As you read these words, photons of light are being absorbed, reflected, and in some cases emitted (if you are reading this on a computer or a tablet) in different spatial and temporal patterns by the materials that make up the physical objects around you. Your brain is then interpreting these patterns as letters and words that in turn communicate the ideas I want to convey through my words. The photons, which one can think of as little discrete packets of energy that travel through space as waves, are absorbed by the retinas in your eyes and the resulting ‘information’ is then sent to your brain where it is processed and interpreted. But what does this all really mean, and how does it actually happen? What does it mean for a photon to be absorbed by your eyes resulting in you being able to see? The goal of this series of three short articles is to introduce in simple terms how the visual system works. In this first article the focus is on the early stages of how light is converted into a neurochemical signal that the brain can understand, which is in effect the ‘language’ the brain uses to take in information from the outside world, process that information internally, and produce decisions, physiological, and behavioral outputs that allow us to interact with our environment and thrive in it. The process of converting light into a neurochemical signal is called phototransduction, because the retina transduces or changes light energy into the brain’s internal language of communication. The second article will discuss what happens to visual signals as they are processed by the rest of the visual system and the brain. And the third article will consider visual information from and an information theoretic perspective. In other words, where is the information in the raw signals that the visual system has access to and how is it used.
Our sensory systems
The visual system is one of our five sensory systems. Collectively, the sensory systems are responsible for taking in information about the world outside our bodies and mind. Their job is to transduce, i.e. change, specific forms of physical energy into neural signals that the brain can understand. In the case of hearing the energy the auditory system transduces are accordion-like compression waves of air molecules vibrating back and forth produced by different physical objects, i.e. sound waves. In the case of touch the energy is mechanical energy in the form of pressure (a force over a given surface area) on our bodies. For taste and smell we are transducing chemical energy (air molecules or molecules in foods). In each case there are specialized neurons called receptor neurons or receptor cells that are specifically designed to carry out this transduction process. For example, the cochlea in the inner ear contains hair cells with tiny finger like projections called cilia that are attached to a specialized membrane called the basilar membrane. As sound waves (vibrating molecules) hit the tympanic membrne in our ear drums the back and forth vibration of the thin membrane covering the oval window causes a tiny bone connected to it to start moving back and forth (the maleus or hammer). This in turn causes the back and forth movement of a second bone and then a third tiny bone which eventually results in the vibration of a membrane at the oval window in the cochlea that eventually results in a vibration of the the basilar membrane inside the cochlea. As the basilar membrane vibrates in turn produces vibrations in the tectorial membrane whose movements lead to the bending of the sterocilia on the hair cells. This bending causes these neurons to ‘fire’ or produce electrical pulses called action potentials that encode the degree and direction of bending, which in turn encodes the physical properties of the sound waves themselves. It is this electrical information that then propagates to the brain to ultimately produce the sensation of hearing. In other words, mechanical energy has been transduced into the electrical impulses that underlie the information processing associated with hearing and the interpretation of that information by the brain.
The retina and photoreceptor neurons
Vision, of course begins, in the eyes. Lining the back of both eyes is the neural sensory retina, a thin tissue paper-like piece of the brain made up of a number of functionally and cellularly distinct layers of highly specialized neurons. The retina is actually a part of the brain itself that extends into the eyes and connects to the rest of the brain via the optic nerves, one for each eye. The optic nerve is formed by the axons of the retinal ganglion cell neurons that form the last layer of the retina. Axons are the long projections that connect a neuron to other down stream neurons. In the retina most neuron types do not have axons because the physical distances over which signals need to travel are short, but ganglion cells have long axons that project to the other parts of the visual system in the brain. All the ganglion cell axons bundled together form the optic nerve, one for each eye. (In reality the optic nerves are misnamed, they are not actually nerves but central white matter tracts because they are part of the brain proper- the term ‘nerve’ is reserved for bundles of axons in the peripheral nervous system.) Photons enter the eye through the pupil and are focused onto the retina by the optics of the eye, the cornea and lens. Incoming photons are ‘absorbed’ by the first layer of specialized retinal neurons called photoreceptor neurons. These are the neurons in which phototransduction takes place, the transduction process of converting light into the language of the brain.
