Download Neural Control Hierarchy Of The Heart Has Not Evolved To Deal

Articles in PresS. Physiol Genomics (May 21, 2013). doi:10.1152/physiolgenomics.00027.2013
1
Neural Control Hierarchy Of The Heart Has Not
2
Evolved To Deal With Myocardial Ischemia
3
G. Kember1 , J. A. Armour2 , M. Zamir3,4,†
1
Department of Engineering Mathematics, Dalhousie University,
P. O. Box 1000, Halifax, Nova Scotia
2
Department of Pharmacology, Quillen College of Medicine,
East Tennessee State University, Johnson City,
Tennessee, 37614-8866.
3
Department of Applied Mathematics
4
Department of Medical Biophysics
The University of Western Ontario
London, Canada, N6A 5B7
†
4
corresponding author: [email protected]
May 20, 2013
1
Copyright © 2013 by the American Physiological Society.
Abstract
1
2
The consequences of myocardial ischemia are examined from the standpoint
3
of the neural control system of the heart- a hierarchy of three neuronal centers
4
residing in central command, intrathoracic ganglia, and intrinsic cardiac gan-
5
glia. The basis of the investigation is the premise that while this hierarchical
6
control system has evolved to deal with “normal” physiological circumstances,
7
its response in the event of myocardial ischemia is unpredictable because the
8
singular circumstances of this event are as yet not part of its evolutionary reper-
9
toire. The results indicate that the harmonious relationship between the three
10
levels of control breaks down- because of a conflict between the priorities which
11
they have evolved to deal with. Essentially, while the main priority in central
12
command is blood demand, the priority at the intrathoracic and cardiac levels
13
is heart rate. As a result of this breakdown, heart rate becomes less predictable
14
and therefore less reliable as a diagnostic guide as to the traumatic state of the
15
heart, which it is commonly used as such following an ischemic event. On
16
the basis of these results it is proposed that under the singular conditions of
17
myocardial ischemia a determination of neural control indices in addition to
18
cardiovascular indices has the potential of enhancing clinical outcome.
19
Keywords: Intrinsic cardiac nervous system. Central command. Neuroplasticity.
20
Neurocardiology. Heart rate.
2
1
1
Introduction
2
Myocardial ischemia is generally considered from the standpoint of its deleterious
3
effect on myocardial tissue and in conjunction with coronary artery disease. The
4
way in which the ischemic event affects the neural control of the heart, however, has
5
received little attention so far. The latter is clearly important because a derangement
6
of the neural control hierarchy of the heart may be ultimately as critical a determinant
7
of the clinical outcome of myocardial ischemia as is the damage of myocardial tissue.
8
In particular, the lack of or reduced blood supply to the ischemic region of the
9
myocardium will also have a direct effect on the viability of neural tissue within that
10
region. In addition, the altered chemical and/or mechanical milieu of the affected
11
cardiac tissue will lead to a greatly altered neural transduction from that region.
12
Both of these effects will impinge on the stable dynamics of the neural control system
13
of the heart in general and heart rate in particular.
14
In addition to direct damage to neural tissue, however, the ischemic event presents
15
the complex hierarchical neural control system of the heart with a set of singular
16
circumstances which the system has not evolved to deal with. The way in which
17
the system responds to these “new” circumstances, new on the evolutionary scale, is
18
currently unknown. The lack of this knowledge renders the interpretation of cardiac
19
indices at the time of myocardial ischemia, particularly heart rate, without a clear
20
physiological basis. The present study is an attempt to explore the physiological
21
pathways which the hierarchical neural control system of the heart may follow in the
22
event of myocardial ischemia, using a simulation of that event.
23
Some important experimental work have been done in the past to explore the neu-
24
rological aspects of myocardial ischemia. In a review by Eckberg (9) it was pointed out
25
that “Vagal and sympathetic cardiovascular control is deranged profoundly in patients
26
with congestive heart failure.” A study by del Rio et al (8) examined the electrical
27
changes induced by myocardial ischemia in dogs, concluding that “Parasympathetic
28
activity during acute coronary occlusion can protect against ischemia-induced ma-
29
lignant arrhythmias; nonetheless, the mechanism mediating this protection remains
3
1
unclear.” These studies illustrate the fact that experimental or clinical observations
2
in the event of myocardial ischemia, particulary of heart rate, provide the manifes-
3
tations of a derangement of the neural control system of the heart but they do not
4
provide the underlying mechanisms. The present study is therefore aimed at comple-
5
menting these observations by exploring the possible mechanisms of a derangement
6
of the neural control system of the heart caused by circumstances which the system
7
has not evolved to deal with.
8
Heart rate anomalies have been independently linked to pathologies including
9
coronary artery disease, myocardial infarction and heart failure (16,24). In particular,
10
elevated heart rate and reduced heart rate variability have been found to correlate
11
with increased risk of both atrial (22) and ventricular (27) arrhythmias. An elevated
12
heart rate in patients experiencing acute myocardial ischemia (AMI) has been linked
13
to an increase in the risk of mortality within one year of the event (2). Reduced heart
14
rate variability has been shown to be a risk factor for a poor outcome on the time
15
scale of several weeks after an AMI (5). Autonomic disturbances have been associated
16
with idiopathic ventricular arrhythmias (10), and disturbances within 30 minutes of
17
an AMI have been associated with increased risk of sudden cardiac death (31).
