From sensation to perception

The cup looks smaller when you move farther away, yet you do not usually conclude that it has shrunk. A shadow changes the light reaching your eyes without necessarily changing the cup's material. Its handle can disappear behind the body while the object remains recognizable. Vision therefore cannot consist simply of copying an image into the brain. The nervous system must extract useful information from signals whose relationship to the outside world is informative but ambiguous.

Three things to keep separate

A physical stimulus is an event or property that can affect a sensory system: light entering the eye, a vibration reaching the ear, or pressure deforming skin. Neural activity is the resulting activity of cells and circuits. Perception concerns what the person experiences and distinguishes through those processes. These are connected, but they are not interchangeable descriptions. A measured light intensity is not a spike rate, and a spike rate is not itself a verbal report of brightness.

The distinction prevents two opposite mistakes. One is to think that perception is a perfectly faithful photograph of the stimulus. The other is to think that because perception involves processing, the outside world places no constraints on it. A cup has a real position and surface, but the information available about those properties depends on illumination, viewpoint, sensory sensitivity, attention, and prior experience. Reliable perception is an achievement under constraints.

For an original optical example, compare a small cup nearby with a larger cup farther away. They can occupy similar angles at the eye. Angular size alone then cannot tell us both actual size and distance. Other information, such as surrounding objects, perspective, movement, and differences between the eyes, can help. The problem is not that the visual signal is useless. It is that one measurement can be compatible with more than one physical arrangement.

We will follow vision because its pathways make the organizational problem clear. This is a selected route, not a claim that hearing, smell, touch, and internal sensation follow identical anatomy. The later course on the eye examines its optical and retinal structures in greater detail. Here the emphasis is what happens when sensory information enters the brain's connected processing systems.

The retina is already neural tissue

Transduction means converting a stimulus into a biological signal. In the retina, light-sensitive cells change their electrical and chemical behavior in response to absorbed light. Retinal circuits transform those changes before output leaves the eye. The retina is therefore not passive film with an optic cable attached. Its organization already contributes to what information is transmitted.

The axons of retinal ganglion cells form the optic nerve. A major route continues through the optic chiasm, the optic tract, and the lateral geniculate nucleus, or LGN, in the thalamus. From the LGN, axons travel in the optic radiations toward primary visual cortex, often called V1, in the occipital lobe. Other retinal projections contribute to functions such as orienting and regulation of responses to light. The LGN route is central to our account, but it is not every visual connection.

A nerve, a tract, a nucleus, and a cortical area are different anatomical categories. The nerve and tract contain axons carrying signals; the nucleus contains organized groups of neuronal cell bodies and their connections; cortex is layered tissue. Reciting their names in sequence is only the beginning. At each synaptic stage, cells transform inputs rather than simply passing an unchanged photograph to the next station.

A retinotopic map preserves an orderly relationship to positions represented on the retina. Neighboring locations in visual space are represented in an organized pattern in parts of the pathway. This is not a miniature screen waiting for an internal spectator. It is a spatial arrangement of neural responses. The reference on visual-field representation connects this organization to the crossing of optic fibers and the mapping in visual cortex.

Left field is not left eye

Draw a vertical line through a point you are looking at. The space on the left of that line is the left visual hemifield; the space on the right is the right hemifield. Much of the central field is visible to both eyes. Consequently, “information from the left side of the scene” and “information from the left eye” are different categories.

At the chiasm, axons from the nasal half of each retina cross, while those from the temporal half remain on the same side. This sorts information so that the right optic tract carries the left visual hemifield and the left optic tract carries the right hemifield. Each tract includes contributions from both eyes. The arrangement is easier to understand by following one point in space than by trying to memorize an X-shaped picture.

Place the cup just to the left of fixation, within the region visible to both eyes. Its light reaches the nasal part of the left retina and the temporal part of the right retina. The first contribution crosses at the chiasm; the second does not. They therefore travel toward the right side of the brain in the pathway under discussion. Move the cup to the other side of fixation and the corresponding logic reverses. This exercise explains why the chiasm is a partial crossing rather than an exchange of two complete eye cables.

The map is also uneven in its allocation of tissue. Central vision receives a disproportionately large representation in early visual cortex. Equal distances on a retinal drawing do not imply equal areas of cortex. A map can preserve order while changing scale, much as a transit map can preserve the order of stations while distorting the distances between them. The comparison explains organization, not the detailed shape of cortical tissue.

For a fictional disconnection exercise, cut the left optic nerve on your paper. That removes the illustrated output from the left eye before the sorting at the chiasm. Now restore it and instead interrupt the right optic tract. That affects the illustrated route for the left visual hemifield from both eyes. These are predictions from an anatomical model, not a guide to interpreting anyone's visual symptoms. Their educational purpose is to distinguish an eye from a field.

A map does not explain recognition

Knowing where a point is represented does not yet explain how a cup is distinguished from a bowl. Neural responses can be sensitive to relationships such as contrast, orientation, motion, and combinations of features. As information is processed through connected visual areas, increasingly complex relationships become available for recognition and action. The system includes parallel routes and feedback, not simply a row of cells each adding one adjective to a finished object.

A receptive field describes the region of sensory space in which a stimulus influences a particular neuron's response under specified conditions. The response can depend on what happens within that region and on wider context. The term therefore identifies an experimentally characterized relationship, not a little window through which the cell consciously looks. A cell responding more to one feature need not be exclusively dedicated to that feature in all circumstances.

Imagine an original model with three units. Unit A responds strongly to a vertical edge, B to a curved boundary, and C to a particular combination of those inputs. The combination may help distinguish two drawings. Yet this model does not explain how a person recognizes a real cup from every viewpoint. It omits illumination changes, occlusion, motion, material, learning, and many other inputs. A small model is useful when its success and its omissions are stated together.

