Lecture 11: Development, Nature and Nurture II

What this lecture covers

This lecture continues the previous one's investigation of nature versus nurture, moving from face perception to the broader question of how the brain's long-range wiring shapes where specialized regions end up. It asks whether pre-existing structural connectivity, rather than experience alone, determines the location of functional brain regions, and tests this with diffusion tractography in humans, a classic rewired-ferret experiment, the human visual word form area, spatial navigation cells in rodents, chicks raised without geometric experience, and what happens to the brain after damage or congenital blindness.

By the end you'll understand how diffusion MRI infers long-range brain connections, why the visual word form area is a rare case where selectivity must be learned, why head-direction and reorientation abilities look innate in animals, and why the question of whether brain function can relocate after damage or altered sensory experience remains unresolved. The lecture explicitly revisits and partly rehabilitates Kant's claim that some structure, particularly for space, is prior to experience.

Key ideas

• Connectivity fingerprints: diffusion tractography can measure a brain region's pattern of long-range connections, and in adults this fingerprint predicts where face-, scene-, and body-selective regions will be located.
• Rewired ferrets: surgically redirecting visual input into what would be primary auditory cortex at birth causes that region to develop visual-cortex-like orientation columns and drives the animal to report seeing, not hearing, the input.
• The visual word form area: because reading is too recent an invention for evolution to have built dedicated circuitry, a brain region's selectivity for written words and letters must come from experience, and its response depends on which alphabets a person has learned to read.
• Connectivity can predict function before it exists: a child's brain connectivity at age five predicts the exact location of their visual word form area at age eight, before they could read.
• Innate spatial cells: head-direction cells are present in rat pups the moment they first leave the nest, before any navigation experience, supporting Kant's claim about an innate representation of space.
• Reorientation without geometric experience: chicks raised in darkness or in a featureless cylindrical cage can still use the shape of a rectangular space to reorient themselves on their very first exposure to it.
• Limited reorganization after brain damage: adults rarely recover lost visual or language functions after brain damage, though children who lose left-hemisphere language regions early can often shift function to the right hemisphere.
• Congenital blindness reshapes cortex: visual cortex in blind people becomes active during language tasks and even causally supports language processing, and some category-selective regions show a weak, contested resemblance to sighted category maps when tested with sound instead of sight.

Walkthrough

Recap and the connectivity hypothesis (0:10)

Kanwisher recaps the previous lecture's conclusion that little about face perception looks strongly innate, then introduces the open question of how face patches consistently land in the same brain location across people. She previews two candidate explanations: an innate bias toward simple visual features like curvature, and pre-existing long-range structural connectivity, which becomes this lecture's main thread.

Diffusion tractography and connectivity fingerprints (4:12)

The lecture explains how diffusion MRI infers the direction of water diffusion along myelinated axons to reconstruct probable long-range fiber pathways, and how this "connectivity fingerprint" of a brain voxel can be used to predict, from diffusion data alone, whether that voxel belongs to a face-, scene-, or body-selective region in adults. The method has clear limitations, such as failing where fiber tracts cross, but its predictions are surprisingly accurate.

The rewired ferret experiment (14:17)

Because rodent- and ferret-like animals are born before their eyes open, researchers can surgically redirect retinal input into the pathway leading to primary auditory cortex. The rewired cortex develops visual-like orientation columns, and behavioral testing shows the animal reports seeing rather than hearing when that region is stimulated. This demonstrates that both connectivity and the experience delivered through it can causally determine what a cortical region does.

The visual word form area as a test case for learning (24:25)

Because writing is a recent cultural invention, any brain region selective for visually presented words cannot be innate. The lecture walks through the evidence: a small left-hemisphere region responds more to words than pictures, responds similarly to unpronounceable consonant strings (ruling out a "word" interpretation in favor of letter processing), and responds more strongly to Hebrew script in people who can read Hebrew than in those who cannot. A follow-up study scanning the same children at ages five and eight shows this region is not word-selective before children learn to read, and that connectivity data collected at age five predicts where the region will develop by age eight.

Innate spatial representations in rodents and chicks (40:34)

Returning to Kant's claim about an innate representation of space, the lecture reviews head-direction, place, border, and grid cells from earlier lectures and shows that head-direction cells are present in rat pups as soon as they first leave the nest, before they have any navigation experience. A separate line of work with newly hatched chicks, raised either in darkness or in a symmetrical cylindrical cage with no geometric cues, shows they can still use the shape of a rectangular test space to reorient themselves on their very first exposure, indicating this ability does not depend on prior experience.

Reorganization after brain damage and blindness (57:39)

The lecture surveys what happens when brain regions are damaged or deprived of their normal input. Adults with brain damage rarely recover lost visual or language functions, following what's called the Kennard principle (early damage recovers better) for language, though the "head principle" argues this only holds when the damaged region isn't a necessary precursor for other functions. A case study of a patient with early bilateral damage to face-processing regions shows persistent, severe difficulty recognizing faces even after a lifetime of exposure. Finally, studies of congenitally blind people show visual cortex becoming active, and even causally necessary, during language tasks, and a contested study finds weak evidence that face-, body-, and scene-related sound categories activate a spatial pattern resembling the sighted visual category map, directly conflicting with the face-deprived monkey findings from the previous lecture.

Before you watch

• Watch "Lecture 10: Development, Nature and Nurture I," since this lecture opens by directly recapping its conclusions.
• Review the earlier lectures on navigation and spatial cells (place cells, grid cells, head-direction cells, and reorientation using the shape of space).
• Be familiar with functional localizers and MVPA as methods, since the connectivity-prediction studies build on that logic.

Check your understanding

1. What does a "connectivity fingerprint" mean, and how was it used to predict brain function in adults?
2. Why does the visual word form area provide unusually strong evidence that a brain region's selectivity is learned rather than innate?
3. What did the rewired ferret experiment reveal about the relative roles of connectivity and experience in determining cortical function?
4. Why do head-direction cells and chick reorientation behavior support Kant's idea of an innate representation of space?
5. How does the evidence on brain reorganization after early damage or congenital blindness complicate a simple story of fixed cortical function?