Vol.3, No.1, 53-56 (2011) Natural Science
http://dx.doi.org/10.4236/ns.2011.31007
Copyright © 2011 SciRes. OPEN ACCESS
Analog-to-digital conversion of information in the retina
Andrey N. Volobuev, Eugeny. S. Petrov
Samara State Medical University, Samara, Russia; volobuev@samaramail.ru
Received 17 November 2010; revised 19 December 2010; accepted 22 December 2010.
ABSTRACT
We considered the physiological mechanisms
of functioning of the retina’s neural network. It
is marked that the primary function of a neural
network is an analog-to-digital conversion of
the receptor potential of photoreceptor into the
pulse-to-digital signal to ganglion cells. We
showed the role of different types of neurons in
the work of analog-to-digital converter. We gave
the equivalent circuit of this converter. We re-
searched the mechanism of the numeric coding
of the receptor potential of the photoreceptor.
Keywords: Analog-to-Digital Converter; A Ganglion
Cell; Oscillator of Clock Frequency; Pulse; Intensity;
Neuron; Action Potential; the Retina; Photoreceptor;
Digital-to-Analog Converter
1. INTRODUCTION
Conversion of information in the retina is carried out
both by the photoreceptor and by the brachiate network
of neurons. Initial perception of the light pulse occurs in
the retinal photoreceptors - in rods and cones. Rods are
responsible for the black and white, twilight vision. They
are evenly distributed on the nervous tunic (the retina) of
eyeball, their total number is about 100 million units in
each retina [1]. Cones, which are 3 million units, are
responsible for the color and daytime vision. They are
more concentrated in the central part (fovea centralis) of
the retina (140000 cones on mm2).
Feature of the photoreceptors is the hyperpolarizing
nature of the receptor potential which is generated in
response to light stimulation. Figure 1 shows the time
dependence of the receptor potential Uin of single cones
that arises in response to short flashes of the light (10 ms)
by three different light intensities [2]. The potential Uin
subsequently proceed to the input of neural network of
retina.
In the darkness on the photoreceptor membrane rest-
ing potential of a little more (modulo) value of –20 mV
(see Figure 1). Curve 1 corresponds to the weakest in-
tensity 1
I
, curve 2—the light intensity 21
4
I
I, curve
3—the intensity 31
16
I
I
. Between the intensity of the
light flash and the receptor potential analog (logarithmic)
dependence was observed.
2. THE EQUIVALENT ELECTRIC
CIRCUIT OF THE NEURAL
NETWORK RETINA
Receptor potential Uin photoreceptor through synaptic
switching proceeds at the input of a bipolar neuron. At
Figure 2 the distribution of electrical signals along the
system retinal neurons is shown. Circuit of the neural
network adopted in accordance with [3], a slight simpli-
fication of concern, in particular, the amount involved in
information processing similar amacrine neurons. These
simplifications we will discuss further.
Neural network retina represents a slightly modified
circuit of the analog-to-digital converter (ADC) consis-
tent accounts [4]. Block diagram of the ADC, the corre-
sponding neural network retina is shown in Figure 3.
Each element of the ADC converts the analog signal
from the photoreceptor PR into digital code, corresponds
to a neuron. In particular, the comparator K corresponds
Uin
t, s
0
-10
-20
-30
-40
0.1
0.2
0.3 0.4
0.5
0.6
I
1
2
3
RP
Figure 1. Time dependence of receptor potential of the rod by
its stimulation with a light flash: RP—the resting potential, I -
moment of light flash, Uin—receptor potential in mV at the
input into the neural network of the retina, t—time in seconds.
A. N. Volobuev et al. / Natural Science 3 (2011) 53-56
Copyright © 2011 SciRes. OPEN ACCESS
54
U2 + RP
P
R
P
R
Uin
U2 + RP
U2 + RP
U1i
U1i
U3
U3
K
1
K
1,2
H
I
B
A
G
AP
Figure 2. Schematic circuit of the neural network retina.
Showing photoreceptor PR and the following neurons:
B—bipolar, A—amacrine, I—interplexiform, G—ganglionic,
H—horizontal. U2 + RP—analog output potential of the I; U1i
—pulse voltage A; U3—analog output voltage of the B; K1,
K2—places the signal is interrupted (keys); AP—the action
potential. The dotted line shows the version of the signal pas-
sage from the I through the body of horizontal neuron H.
