Impact of neuron with two dendrites

in heart behavior

Abstract:
Neurons are the fondemental units of the brain and nervous system. The variable structure model of
neurons is a system of differential equations with variables parameters. Valorizing these parameters
produces a unique model that describes the dynamique behavior of a single neuron. A neural networks
based on neurons with multi dendrites employing activation function with variable structure is introduced.
No prior work, however, has discovered the impact of dendrites position on heavies’ information to
propagate. Can neuron with two dendrites be used as a new tools to model, identify and generate a
dynamic heart behavior?
In this paper, we have discovered a model of heart behavior.
Keywords: Variable Structure Model of Neuron (VSMN), Neural networks, neuron, dendrites, heart
behavior, (ECG) Electrocardiogram.

1. Introduction
The humain brain contains around 100 billion neurons and these neurons have approximately 100
trillion connections each.

A neuron’s fundamental structure consists of a cell body, an axon and dendrites.

Dendrites receive incoming signal impulses from other neurons, between a neuron’s axon and another
neuron’s dendrite there is what we call a synapse, and that’s where the signal transmission happens [11]
Activation functions have a large impact on the performance of artificial neural networks (ANNs). The
design of of ANNs simulates the structure of neural networks of the brain’s central nervous system to a
certain extent,.
An ANNs consists of a number of connected neurons, each act as a parallel distributed processor, which
can perform large-scale computations for data processing and classification[12].
Our VSMNs can also be combined to create dendrite based neural networks which generated promissing
results in different studies[1 and 2].
 In this work, we can present the behavior of heart throught the use of our VSMN.
The electrocardiogram (ECG) is a non-stationary physiological signal that represents a recording of the
electrical activity of the heart. After decades, many artificial neural networks have been proposed to
develop tools that may lead to a better understanding of the behavior of the heart. The ECG has many
applications in various fields, namely medicine, internet of things, cryptography, wearable sensors,
biometric recognition and more. In our approach, we use a neuron with two dendrites inspired from
Bouallegue neural networks [1 and 2].

2. Heart behavior

2.1 Potential application in cardiovascular:

An electrocardiogram (ECG) is a test that detects and records the strength and timing of the
electrical activity in your heart.
The ECG signal consists of P, Q, R, S and T waves as shown in figure 1.

Figure1: Standard ECG Waveform

• This electrical signal is recorded as the P wave.

• The PR Interval is the time, in seconds,from the beginning of the P wave to the beginning of the
QRS interval.
• The PR segment represents the electrical conduction through the atria and the delay of the
electrical impulse in the atrioventricular node.
• The signal travels across the heartˆas ventricles causing them to contract.This signal is recorded
as the QRS waves on the ECG.
• The ventricles then recover to their normal electrical state, shown as the T wave.
 • The ST segment connects the QRS Interval and the T wave and represents the beginning of the
electrical recovery of the ventricles.
• The QT interval represents the time during which the ventricles are stimulated and recover after
the stimulation.

3. Equations:

3.1 Model of neurons with one dendrites [1]:

The variable structure model of neurons with one dendrite is a system of differential equations with
variable parameters.
This model allows the neuron to have any number of dendrites, with variable activation function.
The following equation system describes the dynamics of the neuron with one dendrite [3]:

u̇ = − u /τ + (u + p1) f (βv) f (λ (u + p1)

v̇ = − αv + χ α (u + q1) n1 f2 (λ (u + p1))
(1)
With f (t) = exp – (t2/2) and:
• α, β and λ are positive reels.
• τis constant of time.
• The sign of χ mark the polarity of activation potential. (χ = ±1)
• The value of n proper the behavior of neuron.
• The value of p the position of spike
• The value of q the degradation of spike.

This system can be represented by a graphical symbol as shown by figure 2.

NBDn
ሺp,qሻ

Figure2: Graph of a neuron with one dendrite

Where:
• N: is abbreviation for Neuron
• D=k=the neuron’s polarity
 • n,p,q are the parameters from the system
• B= the behavior of the neuron (stable, hole, degraded, oscillator, harmonic or chaos)

Figure 3 shows a neuron with one dendrite.

Figure3: Neuron with one position of dendrites

Figure 4 displays different potentiel functions generated by neuron with one dendrite.

Figure4: Behavior of neuron with one dendrite

Figure 5 shows a polarity of neuron with one dendrite.

Figure 5: Neuron polarities

3.2 Model of neurons with two dendrites [1]:

The following equation system describes the dynamics of the neuron with two dendrites:

u̇ = − (u) /τ + (u + p1) (u + p2) f (βv)f (λ (u + p1) (u +p2))

v̇ = − αv + χ (u + q1) n1(u + q2) n2α f2 (λ (u + p1) (u +p2))
(2)

Figure 6 shows a neuron model with two dendrites.