18
Studies of congenital heart disease have yielded similar results. An increased risk of
19
lethal ventricular arrhythmias after surgical repair of tetralogy of Fallot (6) and higher
20
incidence of mortality involving atrial tachyarrhythmias after Fontan surgery (7) were
21
observed in patients having reduced heart rate variability and baroreceptor sensitivity.
22
Reduced heart rate variability was also linked to sudden cardiac death in patients
23
with other forms of congenital heart disease (17) whether or not they had undergone
24
surgery. In these studies, and in a seminal review by (23), autonomic disturbance are
25
noted as being associated with an increased risk of atrial and ventricular arrhythmias.
26
There is thus mounting support for the view that derangement of the neural
27
control system of the heart may be ultimately as critical a determinant of the clinical
28
outcome of myocardial ischemia as is the pathological insult on myocardial tissue. In
29
addition, the lack of or reduced blood supply to the ischemic region of the myocardium
30
will have a direct effect on the viability of neural tissue within that region (1) and
4
1
may result in complex effects such as ischemic pre-conditioning (25), and the altered
2
chemical and/or mechanical milieu of the affected cardiac tissue will lead to a greatly
3
altered neural transduction from that region (26). Each of these effects will have a
4
highly variable impact on the dynamics of the neural control system of heart function
5
(13).
6
While the existence of these effects is not in any way questionable, as the above
7
studies indicate, they have not generally been considered so far because of the difficul-
8
ties involved in integrating such effects within the present view of the neural control
9
of heart function, specifically the view that this control resides largely in central com-
10
mand (4,16,19,20). This view is based on a central control system represented by pre-
11
to post-ganglionic efferent neurons, the preganglionic neurons residing in the central
12
nervous system. These central neurons are made up of medullary parasympathetic
13
and spinal cord sympathetic efferent preganglionic neurons. They receive sensory
14
inputs from cardiovascular afferent neurons for feed back of heart status. Hence, all
15
control decisions are assumed to emanate from the central nervous system. In this pa-
16
per we shall refer to this as the “central command” level of control, to be distinguished
17
from two other levels as described below. As such, the peripheral neural components
18
are seen as serving a passive role to either deliver centrally determined inputs to the
19
heart or to return feed back of heart status to the central nervous system.
20
Contrary to this view, it is our underlying hypothesis that peripheral neural com-
21
ponents must be considered in order to understand the full dynamics of neural control
22
of heart function and, for the purpose of the present paper, in order to determine the
23
full consequences of neural derangement during a cardiac ischemic event. Under-
24
standing the relationship between autonomic dysfunction and the onset of cardiac
25
arrhythmias cannot be explained solely within the scope of central neural command
26
theories (23). This is evident from anatomical and functional studies that have un-
27
covered intrathoracic extracardiac and intrinsic cardiac elements of the wider control
28
system of heart function (1, 11, 12).
29
To that end, we use a model of neural control of the heart based on a hierarchy of
30
three populations of neurons, the “top” level residing in central command, the “mid5
1
dle” residing within intrathoracic extracardiac ganglia, and the “bottom” residing
2
within intrinsic cardiac ganglia. At each level, neurons act both individually and in
3
concert with others. The ability of this model to explain heart rate anomalies, specif-
4
ically the presence and absence of heart rate oscillations (Mayer waves), has been
5
demonstrated previously (14). It was also shown that “networking” among neurons
6
endows this neural control system of the heart with a measure of plasticity that leads
7
to heterogeneity in neural behavior consistent with common observations (15). In
8
the present paper we explore the way in which this network plasticity and the inter-
9
play between and within the three levels of control respond to the singular conditions
10
created by myocardial ischemia.
11
2
12
2.1
13
The neural network on which our model is based has 3N neurons equally divided
14
among three levels of control to be referred to as levels 1,2,3 or “bottom”, “middle”,
15
“top”, or “cardiac”, “intrathoracic”, “central”, respectively. Two indices, j, k are
16
used to identify the k th neuron at the j th level. The state of activity (≡ level of
17
discharge) of neuron j, k is denoted by Sj,k which is scaled such that its value ranges
18
between 1.0 when the neuron is most active and 0.0 when it is inactive.
19
2.2
20
Broadly speaking, the neural network receives continuous updates of current demand
21
for blood flow and current heart rate, and processes these to produce an appropriate
22
change in heart rate. The main result of this algorithm, and the chief feature of the
23
model, is that demand for blood flow does not proceed directly to the heart or to
24
central command but to the neural network as a whole. Essential details of how this
25
occurs are given in what follows, more details can be found in (14, 15).
26
METHODS
Neural Network Structure
Heart Rate Control Algorithm
Heart rate is constrained to lie between prescribed maximum and minimum values.
6
1
A scaled heart rate H is used in what follows, such that H = 1.0 when heart rate is
2
maximum and H = 0.0 when heart rate is minimum.
