**3. Construction and analysis of proposed models**

In this section we turn to constructing new models for specific pathologies we would like to describe. All the models we propose below consist of two modified van der Pol models. The system with delayed feedback describes various pathologies observed in the heart action, e.g. WPW syndrome and different types of AVNRT.

A numerical model was created using Dynamics Solver and a program which was written in C++. The explicit Runge-Kutta formula with a variable integration step was used. Results were confirmed and plots were made with Matlab standard dde23 solver.

For the single node model (Eq. (4)) of an electrical conduction system with no feedback, a periodic potential and the corresponding limit cycle (with a transient state) are obtained like in **Figures 1** and **2**. Potential period equals 1.4.

The reference parameters have values which are given below:

$$\mathbf{a}\_1 = \mathbf{a}\_2 = \mathbf{S}, \; \mathbf{d}\_1 = \mathbf{d}\_2 = \mathbf{3}, \; \mathbf{e}\_1 = 7, \; \mathbf{e}\_2 = 4.5, \; \mathbf{f}\_1 = \mathbf{f}\_2 = \mathbf{3}. \tag{6}$$

**Figure 1.** Time series for reference model without feedback and delay, see Eq. (5).

**Figure 2.** Phase portrait for reference model without feedback and delay, see first part of Eq. (5).

The addition of coupling (especially feedback with delay) to the *x*<sup>1</sup> term allows modelling the re-entry wave which causes the exceptional situation when AV node is the master for SA node. Such situation takes place in case of WPW syndrome, this situation is presented by the following equations:

$$\dot{\mathbf{x}}\_1 = y\_1 + \mathbf{k}(\mathbf{x}\_1(\mathbf{t} - \mathbf{r}) - \mathbf{x}\_1)$$

$$\dot{y}\_1 = -a\_1(\mathbf{x}\_1 \mathbf{2} - 1)y\_1 - f\_1 \mathbf{x}\_1(\mathbf{x}\_1 + d\_1)(\mathbf{x}\_1 + e\_1)$$

Parameters have the following values: *k* =2, *τ* =0.25 and other parameters are the same as in the reference system. In the time series of the modified model with feedback, there is a 'delayed impulse'. The result of introducing feedback to the system is the creation of a wave similar to the initial wave—action potential, but delayed. It can be assumed that such a feedback can be treated as an external, periodic excitation of the wave. In **Figure 3** we can observe a kind of tachycardia which is a typical symptom of the Wolff-Parkinson-White syndrome. The period of oscillations is about 0.9 whereas the reference period of physiological action potential is 1.4.

**Figure 3.** Dynamics of action potential during WPW syndrome.

**Figure 1.** Time series for reference model without feedback and delay, see Eq. (5).

80 Numerical Simulation - From Brain Imaging to Turbulent Flows

**Figure 2.** Phase portrait for reference model without feedback and delay, see first part of Eq. (5).

Now we consider the AVNRT pathology. We would like to propose a mechanism which reproduces the dynamics of action potential occurring at the phenomenon of AVNRT. In this case, we have a coexistent conduction by two or more pathways: slow and fast. In this model, we assume that the fast pathway is described by the first van der Pol model, but the slow pathway is treated as an action potential in other characteristic, so we use the second model of van der Pol to describe the pathway.

Next, we focus on such a type of AVNRT in which only one fast pathway with depolariza‐ tion of slow pathway occurs. The conduction is only out of the fast pathway, because de‐ scending depolarization of slow pathway prevents conduction. To describe that situation, we add only one unidirectional coupling to our model. This coupling does not change the main rhythm *x*<sup>1</sup> (see **Figure 4**). In this case, the model could be written in the following form:

