This full-text paper was peer-reviewed and accepted to be presented at the IEEE ICCSP 2015 conference.
A Review on Methods, Issues and Challenges in
Neuromorphic Engineering
Mohammed Riyaz Ahmed, Member IEEE, and B.K. Sujatha, Member, IEEE
Abstract- Biological systems have many key answers for our
current limitations in scaling. Miniaturization and more speed
were the driving forces for VLSI technology in past few decades.
As we are reaching the dead end of Moore's law, now the
paradigm has shifted towards intelligent machines. Many efforts
are made to mimic the commonsense observed in animals.
Automation and smart devices have taken a great momentum in
recent years, but to imitate human intelligence in machines with
existing hardware is not viable. By only developing complex
algorithms and implementing software, intelligence is not
accomplishable. This paper brings out the very basic differences
between brain and computers. The complexity of human brain
and flaws of current computing architecture is discussed.
Neuromorphic Engineering emerges as a realistic solution whose
architecture and circuit components resemble to their biological
counterparts. This paper acts as a primer for Neuromorphic
engineering. Brain functionality being analogous, the limitation
of only digital systems is addressed by the mixed mode operation
of ICs. Modelling of Neuron, various software and hardware
available to realize these morphed architectures are studied. The
gap between the software simulation and Hardware emulation,
FPGA and VLSI implementation is debated. To reach to a larger
audience the paper exposes many limitations of Neuromorphic
engineering and come out with many open research issues.
Index Terms- Brain simulation, Cognition, FPGA vs VLSI,
Learning and Memory, Mixed mode signal processing,
Neuroscience, Neuromorphic VLSI Chips.
I. INTRODUCTION
ONFUCIUS [1] once said, " I hear and I forget, I see and
I remember, I do and I understand", to understand the
� and human is cognition. This cognition is integration of all
�plex functionality of brain, we need to build one. This
goal has been driving many neuroscientists, engineers,
psychologists who have come together in a single umbrella of
Neuromorphic Engineering. Neuromorphic Engineering is an
interdisciplinary domain which takes inspiration from
neuroscience and tries to emulate the functionality, especially
brain. Miniaturized electronic devices built on the brain's
architecture could answer many neural disorders. We can
Mohammed Riyaz Ahmed is a research Scholar in School of Electronics,
JAIN University, Bangalore, Karnataka, 560004 INDIA. He is currently
working as Assistant Professor in school of Electronics and Communication,
REVA University, Rukmini Knowledge Park, Bangalore, Karnataka, 560064
INDIA. (E-mail: riyaz@reva.edu.in).
B.K. Sujatha is with the Department of Telecommunication and
Engineering, M S Ramaiah Institute of Technology and Management,
Bangalore,Karnataka,560054 INDIA (e-mail:bksujatha@msrit.edu).
978-1-4 799-8081-9/15/$3l.00 © 2015 IEEE
0899
understand the brain malfunction and implant artificial vision
systems, auditory systems as in [2-5] etc... , to restore vision,
audibility and many other cognitive behaviour. Study of brain
will give us insights about the structure and functioning of it
which will help us to understand and imitate human behaviour
in intelligent systems.
Though our computers are faster than individual neurons,
still they are no match to a monkey when it comes to skills like
pattern recognition and visual processing. The biggest gap is
architecture on which our modern day computing devices are
built. Section II deals with the details of biological system and
things to be implemented in engineering computing devices.
The neurons communicate between themselves through spikes,
which are membrane potentials and asynchronous. This is a
characteristic of random analog signals, whereas such analog
communication is not possible with synchronous digital
systems. The section II also deals with mixed mode signal
processing in ICs.
Brain modelling on ICs can begin from modelling a
neuron, observing the neuronal dynamics, structure, and
communication between neurons. Later we can proceed to the
networks of neurons. Section III brings out a brief review of
Morphological implementations at all levels i.e.,
component/device level, circuit level, storage and access level
and at last system level. This part of the paper describes neuron
modelling where we take some lessons from biology and
implement in VLSI Chips.
For a given set of input, output is generated. And if
different outputs are generated for same set of input then the
system is adapting itself by using feedback. This adaptation is
decided by learning. The main difference between machines
human behaviour such as attention, perception, learning,
recognition, classification etc. The learning exhibited by human
brain can be modelled in many different ways. Section IV deals
with cognitive implementation with an emphasis on
applications such as vision system, ear, nose and Brain.
Majority of research done and proved is on the basis of
detailed simulation of biophysics, but not physical emulation of
biological neural computation [6]. Simulation results and
hardware results doesn't resemble much. The huge gap and
differences between simulation and hardware implementation
is brought out in section V.
Section VI brings out the recent achievements, advantages,
limitations, and open research issues in the emerging field of
Neuromorphic Engineering. This work acts as a primer for
Neuromorphic Engineering aspirants.
Finally paper ends with concluding remarks and Future
scope in the promising domain of Neuromorphic Engineering.
 This full-text paper was peer-reviewed and accepted to be presented at the IEEE ICCSP 2015 conference.