Photoreceptors come in two classes, rods and cones. Humans have one type of rod but three types of cones. What makes each type of photoreceptor different is the spectral properties of the photopigment in each, how each photopigment interacts with photons of different wavelengths (which we perceive as different colors). Photons, being tiny packets of energy that are both simultaneously waves and particles have a wavelength that determines the amount of energy each packet carries. The photopigment is the actual molecule in photoreceptors that interacts with photons. Each type of photopigment responds maximally to photons of a particular wavelength, and absorbs photons of other wavelengths progressively less as you move away from its peak response, producing a bell shaped like curve called its absorbance spectrum (the amount of absorbance as a function of wavelength). Remember that photons are little packets of wave energy, so the ‘size’ of the waves that make up each photon will vary, measured as the distance from peak to peak of the wave (or more technically correct as the distance between any two points with the same phase). We perceive photons of different wavelengths as different colors using our cone photoreceptors. The rod photopigment, which is called rhodopsin, has a peak absorbance at 498 nanometers (nm). One nm is one billionth of a meter, so these are very tiny wavelengths. Rods can only perceive shades of gray, not color. The cone photopigments have peak absorbances at 420 nm, which is in the violet-blue range of the visual spectrum (the S cones), 534 nm which is green (the M cones), and 564 nm which we perceive as red (the L cones). The tail end of the absorbance spectrum of the S cones at the short (blue) wavelength at about a little less than 350 nm, to the tail end of the L cones at the long end of its spectrum, a little less than 700 nm, defines the range of the human visual spectrum, the spectrum of light (electromagnetic radiation) that we can see and perceive. This is of course a tiny sampling of a much larger set of electromagnetic radiation wavelengths that includes everything from very short high-energy particles such as x-rays and gamma rays to much longer wavelengths such as microwaves and radio waves. The visible spectrum we see is a tiny sliver of a huge electromagnetic radiation continuum.
Phototransduction and the photocurrent
Photoreceptors, both rods and cones, are elongated cells with two distinct anatomical compartments, an inner segment and an outer segment. The inner segment end of the photoreceptor connects to the next layer of neurons in the retina, the bipolar cells. This type of connection is called a chemical synapse, which we will get to a bit later. The outer segment is connected to the inner segment by a thin and fragile hair like piece of the cell called a cilium. It is in the outer segments that one finds the photopigments embedded in a series of overlapping membranes called membrane disks. Think of a very large flat continuous piece of play dough folded onto itself over and over again so that it forms a stack. The photopigment molecules are located in those stacks. In the case of the rods those membrane disks are contained inside the cell membrane that forms the outer segment. In cones, the disks themselves form the outer wall of the cell membrane. For the sake of simplicity we are going to focus on the rod photopigment rhodopsin. However, the rest of this discussion applies equally well to any of the three cone photopigments too. Rhodopsin is a large molecule composed of two parts. The opsin part is the majority of the molecular complex and is a large bulky structure that is embedded in the membrane disks. The second part is a molecule called retinal. You can think of retinal as sitting right in the middle of the opsin portion. It is small and contained within the opsin.
If the human brain were so simple that we could understand it, we would be so simple that we couldn’t.” — Emerson M. Pugh
When a photon of light is absorbed by a rod photoreceptor what that really means is that the photon interacts with the retinal molecule in rhodopsin. This is at the heart of how vision starts. In the absence of light, retinal is in a molecular confirmation called 11-cis retinal. This is a particular three dimensional configuration of the atoms and chemical bonds that hold those atoms together. Retinal, like any molecule, will prefer to be in a state that is thermodynamically stable representing a minimal energy state. When a photon of light hits a molecule of retinal, the energy contained in the photon breaks one very specific chemical bond (the double bond between the carbon atoms in the 11th and 12th positions), which then allows retinal to change its structural configuration from 11-cis to a different more thermodynamically stable state called all-trans. In organic chemistry this is called cis/trans isomerism and it is something organic molecules do all the time. So the interaction of a photon with retinal is called a photoisomerization.
