As I throw a multicolored ball in the air in front of my son, the information makes its way through Kai’s retinas and quickly stimulates brain-stem sites, such as the superior colliculus, that focus his attention toward the object. This brain-stem attention-grabbing circuitry is about as evolutionarily ancient as any primate brain region, since it is found in every vertebrate. From the brain-stem sites, the visual information about the ball travels through Kai’s thalamus and enters his primary visual cortex, which provides him with the first conscious perception of the object. Thus, although the brain-stem activation produced a change in behavior causing Kai to focus on the ball, this processing is beneath the surface of consciousness.
Once the visual information reaches V1, it quickly diverges into two dominant streams: one responsible for processing information about an object’s spatial location and movement (the “where” pathway) and another responsible for processing the form features of the object such as shape, color, and texture. The latter stream is known as the “what” pathway.
The “where” and “what” visual pathways develop and mature at different rates in humans. In primates the brain pathways devoted to processing object motion develop and mature far earlier than those responsible for processing advanced object form information. For example, while Kai was delighted by the appearance of almost any slowly moving object when he was four months old, he is now a ten-month-old clearly in love with bright, primary colors. This is not really a transition from preferring objects that move to preferring objects that have bright colors. Rather, it is the addition of a fondness for bright, primary colors that joins the list of pleasure-inducing forms of stimulation. This sequence is characteristic of all human infants and maps onto the relative maturity of the “where” and “what” pathways.
At four months an infant’s posterior parietal areas that comprise the where pathway are just beginning to undergo a major increase in synaptic pruning. Remember from earlier chapters that during synaptic pruning, experience becomes the crucial instrument in shaping and fine-tuning brain circuitry in the early stages of development. Hence a four-month-old infant’s where pathway is just beginning to enter a phase of synaptic pruning, where its continued development and fine-tuning depend on appropriate stimulation. In this case, appropriate stimulation consists of any experiences that would optimally activate the mature circuit—namely, moving objects.
If we were to design this process in a robot—creating a sensory perception system that depends on experience for fine-tuning—any good engineer would build in a process to increase the probability that the optimal forms of required stimulation are experienced at the right times. Likewise, nature doesn’t rely on the mere dumb luck that a developing infant will just happen to encounter specific forms of stimulation that are required for normal development. Nature has solved this problem by linking the brain circuitry that supports natural reward (primary reinforcing stimuli) with the maturing circuitry from the primary sensory systems. For example, to make the experience of motion perception pleasurable to Kai at four months, the growing circuitry in his posterior parietal lobe begins to establish reciprocal connections to several brain-stem and limbic regions that are involved in natural reward, motivation, and analgesia. Thus the activation of Kai’s posterior parietal lobe circuitry by a slowly moving object such as a ceiling fan begins to be accompanied by pleasurable sensations much like those corresponding to primary reinforcing stimuli (such as sweets).This process ensures that Kai naturally seeks out objects that fill this experience-expectant requirement for the successful fine-tuning of his visual where pathway and the continued refinement of his capacity to discern motion.
Our crawling ten-month-old Kai is now entering a phase where some areas in his what pathway are undergoing extreme synaptic pruning. At about this time regions such as V8, devoted to processing color vision, begin to mature and consequently need proper stimulation for continued growth and refinement. Kai’s emerging attraction to objects that are composed of bright primary colors—reds, greens, and blues—will encourage him to seek out these optimal forms of stimulation that provide the fine-tuning needed in region V8. Indeed, the primary colors correspond to the wavelengths of light that optimally activate distinct classes of brain cells in region V8. If these cells are damaged in an adult, for instance by a stroke or related trauma, the result is a complete loss of color vision with no change in other features associated with visual acuity. The developmental pattern that is seen in infants—object motion pathways maturing before most brain areas involved in visual object recognition—echoes the evolution of vision in primates. Comparative studies suggest that the object motion pathway evolved well before most brain regions that are devoted to object recognition. For instance, the object motion pathway is observed in all mammals, but features that support object recognition such as trichromacy did not appear until the divergence of the primate lineage from other mammals.
The Pleasure of Learning
Vision is no different in terms of general developmental properties than any other sensory system. Genes play a direct role in mapping out the major brain regions dedicated to vision and the general pathways that connect them. Somewhere in the mere twenty-five thousand or so genes that comprise the human genome, there is enough information to ensure that the enormously complicated wiring of the human brain (and the rest of the body, for that matter) is mapped out. Genes do not code specific paths, but rather cause the development of unique molecular markers that are used by growing brain fibers as targets. This process gets the connections approximately right but leaves the remainder of the job—the fine-tuning—to experience.
Fine-tuning the visual system takes a long time. While the neural pathways that mediate motion processing mature fairly early, the circuitry responsible for higher-order processing and detailed visual acuity continues to be fine-tuned by experience well into the toddler period. Like the other sensory systems, brain cells that comprise visual circuitry are not mini-blank slates waiting to be written upon. They come preprogrammed with certain receiver biases from the very beginning. Experiments done in the early 1960s showed that neurons in the primary visual cortex, visual thalamus, and even in the retina itself respond optimally to certain forms of stimulation and are barely activated at all by others. For instance, many cells in V1 tend to respond to straight lines of a particular orientation. If we could record from cells in your primary visual cortex right now, we could perform the following experiment. Imagine I display a completely white screen directly in front of your eyes.While you focus on the screen, I lower a pencil held by its tip until it enters your visual field. Light reflected off the pencil enters your retina, where photoreceptors transduce the light energy into electrical impulses that are then sent along the visual pathways we have been discussing. At each stage in the processing of this image, some cells respond to this particular form of stimulus; however, most remain quiet. Starting with your retina, then your thalamus, and including V1, only a select group of cells have a preferred tuning for this specific orientation of the pencil. Other V1 cells are sensitive to straight lines, but they will only respond when the pencil is rotated to their preferred orientation (for example, horizontal instead of vertical).