P
R
U
in
U
2
+ RP U
1i
U
3
K
1
I
B
A
G
AP
U
3
U
3
K
2
CP DAC
U
1i
GC F
K
Figure 3. Equivalent circuit of a neural network retina: K1, K2
keys.
diaxon B, digital-to-analog converter (DAC)—inter-
plexiform neuron I, pulse counter (PC)—ganglionic
neuron G, generator clock frequency GCF—amacrine
neuron A. Action potentials AP originally generated by
amacrine neurons and in the sequel, according to the
used encoding, through the excitation of ganglionic
neurons are transmitted to fibers of the optic nerve.
Horizontal neurons apparently feebly involved in the
analog-to-digital conversion. Their primary function is
providing contrast boundaries in the field of view visual
image [1]. Analysis of this neural network of the retina
problem in this article is not performed.
The purpose of this study is to examine the transition
of hyperpolarizing receptor potential of photoreceptors
in the pulse sequence of action potentials of ganglion
neurons.
Let us consider in more detail the equivalent circuit of
the ADC retina, shown in Figure 3. Showing items are
required for any variety of ADC. One of the most im-
portant elements is the generator of clock frequency
(GCF), which generates a continuous sequence of pulses,
in the simplest case—with a constant frequency. This
sequence of electrical pulses U1i is shown in Figure 4a.
It is known that amacrine neurons generate action po-
tentials [3], but until now their role is not clear [5].
There are identified by morphological and histochemical
methods about 30 species amacrine neurons [1]. There’s
also stated that amacrine cells “react to the inclusion and
turning off lights, which simply signaling about change
of lighting, regardless of its direction.” This correlates
well with their role GCF in the circuit ADC. A large
variety of amacrine neurons indicates about variability
parameters of ADC retina.
An important element of the ADC is the comparator K.
It does not work with pulsed signals, and compares the
two analog signals—the input to the neural network of
photoreceptor Uin and the output signal from digital-to-
analog converter (DAC), gradually and step increases to
RP
U
in
U
1i
ΔU
a
0
t
AP
t
t
t
b
c
d
U
3
U
2
Figure 4. Graphs of the analog-to-digital conversion of recep-
tor potential in the retina: aa sequence of AP amacrine neu-
rons, bthe accumulation with a step U of potential com-
parison U2+RP interplexiform neuron, RPthe resting poten-
tial of photoreceptors, ca countable sequence (spike) AP of
ganglion neurons; dthe output signal of bipolar neurons.
A. N. Volobuev et al. / Natural Science 3 (2011) 53-56
Copyright © 2011 SciRes. OPEN ACCESS
5555
U2 + RP (see Figure 3). In case of equality Uin = U2 +
RP comparator generates a constant voltage to the U3,
see Figure 4d. This is consistent with the fact that bipo-
lar neurons do not generate action potentials. Signal for
bipolar neurons spreads like electrotonic [2].
DAC is a required element of ADC. It is discrete, with
a step U (Figure 4b), accumulates potential U2 + RP at
its output for comparison with the receptor potential Uin
photoreceptor (see Figure 3).
An interplexiform neuron feedback plays the role of
DAC. On the input of this neuron by its impulse
branches served U1i impulse voltage and on the output
occurs a step analog potential comparison U2 + RP. So
far the role and functioning of these neurons were not
described. Some authors [1] consider that the role of
these neurons is not very important and call them in-
ter-retinal.
Pulse counter in the ADC retinais undoubtedly the
ganglionic neuron G (see Figures 2,3). Ganglionic neu-
rons direct the signal in the form of action potentials to
the central nervous system.
3. INFORMATION CODING BY A
NEURAL NETWORK RETINA
Work ADC retina begins from supply the receptor
potential Uin produced by the photoreceptors to the input
of the comparator K. By the closed keys K1 and K2 (see
Figure 3) the impulses from the GCF served on the
pulse counter, where they come out in the form of a se-
quence of action potentials, and on the DAC, where the
potential comparison accumulate U2 + RP (Figure 4b).
If the conditions arise Uin = U2 + RP comparator gener-
ates a voltage of U3 (Figure 4d), which breaks the out-
puts from the GCF by keys K1 and K2. It is also possible
to supply the signal U3, produced by the comparator K
(bipolar neurons), directly on ganglionic neurons G
(pulse counter PC) to stop the generation of pulses of
that neuron. Synaptic switching from the B on the G is
another analog of the key K1 in Figure 2. Thus, each
signal from the photoreceptor, ganglionic neuron (or
pulse counter) generates a strictly defined number of
pulses n-pulsed digital signal.
