I, Shani N, Insler V. Clinical significance of
sinusoidal heart rate pattern. *Br J Obstet Gynaecol*
1983;**90**:832-6.

46. Reimann HA. *Periodic Diseases.* Philadelphia: F.A. Davis Company,
1963.

47. Glass L, Mackey MC. Pathological conditions resulting from
instabilities in physiological control systems. *Ann NY Acad Sci*
1978;**316**:214-35.

48. Mackey MC, Glass L. Oscillations and chaos in physiological control
systems. *Science* 1977;**197**:287-9.

49. Milton J, Black D. Dynamic diseases in neurology and
psychiatry. *Chaos* 1995;**5**:8-13.

50. Peng CK, Havlin S, Stanley HE, Goldberger AL. Quantification of scaling
exponents and crossover phenomena in nonstationary hearbeat time series.
*Chaos* 1995;**5**:82-7.

51. Nearing BD, Huang AH, Verrier RL. Dynamic tracking of cardiac
vulnerability by complex demodulation of the T wave. *Science*
1991;**252**:437-40.

52. Ho KKL, Moody GB, Peng C-K, Mietus JE, Larson MG, Levy D, Goldberger AL.
Predicting survival in heart failure cases and controls using fully automated
methods for deriving nonlinear and conventional indices of heart rate dynamics.
*Circulation* 1997;**96**:842-8.

53. Mäkikallio TH, Seppänen T, Airaksinen KEJ, Koistinen J, Tulppo
MP, Peng C-K, Goldberger AL, Huikuri HV. Dynamic analysis of heart rate may
predict subsequent ventricular tachycardia after myocardial infarction. *Am
J Cardiol* 1997;**80**:779-83.

54. Mäkikallio TH, Ristimäe T, Airaksinen KEJ, Peng C-K,
Goldberger AL, Huikuri HV. Heart rate dynamics in patients with stable angina
pectoris and utility of fractal and complexity measures. *Am J Cardiol*
1998;**81**:27-31.

55. Mäkikallio TH, Hoiber S, Kober L, Torp-Pedersen C, Peng C-K,
Goldberger AL, Huikuri HV. Fractal analysis of heart rate dynamics as a
predictor of mortality in patients with depressed left ventricular function
after acute myocardial infarction. *Am J Cardiol*
1999;**83**:836-9.

56. Amaral LAN, Goldberger AL, Ivanov PCh, Stanley HE. Scale-independent
measures and pathologic cardiac dynamics. *Phys Rev Lett*
1998;**81**:2388-91.

57. Iyengar N, Peng C-K, Morin R, Goldberger AL, Lipsitz LA. Age-related
alterations in the fractal scaling of cardiac interbeat interval dynamics.
*Am J Physiol* 1996;**271**:1078-84.

58. Hausdorff JM, Mitchell SL, Firtion R, Peng C-K, Cudkowicz ME, Wei JY,
Goldberger AL. Altered fractal dynamics of gait: reduced stride interval
correlations with aging and Huntington's disease. *J Appl Physiol*
1997;**82**:262-9.

59. Pincus SM, Goldberger AL. Physiological time-series analysis: what does
regularity quantify? *Am J Physiol* 1994;**266**(*Heart Circ
Physiol*):H1643-56.

60. Skinner JE, Carpeggiani C, Landesman CE, Fulton KW. The
correlation-dimension of the heartbeat is reduced by myocardial ischemia in
conscious pigs. *Circ Res* 1991;**68**:966-76.

61. Schiff SJ, Jerger K, Duong DH, Chang T, Spano ML, Ditto WL. Controlling
chaos in the brain. *Nature* 1994;**370**:615-20.

62. Regalado A. A gentle scheme for unleashing chaos. *Science*
1995;**268**:1848.

63. Garfinkel A, Spano ML, Ditto WL, Weiss JN. Controlling cardiac chaos.
*Science* 1992;**257**:1230-5.

64. Peng CK, Buldyrev SV, Goldberger AL, et al. Long-range correlations in
nucleotide sequences. *Nature* 1992;**356**:168-70.

65. Buldyrev SV, Goldberger AL, Havlin S, Peng CK, Stanley HE, Simons M.
Fractal landscapes and molecular evolution: modeling the myosin heavy chain
gene family. *Biophys J* 1993;**65**:2673-9.

66. Ossadnik SM, Buldyrev SV, Goldberger AL, et al. Correlation approach to
identify coding regions in DNA sequences. *Biophys J*
1994;**67**:64-70.

67. Ivanov PCh, Amaral LAN, Goldberger AL, Stanley HE. Stochastic feedback
and the regulation of biological rhythms. *Europhys Lett*
1998;**43**:363-8.

68. Ivanov PCh, Rosenblum MG, Peng C-K, Mietus J, Havlin S, Stanley HE,
Goldberger AL. Scaling behavior of heartbeat intervals obtained by
wavelet-based time series analysis. *Nature* 1996;**383**:323-7.

69. Ivanov PCh, Amaral LAN, Goldberger AL, Havlin S, Rosenblum MG, Struzik
Z, Stanley HE. Multifractality in human heartbeat dynamics. *Nature*
1999;**399**:461-5.

70. Peng C-K, Hausdorff JM, Goldberger AL. Fractal mechanisms in neural
control: human heartbeat and gait dynamics in health and disease. In:
Walleczek J, ed. *Self-Organized Biological Dynamics and Nonlinear
Control.* Cambridge: Cambridge University Press, 1999.