II. DISSECTING BRAIN AND COMPUTERS B. Information and Signal Processing

A. Complexity of Brain Implementing neural networks on VLSI was the inception
There are 2 best computing devices on earth: Brain and for the debate between Analog and digital circuits. The real
Computers. We know everything about one and little about world signal being analog, the ears and eyes are processing the
another [7]. We are in a process to realize something whose incoming signal directly without the need of digital
functionality and working is not clear on something which we conversion. Thousands of Analog cells can work in parallel,
have built earlier. The good news is we are a bit successful so leading to high number of operations per second. Analog
far, but the tragedy is the journey has just begun. computation treats transistors as complicated, non-liner
The biggest question in simulating a brain is to simulate devices with many physical considerations, which comply
which part of the brain? And to what degree of fidelity? When with neuro dynamics. Whereas digital computing considers
do we feel "Mission Accomplished"? Are we going to transistors only as a switch.
challenge the millions of year's evolution with 50 years old Analog systems are more efficient than digital if precision is
computers? traded-off. But keen observations suggest that neurons can be
If we take computer analogy of human brain, brain would be perceived as AID Converters. As a neuron collects all
hardware, mind will be software, computations being the incoming spikes and based on the threshold it decides whether
algorithms and memory would be stored data. A child a spike has to be transmitted (logic 1) or not (logic 0). Analog
perceives the data in childhood, develop database by the signals require more circuitry hence more power. Our goal is
Thought and Feeling process, develop mental representations to achieve Low Power consumption. This is where Digital
as data structures and is programmed to work accordingly in systems come into picture.
future. The thing which differentiates child brain and Digital systems are more power efficient, precised, flexible,
computer is, human brain is capable of choosing the best and most importantly immune to noise. They are robust
possible reward in a new, unfamiliar and strange environment systems and doesn't change with temperature, power supply
without being trained. The basic level cognitive tasks of fluctuations and to variations in transistor behaviour.
Perception, rational, learning and memory, information The question is should we go with analog systems or digital
processing and retrieval are the areas to work for. one. One may take inspiration from biology where initial low
Brain has different parts which are specialized in different power analog processing at sensory organs is followed by
aspects, as faculty in a university. The occipital lobe processes digital transmission of information towards brain. Signal
visual information likewise temporal lobe processes auditory processing on VLSI systems is an emerging field, where much
information. Therefore the first thing we have to keep in mind of the attention is given for Mixed Mode signaling.
is Application Oriented design is preferred over general
purpose. Second thing is we can achieve massive parallelism TABLE I
and do high speed computation, but the concern here is low BRAIN VS COMPUTER

power, high fan-out capacity. Therefore "speed is not the

Comparison table
need". Slono
Functionality Brain Computer
Super computers in [3] are 1000 times heavier, 10,000 times
Perception Sense organs Sensors
bulkier and consumes millions of watts as compared to a
brain. The challenge is to make neuromorphic devices 2 Data Spikes Current
compatible, implantable and low power. One of the biophysics
3 Basic component Neuron Transistor
lesson from highly specialized structures, such as ears and
eyes is that the initial computation done by them is in analog 4 Signal processing Analogous/mixed Digital

domain. Most of the processing such as filtering, spectrum 5 Transmission Synaptic activity Wires
analysis and signal compression is done before sending to
6 Hardware Brain Processor
brain. The inner ear uses just 14 microwatts and could run for
15 years [8]. The challenge is to meet this lowest power 7 Software Wetware/ Mind Operating system
consumption.
8 Memory hippocampus Gate capacitance
Despite the fact that, brain doesn't execute coded
instructions, instead uses spikes for communication between 9 Storage Monolithic Modular