3
The dynamics of the neural network unfold piecewise at consecutive time intervals
4
t(n) , n = 1, 2, . . ., where demand for blood flow and current heart rate are used as
5
inputs and an incremental “move” ∆M (n) (t) is produced as output. The latter is
6
based on the mean activity of neurons at the cardiac level
∑N1
F
7
(n)
=
(n)
k=1
S1,k
(1)
N1
such that
∆M (n) = β(F
(n)
− α)
(2)
(n)
8
where α is a reference activity level, β is a constant and S1,k , k = 1, 2, . . . , N1 , is the
9
level of activity of neurons at level 1 (cardiac) of the network at time interval t(n) .
10
Within each time interval, heart rate is a continuous function of time governed by a
11
first order linear system
τH
12
dH (n) (t)
+ H (n) (t) = M (n) (t)
dt
(3)
where τH is a time constant and
M (n) (t) =
i=n−1
∑
∆M (i) (t)
(4)
i=0
13
For simplicity, in what follows the time variable t shall not be shown explicitly but
14
will be implicit within each time interval t(n) .
15
2.3
16
The discharge from a neuron j, k within the network in time interval t(n) is represented
17
by the state of activity Sj,k of that neuron. This state is affected by (i) current blood
18
demand, (ii) current heart rate and, (iii) the level of activity of neighboring neurons.
19
A change in the state of activity the neuron due to these effects shall be denoted
20
respectively by δ1 Sj,k , δ2 Sj,k , δ3 Sj,k , and total change by ∆Sj,k .
Neural Discharge
(n)
(n)
(n)
(n)
(n)
7
1
We distinguish between two types of neurons (29,30): “heart rate neurons” which
2
are affected by only current heart rate and the activity of neighboring neurons, and
3
“blood demand neurons” which are affected by only demand for blood flow and the
4
activity of neighboring neurons. Thus the change in the state of activity of a neuron
5
j, k is given by
(n)
(n)
(n)
∆Sj,k = δ1 Sj,k + δ3 Sj,k
(n)
(n)
(n)
∆Sj,k = δ2 Sj,k + δ3 Sj,k
heart rate neurons
(5)
blood demand neurons
(6)
6
The extent to which heart rate H and blood demand D affect the state of activ-
7
ity of a neuron (j, k) are represented respectively by heart rate and blood demand
8
“sensitivities” hj,k and dj,k , such that
(n)
(n)
(n)
(n)
δ1 Sj,k
=−
hj,k
× H (n−L)
(n)
h
(7)
(n)
9
where H (n−L) is heart rate delayed by a time constant τL = L∆t and h
is the mean
(n)
10
of hj,k over the entire network and, similarly,
(n)
(n)
δ2 Sj,k
11
=
dj,k
(n)
d
×
D(n)
Dmax
(8)
where the demand is scaled with respect to a maximum demand Dmax .
12
It is important to note, as indicated by the (n) superscripts, that the sensitivities
13
hj,k and dj,k can change from one time interval to the next, a feature which plays a
14
critical role following the onset of myocardial ischemia. This feature can be thought
15
of as a measure of “plasticity” in the sensitivities of neurons to heart rate and to
16
blood demand and we shall refer to it as “neural plasticity” to be distinguished
17
from “network plasticity” which represents plasticity in the neighbor weightings PJ,K
18
within the network. To demonstrate the role of neural plasticity following the onset
19
of myocardial ischemia, we shall compare a scenario in which this feature is enabled
20
(“neural plasticity: ON”) and one in which it is disabled (“neural plasticity: OFF”).
(n)
(n)
(n)
8
1
2.4
Neural Networking
2
Another component of change in the activity of a neuron is due to the extent to which
3
it is affected by the activity of other neurons, which we refer to briefly as “networking”
4
and which is a key feature of our model. Networking occurs between a neuron and
5
its neighbors at the same level of the neural network and at the next adjacent level.
6
Thus, a neuron at the cardiac level has neighbors at the cardiac and intrathoracic
7
level, a neuron at central command has neighbors at the central and intrathoracic
8
levels, while a neuron at the intrathoracic level has neighbors at all three levels.
9
The degree of connectivity or “synaptic strength” between a neuron j, k and each
10
of its neighbors is represented by a weighting PJ(i),K(i) , where i = 1 . . . Nb(j,k) identify
11
the neighbors and J, K (upper case) are the positions of these neighbors within the
12
network. Again we note, from the superscript (n) that synaptic strength can change
13
from one time interval to the next, thus providing, in effect, a measure of plasticity
14
within the network as a whole (15).
(n)
15
The way in which the state of activity of a neuron j, k is influenced by the state of
16
activity of its neighboring neurons J, K is what we have referred to as “networking”
17
among neurons and is represented by δ3 in Eqs.5,6. It is a key feature of the model
18
whereby every neuron within the network influences and is influenced by other neu-
19
rons. The extent of networking between a particular neuron j, k and its neighboring
20
neurons J(i), K(i) is determined by the sum of the difference prevailing in time in-
21
terval t(n) between the state of activity of that neuron and the states of activity of
22
the neighboring neurons, specifically
∑ (
Nb(j,k)
(n)
δ3 Sj,k
=
(n)
(n)
SJ(i),K(i) − Sj,k
)
(n)
× PJ(i),K(i)
(9)
i=1
23
2.5
Network Plasticity
24
The concept of plasticity, which has its origin in brain function, is now well accepted
25
as a property of neural networks in general (29, 30). As pointed out in the previous
26
section and demonstrated in (15), plasticity is enabled by allowing the connectivity
27
among neurons PJ(i),K(i) to change in time. The rules by which the change occurs are
(n)
9
1
referred to as “plasticity rules” and there are currently two principal rules which are
2
believed to operate concurrently and which we implement in this paper.
