The Brain Is Less Self-Contained Than We Thought

This week

This week’s neuroscience has a strikingly outward-looking theme. The brain is not simply a sealed organ that develops according to its own internal blueprint, runs on a fixed fuel, and keeps the same resident immune cells for life. Signals from the thalamus help decide which human cortical neurons are born. Cells from the blood may slowly enter and replace part of the brain’s immune population in old age. Gut molecules can influence emotional behavior; peripheral nerve metabolism can shape chronic pain; and a psychedelic compound may protect nerves damaged by chemotherapy.

At the same time, several studies complicate the usual divisions between perception, memory, and action. The hippocampus appears to make vision more precise by suppressing distracting remembered locations. The maps in early visual pathways can be redistributed by lifelong deafness. And even a familiar emotion-regulation technique—putting feelings into words—depends unexpectedly on the language in which those words are spoken.

Development is a conversation, not a construction project

Cerebral organoids have often been treated as miniature, self-organizing versions of developing cortex. A Science study makes a persuasive case that this picture is incomplete. When researchers fused human cortical organoids with thalamic ones, axons arriving from the thalamus increased the production of cortical excitatory neurons, especially the upper-layer neurons that support much of the cortex’s long-range communication. The key was physical contact between thalamic axons and outer radial glia, mediated by thalamic neurexin-1 (NRXN1). Removing NRXN1 reduced that contact and blunted excitatory neurogenesis. [1]

The finding is important not because it makes organoids less useful, but because it makes development more interesting. The cortex is not merely generating cell types on an internal schedule. It is being instructed by an incoming partner while it is still building itself. Another developmental study finds that timing is encoded even within cortical progenitors: stage-specific phosphorylation of transcription factors shifts cells from proliferation toward neuronal maturation and axon growth. [2] Meanwhile, a three-dimensional epigenomic atlas of fetal human glia maps the regulatory elements that distinguish radial glia, oligodendrocyte precursors, and microglia—and identifies schizophrenia-associated variants that may disrupt those programs. [3]

The common message is that neural identity emerges through layers of interaction: genes, molecular modifications, neighboring cells, and projections arriving from elsewhere in the brain.

A brain shaped by the body—and age

One of the week’s most consequential findings challenges a foundational assumption about microglia. These immune cells arise early in development and have long been thought to persist for life, largely separate from the circulating immune system. Using naturally accumulated mutations as lineage markers, a Nature study reports that bone-marrow-derived cells colonize and replace at least some human brain immune cells during ageing. [4]

It remains to be seen precisely how widespread this replacement is, what these incoming cells do, and whether it is helpful, harmful, or both. But the result makes the ageing brain look more porous to the body than conventional accounts allowed. That matters for neurodegenerative disease, where immune activity is increasingly central. In tauopathy-model mice, peripheral dendritic cells primed CD8 T cells outside the brain, enabling their later accumulation in the brain and worsening neurodegeneration. [5] In Parkinson’s disease tissue, transposable-element activation in neurons and microglia tracked with a region-specific interferon response in the substantia nigra and putamen. [6]

There are also hints that resilience may be built into neurons themselves. In multiple-sclerosis tissue and experimental models, neurons with higher intracellular complement factor H were more resistant to inflammatory degeneration. Rather than acting through its familiar immune role outside cells, this protein appears to limit oxidative damage and lipid peroxidation within neurons. [7] “Inflammation” is not one thing, these studies suggest; it is an exchange among immune cells, glia, peripheral organs, and vulnerable—or unusually resilient—neurons.

