Computers in Biology and Medicine 37 (2007) 8 – 20

www.intl.elsevierhealth.com/journals/cobm

Phase response characteristics of sinoatrial node cells

D.G. Tsalikakisa, d , H.G. Zhangb , D.I. Fotiadisa, c, d,∗ , G.P. Kremmydasa , Ł.K. Michalisd, e
a Unit of Medical Technology and Intelligent Information Systems, Department of Computer Science, University of Ioannina, GR 45110 Ioannina, Greece
b University of Manchester, Manchester, UK
c Biomedical Research Institute-FORTH, GR 45110 Ioannina, Greece
d Michaelidion Cardiology Center, GR 45110 Ioannina, Greece
e Department of Cardiology, Medical School, University of Ioannina, GR 45110 Ioannina, Greece

Received 29 December 2004; accepted 20 September 2005

Abstract
In this work, the dynamic response of the sinoatrial node (SAN), the natural pacemaker of the heart, to short external stimuli is investigated
using the Zhang et al. model. The model equations are solved twice for the central cell and for the peripheral cell. A short current pulse is
applied to reset the spontaneous rhythmic activity of the single sinoatrial node cell. Depending on the stimulus timing either a delay or an
advance in the occurrence of next action potential is produced. This resetting behavior is quantified in terms of phase transition curves (PTCs)
for short electrical current pulses of varying amplitude which span the whole period. For low stimulus amplitudes the transition from advance
to delay is smooth, while at higher amplitudes abrupt changes and discontinuities are observed in PTCs. Such discontinuities reveal critical
stimuli, the application of which can result in annihilation of activity in central SAN cells. The detailed analysis of the ionic mechanisms
involved in its resetting behavior of sinoatrial node cell models provides new insight into the dynamics and physiology of excitation of the
sinoatrial node of the heart.
䉷 2005 Elsevier Ltd. All rights reserved.

Keywords: Heart; Sinoatrial node; Mathematical models; Cardiac models; Phase resetting; Three-dimensional phase transition curves; Regional differences

1. Introduction from the alteration of the beat-to-beat interval depends on the

Auto-rhythmic excitation in the cardiac muscle is driven by
the sinoatrial node (SAN), the natural pacemaker of the heart.
Under physiological or pathological conditions during clinical
interventions, such us defibrillation of cardiac muscle or acci-
dental situations like electrocution, the pacemaker of the heart
is subjected to external stimulation [1].
Experimental [2,3] and numerical [4–6] studies indicate that
the application of short current pulses to sinoatrial pacemaker
cells result in changes in the cell’s cycle length which depend
on both the phase and the amplitude of the stimulus [3,7–31].
In most cases, the applied perturbation does not affect the am-
plitude or the configuration of the pacemaker cell action poten-
tial. The magnitude and direction of the phase shift resulting
∗ Corresponding author. Unit of Medical Technology and Intelligent Infor-
mation Systems, Department of Computer Science, University of Ioannina,
GR 45110 Ioannina, Greece. Tel.: +30 6510 98893; fax: +30 6510 98889.
E-mail address: fotiadis@cs.uoi.gr (D.I. Fotiadis).
0010-4825/$ - see front matter 䉷 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.compbiomed.2005.09.011
timing as well as on the strength of the stimulus. In terms of the
resetting effect, the total charge injected into the cell seems to
be the crucial factor. Thus, any combinations of various pulse
amplitudes and durations, which result in the injection of the
same amount of current, are equivalent [5].
The response of biological oscillators to external perturba-
tions has been extensively studied using both theoretical and
experimental approaches (reviewed in [32]). Using topological
arguments, Winfree [33] greatly contributed to the qualitative
understanding of phase resetting responses in biological oscil-
lators and pointed out that, in a spatially extended system like
cardiac tissue, such responses play a crucial role in the initi-
ation of arrhythmias. Stimuli of critical amplitude and phase
might lead to annihilation of normal rhythmic activity and the
initiation of complex spiral wave arrhythmias.
Mathematical models which describe the electrical activity of
SAN have been presented by various research groups [1]. The
first models produced by Yanagihara et al. [34] and Noble and
Noble [35], where the basis for models presented later [36,37].
 D.G. Tsalikakis et al. / Computers in Biology and Medicine 37 (2007) 8 – 20
The analysis of phase response characteristics provides new
insight into the dynamics of the mathematical models, of the
electrical activities of SAN, and an experimentally verifiable
test for the accurate reconstruction of SAN tissue dynamics.
Simulation of the spatial variation in the electrical properties
of SAN cells and their phase response profiles is necessary
for reconstruction of the complex dynamics of the SAN tissue.
Zhang et al. [38] recently presented one-dimensional model
which describes all of the above.
In this work, we investigate and characterize the effects of
external stimulation on central and peripheral SAN cells us-
ing the Zhang et al. model [38]. The equation of the model
along with the appropriate initial conditions are solved using
the Runge–Kutta method. Short (0.5 ms) depolarizing and hy-
perpolarizing electrical current pulses of varying amplitude and
of timing spanning the whole period are applied and the phase
transition curves (PTC) are obtained [22]. Three-dimensional
PTC [4], relating old phase and stimulus amplitude to the new
phase after stimulation, are generated in order to locate critical
stimuli and singularities in the models. Numerical simulations
for the electrical activity of the transitional cells lying between
the center and the periphery of the SAN [38] are carried out.
Simulations are conducted when specific ionic currents vary.
The application of a critical depolarizing stimulus (about
0.4 nA) during the late repolarization phase of action potential
resulted in annihilation of the activity in central SAN cells, re-
vealing the existence of a stable singularity in the correspond-
ing model configuration [22]. Detailed analysis of the phase
response characteristics of the peripheral cells failed to show
a similar singularity and annihilation of normal activity. This
difference in phase response behavior of central and peripheral
cells is portrayed in the structure of the corresponding PTC.
2. Methods
We assume that we apply a short pulse at the SAN and we
obtain the PTC using the model proposed by Zhang et al. [38].
The spatial effects are ignored and the problem is reduced to
a time dependent problem which is described by the set of
ordinary differential equations.
2.1. Model equations
The electrical activity of central and peripheral SAN cells is
simulated using the model proposed by Zhang et al. [38] which
was developed using experimental data. The model includes
formulations for the ionic currents (INa , ICa,L , ICa,T , Ito ),
sensitive sustained current (Isus , IK,r , IK,s , If ) and back-
ground currents (Ib,Na , Ib,Ca , Ib,K , Ip , INaCa ). Table 1 shows
the parameters entering the model. The total ionic current
Itot appearing in the differential equation which models
the membrane potential V is comprised of the currents:
INa , ICa,L , Ia,T , Ito , Isus , IK,r , IK,s , If , Ib,Na , Ib,Ca , Ib,K , INaCa
and Ip . The total membrane potential is modelled as
dV −1
= Itot ,
dt C V for each of the corresponding simulations Fig. 1(b) and (d).
9
where V is the membrane potential, t is the time, C is the
membrane capacitance of SAN and
Itot = (INa + ICa,L + ICa,T + Ito + Isus + IK,r + IK,s
+ If + Ib,Na + Ib,Ca + Ib,K + INaCa + Ip ).
The transition from central to peripheral SAN activity was
modelled as in [38], using a scaling factor which depends on
the distance of each cell from the center of the SAN. The set of
differential Eqs. (1)–(16) given in Tables 2–9 is solved using a
fourth order Runge–Kutta method with time step 0.1 ms. Fur-
ther details for the formulation of the equations described in
Tables 2–9 and the experimental background behind the equa-
tions are sited in [38]. The set of differential Eqs. (1)–(16) are
solved twice: first the set of equations is solved for the central
cell with initial conditions shown in the left column of Table 9
and second for the peripheral cell with initial conditions shown
in the right column in Table 9. The model equations for central,
peripheral and transitional cell configurations are integrated un-
til a stable limit cycle is obtained. The minimum diastolic po-
tential is determined on the stable limit cycle and the vector
of dynamic variables at this point is recorded and used as the
initial condition in the subsequent simulations. For each of the
configurations examined, a single 0.5 ms depolarizing or hyper-
polarizing current stimulus of a given amplitude is applied at
several time steps during the whole period of integration. The
starting reference point is selected at the membrane potential
level of (−10 mV) during the depolarization phase of the action
potential and at each subsequent run the stimulus timing in-
creases by 1 ms. The stimulation current amplitude varies from
0 to 7.5 nA. For a given stimulus amplitude, the corresponding
PTC are obtained [22], for the whole range of stimulus ampli-
tudes, three-dimensional PTC are also obtained [4].
3. Results
The normal electrical activity of central, peripheral, and tran-
sitional SAN cells is shown in Fig. 1. The membrane poten-
tial V of central (dark line—in Fig. 1(a)) and peripheral (dark
line—in Fig. 1(c)) SAN cells is plotted vs time. The gray lines
indicate the electrical activity of the transitional cells. The du-
ration of the simulation is 1 s. The transition between central
and peripheral activity is modelled using a scaling factor (rang-
ing from 0 to 1) which depends on the distance of the cell from
the center of the SAN (see Eqs. (80) and (81) in Ref. [38] and
Ref. [39]). The scaling factor is used in order to calculate the
capacitance and the ionic conductances of the transitional cells
given their corresponding values at the center and the periphery
of the SAN.
For the transitional cell activity shown in Fig. 1 the scaling
factor values (0.04 in Fig. 1(a) and 0.13 in plot Fig. 1(c)) were
selected to indicate the point where a sharp change in mem-
brane potential configuration is evident. Thus, in Fig. 1(c) the
transitional cell appears to have the characteristic spike-and-
dome configuration of peripheral cells.
The total ionic current Itot is plotted vs membrane potential
(1)
10 D.G. Tsalikakis et al. / Computers in Biology and Medicine 37 (2007) 8 – 20