A photoisomerization is only the first step in a biochemical cascade, a series of self-triggered biochemical events, called phototransduction. Following a photoisomeri- zation event, the activated rhodopsin in turn chemically activates a large molecule called tansducin that is floating around in the disk membranes. Transducin belongs to an important class of signaling molecules called G-proteins. G-proteins work by splitting off specific molecular sections or domains of the larger protein once they become activated, in this case triggered by the all-trans form of rhodopsin. The activated part of transducin then floats away inside the disk membranes until it encounters by chance (via Brownian motion and diffusion) a molecule of phosphodiesterase (abbreviated as PDE), which is also floating around in the disk membranes. PDE is an enzyme, so it is designed to catalyze or speed up another chemical reaction. The reaction PDE catalyzes is the conversion of a small nucleotide molecule called cyclic guanosine monophosphate (cGMP) that is floating around this time not in the membrane but in the ‘solution’ inside the outer and inner segments, called the cytoplasm. PDE converts cGMP into a different form that makes it just GMP (i.e. guanosine monophosphate). This conversion of cGMP to GMP has an important consequence on the electrical state of the photoreceptors, which in turn affects how they signal the next layer of neurons in the retina.
Excitable cells such as neurons have a measurable electric potential across their cell membranes referred to as the membrane potential. It is produced by a differential imbalance of chemical charges called ions on either side of the membrane. In most neurons the key charge carriers are ions of potassium and sodium, although other ions such as calcium ions also play important roles. Sodium and potassium ions both have an electric charge of positive one (+1) because of how electrons are distributed across them. If there are an equal number of ions on either side of the cell membrane, then there is no charge imbalance and the electric potential difference is zero. But as is often the case, if the concentration of positive ions is greater on the outside of the membrane than on the inside, then the total number of positive charges will be greater outside the membrane then inside, and the result is a net negative charge on the inside of the membrane. This then results in a membrane electric potential across the membrane that is negative inside relative to outside. (As a technical aside, it is important to note that this charge imbalance exists only right at the membrane itself along a very thin nanometer sized volume along the membrane, and does not affect the bulk concentration of ions inside the cell, which is much greater than the number of ions producing the charge imbalance. In other words, the potential difference between the inside of the neuron and its exterior away from the membrane will under normal conditions be zero.) Photoreceptors have a population of ion channels in the outer segment that is functionally different than the set of ion channels and other related ion pumps in the inner segment. Ion channels are tiny pores that allow the flow of ions between the inside and outside of the cell across the cell membrane which open and close in response to a specific chemical or electrical trigger. The ion channels in the outer segment require cGMP to be chemically bound to them in order for the pore to remain open and allow the flow of ions across the membrane. We say they are cGMP gated ion channels. However, the ion channels and pumps in the inner segment in contrast do not depend on cGMP to remain open. They can operate independent of and with no regard to whether cGMP is there or not. In the dark in the absence of light, the concentration of free floating cGMP in the cytoplasm, i.e. the number of molecules of cGMP, is maximal since there is no PDE being activated through the phototransduction cascade and therefore no conversion of cGMP to GMP. This allows ions and charge to flow into the outer segment through the open cGMP gated channels. This current is called the photocurrent.
