Pediatrics · Part Three · How a Child Develops
Lesson 39 / 57
Vision Develops: How a Child’s Brain Learns to See
Sight is built in the cortex, not delivered by the eye. Deprive one eye during the critical period and the cortical territory reorganizes around the other.
Vision develops in the brain rather than in the eye, and it is shaped by the patterned light a child actually receives. Grating acuity in healthy infants aged 6 to 10 months measured 5.58 cycles per degree behaviorally and 8.97 by evoked potential. The visual critical period closes when perineuronal nets condense in primary visual cortex, and deprivation experiments reorganized ocular dominance columns. The Unified Model of Tone reads vision as a channel whose weighting is learned.
Grating acuity, 6 to 10 months
5.58 cycles per degree
Same infants by evoked potential
8.97 cycles per degree
Healthy infants measured
61
Visual critical period brake
perineuronal nets in visual cortex
Where sight is assembled
Light reaches the retina, and signal travels the optic nerve and crosses partly at the chiasm. It relays in the lateral geniculate nucleus of the thalamus, then arrives at primary visual cortex at the back of the head. The eye collects. The cortex learns what the collection means, and that learning takes years.
Acuity and how it is measured
Visual acuity in an infant is measured in cycles per degree, the finest stripe pattern the child reliably resolves. It can be read from behavior, by watching where a baby looks, or electrically, by recording the cortical response. The two methods do not always agree, which is itself informative.
01What a baby resolves
Infant acuity is measurable, and the method changes the answer
Infant vision can be measured, and the numbers are more interesting than a single figure suggests. Polevoy and colleagues assessed 61 healthy infants aged 6 to 10 months using both Teller acuity cards and sweep visual evoked potentials. The behavioral method gave a threshold of 5.58 cycles per degree. The electrophysiological method gave 8.97 (Polevoy 2017).
The two disagreed in a specific way. The difference fell within one octave for 64 percent of infants, and it shrank as behavioral acuity rose. At the highest behavioral acuity the card method slightly exceeded the evoked-potential method.
A newborn works at close range. The clearest zone in the first weeks sits at about the span between a feeding baby and a parent’s face, and high-contrast edges pull the gaze first. A hairline against a forehead, the dark rim of an eye. That close working range is an arrangement rather than a limitation, because it puts the sharpest part of a baby’s visual world exactly where her people are.
Why the gap is the finding
The cortex was responding to detail the baby was not yet demonstrating. That gap is a fair description of infant vision generally: the hardware registers more than the behavior reveals, and the two converge as the system matures. It is also a caution about any single quoted acuity figure, since the number depends on how it was obtained.
02Findings
What the research shows
The figures below cover what infants actually resolve, and what happens to a visual cortex without balanced input.
03Built by input
Cortical territory follows the input each eye delivers
The visual cortex allocates itself according to what arrives. LeVay, Wiesel and Hubel traced the development of ocular dominance columns in normal and visually deprived monkeys and found the columns developing differently when input was unbalanced (LeVay 1980). Hubel and Wiesel received a Nobel Prize for this line of work.
The principle generalizes past vision. Cortex is not pre-wired to a fixed plan and then switched on. It is allocated competitively, and the competition is decided by the traffic each pathway carries.
Why balanced input matters clinically
This is the mechanism behind taking a persistent inward or outward turn of one eye seriously in a young child. It is not that the eye looks unusual. It is that unequal input during the period of active allocation changes how the cortex divides itself. The window matters more than the appearance.
04The visual window
The critical period closes through a structure you can image
The visual critical period has a physical end. Sigal and colleagues used superresolution imaging to watch perineuronal nets condense around parvalbumin-positive interneurons in primary visual cortex, and describe that net as a molecular brake on plasticity (Sigal 2019). Before it condenses the circuit reorganizes readily. Afterward the same input does far less.
The schedule is not fixed by age alone. In that work total sensory deprivation delayed net maturation, and deletion of MeCP2, the gene behind Rett syndrome, accelerated it. The brake tracks the input history of the circuit rather than the calendar.
How this fits the wider build
Vision is one window among several running on different clocks. Synaptic density peaks near three months in auditory cortex and after fifteen months in the middle frontal gyrus (Huttenlocher 1997). Early sulcal development has been linked to later sensorimotor and cognitive outcome (de Vareilles 2023). The window of plasticity covers the general case.
05How much to trust the eyes
Young children lean on vision and cannot easily stop
Vision is not simply a channel that improves. It is a channel whose weighting has to be learned, and young children weight it heavily. Forssberg and Nashner tested children on a moving platform inside a moving visual surround. Those under about seven and a half years could not suppress visual input when it reported false orientation (Forssberg 1982).
Later work measured the change directly. Four-year-olds, eight-year-olds, twelve-year-olds and adults stood inside an oscillating room, and the adults and older children down-weighted a misleading visual signal more than the younger children did (Polastri 2013).