3
(i) Homeostatic plasticity rule: Neurons that are making only minimal contribu-
4
tion to the overall network activity have their contribution increased while,
5
conversely, neurons that are making large contribution have their contribution
6
decreased.
7
(ii) Hebbian plasticity rule: Neurons that are making only minimal contribution to
8
the overall network activity have their contribution reduced even further while,
9
conversely, neurons that are making large contribution have their contribution
increased even further.
10
11
The rules are implemented at time intervals ∆t = 0.01s. In all cases the change
12
in the activity of a neuron is influenced by a change in the values of its neighbor
13
weightings PJ(i),K(i) .
(n)
14
It has been shown that both rules are required for the stability of a neural network
15
(18, 29, 30). However, the consequences of these rules following the singular event of
16
myocardial ischemia have yet to be tested and they form the main focus of the present
17
paper. For this purpose we examine the dynamics of the network with and without
18
the application of these rules and we refer to these conditions as “Network plasticity:
19
ON” and “Network plasticity: OFF” respectively.
20
2.6
21
The onset of myocardial ischemia is implemented by imposing a 50% step increase
22
in blood demand D(n) at the cardiac level of the network at time t = 500s. This
23
is followed, at times t ≥ 500s, by an incremental increase in the sensitivities dj,k of
24
blood demand neurons and a concurrent decrease in the sensitivities hj,k of heart rate
25
neurons throughout the network. The incremental changes ∆d, ∆h are very small,
26
distributed randomly in the range [0, 10−5 ], but they are implemented continually at
The Onset of Myocadial Ischemia and Neural Plasticity
(n)
(n)
(n)
27
every time step (∆t = 0.01s) up to a preset upper thresholds of dj,k = 0.99 and a
28
preset lower threshold of hj,k which is randomly distributed in the range [0, 0.1].
(n)
10
1
3
2
The aim of the results which we sought was to determine the way in which the three
3
neuronal centers respond to the ischemic event individually and collectively as they are
4
influenced by each other via network and neural plasticities. To that end we examined
5
heart rate and related properties of the neural control system under three conditions:
6
(i) network and neural plasticities OFF, (ii) network plasticity ON, neural plasticity
7
OFF, (iii) network and neural plasticities ON. Under the first condition networking
8
between neurons within and between the three control centers with the connectivity
9
among neurons being based on fixed parameters. Under the second condition these
10
parameters are variable, thereby giving rise to network plasticity. Under the third
11
condition, the sensitivities of neurons to blood demand and to heart rate, which are
12
based on fixed parameters in (i) and (ii), become variable and thereby give rise to
13
neural plasticity. In all cases a 2, 000s run of the system is executed.
14
15
Results
The course of heart rate through the onset of myocardial ischemia and under the
three states of the neural network is shown in Figure 1.
16
Networking among neurons gives rise to two streams of priorities within the net-
17
work which have previously been referred as “demand drive” and “heart drive” (14).
18
The heart drive of a neuron j, k is defined as the sum of the neuron neighbor weight-
19
ings PJ(i),K(i) multiplied by the sensitivity of hJ(i),K(i) of each of the neighbor neurons
20
to current heart rate, i.e.
(n)
(n)
∑
Nb(j,k)
(n)
Hdr (j, k)
=
(n)
(n)
(10)
(n)
(11)
hJ(i),K(i) × PJ(i),K(i)
i=1
21
while the demand drive is defined similarly as
∑
Nb(j,k)
(n)
Ddr (j, k)
=
(n)
dJ(i),K(i) × PJ(i),K(i)
i=1
22
The evolutions of these two drives under the three states of the neural network and
23
through the onset of the ischemic event are shown in Figures 2 and 3 respectively.
24
A measure of the extent to which network plasticity is at play in the dynamic
25
control of heart rate is the accumulated number of implementations of the Hebbian
11
1
and Homeostatic plasticity rules, as described in Section 2.5. The total number of
2
these implementations at the three levels of the network and under the three different
3
states of the network are shown in Figure 4.
4
The key feature of a neural “network” is that neurons within the network do not
5
respond en masse as a single unit but act individually while, at the same time, they
6
are influencing and are being influenced by each other. A mapping of the simultaneous
7
activities of neurons within the network is therefore useful for interpreting the ultimate
8
outcome, namely the pattern of heart rate, in terms of neural activity. Individual
9
neural cell activity or discharge Sj,k , scaled between 0 (inactive) and 1.0 (most active),
10
as described in Section 2.3, are shown for a representative selection of neuron from
11
each of the three levels of the network in Figures 5-7.
