A low power VLSI implementation of the Izhikevich neuron model

Demirkol, Ahmet Şamil; Özoğuz, Serdar

doi:10.1109/newcas.2011.5981282

Cited by 18 publications

(15 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We also present a neuromorphic model of the artificial neuron developed using SPICE and Verilog-A, which reproduces the neuronal persistent firing behavior by integrating somatic Considering there are many existing VLSI implementations of Izhikevich neuron model and its variants [20]- [25], the silicon implementation of the proposed artificial neuron will be straightforward by incorporating the axonal computation unit into the Izhikevich VLSI designs. The persistent firing activity is found in a diverse set of brain regions and organisms and serval in vitro systems, it may represent a very general and fundamental form of brain dynamics [10].…”

Section: Discussionmentioning

confidence: 99%

“…Due to the computational simplicity of our model, it is rather straightforward to implement the proposed neuron model in hardware, either in digital circuits or analog circuits. There have been several circuit implementations of Izhikevich neuron model and its variants [20]- [25] . Thus in our case it is intuitive to add a leaky integrator emulating the axon, as well as the switching devices for selecting one of two sets of a, b, c, d, e parameters, which may be stored in memory devices, e.g.…”

Section: Neuromorphic Model Of the Artificial Neuronmentioning

confidence: 99%

See 1 more Smart Citation

Artificial neuron with somatic and axonal computation units: Mathematical and neuromorphic models of persistent firing neurons

Ning¹,

Huang²,

Shi³

2012

The 2012 International Joint Conference on Neural Networks (IJCNN)

View full text Add to dashboard Cite

Section: Discussionmentioning

confidence: 99%

Section: Neuromorphic Model Of the Artificial Neuronmentioning

confidence: 99%

Artificial neuron with somatic and axonal computation units: Mathematical and neuromorphic models of persistent firing neurons

Ning¹,

Huang²,

Shi³

2012

The 2012 International Joint Conference on Neural Networks (IJCNN)

View full text Add to dashboard Cite

“…In contrast to simple models such as leaky integrate-and-fire type neurons, the Izhikevich neuron is more biologically realistic with voltage reset occurring at the peak of the spike (as opposed to occurring at threshold) and an instantaneous reset of the membrane potential when the membrane voltage reaches a cutoff voltage (Equation 2). Additionally, Izhikevich neurons can be efficiently implemented in hardware [61]. Table 1 reports the parameters used for simulation.…”

Section: Neuron and Synapse Modelmentioning

confidence: 99%

Unsupervised heart-rate estimation in wearables with Liquid states and a probabilistic readout

et al. 2018

View full text Add to dashboard Cite

Heart-rate estimation is a fundamental feature of modern wearable devices. In this paper we propose a machine intelligent approach for heart-rate estimation from electrocardiogram (ECG) data collected using wearable devices. The novelty of our approach lies in (1) encoding spatio-temporal properties of ECG signals directly into spike train and using this to excite recurrently connected spiking neurons in a Liquid State Machine computation model; (2) a novel learning algorithm; and (3) an intelligently designed unsupervised readout based on Fuzzy c-Means clustering of spike responses from a subset of neurons (Liquid states), selected using particle swarm optimization.Our approach differs from existing works by learning directly from ECG signals (allowing personalization), without requiring costly data annotations. Additionally, our approach can be easily implemented on state-of-the-art spiking-based neuromorphic systems, offering high accuracy, yet significantly low energy footprint, leading to an extended battery life of wearable devices. We validated our approach with CARLsim, a GPU accelerated spiking neural network simulator modeling Izhikevich spiking neurons with Spike Timing Dependent Plasticity (STDP) and homeostatic scaling. A range of subjects are considered from in-house clinical trials and public ECG databases.Results show high accuracy and low energy footprint in heart-rate estimation across * Corresponding author

show abstract

“…There is tremendous growth in the study and development of fractional-order neural networks as far as mathematical modelling and numerical simulations are concerned [3][4][5][6][7][8][9][10][11]. Knowing the importance in the development of fractional-order neural networks, various researchers have attempted the circuit implementation of integer as well as fractional-order neuronal models [16][17][18][19][20][21][22][23][24][25][26]. The electronic implementation of neuron models can be used to understand the behaviour of biological systems.…”

Section: Introductionmentioning

confidence: 99%

0.65 V integrable electronic realisation of integer‐ and fractional‐order Hindmarsh–Rose neuron model using companding technique

Khanday

Dar

Kant

et al. 2018

IET Circuits, Devices & Systems

View full text Add to dashboard Cite

Some neurons like neocortical pyramidal neurons adapt with multiple timescales , which is consistent with fractionalorder differentiation. The fractional-order neuron models are therefore believed to portray the firing rate of neurons more accurately than their integer-order models. It has been studied that as the fractional order of differentiator and integrator involved in the neuron model decreases, bursting frequency of the neurons increases. The opposite effect has been observed on increasing the external excitation. In this study, integer-and fractional-order Hindmarsh-Rose (HR) neuron models have been implemented using sinh companding technique. Besides, the application of the HR neuron model in a simple network of two neurons has also been considered. The designs offer a low-voltage and low-power implementation along with the electronic tunability of the performance characteristics. Due to the use of only metal-oxide semiconductor (MOS) transistors and grounded capacitors, the proposed implementation can be integrated in chip form. On comparing with existing implementations, the implemented fractional-order and integer-order models show a better performance in terms of power consumption, supply voltage, order and flexibility. The performance of the circuits has been verified using 130 nm complementary MOS (CMOS) technology process provided by Austrian Micro Systems using HSPICE simulation software.

show abstract

A low power VLSI implementation of the Izhikevich neuron model

Cited by 18 publications

References 19 publications

Artificial neuron with somatic and axonal computation units: Mathematical and neuromorphic models of persistent firing neurons

Artificial neuron with somatic and axonal computation units: Mathematical and neuromorphic models of persistent firing neurons

Unsupervised heart-rate estimation in wearables with Liquid states and a probabilistic readout

0.65 V integrable electronic realisation of integer‐ and fractional‐order Hindmarsh–Rose neuron model using companding technique

Contact Info

Product

Resources

About