25 0.05

0 0.00

Itot
-25
V

-0.05

-50 -0.10

-75 -0.15
0.0 0.2 0.4 0.6 0.8 1.0 -75 -50 -25 0 25
(a) t (b) V

30 0.35

0 0.00

-30 -0.35
V

Itot
-60 -0.70

-90 -1.05
0.0 0.2 0.4 0.6 0.8 1.0 -90 -60 -30 0 30
(c) t (d) V

Fig. 1. Normal electrical activity in central (a,b), peripheral (c,d) and transitional sinoatrial node cells (gray lines in a–d). In figures (b) and (d) the total ionic
current Itot is plotted vs the membrane potential V shown in figures (a) and (c). Figures (a) and (c) indicate the variations of membrane potential V with time t.

40 0.05

T0 T
0 0.00

Itot
V

-40 -0.05

-80 -0.10
0.0 0.4 0.8 1.2 1.6 -75 -50 -25 0 25
(a) t (b) V

Fig. 2. (a) Single pulse perturbation protocol for the central sinoatrial node cell. A depolarizing stimulus of 0.45 nA is applied with a time delay
after the
detection (at V = −10 mV during the depolarization phase) of the second action potential. The gray trace corresponds to the unperturbed activity. (b) The total
ionic current Itot vs the membrane potential V .

The difference in Itot amplitude between central and peripheral
cells is mainly due to the contribution of INa which is not
present in central cells. The Itot –V plots provide a means to
visualize the effects of external perturbation on the SAN cell
activity (see Fig. 2) and can be considered as a representation
of the complete multi-dimensional phase space of the model.
The single pulse perturbation protocol used is shown in
Fig. 2. The gray solid line represents the normal unperturbed
electrical activity of central SAN cell with a period T0 =334 ms.
The onset of the action potential is marked at V = −10 mV (de-
polarization phase). At a time interval
after the onset of the
action potential a short (0.5 ms) depolarizing current pulse is
applied resulting in early depolarization of the cell membrane
(dark line). The phase  of the stimulus is
/T0 while the new
phase of the perturbed oscillator is  =  + 1 − T /T0 where
T is the period of the perturbed cycle [22]. The electrical ac-
tivity resumes immediately after the first perturbed cycle as is
evident from the Itot –V plot (Fig. 2(b)) in which both simula-
tions are shown (gray line corresponds to the unperturbed limit
cycle oscillation of the central SAN cell and the black line cor-
responds to the limit cycle oscillation of the SAN).
The individual ionic currents comprising Itot in both the cen-
tral and peripheral SAN cell models are shown in Fig. 3(a),
(b). Current traces of 1 s of spontaneous activity are plotted.
 D.G. Tsalikakis et al. / Computers in Biology and Medicine 37 (2007) 8 – 20 11