Figure6: Neuron with two positions of dendrites

Figure 7 illustrates the results of different values of n1 and n2 .

 Figure7: Behavior of neuron with two dendrites

4. Implementation and Results:

In this section, we present our successful simulation of the heart’s ECG diagram using our VSMN.
The behavior of the heart generates by neural network build with neurons on cascade:
- The first neuron (N1) with one dendrite generates behavior of the P region,
- the second neuron (N2) wich contains two dendrites generates behavior of QRS region
- and the third neuron ( N3)with one dendrite generates the T region.

The model of neural network becomes as follows:

u̇ = − u /τ + (u + p1) f (βv) f (λ (u + p1)
N1 v̇ = − αv + χ α (u + q1) n1 f2 (λ (u + p1))

u̇ = − (u) /τ + (u + p1) (u + p2)f (βv)f (λ (u + p1) (u + p2))

N2 v̇ = − αv + χ (u + q1) n1(u + q2) n2α f2 (λ (u + p1) (u + p2))

u̇ = − u /τ + (u + p1) f (βv) f (λ (u + p1)

N3 v̇ = − αv + χ α (u + q1) n1 f2 (λ (u + p1))
(3)
Figure 8 shows the heart behavior simulation. This result is an accord with real ECG.

Figure8: Heart behavior Simulation

Figure 9 and 10 shows degraded behavior of ECG. Many disease that affects the heart’s behavior.
This simulation present different perturbations to our neuron’s models.
The variable structure model of neurons, in its different states of one and two dendrites, is showing
promising results. We showed its capacity to produce healthy as disease-affected ECG diagrams.

Figure9: Q Wave Degradation

 Figure10: S Wave Degradation

5. Conclusion:

In this paper, we proposed a new theoretical approach to generate an artificial ECG signals using a
set of neurons with two dendrites, we have discovered model of dynamic heart behavior.
The result show that this work could have a strong impact on the welfare of humanity and can lead to
a care heart disease.

Acknowledgements

Financial support from Electronics and Microelectronics laboratory, Faculty of Science,University of

Monastir,Tunisia

References:

[1] Ghaith Bouallegue, Lotfi Hamrouni, and Kais Bouallegue. “Neuron Model With Two Dendrites and
its Behaviors”. In: 17th International MultiConference on Systems, Signals & Devices (SSD). IEEE.
2020, pp. 597–603.
[2] Kais Bouallegue. “A new class of neural networks and its applications”. In: Neurocomputing 249,
2017, pp. 28–47.
[3] Salma Ben Mamia, William Puech, Kais Bouallegue, “Generation of chaotic attractors using neurons
with multidentrites, “International Journal of Modelling Identification and Control 40(1):92,
[4] Kais Bouallegue. “Chaotic attractors with separated scrolls”. In: Chaos: An Interdisciplinary
Journal of Nonlinear Science 25.7 ,2015, p. 073108.
[5] Kais Bouallegue. “Gallery of chaotic attractors generated by fractal network”. In: International
Journal of Bifurcation and Chaos 25.01, 2015, p. 1530002.
[6] Kais Bouallegue, Abdessattar Chaari, and Ahmed Toumi. “Multi-scroll and multi-wing chaotic
attractor generated with Julia process fractal”. In: Chaos, solitons & fractals 44.1-3 ,2011, pp. 79–85.
[7] Qingdu, Lia & ,QING05 Qingdu, Lia. Xiao- Song,& Yangb. Fangyan, Yanga.” Hyperchaos in
Hopfield-type neural networks,” Int. J. Neuro computing 67, 2005, pp.275-280
[8] Abdon Atangana,Ghaith Bouallegue, Kais Bouallegue
New multi-scroll attractors obtained via Julia set mapping
Chaos, Solitons and Fractals 134 (2020) 109722.
[9] G. Lee and N.H. Farhat(2001)]GLE01G. Lee and N.H. Farhat.”The Bifurcating Neuron
Network,”Neural Networks, 2001, 14, pp.115-131.
[10] Albert Gidon, Timothy Adam Zolnik, Pawel Fidzniski, Felix Bolduan, Athanasia Papousti,
Panayiota Poirazi, Martin Holtkamp, Imre Vida, and Matthew Evan Larkum. Dendritic action
potentials and computation in human layer 2/3 cortical neurons, SCIENCE, Vol.367,
Issue6473,2020,pp.83-87..
[11] Zheng Tang, Hiroki Tamura, Makoto Kuratu, Okihiko Ishizuka, and Koichi Tanno. A model of the
neuron based on dendrite mechanisms, Electronics and Communications in Japan. Vol.84, Issue 8, 2001,
pp..11-24
[12] Jayvadan Patel and Anita Patel. Artificial neural networking in controlled drug delivery, in Artificial
Neural Network for Drug Design, Delivery and Disposition, 2016, pp.195-218.