$$\begin{cases} \dot{\mathbf{x}}\_1 = \mathbf{y}\_1 \\ \dot{\mathbf{y}}\_1 = a\_1 \left( \mathbf{x}\_1^2 - \mathbf{1} \right) \mathbf{y}\_1 - f\_1 \mathbf{x}\_1 \left( \mathbf{x}\_1 + d\_1 \right) \left( \mathbf{x}\_1 + e\_1 \right) \\ \dot{\mathbf{x}}\_2 = \mathbf{y}\_2 + k \mathbf{x}\_1 \\ \dot{\mathbf{y}}\_2 = a\_2 \left( \mathbf{x}\_2^2 - \mathbf{1} \right) \mathbf{y}\_2 - f\_2 \mathbf{x}\_2 \left( \mathbf{x}\_2 + d\_2 \right) \left( \mathbf{x}\_2 + e\_2 \right) \end{cases} \tag{7}$$

where *k* is the coupled coefficient, *τ* is the delay and other parameters are fixed as *a* =5, *f* =3, *d* =3, *e* =7.

**Figure 4.** Time series for fast type of AVNRT.

Now, we consider a situation where the difference in the refractive state of both pathways leads to an excitation of one of them causing the re-entry wave [20]. In **Figure 5**, the result for a typical AVNRT (fast/slow) is presented. This type of AVNRT is described by the following system:

case, we have a coexistent conduction by two or more pathways: slow and fast. In this model, we assume that the fast pathway is described by the first van der Pol model, but the slow pathway is treated as an action potential in other characteristic, so we use the second model

Next, we focus on such a type of AVNRT in which only one fast pathway with depolariza‐ tion of slow pathway occurs. The conduction is only out of the fast pathway, because de‐ scending depolarization of slow pathway prevents conduction. To describe that situation, we add only one unidirectional coupling to our model. This coupling does not change the main rhythm *x*<sup>1</sup> (see **Figure 4**). In this case, the model could be written in the following

( ) ( )( )

1

(7)

*x y y a x y fx x d x e x y kx y a x y fx x d x e*

2

<sup>ì</sup> <sup>=</sup> <sup>ï</sup>

2

= + ï

1

&

1 1

1 1 1 1 11 1 1 1 1 22 1

ï = -- + +

&

<sup>ï</sup> = -- + + <sup>î</sup>

2 2 2 2 22 2 2 2 2

( ) ( )( )

where *k* is the coupled coefficient, *τ* is the delay and other parameters are fixed as

of van der Pol to describe the pathway.

82 Numerical Simulation - From Brain Imaging to Turbulent Flows

í

&

&

form:

*a* =5, *f* =3, *d* =3, *e* =7.

**Figure 4.** Time series for fast type of AVNRT.

$$\begin{cases} \dot{\mathbf{x}}\_{1} = \mathbf{y}\_{1} - k\left(\mathbf{x}\_{1}\left(t - \tau\right) - \mathbf{x}\_{1}\right) \\\\ \dot{\mathbf{y}}\_{1} = a\_{1}\left(\mathbf{x}\_{1}^{2} - 1\right)\mathbf{y}\_{1} - f\_{1}\mathbf{x}\_{1}\left(\mathbf{x}\_{1} + d\_{1}\right)\left(\mathbf{x}\_{1} + e\_{1}\right) \\\\ \dot{\mathbf{x}}\_{2} = \mathbf{y}\_{2} \\\\ \dot{\mathbf{y}}\_{2} = a\_{2}\left(\mathbf{x}\_{2}^{2} - 1\right)\mathbf{y}\_{2} - f\_{2}\mathbf{x}\_{2}\left(\mathbf{x}\_{2} + d\_{2}\right)\left(\mathbf{x}\_{2} + e\_{2}\right) \end{cases} \tag{8}$$

**Figure 5.** Dynamics of action potential for a typical AVNRT, given in Eq. (8).

The mechanism of the typical AVNRT is similar to this which is observed in the WPW syn‐ drome. Also, we have a feedback with delay in the first part of the given model.