0.04 1
Itot

Itot
-0.10 -5
0.02 0.0
ICa.L

ICa.L
-0.10
-1.4
0.000 0.00

ICa.T
ICa.T

-0.004 -0.08
0.001 0.012
IK.s

IK.s
0.000 -0.002

0.004 0.004
INaCa

INaCa
-0.002 -0.010

0.008 0.026
INaCa

INaK
0.005 0.012
0.00 0.00
Ib

Ib

-0.01 -0.04

0.001 0.03
If

If

-0.002 -0.05
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
(a) t (b) t

Fig. 3. (a) Ionic currents for central sinoatrial node cell model; (b) Ionic currents for peripheral sinoatrial node cell model.

A comparison of the two plots reveals the contribution of each
ionic current in Itot . The sodium current INa is the principal
component of Itot in peripheral while it is absent in central
SAN cells.
In the central cell model, the maximum upstroke velocity of
the action potential is 6 V s−1 , the minimum diastolic potential
is −68 mV and the maximum systolic potential is 22 mV, while
in the peripheral cell model the maximum upstroke velocity of
the action potential is 86 V s−1 , the minimum diastolic potential
is −81 mV and the maximum systolic potential is 25 mV.
For each one of the experiments shown in Fig. 1(a) and
(c), we repeat those with
from 0 to T0 (correspondingly
0 <  < 1) and current stimulus amplitudes from 0 to 7.5 nA.
The results are shown in Fig. 4 where representative PTC are
depicted. Fig. 4(a)–(d) correspond to low amplitude stimuli
(−0.3 nA) while Fig. 4(e)–(g),(j) traces correspond to high am-
plitude stimuli. Fig. 4(a)–(d), (j) correspond to central and pe-
ripheral cells while the Fig. 4(b), (c), (f), (g) correspond to the
transitional cell configurations shown in Fig. 1 (gray lines in
Fig. 1(a) and (c), respectively).
The advance or delay in the occurrence of the first action
potential after perturbation for a set of experiments for which
the stimulus phase varies from 0 to 1 can be visualized by
plotting the ratio of the corresponding period T1 (during which
the stimulus was applied) to the period T0 of the unperturbed
cycle over the stimulus phase . Such plots were obtained for
the central and peripheral SAN (Fig. 5).
Low amplitude stimulation shown in Fig. 5(a), (b), in both
the central and peripheral SAN cells, produces only minimal
advances (T1 /T0 < 1) or delays (T1 /T0 > 1) for most of the cy-
cle excluding a small region around  = 0.4 where a significant
advance is observed. At this region the membrane is in the ab-
solute refractory period and the stimulus amplitude is sufficient
to initiate an action potential and thus to shorten T1 . Increasing
the stimulation amplitude extends the  region where an ad-
vance is observed. In central cells the delay in the occurrence
of the action potential at intermediate stimulus levels observed
close to  = 0.6 is associated with the approximation of a sin-
gularity in the phase space. Around this singularity T1 becomes
close to 2 × T0 (Fig. 5(c)) while for 0.4 <  < 0.6 approach-
ing T1 /T0 = 2, there are missing points which correspond to
T1 /T0 → ∞.
The PTC shown in Fig. 4 indicate that, in central, periph-
eral, and transitional SAN cells, both Type 0 (Fig. 4(a)–(d)) and
Type 1 (Fig. 4(e)–(j)) phase resetting is obtained. Low ampli-
tude stimuli produce continuous PTC since the effect of the
external perturbation on the phase of the limit cycle oscillator
is less pronounced. The discontinuities of the PTC for stronger
12 D.G. Tsalikakis et al. / Computers in Biology and Medicine 37 (2007) 8 – 20

1 1

φ'

φ'
0 0
0 1 0 1
(a) φ (b) φ

1 1

φ'
φ'

0 0
0 1 0 1
(c) φ (d) φ

1 1
φ'
φ'

0 0
0 1 0 1
(e) φ (f) φ

1 1
φ'
φ'

0 0
0 1 0 1
(g) φ (h) φ

Fig. 4. Phase transition curves for central ((a), (e)), peripheral ((d), (j)), and transitional cells ((b), (c), (f), (g)).

stimuli (Type 1 phase resetting) are indicative of the existence
of a singularity in the limit cycle.
Each of the PTC in Fig. 4 represents a set of experiments like
the one shown in Fig. 2 where the stimulus amplitude is con-
stant while the interval
varies. A set of such curves for var-
ious stimulus amplitudes produces a three-dimensional phase
transition surface [4] relating both  and stimulus amplitude
to the new phase  of the perturbed oscillator. Such three-
dimensional phase transition plots allow further investigation
of the phase resetting behavior of an oscillator. Moreover, such
PTC can help in finding the exact stimulus phase and amplitude
combination which brings the oscillator close to the singularity
range.
The three-dimensional phase transition plots for the cen-
tral, transitional and peripheral SAN cells are shown in
Fig. 6 as contour plots with the new phase  is color-coded
 D.G. Tsalikakis et al. / Computers in Biology and Medicine 37 (2007) 8 – 20 13

2.0 2.0
1.8 1.8
1.6 1.6
1.4 1.4
1.2 1.2
T1
T0

1.0 1.0
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
(a)  (b) 
2.0 2.0
1.8 1.8
1.6 1.6
1.4 1.4
1.2 1.2
T1
T0

1.0 1.0
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
(c)  (d) 
2.0 2.0
1.8 1.8
1.6 1.6
1.4 1.4
1.2 1.2
T1
T0

1.0 1.0
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
(e)  (f) 
2.0
1.8
1.6
1.4
1.2
T1
T0