synapses, collectively brain seems to be faster. A staggering 10 Logic Fuzzy Digital 0/1
amount of 10 quadrillion (1016) synaptic activity per second.
11 Connectivity High Poorly connected
The challenge is to achieve such a high connectivity and high
fan-out capability. 12 Fan-out High Very low
Human brain is fault tolerant, and is only vulnerable during Individually slow Individually fast
13 Speed
accidents and aging. It is said that von Neumann architecture Collectively fast collectively slow
Power
mimics left brain, and we are in pursuit of Right part of brain 14 Low High
consumption
which is responsible for being fault tolerant, reconfigurable
15 Reliabilty Redundant Fault-sensitive
and event driven. Learning is exhibited by this part of the
brain [9-11]. A comparison of brain's equivalent counterpart 16 Cognition Exhibited Yet to exhibit
of computers are shown in Table I.

0900
 This full-text paper was peer-reviewed and accepted to be presented at the IEEE ICCSP 2015 conference.

III. IMPLEMENTATlON LEVELS OF BRAIN MODELS which it may be expensive, and the models become too
complex. In Bottom-up approach as shown in Fig. 1, we
Modelling is a reductionist approach. Models reduce
generalize one model of neuron and climb up to behavior level
complexity and provide simplified representation of real
where most of the time the models fail to replicate the
systems. All neuronal dynamics is not yet understood, hence
biological counterpart. This ambiguity whether to choose top
creating an exact replica is not under our grasp. Abstraction is
down approach (complex biological implementation, Biology
done at every level to make our life simple.
has the upper hand) or Bottom-up approach (abstract level
Brain being the most complex system, its implementation
implementations, existing engineering technology has the
can be studied in a hierarchy in Table II. The interactions of
upper hand) leads us to the Valley of Death.
different neuronal components give rise to behaviour. Since
behaviour is a collective function of different neural
TABLE IT
components, its abstraction is placed above system level.
LEVELS OF IMPLEMENTATION
Protein/Genetic level describes the genes structure. It's a
neglected/not yet explored field in Neuromorphic Engineering. Neuromorphic Correspondence for Implementation Hierarchy
At Membrane Level, electrical and ion exchange happens. levels
Hierarchy Neuroscience Electrical science
Transistor is often abstracted as switch: ON or OFF. But the
7 Behavior Level Mind Architecture
V-I characteristics shows that current flowing is smooth and
steep function of applied Voltage. Transistors work in sub 6 System Level Brain system Macro Block
threshold region where the V-I characteristics resemble the Local Neuronal
5 Circuit Level Block/Cell
current voltage relationships in molecular structures on surface population
Component
of brain cells. Hence membrane level abstractions and 4 Single Neuron Perceptron
Level
implementations are done at transistor level. CMOS/
3 Device Level Synapses
Intra-neuron communication happens through synapses. Memristors
Synaptic plasticity is the reason for the emergent animal Membrrane
2 Channel Ions Transistor
Level
behaviors: adaptation and Learning. In recent years the
learning process has dragged more attention. Selective Protein! Genetic Genes

attention is the main functionality of this abstraction level.

The synaptic plasticity can be implemented using CMOS The two approaches shown in Fig. 1, are in fact 2 faces of
circuits. Memristors are an alternative, which also has same coin, no matter which approach we inculcate, the million
answers to memory modelling. Large scale memristive fabric dollar question is : "Is the machine Intelligent ?" finding the
is yet to be realized. Selective attention, Efficient learning and right trade off and compromises to make it more real (to
memory modelling is the biggest open research issue at humans) and realizable is the open research issue.
synaptic level.
Perceptron's implementation in 1950's was the pioneering
work, which was foundation for implementation of Artificial
3;:
Neural Networks on hardware. Perceptron is computational 0
@ Behaviour Level
model of Neuron. Hodgkin-Huxley and Morris-Lecar [12, 13] ::!
(1)

models are conductance based which has high biological !:l

5' iO-.
precision but comes with huge cost. Another type (Type II) are ::! System Level ::J
(1) oJ)
t:
spike based such as Integrate and fire model which describe .t:
3
'" ;;
.., () 0
temporal behaviour of spikes which are earliest and simplest 0 ,, ' Circuit Level E
"II
()
::! ""
models [14] . .,
t:
= (1)