In the dark, the amount of charge flowing in through the outer segment is balanced by the amount of charge flowing out through the different population of ion channels and pumps in the inner segment. However, in the presence of light when photons start activating the phototransduction cascade, the result is a reduction in the concentration of available cGMP molecules. This in turn reduces the number of cGMP available for binding to cGMP gated ion channels in the outer segment. As the concentration of cGMP decreases there is not be enough cGMP molecules available to keep all the cGMP ion channels open, and so some of the cGMP gated ion channels begin to close. The greater the amount or intensity of light, which is the same as saying the greater the number of photons that hit the retina, the more cGMP gated ion channels will close. This reduces the flow of current into the photoreceptor in the outer segment, but does not affect the flow of current out through the inner segment. The net result is that more positive charges are moved out of the photoreceptor in the presence of light compared to the absence of light in the dark. Moving positive charges across the membrane from inside the cell to the outside results in making the membrane of the photoreceptor more negative inside, as dis- cussed above. We say the neuron is being hyperpolairzed. So the greater the light intensity, the more hyperpolarized the photoreceptor becomes. As the flux (number of photons per second per unit area) decreases as light intensity is reduced, the concentration of cGMP recovers through a different enzymatic pathway, the number of open cGMP gated ion channels goes up, and the photocurrent is restored as the inside of the membrane becomes more positive again. The neuron in this case is being depolarized. The effect of a hyperpolerizing photocurrent is a decrease in the amount of neurotransmitters being secreted by the photoreceptors at the chemical connection, the synapse, with the bipolar cell layer of neurons they are in contact with.
Signaling from photoreceptors to bipolar cells in the retina
Neurotransmitters are signaling molecules that are packaged into discrete structures called vesicles. These vesicles are little packages formed by their own membrane similar to the cell membrane of the neuron itself, accept that they are inside the cell, typically at one end of the neuron called the synaptic terminal. The neurotransmitters that are the contents of the vesicles can be many different types and classes of molecules. In fact, the term ‘neurotransmitter’ does not refer to a particular class of molecule, but to any molecule that is used by a neuron to signal another neuron. It is a context dependent term. For example, one important class of neurotransmitters are the catecholamines, which include epinephrine (adrenaline) and norepinephrine (noradrenaline). But when the very same molecules of epinephrine and norepinephrine are produced by the adrenal glands that sit on top of the kidneys, are secreted into the blood stream, and have a signaling effect on down stream cells in other non-neural tissues, in this context they are not acting as neurotransmitters they are acting as hormones. Neurons, photoreceptors included, secrete neurotransmitters when they become depolarized. Synaptic vesicles are docked at the synaptic terminal in a kind of standby waiting mode. When the cell membrane becomes depolarized a series of intricate and involved biochemical steps and molecular machinery move the front row of vesicles so that they touch the inside of the cell membrane at the synaptic terminal. This causes the fusion of the vesicles’ membranes with the cell membrane, essentially turning the inside of the vesicle into a little pocket of the cell membrane itself and ex- posing its contents to the outside of the neuron. The neurotransmitter molecules then diffuse across a small gap called the synaptic cleft between the neuron doing the signaling, the presynaptic neuron, and the neuron being signaled, called the post synaptic neuron. Molecular binding of the neurotansmistters with the receptors on a specialized part of the membrane of the postsynaptic neuron eventually causes a change in its cell membrane potential. This is the basics of how chemical signaling between neurons occurs. In the case of photoreceptors, the neurotransimitter being secreted is the amino acid glutamate, which are docked in vesicles at a special type of synapse in the photoreceptor called a ribbon synapse that is designed to allow rapid release of many vesicles. This in turn activates bipolar cell neurons. This is the eventual end result of phototransduction and is the process by which photoreceptor neurons transduce, or change, light energy into a neurochemcial signal.
Why we need a phototransduction cascade to see
Before leaving phototransduction, let’s touch on two additional points. First, the above description leads to an interesting and somewhat counter intuitive consequence. Typically, we associate the excitability or activation of a neuron with the signaling of an incoming neural message. A signal from an upstream presynaptic neuron electrically excites a downstream postsynaptic neuron as a signal is communicated from one to the other. This excitation is associated with a depolarization that in turn results in the secretion of neurotransmitters. So a positive signaling event is associated with an increase in neurotransmitter release. However, this is reversed in photoreceptors because of the way the photocurrent works. Photoreceptors are maximally depolarized in the dark in the absence of light, which in this case is the stimulus. This is because in the absence of light there is no phototranduction occurring and therefore the concentration of cGMP is maximal resulting in all cGMP gated ion channels being open and therefore a maintained depolarized state for the photoreceptor. As the intensity of the light increases, which is the same as saying the flux of photons hitting the retina increases, the fraction of cGMP gated channels proportionate to the light intensity close down and the photoreceptor becomes hyperpolarized, resulting in a decrease in glutamate release at the bipolar cell synapse. In other words, in photoreceptors, more stimuli produces less signal and not the other way around as is typically the case. This then also implies that there is a maximum light intensity, a maximum flash intensity, that completely shuts down the entire population of cGMP gated ion channels in the outer segment for a brief period of time. Beyond that light intensity we are not capable of distinguishing brighter flashes of light. And because there is a finite period of time needed for the concentration of cGMP and ultimately the photocurrent to recover, this is the basis for why a bright flash, such a strobe flash or bright camera flash, is temporarily blinding.