1.0
0.8
0.6
0.4
0.2
0.0 0.2 0.4 0.6 0.8 1.0
(g) 

Fig. 5. Alteration of the timing of the first action potential after stimulation is plotted T1 /T0 vs  for various stimulation amplitudes. (a) Peripheral cell at
−0.5 nA, (c) peripheral cell at −2.5 nA, (e) peripheral cell at −5 nA, (g) peripheral cell at −6.5 nA and (b) central cell at −0.2 nA, (d) central cell at −0.4 nA,
(f) central cell at −0.6 nA (T0,central = 320 ms, T0,peripheral = 160 ms).

on a gray scale from 0 to 1. The ordering of the plots is simi-
lar to that of Fig. 4 and the corresponding PTC of Fig. 4 can
be considered as horizontal “slices” of the data at the corre-
sponding current stimulus amplitude (Istim ). The difference
in the stimulus amplitude for the corresponding plots is due
to the difference in the amplitude of the corresponding total
membrane current (a stronger stimulus is required to produce
Type 1 phase resetting behavior of Fig. 1).
In the three-dimensional phase transition plots shown in
Fig. 6, there exist regions where the new phase  is very dense
14 D.G. Tsalikakis et al. / Computers in Biology and Medicine 37 (2007) 8 – 20

Central Peripheral
0.0
0.1
-0.3 -1.5
0.2
0.3
-0.6 -3.0
0.4
Istim

0.5
-0.9 -4.5
0.6
0.7
-1.2 -6.0
0.8
0.9
-1.5 -7.5
1.0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
 (b) 
(a)

Transitional Transitional
0.0
0.1
-0.4
-1 0.2
-0.8 0.3
-2 0.4
-1.2
Istim

0.5
-3 0.6
-1.6
0.7
-2.0 -4 0.8
0.9
-2.4
-5 1.0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
(c)  (d) 

Fig. 6. Three-dimensional contour plots of the new phase vs old phase and amplitude: for (a) central, (b) peripheral, and (c), (d) transitional SAN cell.

60

30
V

-30

-60
0.0 0.5 1.0 1.5 2.0
t

Fig. 7. A critical stimulus of −0.4 nA applied at t = 0.53 s resulting in permanent annihilation of sinoatrial node repetitive activity.

(e.g. for  = 0.55 and Istim = −0.4 nA). Such regions indicate Further experimentation with finer values for
and Istim
the existence of a singularity and should be further investigated within those regions revealed the existence of a stable singular-
by fine-tuning of both
interval and Istim amplitude. ity for the central SAN cell. In the simulation shown in Fig. 7 a
 D.G. Tsalikakis et al. / Computers in Biology and Medicine 37 (2007) 8 – 20
critical stimulus of −0.4 nA is applied at t = 0.53 s resulting in
permanent annihilation of the normal repetitive activity. Tran-
sitional and the peripheral cells do not exhibit similar behavior.
The application of stimulus of appropriate amplitude and
phase (e.g.  = 0.4 and Istim = −5.4) results in a series of low
amplitude oscillations of the membrane potential with normal
limit cycle behavior resuming after 2–3 cycles.
4. Discussion
Investigation of the dynamic behavior of biological oscil-
lators and their response to external perturbation is of great
importance in biological research since biological oscillations
are involved in many vital processes in living systems [32,33].
Understanding the dynamic response of the SAN to external
perturbations is important in elucidating its behavior under nor-
mal and pathological conditions [33]. Studies in the past have
concentrated on the elucidation of the phase resetting dynam-
ics of the SAN using biophysical ionic models [15,29,5] but
did not address the issue of regional differences between cen-
tral and peripheral node cells (the corresponding ionic models
described central cell behavior).
The SAN is a complex structure and regional differences in
the electrical activity of regional cells seem to be important in
the behavior of the natural pacemaker of the heart [39,40,30].
The Zhang et al. model [38] used in this study accounts for
regional differences and can be used to model the electrical
activity of central, peripheral, and transitional cells.
In this work, the analysis of phase resetting characteristics
of each type of SAN cell by means of three-dimensional phase
transition plots, revealed important differences in their response
to external electrical stimuli (see Fig. 6). Although all types of
cells display both Type 0 and Type 1 phase resetting behavior,
the stimulus amplitude at which Type 1 phase resetting behavior
first appears varies. This is a result of the regional differences
in Itot amplitude (see Fig. 1): Itot is higher in the periphery of
SAN (mainly due to the dominance of INa ) and thus a higher
amplitude of external stimulus is required in order to signifi-
cantly affect the normal limit cycle behavior. We have produced
Itot –V plots to provide a means to visualize the effects of ex-
ternal perturbation on the SAN cell electrical activity (Fig. 2).
Our simulations show the existence of a critical stimulus
amplitude-phase combination which is capable of annihilating
the electrical activity of the central SAN cells (Fig. 7). So far,
there is no reference in the literature describing and proofing
the existence of such a combination in Zhang model. It is be-
lieved that such critical stimuli are involved in the generation
of certain cardiac arrhythmias [33] and gives new insights into
the investigation of cardiac models.
In a complex structure like the SAN, regional differences
and spatiotemporal interactions could play an important role in
its pathophysiological response to external perturbations and
should be further investigated. This must be included in our next
set of experiments targeting one-dimensional cells. However,
the use of the finite element method might help in the analysis of
two- and three-dimensional structures. The comparison of phase
resetting characteristics of other models will be included too.
15
Investigation of the effects of increases of various concentration
on annihilation of SAN activity must be further studied.
5. Summary
In the present work, we use the model proposed from Zhang
et al. [38] and we examine the dynamic response of the SAN to
external stimulus. We apply a short current to reset the sponta-
neous rhythmic activity of the single SAN. This application of
the external stimulus results in the prolongation of the normal
activity and a new phase is obtained. This resetting behavior is
quantified in terms of phase transition curves (PTCs) for short
electrical current pulses of varying amplitude which span the
whole period. Discontinuities, as we show in the figures, re-
veal critical stimulus. The application of this critical stimulus
can result in annihilation of activity on SAN cells. Detailed
analysis of the ionic mechanisms which involved in the reset-
ting behavior could be useful and provide new insight into the
physiology of the excitation in SAN cell. Our work, is a step
forward to a future work on higher dimensionality model were
we believe the resetting technic can be more significant.
Acknowledgements
This work is partially funded by the Greek Secretariat for Re-
search and Technology PENED-01E
511 to D.G. Tsalikakis.
Appendix
The parameters used in the model are given in Table 1. The
set of differential equations (1)–(16) given in Tables 2–9 was
solved using a fourth order Runge–Kutta method with time step
0.1 ms.
Table 1
Parameters of the model
4-AP 4-Aminopyridine
AM Atrial muscle
APD Action potential duration
CL Spontaneous cycle length
Cm Cell capacitance
Cma (x), C s (x)
m Capacitance of atrial muscle cell or SA node
cell in one-dimensional model of intact
SA node at distance x from center of SA node
dL , dT Activation variables for ICa,L and ICa,T
dNa,Ca Denominator constant for INa,Ca
dV /dtmax Maximum upstroke velocity of action potential
Da , Ds Diffusion coefficient between atrial muscle cells
or SA node cells in one-dimensional
model of the intact SA node
EK,s Reversal potential for IK,s
ENa , ECa , EK Equilibrium potentials for Na+ , Ca2+ , and K+
ECa,L , ECa,T Reversal potentials for ICa,L and ICa,T
F Faraday’s constant
FK,r Fraction of activation of IK,r that occurs slowly
FNa Fraction of inactivation of INa that occurs slowly
16 D.G. Tsalikakis et al. / Computers in Biology and Medicine 37 (2007) 8 – 20