Spikes are used in nervous system for information �

�
Vl

Component Level ;:;

� ,.
transmission. Intra neuron Communication is important aspect �
(;"
at system level. Address event representation protocol -<
(1) 0;
F Device Level .,
proposed in 1991 to mimic information coding of brain. AER (Jq Vl

0 t:
assigns a fixed address to every neuron, by using which S'
0
(Jq 'B
.,
neurons continuously update their central system about their 0.- Membrane Level
0 tl
excitation levels. This updated info is sent to upper/higher � �
oJ)
layers. AER is a communication protocol for spiking neurons t:
Genetic Level :Q
between different layers [15-18].This field has attracted a huge .,
f-;
community of researchers who are engineering various
protocols for inter and intra chip communication. AER scheme Fig, 1. Implementation Approaches

resembles to the Internet Protocol (IP) addressing in computer

networking where information is routed to individual host IV. COGNITION
corresponding to the IP address. A. Brain Cognition
Selecting the appropriate level of abstraction is very The goal of Neuromorphic engineering is not only to mimic
important [2]. We have choice of Top-down approach, where the architecture of biological nervous system but also to
we arrive to the neuron model keeping behaviour in our mind. exploit properties of biological systems and mimic the human
Here we intend to reflect all biological components due to behaviour. By the existing systems we are unable to

0901
 This full-text paper was peer-reviewed and accepted to be presented at the IEEE ICCSP 2015 conference.

implement simple tasks performed by biological systems. It is possible to simulate these circuits in software. [31-42].
Neuromorphic engineering applies computational cognitive The high computational load can be taken care by parallel
Neuroscience principles discovered in living organisms to processing, but the real time environment needed for
implement such tasks in Machines using CMOS VLSI experimentation is not in the grasp of simulation.
Technology. FPGA has been popular in industry, education and research.
What's cognitive about Neuromorphic engineering? When The reasons for their overwhelming popularity is the
do we say neuromorphic systems are cognitive? What are the flexibility for fast prototyping for implementation of digital
different cognitive tasks of our interest? Neuromorphic designs without altering the hardware. Chip design demands
engineering bridges the gap between existing engineering and large waiting period and puts a hole in our pocket. FPGA is
Cognitive Neuroscience. Following are the humble attempts of continuously evolving and are available for a reasonable price.
it towards filling the gap including Neurolinguistics, vision Xilinx is one of the manufacturer. Achronix semiconductors
and auditory systems, low power devices, locomotion, are providing new class of FPGAs with speed up to 1.5 GHz.
implantable chips(biomedical), which exhibit cognitive The key change is the internal logic is asynchronous [43]. A
behaviour such as information processing, pattern direct transition from Simulink to FPGA circuit synthesis is
classification, learning and memory, decision making etc . . . possible as in [44]. To exploit the analog computation i.e.
B. Achievements ofNeuromorphic Engineering Simulink to Analog hardware conversion is possible, now the
1) Vision system: This is the most explored cognitive tasks arrays are FPAA [45, 46].
which has much interested audience not only from More often Neuromorphic engineers rely on garden-variety
medicine but also robotics and automobile field. This huge VLSI, CMOS technology to experiment as in [47] there is
vested interest is because they have immediate much resemblance between VLSI hardware and neural
applications in sensors and sensory systems. The earliest wetware.
vision systems was proposed by Mahowald and Mead
[19]. Neuromorphic engineers have made significant
VI. CONCLUSION
efforts in copying the functionality of retina by designing
analog silicon circuits which has led to wearable and It seems that we have taken a big leap, but much remains to
implantable chips [2, 3, 6, 20-24]. be achieved. Any Engineer who want to jump into research of
2) Auditory system: In 1997 researchers at Loughborough Brain Computer Interaction, Intelligent machines,
University designed a VHDL-based pitch detection Neuromorphic engineering has to know many trade-offs. This
system, and many implementations of bionic ears have paper has done a brief review of various tradeoffs to be
been carried out in [5, 6, 25]. considered during Hardware implementation of Bio-inspired
3) Microelectronic Nose: Mice and Bees demonstrate Computing Architectures. Levels of Implementation is an
powerful and efficient odor discrimination capabilities. In important contribution of this paper. Implementation
[26], a microcontroller based electronic nose is explained. approaches provide a clear idea for a researcher to start at
This is implemented using time domain encoding schemes what level and what to expect from the upper layer and what
to achieve gas identification. This pioneers translation of
services are necessary to the lower layer. This paper brings out
neurophysiological phenomena such as simple recognition,
many open research issues pertaining to morphed circuits.
easy calibration, and training into hardware. The power
Some milestones achieved are mentioned, and the things yet to
consumption by sensor is of concern.
be achieved are discussed. This has also discussed the gap
4) Brain: In [27] [28] the most ambitious project of "Human
between the software simulation results and Hardware
Brain Project" funded by European countries is aiming
emulation results.
towards many aspects of Neuromorphic Engineering such
Future work would be focusing on Learning and memory
as: Neuroinformatics, Brain simulation, High speed
computing, medical informatics, neuromorphic computing issues in implementation on Neuromorphic VLSI Chips.
and neurorobotics. SYNAPSE of DARPA [29, 30] aims to Reinforcement learning would be the focused field in
rebuild brain. Learning.
e. Open Research Issues inNeuromorphic Engineering
The modern day simulations are slow, we not only have to ACKNOWLEDGMENT