Interestingly, our V1 cells are not ordered haphazardly, but have a strict anatomical organization that is related to the degree of angular rotation of a viewed edge. All of the cells can be activated by a straight line, but the line has to have the correct orientation to excite a given cell and get it communicating with other neurons. The functional consequence of this physiological arrangement is edge detection, a capacity that is critical for many aspects of vision, such as identifying the natural boundaries of an object. Many brain theorists envision a scenario where information from multiple edge detector cells is integrated at higher cortical areas to form a representation of the entire object. Experimental evidence shows that multiple cells from V1 with different orientation tuning converge on the same neurons at higher cortical areas such as V2, so this theoretical position has anatomical support.
Edge detection by V1 cells tuned to straight lines of a particular orientation is just one example of preexisting biases that are built into brain cells from the earliest period of development. Tuning biases like this have been found in newborn brain cells in virtually every mammalian species tested. Interestingly, although most cells in V1 have a signal preference shortly after birth, experience plays a critical role in shaping the tuning to match the particular ecological niche that an organism inhabits. In most species, a critical period exists in the earliest stages of visual development where if V1 cells are denied stimuli that match their preferred orientation, they may die or be retuned to another orientation. If the cells are stimulated by the appropriate signals during this period, however, the tuning of the cell becomes increasingly sharpened and specific to the original bias. This has two important effects. First, the increased tuning makes information transfer less noisy, since variation in terms of what kinds of stimuli may excite a cell naturally decreases. A second consequence of this process is that while some signals may be detected quite easily with minimal stimulation, other signals that are slightly different from the preferred tuning will be completely missed. Hence, existing biases that are in place at or near birth can become magnified with experience, while others may die out or even be replaced. Individual brain cells are far from being blank slates at birth.
The critical periods for visual system tuning—like the other sensory systems we have encountered—occur when that particular circuit is undergoing synaptic pruning (see chapter 3). During this period brain cells increase their sensitivity to some forms of stimulation and necessarily lose their responsiveness to others. The sharpened tuning of cortical cells that support visual perception results in increased visual acuity for some features and a decrease for others.
Experiments since the 1960s have demonstrated that cats and monkeys who are denied visual stimulation in a particular eye during this period of extreme plasticity have marked visual deficits as adults. Moreover, the primary visual cortex (and other visual areas) of the deprived animals looks very different from that of normally reared controls. Usually there is an approximately equal portion of visual cortical area in V1 devoted to processing information from each eye. If one eye is covered during the critical period, the portions of V1 that receive information from the competing eye expand and take over the areas that would have been associated with the covered eye. As we have seen in earlier chapters, synaptic pruning is a process that is ruled by competition.Two synaptic connections vying for the same space will each struggle to stabilize into a mature circuit, but the one that is activated by visual experiences the most usually wins. The old adage “Use it or lose it” rings true in this case. With the appropriate stimulation (nurture), the initial visual circuitry laid down by nature is further shaped to match environmental contingencies.
When cats or monkeys are reared in a carefully controlled environment where they only experience lines of a single orientation (for example, all vertical or all horizontal) during the critical period, their V1 cells stop responding to other orientations and retune to fire only at the experienced orientation. Later, as adults, these animals show poor visual acuity for detecting edges at novel orientations relative to control animals.
Humans also show signs of this effect. For instance, at least one study has demonstrated that North American Indians reared in traditional teepee-shaped dwellings have better visual acuity for oblique or diagonal angles when compared to people raised in “carpentered” environments (that is, house and apartments) that are filled predominantly with vertical and horizontal orientations.We begin with a set of preexisting preferences for visual scenes, but there is considerable latitude in how early exposure can retune brain cells that support vision and the connections between them.
Neural Bootstrapping
At birth, humans have a visual acuity of about 20/600, which is roughly thirty times poorer than that of normal adults. The attraction babies have for high-contrast objects and faces provides just enough stimulation for growing visual cortex cells to continue to mature at a reasonable pace. In the first three months, visual acuity steadily increases. Infants become more attracted to even finer gradations of contrast and are particularly fond of contrasting patterns that have pronounced lateral symmetry. It is not until their ability to experience these more nuanced visual patterns occurs that a second period of tremendous growth and synaptic pruning kicks into gear in V1 and higher cortical areas that are responsible for so-called hyperacuity.
Theoretical calculations of visual acuity based on the actual physical size and density of photoreceptors suggest that we should not be able to see as well as we do. Higher cortical areas, however, have a rich bag of tricks for solving visual problems like completing object patterns from partial or obscured inputs. These mechanisms radically improve visual acuity beyond the expected theoretical limits.
Interestingly, the development of hyperacuity depends more on experience in the second six months than the first. This is because humans must first experience simple visual patterns that promote the development of subcortical areas and V1. Once this circuitry matures in the first six months, babies become increasingly attracted to richer patterns of visual stimulation, such as scenes with more subtle contrasts and strong lateral symmetry. These experiences are, in turn, needed for the stimulation and normal maturation of higher cortical regions during their growth spurt in the second six months that support hyperacuity. This developmental pattern is so lawful that pediatricians often use visual tests of hyperacuity as an indicator of normal brain growth at twelve months.