Table 1 (continued) Table 1 (continued)

fL , fT Inactivation variables for ICa,L and ICa,T x Distance from center of SA node
gp , gc Conductance of a current in peripheral or central in one-dimensional model of intact SA node
SA node cell models xs Activation variable for IK,s
g a (x), g s (x) Conductance of a current in atrial muscle cell y Activation variable of If
or SA node cell in one-dimensional model z Valency of ion
of intact SA node at distance x from center of SA [Na+ ]i , [Ca2+ ]i Intracellular Na+ , Ca2+ , concentrations
node an Voltage-dependent opening rate constant of n
gNa Conductance of INa bn Voltage-dependent closing rate constant of n
gCa,L , gCa,T Conductance of ICa,L and gCa,T NaCa Position of Erying rate theory energy
gto , gsus Conductance of Ito and Isus barrier controlling voltage dependence of
gK,r , gK,s Conductance of IK,r and IK,s INaCa
gf,Na , gf,K Conductance of Na+ and K+ n Time constant of n
components of if  Space constant
gb,Na , gb,Ca , gb,K Conductance of Ib,Na , Ib,Ca and Ib,K
h1 , h2 Fast and slow inactivation variables for INa
h Net fractional availability of INa
INa TTX-sensitive Na+ current
ICa,L , iCa,T L- and T-type Ca2+ currents Table 2
Ito , Isus Transient and sustained components of 4-AP- Formulation of the equations of Na+ current (INa )
sensitive current
IK,r , IK,s Rapid and slow delayed rectifying K+ currents The sodium current INa is given by the equation
IK Sum of IK,r and IK,s
If Hyperpolarization-activated current (F )2.0 (e(V −ENa )F /RT − 1.0)
INa = gNa (m)3 h[Na+ ]o V
If,Na , If,K Na+ and K+ components of If RT (eVF/RT − 1.0)
Ib,Na , Ib,Ca , Ib,K Background Na+ , Ca2+ , and K+ currents
INaCa Na+ /Ca2+ exchanger current
Ip Na+ –K+ pump current Sodiumcurrent m gate
Ip Maximum Ip
Itot Total ionic current in a cell dm (m∞ − m)
= (2)
a (x), I s (x)
Itot tot Total ionic current in atrial dt m
muscle cell or SA node cell in one-dimensional
model of intact SA node at distance x
1.0/3.0
1.0
from center of SA node m∞ =
(1.0 + e−V /5.46 )
Ist Sustained current
IK,ACh ACh-activated K+ current

0.0006247
IK,AP ATP-sensitive K+ current m =
kNa,Ca Scaling factor for INa,Ca (0.832e−0.335(V +56.7) + 0.627e0.082(V +65.01) )
Km,Na , Km,K Dissociation constants for Na+ and K+ activation +0.00004)
of Ip
L Length of string of SA node and atrial
tissue in one-dimensional Sodium current h gate
model of intact SA node
h = ((1.0 − FNa )h1 + FNa h2 )
Ls Length of string of SA node tissue
in one-dimensional model of intact SA node
m Activation variable for iNa dh1 (h1,∞ − h1 )
= (3)
MDP Maximum diastolic potential dt  h1
n∞ Steady-state value of n
pa General activation variable for IK,r dh2 (h2,∞ − h2 )
= (4)
pa,f , pa,s Fast and slow activation variables forIK,r dt h 2
pi Inactivation variable for IK,r
Q10 Fractional change in a variable 1.0
with a 10 ◦ C increase in temperature h1,∞ =
(1.0 + e(V +66.1)/6.4 )
r Activation variable for Ito
R Universal gas constant
h2,∞ = h1,∞
q Inactivation variable for Ito
 
SA node, SAN Sinoatrial node 0.000003171e−0.2815(V +17.11)
t Time  h1 = + 0.0005977
(1.0 + 0.003732e−0.3426(V +37.76) )
T Absolute temperature
TOP Takeoff potential
V Membrane potential
V a, V s Membrane potential of atrial muscle  
cell or SA node cell in one-dimensional model of 0.00000003186e−0.6219(V +18.8)
 h2 = + 0.003556
intact SA node (1.0 + 0.00007189e−0.6683(V +34.07) )
V a (x), V s (x) Membrane potential of atrial muscle
cell or SA node cell in one-dimensional model of
intact SA node at distance x from center of SA node
 D.G. Tsalikakis et al. / Computers in Biology and Medicine 37 (2007) 8 – 20 17