implement behaviour also need to control them. Intelligent
behaviour is about adaptation to any given circumstances with
maximum rewards in long term. Humans are active receivers
of information i.e. perception is not always data driven, it is
Knowledge driven too.
The authors would like to thank MSRIT, REVA University
and JAIN University for providing necessary support and
Infrastructure to carry out research work.
REFERENCES

[I] Confucius. (n.d.). BrainyQuote.com. Retrieved January 21, from

V. IMPLEMENTAnON ISSUES BrainyQuote.com Web site:
http://www.brainyquote.com/quotes/quotes/c/confuciusI36802.html.
Hardware is more near to real time application but too
[2] 1. V. Arthur and K. Boahen, "Silicon-Neuron Design: A Dynamical
tough for carrying research and implementing. Software has Systems Approach," iEEE Trans Circuits Sysli Regul Pap, vol. 58, pp.
many assumptions and there is always a large difference 1034-1043,2011.
between hardware and software simulation results. Here we [3] K. Boahen, "Neuromorphic Microchips," Sci Am, vol. 292, pp. 56-63,
May 2005.
compare Hardware and Software implementations, also FPGA
[4] T. Delbruck, "Silicon retina with correlation-based, velocity-tuned
and VLSI implementations. pixels," iEEE Trans Neural Netw, vol. 4,pp. 529-41, 1993.