The atypical AVNRT (slow/fast type of AVNRT) is described in the following way:

$$\begin{cases} \dot{\mathbf{x}}\_{1} = \mathbf{y}\_{1} - k\_{1} \left( \mathbf{x}\_{2} \left( t - \tau \right) - \mathbf{x}\_{1} \right) \\\\ \dot{\mathbf{y}}\_{1} = a\_{1} \left( \mathbf{x}\_{1}^{2} - 1 \right) \mathbf{y}\_{1} - f\_{1} \mathbf{x}\_{1} \left( \mathbf{x}\_{1} + d\_{1} \right) \left( \mathbf{x}\_{1} + e\_{1} \right) \\\\ \dot{\mathbf{x}}\_{2} = \mathbf{y}\_{2} - k\_{2} \left( \mathbf{x}\_{1} - \mathbf{x}\_{2} \left( t - \tau \right) \right) \\\\ \dot{\mathbf{y}}\_{2} = a\_{2} \left( \mathbf{x}\_{2}^{2} - 1 \right) \mathbf{y}\_{2} - f\_{2} \mathbf{x}\_{2} \left( \mathbf{x}\_{2} + d\_{2} \right) \left( \mathbf{x}\_{2} + e\_{2} \right) \end{cases} \tag{9}$$

In both of those types of AVNRT (fast/slow and slow/fast) we obtain regular fast rhythm, which is a typical behaviour for this kind of pathology (during this type of tachycardia, the rhythm of the heart is about 35% more frequent than normal rhythm, which is in accordance with our results).

The last case is for the situation where there are two re-entry waves but both go through slow pathways (slow/slow type of AVNRT), which we reflect by the following system of equations:

$$\begin{cases}
\dot{\mathbf{x}}\_{1} = \mathbf{y}\_{1} - k\_{1} \left(\mathbf{x}\_{2} \left(t - \tau\right) - \mathbf{x}\_{1}\right) \\
\dot{\mathbf{y}}\_{1} = a\_{1} \left(\mathbf{x}\_{1}^{2} - 1\right) \mathbf{y}\_{1} - f\_{1} \mathbf{x}\_{1} \left(\mathbf{x}\_{1} + d\_{1}\right) \left(\mathbf{x}\_{1} + \mathbf{e}\_{1}\right) \\
\dot{\mathbf{x}}\_{2} = \mathbf{y}\_{2} - k\_{2} \left(\mathbf{x}\_{1} - \mathbf{x}\_{2} \left(t - \tau\_{1}\right)\right) - k\_{3} \left(\mathbf{x}\_{1} - \mathbf{x}\_{2} \left(t - \tau\_{2}\right)\right) \\
\dot{\mathbf{y}}\_{2} = a\_{2} \left(\mathbf{x}\_{2}^{2} - 1\right) \mathbf{y}\_{2} - f\_{2} \mathbf{x}\_{2} \left(\mathbf{x}\_{2} + d\_{2}\right) \left(\mathbf{x}\_{2} + \mathbf{e}\_{2}\right)
\end{cases} \tag{10}$$

In this variant, we also observe shortening (like in **Figure 5**) of the period of oscillations. In this pathology, the presence of more than two conduction paths is possible which is associated with more re-entry waves. Increasing the number of feedbacks modelling re-entry waves of slow pathways causes a progressive shortening of the period of oscillation, while the rhythm remains regular. The conclusion is that a regular excitation may evoke a regular answer.

After examining the influence of excitation with modified van der Pol oscillator wave, we have analysed behaviour of the modified van der Pol oscillator with feedback excited with a single rectangular pulse. In this way, two excitations co-exist in the system—the wave brought by the oscillator's feedback and the one from the external pulse. The pulse is defined by a step function H for the amplitude *A*=4 and the length of pulse *L* =0.05 and it is applied to the oscillator with *k* =1 and *T* =1. The function H is defined as:

$$H\_{\mathcal{A},L}\left(t\right) = \begin{cases} \mathcal{A}, \ 0 < t < L\\ 0, \text{ otherwise} \end{cases}$$

