Visual processing begins in the retina with photoreceptors (rods and cones), passes through retinal ganglion cells whose axons form the optic nerve, crosses at the optic chiasm (nasal fibers decussate), and projects to the lateral geniculate nucleus of the thalamus before reaching primary visual cortex (V1). From V1, processing bifurcates: the dorsal 'where/how' stream (to parietal cortex) handles spatial location and visually guided action; the ventral 'what' stream (to temporal cortex) supports object recognition and face processing. This hierarchical organization means damage at any stage produces predictable, localized deficits.
Draw the full pathway from eye to cortex, including the chiasm crossing and what this means for visual field representation. Then work through specific deficits — hemianopia, prosopagnosia (ventral stream), optic ataxia (dorsal stream) — to test understanding.
You already know that sensory systems convert physical energy into neural signals and route them through hierarchical pathways to cortex. Vision follows this logic, but with an unusual detour: the routing is organized by location in visual space, not by which eye is looking. The process begins in the retina — a sheet of neural tissue at the back of the eye. Photoreceptors (rods for low-light/peripheral vision, cones for color and acuity) transduce light into graded potentials, which are processed locally by bipolar and amacrine cells before converging on retinal ganglion cells (RGCs). The axons of RGCs collect into the optic nerve — the only output channel from the eye to the brain.
The two optic nerves meet at the optic chiasm beneath the hypothalamus, where a partial crossing occurs. Fibers from the *nasal* half of each retina (which see the temporal visual field) cross to the opposite hemisphere; fibers from the *temporal* retina stay ipsilateral. The result is that all visual information from your left visual field — regardless of which eye it entered — ends up in your right hemisphere, and vice versa. This is the key organizational principle: the brain maps visual *space*, not visual *organs*. After the chiasm, the optic tracts continue to the lateral geniculate nucleus (LGN) of the thalamus — the relay station that gates and preprocesses signals before passing them to cortex.
From the LGN, projections reach primary visual cortex (V1) in the occipital lobe. V1 neurons are tuned to local features: edge orientation, spatial frequency, direction of motion, and binocular disparity for depth. V1 does not "see" objects — it extracts oriented contrasts. Object perception is assembled across many subsequent cortical areas. From V1, processing bifurcates into two streams. The dorsal "where/how" stream projects toward parietal cortex, handling spatial localization, depth, motion, and visually guided action. The ventral "what" stream projects toward inferotemporal cortex, supporting object recognition, face perception, and visual memory. These two streams are computationally distinct: one answers "where is it and how do I act on it?" while the other answers "what is it?"
Because the pathway is anatomically well-mapped and hierarchical, damage at any stage produces a predictable, localizable deficit. Cutting the optic nerve before the chiasm causes monocular blindness. Cutting the chiasm itself (a classic consequence of pituitary tumors pressing upward) severs the crossing nasal fibers, causing bitemporal hemianopia — loss of the peripheral visual field in both eyes. Damage to one optic tract causes homonymous hemianopia — loss of the same visual field half in both eyes. Selective damage to the ventral stream causes prosopagnosia (intact acuity, but inability to recognize faces) without blindness, while dorsal stream damage causes optic ataxia (inaccurate reaching) despite normal object recognition. This lesion logic is not just a memorization exercise — it is the pathway's anatomy making predictions, and the predictions are confirmed by the clinical record.