0902
 This full-text paper was peer-reviewed and accepted to be presented at the IEEE ICCSP 2015 conference.
[5] S. Jones, R. Meddis, S. C. Lim, and A. R. Temple, "Toward a digital
neuromorphic pitch extraction system," iEEE Trans Neural Netw, vol.
II,pp. 978-87,2000.
[6] S. C. Liu and T. Delbruck, "Neuromorphic sensory systems," Curr Opin
Neurobiol, vol. 20,pp. 288-95,Jun 2010.
[7] C. Mead, "Neuromorphic electronic systems," Proceedings of the iEEE,
vol. 78,pp. 1629-1636, 1990.
[8] R. Sarpeshkar, "Brain power-borrowing from biology makes for low
power computing [bionic ear]," Spectrum, iEEE, vol. 43, pp. 24-29,
2006.
[9] A. M. Galaburda,M. LeMay,T. L. Kemper,and N. Geschwind, "Right­
left asymmetrics in the brain," Science, vol. 199, pp. 852-856,1978.
[10] N. A. Badian, "Nonverbal disorders of learning: The reverse of
dyslexia?," Ann Dyslexia, vol. 36,pp. 253-69,Jan 1986.
[11] T. Okusa, R. Kakigi, and N. Osaka, "Cortical activity related to cue­
invariant shape perception in humans," Neuroscience, vol. 98, pp. 615-
24,2000.
[12] A. L. Hodgkin and A. F. Huxley, "A quantitative description of
membrane current and its application to conduction and excitation in
nerve," J Physiol, vol. 117,pp. 500-44, Aug 1952.
[13] c. Morris and H. Lecar, "Voltage oscillations in the barnacle giant
muscle fiber," BiophysJ, vol. 35,pp. 193-213,Jul 1981.
[14] E. M. lzhikevich, "Simple model of spiking neurons," iEEE Trans
Neural Netw, vol. 14,pp. 1569-72,2003.
[15] R. T. Davis and J. A. Ferraro, "Comparison between AER and behavioral
thresholds in normally and abnormally hearing chinchillas," Ear Hear,
vol. 5,pp. 153-9,May-Jun 1984.
[16] A. Linares-Barranco, G. Jimenez-Moreno, B. Linares-Barranco, and A.
Civit-Balcells, "On algorithmic rate-coded AER generation," IEEE
Trans Neural Netw, vol. 17,pp. 771-88,May 2006.
[17] C. Zamarreno-Ramos, T. Serrano-Gotarredona, and B. Linares­
Barranco, "A 0.35 mum sub-ns wake-up time ON-OFF switchable
LVDS driver-receiver chip VO pad pair for rate-dependent power saving
in AER bit-serial links," IEEE Trans Biomed Circuits Syst, vol. 6, pp.
486-97,Oct 2012.
[18] B. Zhao, R. Ding, S. Chen, B. Linares-Barranco, and H. Tang,
"Feedforward Categorization on AER Motion Events Using Cortex-Like
Features in a Spiking Neural Network," IEEE Trans Neural Netw Learn
Syst, Oct 22 2014.
[19] M. A. Mahowald and C. Mead, "The silicon retina," Sci Am, vol. 264,
pp. 76-82,May 1991.
[20] K. A. Zaghloul and K. Boahen, "A silicon retina that reproduces signals
in the optic nerve,"J Neural Eng, vol. 3,pp. 257-67,Dec 2006.
[21] T. Delbruck, A. van Schaik, and 1. Hasler, "Research topic:
neuromorphic engineering systems and applications. A snapshot of
neuromorphic systems engineering," Front Neurosci, vol. 8, p. 424,
2014.
[22] G. Indiveri, B. Linares-Barranco, T. J. Hamilton, A. van Schaik, R.
Etienne-Cummings, T. Delbruck, et aI., "Neuromorphic silicon neuron
circuits," Front Neurosci, vol. 5,p. 73,2011.
[23] R. Serrano-Gotarredona,M. Oster,P. Lichtsteiner,A. Linares-Barranco,
R. Paz-Vicente, F. Gomez-Rodriguez, et aI., "CAVIAR: a 45k neuron,
5M synapse, 12G connects/s AER hardware sensory-processing­
learning-actuating system for high-speed visual object recognition and
tracking," iEEE Trans Neural Netw, vol. 20,pp. 1417-38,Sep 2009.
[24] F. Zhengming, T. Delbruck, P. Lichtsteiner, and E. Culurciello, "An
address-event fall detector for assisted living applications," IEEE Trans
Biomed Circuits Syst, vol. 2,pp. 88-96,Jun 2008.
[25] P. K. Park, H. Ryu, J. H. Lee, C. W. Shin, K. B. Lee, J. Woo, et aI.,
"Fast neuromorphic sound localization for binaural hearing aids," Conf
Proc IEEE Eng Med Bioi Soc, vol. 2013,pp. 5275-8,2013.
[26] C. Hung Tat, N. Kwan Ting, A. Bermak, M. K. Law, and D. Martinez,
"Spike latency coding in biologically inspired microelectronic nose,"
iEEE Trans Biomed Circuits Syst, vol. 5,pp. 160-8,Apr 20II.
[27] A. Calimera, E. Macii, and M. Poncino, "The Human Brain Project and
neuromorphic computing," Funct Neurol, vol. 28, pp. 191-6, Jul-Sep
2013.
[28] A. Schwartz, "First volley in the brain race? Europe's human brain