Influence of the pulse is treated as an appearance of the accessory current in the system. For this reason, this pulse is introduced to the equation describing the potential change (the *x* variable):

$$\begin{aligned} \dot{\mathbf{x}} &= \mathbf{y}\left(t\right) + k\left(\mathbf{x}\left(t - T\right) - \mathbf{x}\left(t\right)\right) + H\_{A,L}\left(t - t\_p\right), \\ \dot{\mathbf{y}} &= -a\left(\mathbf{x}^2\left(t\right) - \mathbf{l}\right)\mathbf{y}\left(t\right) - f\mathbf{x}\left(t\right)\left(\mathbf{x}\left(t\right) + d\right)\left(\mathbf{x}\left(t\right) + e\right) \end{aligned}$$

where *tp* denotes the moment in time when the pulse starts.

( ( ) )

t

t

(9)

(10)

( ) ( )( ) ( ( ))

1 1 12 1

*x y kxt x y a x y fx x d x e x y kx xt y a x y fx x d x e*

1 1 1 1 11 1 1 1 1

2 2 2 2 22 2 2 2 2

2 2 21 2

<sup>ï</sup> = -- + + <sup>î</sup>

ï = -- + + ï

2

&

&

1

ï =- - -

<sup>ì</sup> =- -- <sup>ï</sup>

1

2

í

84 Numerical Simulation - From Brain Imaging to Turbulent Flows

&

ï

ì ï ï ï í ï ï ïî

&

& &

results).

variable):

&

( ) ( )( )

In both of those types of AVNRT (fast/slow and slow/fast) we obtain regular fast rhythm, which is a typical behaviour for this kind of pathology (during this type of tachycardia, the rhythm of the heart is about 35% more frequent than normal rhythm, which is in accordance with our

The last case is for the situation where there are two re-entry waves but both go through slow pathways (slow/slow type of AVNRT), which we reflect by the following system of equations:

> ( ( ) ) ( ) ( )( )

t

1 1 12 1

*x y kxt x y a x y fx x d x e x y kx xt kx xt y a x y fx x d x e*

=- -- = -- + + =- - - - - - = -- + +

1 1 1 1 11 1 1 1 1 2 2 21 2 1 31 2 2

2 2 2 2 22 2 2 2 2

t

2

&

oscillator with *k* =1 and *T* =1. The function H is defined as:

1

1

2

( ) ( )( )

In this variant, we also observe shortening (like in **Figure 5**) of the period of oscillations. In this pathology, the presence of more than two conduction paths is possible which is associated with more re-entry waves. Increasing the number of feedbacks modelling re-entry waves of slow pathways causes a progressive shortening of the period of oscillation, while the rhythm remains regular. The conclusion is that a regular excitation may evoke a regular answer.

After examining the influence of excitation with modified van der Pol oscillator wave, we have analysed behaviour of the modified van der Pol oscillator with feedback excited with a single rectangular pulse. In this way, two excitations co-exist in the system—the wave brought by the oscillator's feedback and the one from the external pulse. The pulse is defined by a step function H for the amplitude *A*=4 and the length of pulse *L* =0.05 and it is applied to the

> , ( ) , 0 0, *A L*

Influence of the pulse is treated as an appearance of the accessory current in the system. For this reason, this pulse is introduced to the equation describing the potential change (the *x*

*H t*

*A tL*

*otherwise* <sup>ì</sup> < < <sup>=</sup> <sup>í</sup> î

( ( )) ( ( ))

 t The influence of the pulse, which results in the period length change, depends on in which phase of oscillations it is applied. As mentioned before, the excitation influences the system having the regular behaviour.