12
4
13
Key elements of discussion of the overall neural control of the cardiovascular system
14
include the role of baroreflexes in the control of blood pressure and the role of sym-
15
pathetic and parasympathetic drives in the control of heart rate. A discussion of
16
one or the other of these elements in isolation is an oversimplification because these
17
elements act interdependently rather than independently from each other. Because
18
of the complexity of the integrated system, however, many animal and human studies
19
have focused simply on the apparent effects of different stressors on heart rate (3, 8)
20
and heart rate variability (3, 9, 21, 28). In each case the intervening mechanism is
21
not known and is likely multifaceted, involving an interplay between baroreceptor
22
feedback and sympathetic/parasympathetic drive, followed by final implementation
23
by the hierarchical neural control system of the heart. To place our study and its
24
limitations in proper perspective, therefore, the focus of the study is on this final
25
stage of the process, namely on how the hierarchical neural control system of the
26
heart ultimately implements the changes in heart rate.
(n)
Discussion and Conclusions
27
The main purpose of this study was to investigate the response of the hierarchical
28
neural control network of the heart following the onset of myocardial ischemia. The
12
1
control network, consisting of three populations of neurons residing at the cardiac,
2
intrathoracic, and central command levels, has been shown previously (14) to produce
3
results that are more consistent with physiological observation of heart rate oscilla-
4
tions and variability. The basis of our investigation was the premise that while this
5
hierarchical control system has evolved to deal with “normal” physiological circum-
6
stances, its response in the event of myocardial ischemia is unpredictable because the
7
singular circumstances of this event are as yet not part of its evolutionary repertoire.
8
The use of a model to probe the question of what this response might be is therefore
9
not only apt but necessary.
10
In the event of myocardial ischemia, heart rate is used in the clinical setting both
11
as a diagnostic marker of the traumatic state of the heart as well as a guide to
12
appropriate action. The main focus is usually on stabilizing heart rate and restoring
13
blood flow to myocardial tissue. Our results suggest that, at the same time, attention
14
should also be directed towards the stability of the hierarchical neural control system
15
of the heart.
16
More specifically, the results indicate that the harmonious relationship within and
17
between the three levels of neural control of the heart breaks down because of a clash
18
of the priorities which the three levels of control have evolved to deal with. The
19
clash occurs, essentially, because these priorities are inappropriate in the singular
20
circumstances of myocardial ischemia.
21
In particular, while the main priority at central command is to respond to demands
22
for blood flow, the main priority at the cardiac and intrathoracic levels, as has been
23
demonstrated previously (14), is to process these demands before they actually reach
24
the heart and thereby provide a protective measure against excessive drive of the
25
heart. In other words, and in short, while the main priority in central command
26
is blood demand, the priority at the intrathoracic and cardiac levels is heart rate.
27
While these disparate priorities coexist in harmony under “normal” physiological
28
circumstances, presumably because they have evolved to do so, our results indicate
29
that this harmony breaks down under the singular conditions of myocardial ischemia.
30
An important outcome of this breakdown is that heart rate becomes less pre13
1
dictable and therefore less reliable as a diagnostic marker of the functional state of
2
the heart. Our results indicate that following the onset of myocardial ischemia, heart
3
rate actually becomes an emergent property dependent heavily on the state of the hi-
4
erarchical neural control system of the heart. On that basis we propose, in conclusion,
5
that under the singular conditions of myocardial ischemia, a determination of neural
6
control indices in addition to cardiovascular indices has the potential of enhancing
7
clinical outcome.
8
Summary: Under the singular conditions of myocardial ischemia, the neural con-
9
trol hierarchy of the heart breaks down and, consequently, heart rate becomes an
10
unpredictable property of that breakdown rather than an index of the traumatic
11
state of the heart caused by the ischemic insult on myocardial tissue. Under these
12
conditions, therefore, the monitoring of heart rate alone in the clinical setting may
13
not be a reliable guide to appropriate action.
14
5
15
This work was supported by the Natural Science and Engineering Research Council
16
of Canada.
17
References
18
19
Acknowledgment
1. Armour J. Cardiac neuronal hierarchy in health and disease. Am J Physiol
287:262-271, 2004.
20
2. Berton G, Cordiano R, Palmieri R, Gheno G, Mormino P, Palatini P. Heart rate
21
during myocardial infarction: relationship with one-year global mortality in men
22
and women. Can J Cardiol 18(5):495-502, 2002.
23
24
3. Billman GE. Heart rate variability
a historical perspective. Front Physiol,
Doi:10.3389/fphys.2011.00086, 2011.
14
1
4. Cacioppo J, Berntson G, Binkley P, Quigley K, Uchino B, Fieldstone A. Auto-
2
nomic cardiac control. ii. noninvasive indices and basal response as revealed by
3
autonomic blockades. Psychophysiology 31:586-598, 1994.
4
5. Casolo G, Stroder P, Signorini C, Calzolari F, Zucchini M, Balli E, Sulla A,
5
Lazzerini S. Heart rate variability during the acute phase of myocardial infarction.
6
Circulation 85(5):2073-9, 1992.
7
6. Davos CH, Davlouros PA, Wensel R, Francis D, Davies LC, Kilner PJ, Coats
8
AJ, Piepoli M, Gatzoulis MA. Global impairment of cardiac autonomic nervous
9
activity late after repair of tetralogy of fallot. Circulation 106(suppl)I69-75, 2002.
10
7. Davos CH, Francis DP, Leenarts MF, Yap SC, Li W, Davlouros PA, Wensel
11
R, Coats AJ, Piepoli M, Sreeram N, Gatzoulis MA. Global impairment of car-
12
diac autonomic nervous activity late after the fontan operation. Circulation
13
108(suppl)II180-5, 2003.