Consequently, the neural network retina carries nu-
meric coding of the analog signal at the output from a
photoreceptor with the help of a certain quantity of
pulses. In this case, the quantity of pulses generated by
the ganglion neurons is proportional to the amplitude of
the analog receptor potential, n ~ Uin.
If we consider that the duration of the signal rise Uin
(modulo) of approximately 0.1 s (see Figure 1) and the
duration of the action potential of ganglionic neuron is 1
ms (the frequency of pulsation in the optic nerve comes
up to 1000 pulses per second [6]), then to one receptor
potential of photoreceptor a ganglionic neuron generates
a spike, consisting of approximately 100 action poten-
tials. Figure 4c shows the conditional six pulses.
It is noteworthy that the signal passage U2 + RP along
the horizontal branches of a neuron H (see Figure 2).
Some bipolar neurons B (comparators) receive a signal
comparison U2 + RP from DAC (interplexiform neurons)
not directly but through the body of horizontal neurons
[3] (see Figure 2). This fact indicates the involvement of
horizontal neurons in work of the ADC of the retina, or
vice versa, the involvement of the ADC in the work of
horizontal neurons, there is a mutual influence of the
analog-to-digital conversion in the retina and the forma-
tion process of the contrast of the visual image. The in-
volvement of interplexiform neurons is in work of hori-
zontal neurons also noted in [1]. It is possible that the
signal, which passage from the I through the branches of
body H, need (or energetically favorable) to add the
stepwise changing of the voltage U2 to resting potential
H. However, the cause of the signal passage from the I
through the branches or body H requires further detailed
investigation.
The task of providing information about contrast
boundaries in the total amount of information sent to the
central nervous system is crucial. Early H. Helmholtz
pointed out the imperfections of the optical system of the
eye, and as a consequence, the poor quality of the image
on the retina [7]. Formation of a contrast boundary in the
field of view carry out by submitting a brake signal from
the horizontal neurons to bipolar neurons, in case, if the
photoreceptors, that are connected with these bipolar
neurons, provide a weak receptor potential. In this case,
approximate and strongly illuminated photoreceptors can
generate a large receptor potential and the brake signal
from the horizontal neurons doesn’t send on the bipolar
neurons, connected with them. The necessity for infor-
mation that weak receptor signals should not be involved
in the generation of potential comparison by interplexi-
form neurons may also be a cause of the signal passage
from the I through the branches or body H.
It should be noted that the circuit of the neural net-
work retina [3] is somewhat more complicated than that
is shown in Figure 2. In particular, the generator of
clock frequency GCF as an amacrine cells A is duplicate,
apparently, to give to ADC a greater reliability. There is
a feedback of pulse branch of interplxsiform (DAC) and
amacrine neurons (GCF) which is not shown in Figures
2,3. Perhaps it is necessary to stop the pulsation of GCF
by achieving equality Uin = U2 + RP, since otherwise, by
continuous work of GCF happen depletion amacrine
neurons. Perhaps this branch of interplexiform neuron
provides a more rigid circular connection of two dupli-
cated amacrine neurons in the retinal neural network.
A. N. Volobuev et al. / Natural Science 3 (2011) 53-56
Copyright © 2011 SciRes. OPEN ACCESS
56
4. CONCLUSIONS
Encoding an analog signal at the output from a photo-
receptor to a digital code is apparently one of the main
functions of the neural network retina. Retinal neurons
perform well-defined functions: a comparison of analog
signals, the generation of clock frequency, digital-to-
analog conversion, counting pulses. At the same time,
horizontal neurons in the retina carry out other important
tasks, such as staining of the visual image. Horizontal
neurons are involved in the work of an analog-to-digital
converter retina.
Fulfilled analysis indicates the importance of stimu-
lating the development of electropulse physiotherapy
procedures for diseases of the neural network retina.
REFERENCES
[1] Guyton, A.K. and Hall, D.E. (2008) Medical physiology.
Logosphere, Moscow.
[2] Samoilov, V.O. (2004) Medical Biophysics. SpetsLit,
S-Peterburg.
[3] Kamkin, A.G. and Kamensky, A.A. (2004) Fundamental
and Clinical Physiology. Academia, Moscow.
[4] Pryanishnikov, V.A. (2009) Electronics: A complete
course of lectures. Corona-Century, S-Peterburg.
[5] Hubel, D. (1990) Eye, Brain, Vision. Mir, Moscow.
[6] Kositsky, G.I. (1985) Human Physiology. Medicine,
Moscow.
[7] Schmidt, R. and Thews, G. (1996) Human Physiology.
Mir, Moscow.