**Figure 6.** Numerical solution of the system with feedback (*k* =1, *T* =0.5), excited with the rectangular pulse.

Numerical solution of the system with feedback (*k* =1, *T* =0.5), excited with the rectangular pulse is presented in **Figure 6**. Red colour presents modified van der Pol model with feed‐ back, whereas blue one shows the modified van der Pol model with feedback, which is addi‐ tionally excited with the rectangular pulse. In this case application of the external pulse does not influence oscillator's period length, oscillations are stable. There are phase ranges for which oscillations are sensitive to a disturbance. Phase response curves for modified van der Pol system without feedback and for two cases with feedback (*k* =1, *T* =1 and *k* =1, *T* =0.5) which are excited with the rectangular pulse (*A*=4, *L* =0.05) are shown in **Figure 7**. A phase response curve is a graphic presentation of the influence of excitation for oscillation phases of modified van der Pol system. Its shape depends on in which phase of oscillation the exci‐ tation is exerted. This curve describes the dependence between the change of length of one cycle of oscillation and the excitation phase *φ*. For example, the application of the excitation in a moment when the system abides in the refraction state does not change the behaviour of the system.

**Figure 7.** Phase response curves for models with rectangular pulse.

The system is phase-sensitive, and that, depending on the phase, excitation may change the potential period length. Models with feedback are much less phase-sensitive, **Figure 7**. The feedback strongly modifies the properties of the modified van der Pol model.

The human heart rhythm with feedback about AV node is low-variable. The results of this study were consulted with the medical doctor from the Institute of Cardiology in Anin near Warsaw, who confirmed that the very stable heart rhythm is typical for patients with AV nodal re-entrant tachycardia.

One of the most important features of oscillators is synchronization. The behaviour of cardiac pacemaker cells resembles those relaxation oscillators. A characteristic property of relaxation oscillators is that they may be synchronized by an external signal, if the latter has a periodicity similar to spontaneous frequency of the oscillator [14]. We investigate a phenomenological model for the heartbeat which consists of two coupled van der Pol oscillators. The coupling between these oscillator (action potential) can be either unidirectional or bidirectional, with or without feedback. Below, we present our model with possible couplings and feedbacks and try to analyse all of them for behaviour of given system and its synchronization.

$$\begin{cases} \dot{\mathbf{x}}\_{1} = \mathbf{y}\_{1} - k\_{1} \left( \mathbf{x}\_{2} \left( t - \tau \right) - \mathbf{x}\_{1} \right) \\\\ \dot{\mathbf{y}}\_{1} = a\_{1} \left( \mathbf{x}\_{1}^{2} - 1 \right) \mathbf{y}\_{1} - f\_{1} \mathbf{x}\_{1} \left( \mathbf{x}\_{1} + d\_{1} \right) \left( \mathbf{x}\_{1} + \mathbf{e}\_{1} \right) + \mathbf{s}\_{1} \left( \mathbf{x}\_{2} - \mathbf{x}\_{1} \right) \\\\ \dot{\mathbf{x}}\_{2} = \mathbf{y}\_{2} - k\_{2} \left( \mathbf{x}\_{1} - \mathbf{x}\_{2} \left( t - \tau \right) \right) \\\\ \dot{\mathbf{y}}\_{2} = a\_{2} \left( \mathbf{x}\_{2}^{2} - 1 \right) \mathbf{y}\_{2} - f\_{2} \mathbf{x}\_{2} \left( \mathbf{x}\_{2} + d\_{2} \right) \left( \mathbf{x}\_{2} + \mathbf{e}\_{2} \right) + \mathbf{s}\_{2} \left( \mathbf{x}\_{1} - \mathbf{x}\_{2} \right) \end{cases} \tag{11}$$

If we consider physiological coupling between nodes, then the *s*2 coupling is introduced to our system. It means that the SA node directs AV node. For small value of *s*2 (in **Figure 8**, *s*<sup>2</sup> =1, *k*<sup>1</sup> =0, *s*<sup>1</sup> =0, *k*<sup>2</sup> =0) the result is similar to ref. [1], **Figure 8**.