14
8. del Rio CL, Dawson TA, Clymer BD, Paterson DJ, Billman GE. Effects of acute
15
vagal nerve stimulation on the early passive electrical changes induced by myocar-
16
dial ischaemia in dogs: heart rate-mediated attenuation. Exp. Physiol. 93:931-944,
17
2008.
18
19
9. Eckberg DL. Baroreflexes and the failing human heart. Circulation 96:4133-4137,
1997.
20
10. Fei L, Statters D, Hnatkova K, Poloniecki J, Malik M, Camm J. Change of au-
21
tonomic influence on the heart immediately before the onset of spontaneous id-
22
iopathic ventricular tachycardia. J Am Coll Cardiol 24(6):1515-1522, 1994.
23
11. Horackova M, Armour J, Byczko Z. Distribution of intrinsic cardiac neurons in
24
whole-mount guinea pig atria identified by multiple neurochemical coding. A
25
confocal microscope study. Cell and Tissue Res 297:409-421, 1999.
26
12. Horackova M, Croll R, Hopkins D, Losier A, Armour J. Morphological and im-
27
munohistochemical properties of primary longterm cultures of adult guinea-14
15
1
pig ventricular cardiomyocytes with peripheral cardiac neurons. Tissue and Cell
2
28:411-425, 1996.
3
13. Huang M, Ardell J, Hanna B, Wolf S, Armour J. Effects of transient coronary
4
artery occlusion on canine intrinsic cardiac neuronal activity. Integ Physiol Behav
5
Sci 28:5-21, 1993.
6
7
8
9
10
11
14. Kember G, Armour J, Zamir M. Neural control of heart rate: The role of neuronal
networking. J Theoret Biol 277:41-47, 2011.
15. Kember G, Armour J, Zamir M. Dynamic neural networking as a basis for plasticity in the control of heart rate. J Theoret Biol 317:39-46, 2012.
16. La Rovere M. Heart rate and arrhythmic risk: old markers never die. Europace
12:155-157, 2010.
12
17. Lammers A, Kaemmerer H, Hollweck R, Schneider R, Barthel P, Braun S, Wacker
13
A, Brodherr-Heberlein S, Hauser M, Eicken A, Schmidt G, Hess J. Impaired
14
cardiac autonomic nervous activity predicts sudden cardiac death in patients
15
with operated and unoperated congenital cardiac disease. J Thorac Cardiovasc
16
Surg 132:647-655, 2006.
17
18
19
20
18. Maffei A, Bucher D, Fontanini A. Homeostatic plasticity in the nervous system:
Editorial. Neural Plasticity, doi:10.1155/2012/913472, 2012.
19. Malliani A, Schwartz P, Zanchetti A. Neural mechanisms in life-threatening arrhythmias. Am Heart J 100(5):705-15, 1980.
21
20. Matsukawa K. Central command: control of cardiac sympathetic and vagal effer-
22
ent nerve activity and the arterial baroreflex during spontaneous motor behaviour
23
in animals. Exp Physiol 97:20-28, 2011.
24
25
21. Monahan KD. A new answer to an old question: does ageing modify baroreflex
control of vascular sympathetic outflow in humans? J Physiol 587:1857, 2009.
16
1
22. Okin P, Wachtell K, Kjedsen S, Julius S, Lindholm L, Dahlof B, Hille DA, Niem-
2
inen MS, Edelman JM, Devereux RB. Incidence of atrial fibrillation in relation
3
to changing heart rate over time in hypertensive patients: the life study. Circ
4
Arrhythmia Electrophysiol 1, 337-43, 2008.
5
23. Paton J, Boscan P, Pickering A, Nalivaiko E. The yin and yang of cardiac au-
6
tonomic control: Vago-sympathetic interactions revisited. Brain Res Rev 49:555-
7
565, 2005.
8
24. Pokorny J, Stanek V, Vrana M. Sudden cardiac death thirty years ago and at
9
present. the role of autonomic disturbances in acute myocardial infarction revis-
10
11
12
ited. Physiol Res 60:715-728, 2011.
25. Sanada S, Kitakaze M. Ischemic preconditioning: emerging evidence, controversy,
and translational trials. Intl J Cardiol 97:263-276, 2004.
13
26. Seyedi N, Maruyama R, Levi R. Bradykinin activates a cross-signaling pathway
14
between sensory and adrenergic nerve endings in the heart: A novel mechanism
15
of ischemic norepinephrine release? J Pharmacol Exp Ther 290:656-663, 1999.
16
27. Soliman E, Elsalam M, Li Y. The relationship between high resting heart rate
17
and ventricular arrhythmogenesis in patients referred to ambulatory 24 h electro-
18
cardiographic recording. Europace 12:261-265, 2010.
19
20
21
22
23
24
25
26
28. Studinger P, Goldstein R, Taylor JA. Age- and fitness-related alterations in vascular sympathetic control. J Physiol 587:2049-2057, 2009.
29. Turrigiano G, Nelson S. Hebb and homeostasis in neuronal plasticity. Current
Opinion in Neur 10:358-364, 2000.
30. Turrigiano G, Nelson S. Homeostatic plasticity in the developing nervous system.
Nature Reviews Neuroscience 5(2):97-107, 2004.