The second footnote has to do with why have a phototransduction cascade in the first place. Why not simply have some molecular mechanism that goes straight from a photoisomerization event directly to a closure of a cGMP ion channel? Why go through the trouble, and use of valuable cellular energy, of the multiple biochemical steps in the phototransduction cascade? The answer in one word is amplification. Psychophysically we are able to perceive just a handful of photons hitting the retina. Depending on who you ask this threshold of perception is somewhere in the neighborhood of five to a few tens of photons. The photoreceptors themselves can actually respond to individual photons. This is possible because of the amplifying nature of phototransduction, whereby a single photon can have an appreciable effect on the resultant change in the photocurrent. One single photoisomerization event can activate on the order of 800 or so transducin G-proteins. This is because the activated rhodopsin stays activated long enough to have multiple stochastic interactions with transducins that happen to float by and bump into the rhodopsin in the membrane disks. There is no amplification in the activation of PDE by transducin, but there is an even larger amplification step in the number of cGMP molecules that are catalyzed to GMP by activated PDE. The approximate 800 activated PDE due to the single photoisomerization results in about 4800 cGMP to GMP conversions. This in turn closes down roughly 200 or so cGMP gated ion channels, which equates to an approximate decrease of the photocurrent by about 1pA of current. This decrease in the photocurrent hyperpolarizes the photoreceptor by about 2 mV.
And that is it. That is only thing that photons do. They provide the energy needed to break a chemical bond. Everything that you see and visually perceive, the entire richness of the visual world around you, all of the raw information that enters your eyes that your brain has to work with, begins as a highly repetitive and boring stereotyped chemical event. And that, is amazing.
John E. Dowling (2012) “The retina: An approachable part of the brain”. Harvard Uni- versity Press ISBN: 9780674061545.
This is one of the best and clearest introductions to the retina and visual system ever written. This is a technically rich book but generally accessible to anyone. This book is a classic written by one of the pioneers of visual neuroscience.
Clyde W. Oyster (1999) “The human eye: Structure and function”. Sinauer Associates Inc. ISBN: 0878936440.
This is another superbly written book that is technically very rich but a relatively easy accessible read by anyone. It covers the entire visual system in detail.
Martin J. Tovee (2008) “An introduction to the visual system”. Cambridge University Press. ISBN: 0521709644.
This book is fits into the same category as the two previous two and is another excellent introduction. In particular this book assumes very little to no prior knowledge.
Eric Kandel, James Schwartz, and Thomas Jessell (2012- 5th edition) “Principles of neural science”. McGraw-Hill Medical. ISBN: 0071390111. This text book is one of the most iconic and widely read neuroscience texts ever. It is a must read for any serious individual interested in the brain, including professional neuroscientists. It provides a detailed introduction to not just the retina and visual system but the entire nervous system, and ranges from basic molecular and cellular neurobiology to cognitive neuroscience.
http://webvision.med.utah.edu/. WebVision is an online comprehensive (free) text book on the retina and visual system written by a number of experts maintained by the University of Utah. It was started by Dr. Helga Kolb and colleagues in 1994. Unlike print texts, because it is online WebVision is regularly maintained by its curators and up to date on some of the most recent advances in visual neuroscience.
Edmunt T. Rolls (2002) “Computational Neuroscience of Vision”. Cambridge University Press. ISBN: 0198524889.
Robert Snowden (2006) “Basic vision: An introduction to visual perception”. Oxford University Press. ISBN: 0199286701.