Table 3 Table 4 (continued)

Formulation of the equations of L-type Ca2+ current, ICa,L
1.0
L-type Ca channel ICa,L d,T =
(d,T + d,T )


0.006
ICa,L = gCa,L fL dL + −(V +14.1)/6.0
(V − ECa,L )
(1.0 + e )
1.0
dT,∞ =
(1.0 + e−(V +37.0)/6.8 )
L-type Ca channel d gate

ddL (dL,∞ − dL ) T-type Ca channel f gate

= (5)
dt d,L
dfT (fT,∞ − fT )
= (8)

 dt f,T
(V + 35.0) 42.45V
d,L = −14.19 − −0.208V
(e−(V +35.0)/2.5 − 1.0) (e − 1.0)
f,T = 15.3e−(V +71.7)/83.3
5.71(V − 5.0)
d,L = 0.4(V −5.0)
(e − 1.0)
f,T = 15.0e(V +71.7)/15.38
1.0
d,L = 1.0
(d,L + d,L ) f,T =
(f,T + f,T )
1.0 1.0
dL,∞ = fT,∞ =
(1.0 + e−(V +23.1)/6.0 ) (1.0 + e(V +71.0)/9.0 )
L-type Ca channel f gate

dfL (fL,∞ − fL )
= (6)
dt f,L
Table 5
Formulation of the equations for sensitive currents (Ito and Isus ) and delayed
3.12(V + 28.0)
f,L = rectifying current (IK,r )
(e(V +28.0)/4.0 − 1.0)
Ito
25.0
f,L = Ito = gto qr(V − EK )
(1.0 + e−(V +28.0)/4.0 )
Isus
Isus = gsus r(V − EK )
1.0
f,L =
(f,L + f,L ) q gate

dq (q∞ − q)
1.0 = (9)
fL,∞ = dt q
(1.0 + e(V +45.0)/5.0 )

1.0
q∞ =
(1.0 + e(V +59.37)/13.1 )


0.06517
Table 4 q = 0.0101 + + 0.000024e0.1(V +50.93)
0.57e−0.08(V +49.0)
Formulation of the equations of T-type Ca2+ current (ICa,T )
T-type Ca channel ICa,T 4-AP sensitive currents r gate

ICa,T = gCa,T dT fT (V − ECa,T ) dr (r∞ − r)

= (10)
dt r

T-type Ca channel d gate 1.0

r∞ =
ddT (dT,∞ − dT ) (1.0 + e−(V −10.93)/19.7 )
= (7)
dt d,T


0.01559
d,T = 1068.0e(V +26.3)/30.0 r = 0.00298 +
(1.037e0.09(V +30.61)+ 0.369e−0.12(V +23.84) )

d,T = 1068.0e−(V +26.3)/30.0

References
18 D.G. Tsalikakis et al. / Computers in Biology and Medicine 37 (2007) 8 – 20

Table 5 (continued) Table 6 (continued)

Rapid delayed rectifying potassium current IK,r
14.0
x s =
IK,r = gK,r pa pi (V − EK ) (1.0 + e−(V −40.0)/9.0 )

pa = ((1.0 − FK,r )pa,f + FK,r pa,s )

xs = e−V /45.0

Rapid delayed rectifying potassium current pa,f gate

dpa,f (pa,f,∞ − pa,f )

= (11)
dt pa,f xs
xs∞ =
(xs + xs )

1.0
pa,f,∞ =
(1.0 + e−(V +14.2)/10.6 ) 1.0
xs =
(xs + xs )

1.0
pa,f = Hyperpolarization activated current If
(37.2e(V −9.0)/15.9 + 0.96e−(V −9.0)/22.5 )
If = (If,Na + If,K )

Rapid delayed rectifying potassium current pa,s gate

If,Na
dpa,s (pa,s,∞ − pa,s )
= (12)
dt pa,s If,Na = gf,Na y(V − ENa )

If,K
pa,s,∞ = pa,f,∞ (13)
If,K = gf,K y(V − EK )

1.0 Hyperpolarization activated current y gate

pa,s =
(4.2e(V −9.0)/17.0 + 0.15e−(V −9.0)/21.6 )
dy (y∞ − y)
= (16)
dt y
Rapid delayed rectifying potassium pi gate

dpi (pi,∞ − pi )
= (14)
dt p i y = e−(V +78.91)/26.62
1.0
pi,∞ =
(1.0 + e(V +18.6)/10.1 )
y = e(V +75.13)/21.25

y
y∞ =
(y + y )

Table 6
Formulation of the equations for slow delayed rectifying current (IK,s ) and 1.0
y =
hyperpolarization current (If ) (y + y )

Slow delayed rectifying potassium current IK,s

IK,s = gK,s (xs )2.0 (V − EK,s )

Slowdelayed rectifying potassium current xs gate

dxs (xs∞ − xs )
= (15)
dt xs
 D.G. Tsalikakis et al. / Computers in Biology and Medicine 37 (2007) 8 – 20 19

[1] in: D. Zipes, J. Jalife (Eds.), Cardiac Electrophysiology—From Cell to

Table 7
Bedside, second ed., W.B. Saunders, Philadelphia, 1995.
Formulation of the equations for background, pump and exchanger currents
[2] W. Zeng, L. Glass, A. Shrier, The topology of phase response curves
Sodium background current Ib,Na induced by single and paired stimuli in spontaneously oscillating chick
heart cell aggregates, J. Biol. Rhythms. 7 (2) (1992) 89–104.
Ib,Na = gb,Na (V − ENa )
[3] J.M. Shumaker, J.W. Clark, W.R. Giles, A model of the phase-sensitivity
Potassium background current Ib,K of the pacemaking cell in the bullfrog heart, J. Theor. Biol. 151 (2)
Ib,K = gb,K (V − EK ) (1991) 193–230.
Calcium background current Ib,Ca [4] G.P. Kremmydas, A.V. Holden, A. Bezerianos, T. Bountis, Representation
Ib,Ca = gb,Ca (V − ECa ) of sino-atrial node dynamics by circle maps, Internat. J. Bifurcation
Chaos 6 (10) (1996) 1799–1805.
Sodium calcium exchanger INaCa [5] A.C. Coster, B.G. Celler, Phase response of model sinoatrial node cells,
INaCa = KNaCa Ann. Biomed. Eng. 31 (3) (2003) 271–283.
(([Na+ ]i )3.0 [Ca2+ ]o e0.03743V NaCa −([Na+ ]o )3.0 [Ca2+ ]i e0.0374V (NaCa −1.0) ) [6] S. Abramovich-Sivan, S. Akselrod, A prc based model of a pacemaker
(1.0+dNaCa ([Ca2+ ]i ([Na+ ]o )3.0 +[Ca2+ ]o ([Na+ ]i )3.0 )) cell: effect of vagal activity and investigation of the respiratory sinus
Sodium potassium pump Ip arrhythmia, J. Theor. Biol. 192 (2) (1998) 219–234.