What that explains
A young child in a visually busy environment is receiving a signal the nervous system cannot yet discount. That is a real explanation for why some children are unsteady in crowds, on patterned floors, or in the dark. Sensory integration takes the weighting question in full.
06Looking and moving
What a child sees depends on how a child moves
Vision is fed by movement, because moving changes the visual stream. Among 47 fourteen-month-olds, walkers outperformed crawlers on a visual search task, and effortful search was stronger in expert walkers than in novices (Mulder 2022). In a second sample of 913 two-year-olds, earlier walking onset tracked better visual search.
The mechanism is the change in vantage point and in what a child can approach. An upright child sees further, moves faster, and chooses more distant targets, so the visual system is trained on a different diet.
The traffic runs both ways. As a baby begins to reach, the eyes guide the hand, and the hand teaches the eyes about distance and size. Sitting, then crawling, then walking each open a fresh vantage point, and the finer achievements of color and depth are built on top of that foundation. This is why the whole moving child is worth watching rather than the eye alone.
The structure doing the work
The comparison of what the eyes report against what the body did runs through the cerebellum, which increases 240 percent in volume across the first year (Knickmeyer 2008). Movement is the engine and the vestibular system cover the other two channels in that comparison.
07Gaze and the neck
Steady eyes are built on a steady head
Seeing well while moving is a different problem from seeing well while still. Three cranial nerves move each eye, the oculomotor, the trochlear and the abducens, and they are coordinated through the brainstem and the cerebellum. Following a slow target across the midline comes first. Reliable teaming of the two eyes, and the depth perception that depends on it, follows over the months after that.
Early roving, or a brief crossing that comes and goes, is ordinary while that yoking is still tuning. A constant inward or outward turn of one eye is a different matter and belongs with an eye specialist promptly.
Why the upper neck belongs to vision
Holding an image still while the head moves is the work of the vestibulo-ocular reflex, and it runs on head-motion information. Part of that comes from the vestibular apparatus. Part comes from the neck itself. The nervous system weighs proprioceptive, vestibular and visual input against one another continuously, and early on it does that poorly. Children under about seven and a half years could not suppress vision reporting false orientation (Forssberg 1982).
The deep suboccipital muscles are among the most position-rich tissues in the body, built to report position rather than to produce force. The upper neck in delivery sets out that anatomy in full. The model holds that the fidelity of head-position information sets how well gaze is stabilized during movement. On that reading a head habitually held to one side samples a narrower visual world, and the eyes are working against a reference that is off.
What an examination involves
The looking is read first: how a baby fixes on a face, whether she follows a slow target across the midline, and whether the two eyes move as a pair. Then the head and the neck. Resting head position, rotation to each side, and palpation of the occiput, the upper cervical tissues and the cranial base.
The contact that follows is a sustained light hold, no more than the pressure a person can comfortably rest on a closed eyelid. Force is graded to the age and size of the child. There is no twist, no thrust and no audible release, and a settled baby very often stays asleep through the visit.
Referral runs alongside the reading. A constant eye turn, an eye that does not follow, or any concern about how a child sees belongs with an ophthalmologist or optometrist. The closing of the visual window is exactly why that referral is time-sensitive rather than optional.
08The model’s claim
What the Unified Model of Tone predicts about vision
Everything above is established science. What follows is this model’s reading of it, stated as ours rather than drawn from the papers cited.
Tone is the integrated organization of the body’s interacting state, and the model holds that perception is how that organization reads the world’s frequencies. Your eye is a frequency meter and color is its readout. Vision is therefore not a separate faculty on our reading. It is one measuring instrument among three whose weight in the organization is set by experience, since the nervous system reweights proprioception, vestibular input and vision continuously by their reliability. The reweighting data are the evidence: a five-year-old and an adult receive comparable visual information and treat it very differently, because the weighting is part of the organization rather than part of the eye.
That is also why the model expects skilled contact at the head and neck to reach the way a child looks. An input does not create an outcome on its own. An input meeting a particular tone creates one, so the same sustained hold offered to two children is two different events. What the contact offers is information, delivered at the tissue carrying the densest supply of position information in the body. The model holds that better information yields better organization, and that gaze is among the first places the improvement shows.
The prediction
From that reading the model predicts that vision will rarely behave independently of posture and movement in a young child. It predicts that changing the quality of head-position and movement information can change measurable visual behavior, because the three channels are resolved together rather than separately, with compensation deciding how much of the change the visual channel shows.
The Mulder finding is the shape of evidence the model expects, since earlier walking onset tracked better visual search across 913 two-year-olds. This is a claim about how development is organized rather than about what treatment does. A persistent eye turn belongs with an ophthalmologist or optometrist, and the critical period is exactly why that referral is time-sensitive.
What the model claims is that better-organized tone yields greater adaptive capacity, whichever appropriate input delivered the useful information. If visual acuity development, sensory reweighting, postural control and autonomic regulation are shown to move together, the unification claim is confirmed.