**Figure 8.** Time series: physiological coupling between nodes, the parameters are given in Eq. (9).

**Figure 9.** Time series: physiological coupling for *s*<sup>2</sup> =5.

**Figure 7.** Phase response curves for models with rectangular pulse.

86 Numerical Simulation - From Brain Imaging to Turbulent Flows

2

&

&

ì ï ï ï í ï ï ïî

&

&

2

re-entrant tachycardia.

The system is phase-sensitive, and that, depending on the phase, excitation may change the potential period length. Models with feedback are much less phase-sensitive, **Figure 7**. The

The human heart rhythm with feedback about AV node is low-variable. The results of this study were consulted with the medical doctor from the Institute of Cardiology in Anin near Warsaw, who confirmed that the very stable heart rhythm is typical for patients with AV nodal

One of the most important features of oscillators is synchronization. The behaviour of cardiac pacemaker cells resembles those relaxation oscillators. A characteristic property of relaxation oscillators is that they may be synchronized by an external signal, if the latter has a periodicity similar to spontaneous frequency of the oscillator [14]. We investigate a phenomenological model for the heartbeat which consists of two coupled van der Pol oscillators. The coupling between these oscillator (action potential) can be either unidirectional or bidirectional, with or without feedback. Below, we present our model with possible couplings and feedbacks and

( ( ) )

t

1 sx x

( ( )) ( ) ( )( ) ( )

t

(11)

1 sx x

( ) ( )( ) ( )

feedback strongly modifies the properties of the modified van der Pol model.

try to analyse all of them for behaviour of given system and its synchronization.

1 1 12 1

*x y kxt x*

=- -- = - - + ++ - =- - - = - - + ++ -

1 1 1 1 11 1 1 1 1 1 2 1

2 2 2 2 22 2 2 2 2 2 1 2

2 2 21 2

*x y kx xt*

*y a x y fx x d x e*

*y a x y fx x d x e*

But for bigger value (e.g. 10 or 100), parts of amplitudes are synchronized in-phase with *x*<sup>1</sup> and aperiodic behaviour appears, which is presented in **Figures 9** and **10**.

**Figure 10.** Time series: physiological coupling for *s*<sup>2</sup> =100.

**Figure 11.** Projection of the phase space for *s*1=1.

This causes the attempt of fitting the AV frequency to SA which slows down the heart rate.

We study here the synchronization properties of such an association with respect to the nature and intensity of coupling.

**Figure 12.** Projection of the phase space for s1=10.

But for bigger value (e.g. 10 or 100), parts of amplitudes are synchronized in-phase with *x*<sup>1</sup> and

aperiodic behaviour appears, which is presented in **Figures 9** and **10**.

88 Numerical Simulation - From Brain Imaging to Turbulent Flows

**Figure 10.** Time series: physiological coupling for *s*<sup>2</sup> =100.

**Figure 11.** Projection of the phase space for *s*1=1.

**Figure 13.** Projection of the phase space for s1=1000.

In **Figures 11**–**13**, we can observe an influence of unidirectional coupling for strength of synchronization of given oscillators (Eq. (11)). We must remember that complete synchroni‐ zation of oscillators, characterized by *x*<sup>1</sup> (*t*)= *x*2(*t*) and *y*<sup>1</sup> (*t*)= *y*2(*t*), is unlikely to be observed in practice, due to the non-similarity of the coupled oscillators [21]. But oscillators, which are considered, are similar. In case when there is added unidirectional coupling to the system (*s*<sup>1</sup> ≠0 *or s*<sup>2</sup> ≠0), it is assumed that the coupling coefficient is a variable parameter in the analysed system. This parameter denotes the strength of the unidirectional coupling, all the remaining parameters being held constant. For small *s*<sup>1</sup> there is no synchronization, but for large *s*1 there is complete synchronization, which is shown in **Figure 13**. Similar results are obtained for *s*2. When bidirectional coupling occurs in the system (*s*<sup>1</sup> ≠0 *and s*<sup>2</sup> ≠0), it can also obtain complete synchronization but there must be big asymmetry of value of coupling coefficients: *s*1≪*s*<sup>2</sup> *or s*1≫*s*2.