Now rotate the model cup. The handle's visible contour changes, but the object remains the same. A useful recognition system must tolerate some changes while remaining sensitive to differences that matter. Too little tolerance would make every viewpoint look like a new object. Too much tolerance would confuse different objects. The balance between stability and discrimination is a real computational problem, not an extra label to add after describing V1.

Visual processing also contributes to action. The same scene can support identifying the cup, locating its handle, estimating how to reach around another object, and updating movement as the hand approaches. Ventral visual pathways are strongly associated with object-related processing, while dorsal pathways contribute to spatial and action-related processing. These are interacting specializations, not two isolated systems called “what” and “where” that contain every relevant operation. The cup task requires communication among them and with nonvisual networks.

Attention changes the task

You can look at a crowded shelf while searching for a blue cup or while counting the handles. The light pattern can be similar while the task differs. Attention describes processes that prioritize some information and actions over others. It does not require the eyes to move every time the priority changes. Nor is it an unlimited spotlight that makes everything within its beam equally available.

An original classroom model makes the difference concrete. Suppose a display contains twenty objects, five of them blue and eight with visible handles. A participant is asked either to count blue objects or to count handles. A missed object could reflect difficulty distinguishing its feature, a lapse in prioritizing it, a counting error, or forgetting the total before reporting. The final answer alone does not identify which process failed. A useful experiment asks additional questions or changes one part of the task.

For example, shortening the delay before the answer may reduce a memory demand while leaving the visual display unchanged. Giving one object at a time changes selection demands. Asking for detection of a single handle changes the requirement to maintain a running count. Each variation helps separate explanations, although none turns the task into a measure of only one pure mental faculty. This is why the instructions given to a participant are part of the evidence.

Experience can also shape what is readily recognized. A familiar cup may be identified from a partial outline that would be uninformative for an unfamiliar object. This does not mean expectation can override every sensory constraint. It means interpretation uses information at more than one timescale. Present input, recent context, and established knowledge can contribute differently, and their relative influence can be tested rather than assumed.

The word prediction is often used in accounts of this interaction. At an introductory level, it means that a system's existing state or model can influence how new input is processed. It does not require the person to consciously guess the next image, and it does not settle one complete theory of perception. Different scientific models can agree that feedback matters while disagreeing about its exact implementation.

Detection is not the same as a yes response

Consider a fictional experiment with one hundred trials. A faint mark is present on fifty and absent on fifty. The participant reports seeing it on forty of the present trials and on ten of the absent trials. There are forty hits, ten misses, ten false alarms, and forty correct rejections. Reporting “seen” fifty times is not fifty correct detections; ten of those reports occurred when the mark was absent.

Now imagine that the instructions strongly encourage saying yes whenever uncertain. A different response policy could increase both hits and false alarms without making the visual system better at separating present from absent trials. This is the distinction between sensitivity and a decision criterion, the policy governing which evidence leads to a response. Our numbers illustrate the distinction; they are not data from a published study or a measurement of any reader's vision.

The distinction matters when claims are made about awareness. A verbal report is valuable evidence, but it also involves memory, language, confidence, and a response rule. Conversely, a successful forced-choice response does not by itself establish the full character of someone's experience. Careful research combines behavioral measures and neural evidence while acknowledging which aspects remain difficult to infer. “The brain responded” and “the participant consciously saw it” are not equivalent observations.

This does not make subjective reports worthless. If the research question concerns experience, reports may be indispensable. The task is to understand what produced the report and what other measurements add. Rejecting all reports would discard relevant evidence; treating them as an exact printout of the stimulus would ignore the processing we have just examined.

From a seen cup to a usable world

We can now return to the original scene with a more precise chain. Light interacts with retinal cells; retinal circuits generate output; a major pathway organizes information by visual field and carries it through the thalamus to cortex; connected networks support distinctions useful for recognition and action. Task and context influence which information is prioritized. A person's description adds another level of evidence rather than replacing the earlier mechanisms.

The chain includes several opportunities for different outcomes. An optical change can alter the stimulus reaching the retina. A pathway interruption can change which information reaches a cortical region. A task change can alter selection while the image remains similar. A response-rule change can alter reported detection without an equivalent change in sensory discrimination. Calling all four “perception changed” may be broadly true, but it does not explain the difference among them.

A strong account names the level at which a claim is made and the evidence that connects it to another level. A visual-field diagram supports a prediction about pathways. A carefully controlled behavior task supports a claim about discrimination or reporting. A neural measurement can add information about activity under those task conditions. None of those observations, alone, is a complete explanation of conscious visual experience. There is substantial established anatomy without a finished theory of everything it enables.

Check your understanding: Why would a change in the number of “I saw it” responses fail, by itself, to show improved visual sensitivity?

Expected answer: The person may have changed their decision criterion, producing more hits and more false alarms. The interpretation requires present and absent trials, correct and incorrect responses, and attention to task demands. A report is connected to sensory processing but also depends on how evidence is used to answer.

Application

Allow 15–20 minutes. Draw the two eyes and the partial crossing at the chiasm. Follow a point just left of fixation from both retinas into the right optic tract; use one color for the point rather than one color for each eye. Write the difference between cutting one illustrated optic nerve and one illustrated optic tract.

Then reproduce the hundred-trial table. Invent a second set of counts with more hits and more false alarms, keeping fifty present and fifty absent trials. Explain why the new counts alone do not establish better sensitivity. Finish by giving one way to reduce the memory demand in the shelf-counting task without changing the objects displayed.