09The tone reading
How vision expresses tone
Every topic in this library expresses all of tone. In vision three aspects carry the signature, because a channel measured at 5.58 cycles per degree behaviorally and 8.97 electrically is being weighted, not simply received.
Gain
How heavily vision counts is learned. Children under seven cannot discount a visual signal that is reporting falsely to them.
Input quality
Cortical territory follows the traffic each eye delivers, so unequal input during the open window changes how the cortex divides itself.
Time course
The visual window closes through perineuronal nets condensing in visual cortex, and deprivation shifts when that brake arrives.
The remaining foundations run through vision as well. Prediction: stable sight during head movement depends on predicting the visual consequence of the turn. Coupling: vision, balance and proprioception are resolved into one estimate rather than three. Set point: gaze is held against a postural baseline the body maintains. Constraint: a child who cannot turn the head freely samples a narrower visual world. Load: the visual system is metabolically expensive during the years it is being built. Oscillation: sleep and alertness change what a visual cortex does with the same scene. These are readings of one organization rather than separate systems, which is the core claim of the Unified Model of Tone.
10Across the library
How this page relates to the rest of the library
Vision is one of three channels the body uses to know where it is.
The channel vision is weighed against, and the one built to be working first.
How the channels are weighted together, and what the term actually means when it is measured.
The general case of the critical period, and the same molecular brake seen across circuits.
The other distance sense, on its own schedule, with the earliest cortical peak of any region measured.
Why how a child moves changes what a child sees, and where that signal is compared.
The adult mechanics of moving the eyes, and what each system is built to do.
Sensory function as a measurable state, with the instruments used to read it.
11Frequently asked
Questions families ask about visual development
How well can my newborn actually see?
Less sharply than an adult, and the exact figure depends on how it is measured. In 61 healthy infants aged 6 to 10 months, acuity cards gave 5.58 cycles per degree while cortical recording gave 8.97. The visual cortex was responding to detail the babies were not yet demonstrating. Newborn acuity is lower still, and the clearest zone sits at about the span between a feeding baby and a parent’s face. That close working range puts the sharpest part of her world exactly where you are.
Does vision develop in the eye or in the brain?
Overwhelmingly in the brain. The eye collects light from the start, but the cortex has to learn what the collection means, and that learning is competitive. Studies of ocular dominance columns in normal and visually deprived monkeys showed cortical territory developing according to the input each eye actually delivered. Vision is assembled from experience rather than installed and switched on. This is why ordinary looking, reaching and unhurried face-to-face time are not idle moments. They are the experience the visual brain is written from.
Why does a turned eye need attention early?
Because the cortex is dividing its territory during a window that closes. Unequal input while that allocation is happening changes how the cortex apportions itself between the two eyes, which is a different problem from how the eye looks. That window closes through perineuronal nets condensing in visual cortex, so timing genuinely matters. Early roving, or a brief crossing that comes and goes, is ordinary in the first weeks. A constant inward or outward turn belongs with an eye specialist promptly rather than at the next routine visit.
When should my baby start tracking and using both eyes together?
Following comes before teaming. In the early months most babies fix on a face and follow a slow target across the midline, coordinated by the oculomotor, trochlear and abducens nerves through the brainstem and cerebellum. Reliable convergence on a near target, and the depth perception that depends on it, arrive after that. Roving or brief crossing while the system is still yoking the two eyes is ordinary. A turn that is constant rather than occasional is worth a specialist’s look.
Why is my young child unsteady in busy or patterned places?
Because young children weight vision heavily and cannot easily discount it. Children under about seven and a half years could not suppress visual input that was reporting false orientation, even with adult-like automatic postural responses. When four-, eight- and twelve-year-olds stood inside an oscillating room, the older children and adults down-weighted the misleading visual signal more than the younger ones did. A busy floor pattern or a dim room is therefore genuinely harder for a young nervous system, and it is not clumsiness.
What does a visit for a baby involve, and what has it to do with vision?
The looking is read first: how she fixes on a face, whether she follows a slow target, and whether the two eyes move as a pair. Then resting head position, rotation to each side, and palpation of the occiput and the cranial base. The neck belongs to vision because holding an image still while the head moves runs on head-position information. The contact is a sustained light hold, no heavier than what you could rest on a closed eyelid. There is no twist and no thrust.
What does the Unified Model of Tone say about vision?
That vision is one channel whose weight in the whole organization is set by experience, rather than a separate faculty. The reweighting data are the evidence: a five-year-old and an adult receive comparable visual information and treat it very differently. From this the model predicts that changing the quality of head-position and movement information can change measurable visual behavior, because the channels are resolved together rather than separately. That is a claim about how looking is organized, and it is testable in a single cohort.
12The sources
References
9 primary sources, each linked to its PubMed record. Figures quoted on this page were checked against the published abstract.
Related evidence