Finally, the influence of the bidirectional coupling for behaviour of the model was studied. It would seem that the bidirectional coupling should operate so that when we enter it in relation 2:1 between oscillators, the rhythm of the fast pathway should be slowed down, while the rhythm of the fast pathway should accelerate.

In the model without feedback, periodic potential *x*1 and *x*<sup>2</sup> are obtained, which is a dynamics of a physiological rhythm. But there is no synchronization for these parameters. When coupling coefficients are added to the system: *k*<sup>1</sup> ≠0 *and k*<sup>2</sup> ≠0, we obtain a set of projections of the phase space for different values of couplings, **Figures 14**–**16**.

**Figure 14.** Projection of the phase space for k1=2 and k2=1.

Numerical Simulations of Dynamics Behaviour of the Action Potential of the Human Heart's Conduction System http://dx.doi.org/10.5772/63017 91

**Figure 15.** Projection of the phase space for k1=5 and k2=10..

In **Figures 11**–**13**, we can observe an influence of unidirectional coupling for strength of synchronization of given oscillators (Eq. (11)). We must remember that complete synchroni‐

(*t*)= *x*2(*t*) and *y*<sup>1</sup>

practice, due to the non-similarity of the coupled oscillators [21]. But oscillators, which are considered, are similar. In case when there is added unidirectional coupling to the system (*s*<sup>1</sup> ≠0 *or s*<sup>2</sup> ≠0), it is assumed that the coupling coefficient is a variable parameter in the analysed system. This parameter denotes the strength of the unidirectional coupling, all the remaining parameters being held constant. For small *s*<sup>1</sup> there is no synchronization, but for large *s*1 there is complete synchronization, which is shown in **Figure 13**. Similar results are obtained for *s*2. When bidirectional coupling occurs in the system (*s*<sup>1</sup> ≠0 *and s*<sup>2</sup> ≠0), it can also obtain complete synchronization but there must be big asymmetry of value of coupling coefficients:

Finally, the influence of the bidirectional coupling for behaviour of the model was studied. It would seem that the bidirectional coupling should operate so that when we enter it in relation 2:1 between oscillators, the rhythm of the fast pathway should be slowed down, while the

In the model without feedback, periodic potential *x*1 and *x*<sup>2</sup> are obtained, which is a dynamics of a physiological rhythm. But there is no synchronization for these parameters. When coupling coefficients are added to the system: *k*<sup>1</sup> ≠0 *and k*<sup>2</sup> ≠0, we obtain a set of projections of

(*t*)= *y*2(*t*), is unlikely to be observed in

zation of oscillators, characterized by *x*<sup>1</sup>

90 Numerical Simulation - From Brain Imaging to Turbulent Flows

rhythm of the fast pathway should accelerate.

**Figure 14.** Projection of the phase space for k1=2 and k2=1.

the phase space for different values of couplings, **Figures 14**–**16**.

*s*1≪*s*<sup>2</sup> *or s*1≫*s*2.

**Figure 16.** Projection of the phase space for k1=10 and k2=10.

However, with bigger values of *k*1, *k*2, a partial synchronization can be found in the system, **Figure 15**. Further increasing the value of the coefficients causes, however, the most significant changes in the portrait of the phase space. We can observe partial synchronization, but this synchronization has asymmetric character, **Figure 16**.

We investigated the synchronization between oscillators themselves despite that the genera‐ tion of cardiac dysrhythmias are associated with a lack of synchronization between autono‐ mous pacemakers [22]. But we showed that we can obtain certain types of synchronization: phase synchronization, partial synchronization or lag synchronization. This depends on the strength of coupling coefficients. Corresponding large coefficient may cause synchronization, but often setpoints are not the values of the physiological range.