3.0
2.0 [7] E.E. Verheijck, R. Wilders, R.W. Joyner, D.A. Golod, R. Kumar,
[Na+ ]i [K+ ]o
Ip =Ip,max H.J. Jongsma, L.N. Bouman, A.C.G. van Ginneken, Pacemaker
(Km,Na + [Na+ ]i ) (Km,K + [K+ ]o ) synchronization of electrically coupled rabbit sinoatrial node cells,
1.6
× J. Gen. Physiol. 111 (1) (1998) 95–112.
(1.5 + e−(V +60.0)/40.0 ) [8] S.S. Demir, J.W. Clark, W.R. Giles, Parasympathetic modulation of
sinoatrial node pacemaker activity in rabbit heart: a unifying model,
Am. J. Physiol. Heart Circ. Physiol. 276 (6) (1999) H2221–H2244.
[9] R.F. Gilmour, Phase-resetting of circus movement reentry in isolated
canine purkinje-muscle preparations, Circulation 76 (4) (1987) 112.
Table 8 [10] V.S. Reiner, C. Antzelevitch, Phase resetting and annihilation in a
Formulation of the equations for reversal and equilibrium potentials mathematical-model of sinus node, Am. J. Physiol. 249 (6) (1985) 1143
–1153.
Reversal and equilibrium potentials [11] T.R. Chay, Y.S. Lee, Phase resetting and bifurcation in the ventricular
myocardium, Biophys. J. 47 (5) (1985) 641–651.
RT [Na+ ]o [12] J.M.B. Anumonwo, M. Delmar, A. Vinet, D.C. Michaels, J. Jalife, Phase
ENa = ln
F [Na+ ]i resetting and entrainment of pacemaker activity in single sinus nodal
cells, Circ. Res. 68 (4) (1991) 1138–1153.
RT [K+ ]o
EK = ln + [13] A.M. Kunysz, A.A. Munk, A. Shrier, Phase resetting and dynamics in
F [K ]i isolated atrioventricular nodal cell clusters, Chaos 5 (1) (1995) 184–192.
[14] M.R. Guevara, A. Shrier, Phase resetting in a model of cardiac purkinje-
RT [Ca2+ ]o
ECa = ln fiber, Biophys. J. 52 (2) (1987) 165–175.
2.0F [Ca2+ ]i [15] M.R. Guevara, H.J. Jongsma, Phase resetting in a model of sinoatrial
RT ([K+ ]o + 0.12[Na+ ]o ) nodal membrane—ionic and topological aspects, Am. J. Physiol. 258 (3)
EK,s = ln (1990) H734–H747.
F ([K+ ]i + 0.12[Na+ ]i )
[16] J.R. Clay, R.M. Brochu, A. Shrier, Phase resetting of embryonic chick
atrial heart cell aggregates—experiment and theory, Biophys. J. 58 (3)
(1990) 609–621.
[17] M.R. Guevara, A. Shrier, L. Glass, Phase resetting of spontaneously
beating embryonic ventricular heart cell aggregates, Am. J. Physiol. 251
Table 9
(6) (1986) H1298–H1305.
Initial conditions for central and peripheral cell
[18] J.R. Clay, M.R. Guevara, A. Shrier, Phase resetting of the rhythmic
Variable Value activity of embryonic heart cell aggregates—experiment and theory,
Biophys. J. 45 (4) (1984) 699–714.
Central Peripheral [19] R.M. Luceri, A. Castellanos, D.S. Kayden, F. Moleiro, R.G. Trohman,
V −55.4434 −78.4021 J.T. Horgan, R.J. Myerburg, Phase resetting of ventricular parasystolic
m 0.158763 0.0531058 rhythms produced by extrinsic perturbations, Pace Pacing Clin.
h1 0.0365616 0.341077 Electrophys. 7 (3) (1984) 461.
h2 0.0515223 0.0207796 [20] J. Jalife, A.J. Hamilton, V.R. Lamanna, G.K. Moe, Effects of current
dL 0.00328396 0.000109311 flow on pacemaker activity of the isolated kitten sinoatrial node,
fL 0.853578 0.97761 Am. J. Physiol. 238 (3) (1980) H307–H316.
dT 0.0622561 0.00226369 [21] J. Jalife, V.A.J. Slenter, J.J. Salata, D.C. Michaels, Dynamic vagal control
fT 0.129637 0.278358 of pacemaker activity in the mammalian sinoatrial node, Circ. Res. 52
y 0.0230307 0.0215392 (6) (1983) 642–656.
r 0.0334583 0.0110545 [22] L. Glass, A.T. Winfree, Discontinuities in phase-resetting experiments,
q 0.288777 0.435506 Am. J. Physiol. 246 (2) (1984) R251–R258.
xs 0.137793 0.0950833 [23] D.C. Michaels, D.R. Chialvo, E.P. Matyas, J. Jalife, Chaotic activity in
pa,f 0.29045 0.313397 a mathematical model of the vagally driven sinoatrial node, Circ. Res.
pa,s 0.561665 0.3783 65 (5) (1989) 1350–1360.
pii 0.973981 0.991546 [24] V.I. Arnold, Cardiac arrhythmias and circle mappings, Chaos 1 (1) (1991)
20–24.
[25] L. Glass, Cardiac arrhythmias and circle maps—a classical problem,
Chaos 1 (1) (1991) 13–19.
20 D.G. Tsalikakis et al. / Computers in Biology and Medicine 37 (2007) 8 – 20
[26] A.T. Winfree, Biological rhythms and the behavior of populations of
coupled oscillators, J. Theor. Biol. 16 (1) (1967) 15–42.
[27] W.P.M. Vanmeerwijk, G. Debruin, A.C.G. Vanginneken, J. Vanhartevelt,
H.J. Jongsma, E.W. Kruyt, S.S. Scott, D.L. Ypey, Phase resetting
properties of cardiac-pacemaker cells, J. Gen. Physiol. 83 (4) (1984)
613–629.
[28] M. Kawato, R. Suzuki, Biological oscillators can be stopped–topological
study of a phase response care, Biol. Cybern. 30 (4) (1978) 241–248.
[29] D.C. Michaels, E.P. Matyas, J. Jalife, A mathematical model of the
effects of acetylcholine pulses on sinoatrial pacemaker activity, Circ.
Res. 55 (1) (1984) 89–101.