project the first to the starting line--will U.S. brain initiative catch up?,"
Ann Neurol, vol. 73,p. A7,Jun 2013.
0903
[29] G. Ling, "Newsmaker interview: Geoffrey Ling. DARPA aims to
rebuild brains. Interview by Emily Underwood," Science, vol. 342, pp.
1029-30,Nov 29 2013.
[30] N. V. Thakor, D. F. Moore, R. A. Miranda, and G. S. Ling, "Special
issue of DARPA NEST proceedings," IEEE Trans Neural Syst Rehabil
Eng, vol. 20,pp. 113-6,Mar 2012.
[31] D. Pecevski, D. Kappel, and Z. Jonke, "NEVESIM: event-driven neural
simulation framework with a Python interface," Front Neuroinform, vol.
8,p. 70,2014.
[32] O. Gurcan, K. S. Turker, J. P. Mano, C. Bernon, O. Dikenelli, and P.
Glize, "Mimicking human neuronal pathways in silico: an emergent
model on the effective connectivity," J Comput Neurosci, vol. 36, pp.
235-57,Apr 2014.
[33] O. E. Hernandez and E. E. Zurek, "Teaching and learning the Hodgkin­
Huxley model based on software developed in NEURON's
programming language hoc," BMC Med Educ, vol. 13,p. 70,2013.
[34] E. Arguello, R. Silva, C. Castillo, and M. Huerta, "The Neuroid: A
novel and simplified neuron-model," Conf Proc iEEE Eng Med Bioi
Soc, vol. 2012,pp. 1234-7,2012.
[35] T. Yamazaki,H. Ikeno, Y. Okumura,S. Satoh,Y. Kamiyama,Y. Hirata,
et aI., "Reprint of: Simulation Platform: a cloud-based online simulation
environment," Neural Netw, vol. 24,pp. 927-32,Nov 2011.
[36] M. L. Hines, A. P. Davison, and E. Muller, "NEURON and Python,"
Front Neuroinform, vol. 3,p. 1,2009.
[37] A. P. Davison, D. Bruderle, J. Eppler, 1. Kremkow, E. Muller, D.
Pecevski, et aI., "PyNN: A Common Interface for Neuronal Network
Simulators," Front Neuroinform, vol. 2,p. 11,2008.
[38] D. Goodman and R. Brette, "Brian: a simulator for spiking neural
networks in python," Front Neuroinform, vol. 2,p. 5,2008.
[39] C. Moulin, A. Gliere, D. Barbier, S. Joucla, B. Yvert, P. Mailley, et aI.,
"A new 3-D finite-element model based on thin-film approximation for
microelectrode array recording of extracellular action potential," 1EEE
Trans Biomed Eng, vol. 55,pp. 683-92,Feb 2008.
[40] Y. Hayashida, "[Computer simulation using the "Neuron ": a kinetic
model for channels based on the Markov state transition process],"
Nihon Seirigaku Zasshi, vol. 68,pp. 319-24,2006.
[41] J. F. Vibert, K. Pakdaman, E. Boussard, and E. Av-Ron, "XNBC: a
simulation tool. Application to the study of neural coding using hybrid
networks," Biosystems, vol. 40,pp. 211-8,1997.
[42] B. Ans, 1. C. Gilhodes, and J. Herault, "[Simulation of neuronal
networks (SIRENE). 11. Hypothesis for decoding the message of
movement carried by spindle afferences IA and II by a mechanism of
synaptic plasticity]," C R Seances Acad Sci iii, vol. 297, pp. 419-22,
1983.
[43] R. Paz-Vicente, E. Cerezuela-Escudero, M. Dominguez-Morales, A.
Jimenez-Fernandez, A. Linares-Barranco, and G. Jimenez-Moreno, "A
perfomance comparison study between synchronous and asynchronous
FPGA for spike based systems. Under the AER synthetic generation," in
Performance Evaluation of Computer & Telecommunication Systems
(SPECTS), 2011 international Symposium on, 2011,pp. 38-45.
[44] B. Sbarcea and D. Nicula, "Automatic conversion of matlab/simulink
models to hdl models," in international Conference on Optimization of
Electrical and Electronic Equipment, 2004,pp. 67-70.
[45] F. Baskaya, S. Reddy, S. K. Lim, and D. V. Anderson, "Placement for
large-scale floating-gate field-programable analog arrays," Very Large
Scale integration (VLSi) Systems, IEEE Transactions on, vol. 14, pp.
906-910,2006.
[46] F. Baskaya, B. Gestner, C. Twigg, S. K. Lim, D. V. Anderson, and P.
Hasler, "Rapid prototyping of large-scale analog circuits with field
programmable analog array," in Field-Programmable Custom
Computing Machines, 2007. FCCM 2007. 15th Annual IEEE
Symposium on, 2007,pp. 319-320.
[47] L. Jihong and W. Chengyuan, "A Survey of Neuromorphic Engineering­
-Biological Nervous Systems Realized on Silicon," in Testing and
Diagnosis, 2009. ICTD 2009. iEEE Circuits and Systems international
Conference on, 2009,pp. 1-4.