31. Webb S, Adgey A, Pantridge J. Autonomic disturbance at onset of acute myocardial infarction. Br Med J 3:89-92, 1972.
17
1
Figure Captions
2
Figure 1: Heart rate as impacted by the onset of myocardial ischemia and increased
3
blood demand at t = 500s under three different states of the neural network as
4
indicated in each panel. In the absence of network plasticity and neural plasticity,
5
heart rate simply moves from one state of oscillation to another, with a small change in
6
the mean. With network plasticity ON, these oscillations are promptly extinguished
7
as has been shown previously (14). With the addition of neural plasticity, a dramatic
8
effect emerges as seen in the large panel. There is a clash between network plasticity
9
and neural plasticity, which in this paper is being referred to as “derangement” of the
10
control hierarchy of the neural network as discussed further in the text.
11
Figure 2: Heart drives, as defined in Eq.10, under the three states of the neural net-
12
work as indicated in the individual panels. With both network plasticity and neural
13
plasticity OFF, heart drives are constant at all three levels of the neural network:
14
cardiac (red), intrathoracic (green), central command (blue), because both compo-
15
nents of the drive are constant. With network plasticity ON, PJ,K become variable
16
which results in the initial rearrangement of heart drive at the three levels of the
17
network, followed by constant levels which are little affected by the onset of myocar-
18
dial ischemia at t = 500s. With the addition of neural plasticity, however, as hJ,K
19
become variable, heart drive drops sharply at the onset of myocardial ischemia, as the
20
demand for blood flow becomes a higher priority than heart rate particularly within
21
central command (blue).
22
Figure 3: Demand drives, as defined in Eq.11, under the three states of the neural
23
network as indicated in the individual panels. Results in the top two panels are similar
24
to those for heart drive described in Figure 2, but with both network plasticity and
25
neural plasticity ON (large panel), the large increases in demand drive following the
26
onset of myocardial ischemia at t = 500s are seen to be a clear counterpart to the
27
results for heart drive in Figure 2.
28
Figure 4: An accumulated count of implementations of the Hebbian (solid lines) and
18
1
Homeostatic (dashed) plasticity rules, the number of such implementations being a
2
measure of the extent to which network plasticity is at play in the dynamic control
3
of heart rate. As seen in the top two panels, these effects are absent when network
4
plasticity is OFF and are fairly constant when network plasticity is ON, very little
5
affected by the onset of myocardial ischemia. However, this picture changes drastically
6
with the addition of neural plasticity (lower panel). The number of implementations
7
rise at all three levels of the network (red: cardiac, green: intrathoracic, blue: central),
8
in piecewise fashion where the turning points coincide with those in the time course
9
of the heart rate seen in Figure 1.
10
Figure 5: Neural cell activity of a sample of 30 neurons at each of the three levels
11
of the network: cardiac (red), intrathoracic (green), blue (central command), with
12
network and neural plasticities OFF. Constant activity of a single neuron would
13
appear as a horizontal line at an activity level between 0 (inactive) and 1.0 (most
14
active). Oscillatory activity appears as lines moving up and down which on the large
15
time scale of the figure merge as a solid block. the figure thus shows oscillatory
16
cell activity before and after the onset of myocardial ischemia at t = 500s, which is
17
consistent with the oscillatory heart rate seen in the top left panel of Figure 1. The
18
present figure shows further that cell activity before the ischemic event is dominated
19
by the central command level, while after that event there are significant contributions
20
from the cardiac and intrrathoracic levels.
21
Figure 6: Neural cell activity as in Figure 5 but with network plasticity ON. Oscil-
22
latory behavior is promptly extinguished by the network plasticity, consistent with
23
the corresponding behavior of heart rate seen in the top right panel of Figure 1. The
24
onset of myocardial ischemia at t = 500s causes only a small “blip”, again consistent
25
with the corresonding behavior of heart rate.
26
Figure 7: Neural cell activity as in Figure 5 but with both network and neural
27
plasticities ON. Behavior before the onset of the ischemic event at t = 500s is seen to
28
be similar to that seen in Figure 6, but following that event a dramatic clash is seen
29
to unfold between the network plasticity and neural plasticity at the three levels of
19
1
control. The result is an alternating emergence and extinction of oscillatory behavior
2
which ultimately abates, again, consistent with the corresponding behavior seen with
3
heart rate in the lower panel of Figure 1. This is what in the present paper is being
4
referred to as “derangement” of the control hierarchy of heart rate.
20
0.4
0.4
Network plasticity OFF
Neural plasticity OFF
0.3
heart rate
heart rate
0.3
0.2
Network plasticity ON
Neural plasticity OFF
0.2
0.1
0.1
0
0
0
0
1000
500
time (s)
1000
500
time (s)
0.4
0.35
Network plasticity ON
Neural plasticity ON
heart rate
0.3
0.25
0.2
0.15
0.1
0.05
0
0
500
1000
time (s)
1500
2000
Figure 1: Heart rate as impacted by the onset of myocardial ischemia and increased
blood demand at t = 500s under three different states of the neural network as
indicated in each panel. In the absence of network plasticity and neural plasticity,
heart rate simply moves from one state of oscillation to another, with a small change in
the mean. With network plasticity ON, these oscillations are promptly extinguished
as has been shown previously (14). With the addition of neural plasticity, a dramatic
effect emerges as seen in the large panel. There is a clash between network plasticity
and neural plasticity, which in this paper is being referred to as “derangement” of the
control hierarchy of the neural network as discussed further in the text.