[30] M.R. Boyett, H. Honjo, I. Kodama, The sinoatrial node, a heterogeneous
pacemaker structure, Cardiovasc. Res. 47 (4) (2000) 658–687.
[31] M.R. Boyett, H. Zhang, A. Garny, A.V. Holden, Control of the pacemaker
activity of the sinoatrial node by intracellular Ca2+ . experiments and
modelling, Philos. Trans. Soc. London Ser. A Math. Phys. Eng. Sci. 359
(1783) (2001) 1091–1110.
[32] A.T. Winfree, The Geometry of Biological Time, Springer, New York,
1980.
[33] A.T. Winfree, When Time Breaks Down, Princeton University Press,
Princeton, 1987.
[34] K. Yanagihara, A. Noma, H. Irisawa, Reconstruction of sinoatrial node
pacemaker potential based on the voltage-clamp experiments, Jpn.
J. Physiol. 30 (1980) 841–857.
[35] D. Noble, S. Noble, A model of sino-atrial node electrical activity based
on a modification of the difrancesco-noble equations, Proc. R. Soc.
London B Biol. Sci. 222 (1984) 295–304.
[36] S. Demir, J. Clark, C. Murphey, W. Giles, A mathematical model of a
rabbit sinoartial node cell, Am. J. Physiol. Cell Physiol. 266 (3) (1994)
832–852.
[37] S. Dokos, B. Celler, N. Lovell, Modification of difrancesco-noble
equations to simulate the effects of vagal stimulation on vivo mammalian
sinoatrial node electrical activity, Ann. Biomed. Eng. 21 (1999)
321–335.
[38] H. Zhang, A.V. Holden, I. Kodama, H. Honjo, M. Lei, T. Varghese, M.R.
Boyett, Mathematical models of action potentials in the periphery and
center of the rabbit sinoatrial node, Am. J. Physiol. Heart Circ. Physiol.
279 (1) (2000) H397–H588.
[39] A. Garny, P. Kohl, P.J. Hunter, M.R. Boyett, D. Noble, One-dimensional
rabbit sinoatrial node models: benefits and limitations, J. Cardiovasc.
Electrophysiol. 14 (s10) (2003) S121.
[40] H. Zhang, A.V. Holden, M.R. Boyett, Gradient model versus mosaic
model of the sinoatrial node, Circulation 103 (4) (2001) 584–588.
Dimitrios G. Tsalikakis was born in Thessaloniki, Greece, in 1977. He
received the Diploma degree in Mathematics from the University of Ioannina
and he is a PhD candidate in the Department of Cardiology since 2002.
His research interests include cardiac modeling, monophasic action potential
analysis, automated systems analysis, virtual instrumentation, and scientific
computing. He is member of the Unit of Medical Technology and Intelligent
Information Systems.
Henggui Zhang, born in 1964 in China, educated in Physics (BSc, 1985),
Computer Science (MSc, 1985–1988), Non-linear Science (PhD, 1988–1990),
and Biomedical Science (PhD, 1991–1994). Since 1991, he has been working
as a research assistant (1991–1994, Leeds, funding from MRC), postdoctoral
research fellow (1994–1995, JHU, funding from NSF, USA; 1995–1996,
Leeds; funding from Welcome Trust; 1996–2000, Leeds; funding from the
British Heart Foundation), and then a senior research fellow (2000–2001,
Leeds, funding from British Heart Foundation). In October 2001, he moved
to UMIST to take up a lectureship. Currently he is a senior lecturer in
Biological Physics Group, School of Physics and Astronomy, the University
of Manchester.
Dimitrios I. Fotiadis was born in Ioannina, Greece, in 1961. He received
the Diploma degree in Chemical Engineering from the National Technical
University of Athens and the PhD degree in Chemical Engineering from the
University of Minnesota, USA. Since 1995, he has been with the Department
of Computer Science, University of Ioannina, Greece, where he is currently
an Associate Professor. He is the director of the Unit of Medical Technology
and Intelligent Information Systems. His research interests include biomedical
technology, biomechanics, scientific computing and intelligent information
systems.
Georgios P. Kremmydas, born in 1972 in Argostoli, Kefalonia Island, Greece.
He received his BSc (Hon) Degree in Physiology in 1994, University of Leeds,
UK and a PhD in 2000, Department of Biomedical Sciences, University
of Leeds, UK. He is currently the Director of Research Technology and
Development Department of L.U.M.C. of Kefalonia and Ithaki and a visiting
researcher of the Computer Science Department, University of Ioannina,
Greece.
Lampros K. Michalis was born in Arta, Greece, in 1960. He received the
MD degree with distinction from the Medical School, University of Athens,
Greece, in 1984. Since 1995, he has been with the Medical School, University
of Ioannina, Greece, where he is currently a Professor in Cardiology. He is in
charge of the coronary care unit and the catheter laboratory of the University
Hospital of Medical School, Ioannina, Greece. His research interests focus
on bioengineering, interventional cardiology, and electrophysiology.