21
1
1
Network plasticity ON
0.8 Neural plasticity OFF
heart drive
heart drive
Network plasticity OFF
0.8 Neural plasticity OFF
0.6
0.4
0.2
0.6
0.4
0.2
0
0
500
time (s)
0
0
1000
500
time (s)
1000
1
heart drive
0.8
Network plasticity ON
Neural plasticity ON
0.6
0.4
0.2
0
0
500
1000
time (s)
1500
2000
Figure 2: Heart drives, as defined in Eq.10, under the three states of the neural network as indicated in the individual panels. With both network plasticity and neural
plasticity OFF, heart drives are constant at all three levels of the neural network:
cardiac (red), intrathoracic (green), central command (blue), because both components of the drive are constant. With network plasticity ON, PJ,K become variable
which results in the initial rearrangement of heart drive at the three levels of the
network, followed by constant levels which are little affected by the onset of myocardial ischemia at t = 500s. With the addition of neural plasticity, however, as hJ,K
become variable, heart drive drops sharply at the onset of myocardial ischemia, as the
demand for blood flow becomes a higher priority than heart rate particularly within
central command (blue).
22
1.5
2
Network plasticity OFF
Neural plasticity OFF
demand drive
demand drive
2
1
0.5
0
0
500
time (s)
1.5
1
0.5
0
0
1000
Network plasticity ON
Neural plasticity OFF
500
time (s)
1000
2
Network plasticity ON
Neural plasticity ON
demand drive
1.5
1
0.5
0
0
500
1000
time (s)
1500
2000
Figure 3: Demand drives, as defined in Eq.11, under the three states of the neural
network as indicated in the individual panels. Results in the top two panels are similar
to those for heart drive described in Figure 2, but with both network plasticity and
neural plasticity ON (large panel), the large increases in demand drive following the
onset of myocardial ischemia at t = 500s are seen to be a clear counterpart to the
results for heart drive in Figure 2.
23
0.5
5
Network plasticity OFF
Neural plasticity OFF
Hebb/Homeo count
Hebb/Homeo count
1
0
−0.5
−1
0
500
time (s)
x 10
Network plasticity ON
4 Neural plasticity OFF
3
2
1
0
0
1000
7
500
time (s)
1000
7
5
x 10
Hebb/Homeo count
4
Network plasticity ON
Neural plasticity ON
3
2
1
0
0
500
1000
time (s)
1500
2000
Figure 4: An accumulated count of implementations of the Hebbian (solid lines) and
Homeostatic (dashed) plasticity rules, the number of such implementations being a
measure of the extent to which network plasticity is at play in the dynamic control
of heart rate. As seen in the top two panels, these effects are absent when network
plasticity is OFF and are fairly constant when network plasticity is ON, very little
affected by the onset of myocardial ischemia. However, this picture changes drastically
with the addition of neural plasticity (lower panel). The number of implementations
rise at all three levels of the network (red: cardiac, green: intrathoracic, blue: central),
in piecewise fashion where the turning points coincide with those in the time course
of the heart rate seen in Figure 1.
24
1
cell activity
0.8
0.6
0.4
0.2
0
0
200
400
600
time (s)
800
1000
Figure 5: Neural cell activity of a sample of 30 neurons at each of the three levels of the
network: cardiac (red), intrathoracic (green), blue (central command), with network
and neural plasticities OFF. Constant activity of a single neuron would appear as
a horizontal line at an activity level between 0 (inactive) and 1.0 (most active).
Oscillatory activity appears as lines moving up and down which on the large time
scale of the figure merge as a solid block. the figure thus shows oscillatory cell activity
before and after the onset of myocardial ischemia at t = 500s, which is consistent
with the oscillatory heart rate seen in the top left panel of Figure 1. The present
figure shows further that cell activity before the ischemic event is dominated by the
central command level, while after that event there are significant contributions from
the cardiac and intrrathoracic levels.
25
1
cell activity
0.8
0.6
0.4
0.2
0
0
200
400
600
time (s)
800
1000
Figure 6: Neural cell activity as in Figure 5 but with network plasticity ON. Oscillatory behavior is promptly extinguished by the network plasticity, consistent with
the corresponding behavior of heart rate seen in the top right panel of Figure 1. The
onset of myocardial ischemia at t = 500s causes only a small “blip”, again consistent
with the corresonding behavior of heart rate.
26
1
cell activity
0.8
0.6
0.4
0.2
0
0
500
1000
time (s)
1500
2000
Figure 7: Neural cell activity as in Figure 5 but with both network and neural plasticities ON. Behavior before the onset of the ischemic event at t = 500s is seen to
be similar to that seen in Figure 6, but following that event a dramatic clash is seen
to unfold between the network plasticity and neural plasticity at the three levels of
control. The result is an alternating emergence and extinction of oscillatory behavior
which ultimately abates, again, consistent with the corresponding behavior seen with
heart rate in the lower panel of Figure 1. This is what in the present paper is being
referred to as “derangement” of the control hierarchy of heart rate.
27