Metabolism becomes a nervous-system story

The most unexpected clinical prospect this week comes from psilocybin. In mouse models, just two doses given before platinum- or taxane-based chemotherapy prevented chemotherapy-induced peripheral neuropathy without reducing anti-tumor effects. The proposed mechanism is not an alteration in perception or mood, but preservation of mitochondrial trafficking within peripheral axons. Psilocybin kept mitochondria distributed where nerves need energy, protected sensory endings in the skin, and also normalized cortical activity altered by chemotherapy. [8]

A few weeks back we looked at psilocybin making human brain dynamics more responsive to immediate context and setting. This result is a striking mechanistic rhyme: the same compound is now implicated in protecting the physical logistics of axons. It is still a mouse finding, not a reason for cancer patients to self-medicate. But given the scarcity of effective preventions for this often irreversible toxicity, it provides a compelling rationale for clinical trials.

Other papers made metabolism similarly hard to relegate to background housekeeping. After peripheral injury in mice, spinal oligodendrocytes diverted resources away from myelin-related programs and toward lipid synthesis. The resulting disruption to axons promoted neuronal amyloid-β42 accumulation during the window in which acute pain becomes chronic; blocking its production prevented persistent pain behavior. [9] A complementary study found that nerve injury and spinal disinhibition awaken a large population of previously silent dorsal-horn neurons, turning them into broad sensory responders and potentially creating the circuitry of touch-evoked pain. [10]

Even the brain’s fuel supply can be surprisingly flexible. Frogs emerging from hibernation can sustain neural activity while largely abandoning glucose metabolism, using ketone bodies produced by astrocytes within the brain itself. [11] The lesson is not that human brains can casually do the same. It is that neural function depends on metabolic arrangements far more inventive than the standard neuron-centered picture suggests.

Memory chooses what perception will see

Memory is often described as a record of the past. This week, it looked more like a tool for filtering the present. In a human memory-guided visual-search task, participants whose hippocampal representations of competing memories were more distinct made more accurate eye movements toward a remembered target. Their visual cortex also showed less preparatory activation for a distracting, competing location. [12] The hippocampus, in other words, may help attention not simply by retrieving the right memory, but by making rival memories less able to colonize perception.

That framework gives added weight to a clinical epilepsy study. Among 20 people undergoing invasive recordings for focal epilepsy, the cumulative burden of seizures and post-seizure dysfunction in the language-dominant hippocampus predicted poorer verbal memory a week later—better than routine seizure counts or structural scans did. [13] Five weeks ago, we covered evidence linking hippocampal-cortical alpha hypersynchrony in temporal-lobe epilepsy to worse delayed memory. This new work sharpens the problem: it may be the accumulated physiological disruption of the hippocampus, not simply how many seizures someone has, that erodes long-term consolidation.

Memory is therefore vulnerable not only when information fails to enter the brain, but when the machinery that separates, stabilizes, and prioritizes representations is repeatedly interrupted.

Experience redraws maps—and words carry history

Early deafness is associated with exceptional peripheral visual sensitivity. Functional imaging now shows a corresponding redistribution of neural territory: deaf adults had a larger representation of far-peripheral space in primary visual cortex and the lateral geniculate nucleus, at the expense of central visual-field representation, without an overall enlargement of either structure. [14] Sensory deprivation does not merely recruit spare cortical capacity. It can apparently reallocate the spatial map at remarkably early stages of the visual system.

Experience also shapes emotion regulation in subtler ways. Labeling a feeling—saying “I’m sad”—reduced amygdala activity in unbalanced Spanish-English and Spanish-Catalan bilinguals only when they used their dominant native language. Balanced bilinguals, unexpectedly, showed no such amygdala reduction in either language. [15] The finding does not mean a second language cannot be emotionally meaningful. But it suggests affect labeling depends on learned linguistic fluency and perhaps on the depth of emotional associations built in a language over a lifetime.

Looking ahead

The week’s studies resist tidy boundaries: between cortex and thalamus, brain and body, neuron and glia, memory and perception, experience and anatomy. That is not merely a philosophical shift. It changes where researchers might look for treatments—at immune cells primed in the periphery, mitochondrial transport in sensory axons, metabolic switches in glia, or the specific physiological burden of seizures.

The larger challenge will be translating these findings without losing their complexity. A psychedelic that protects mouse nerves is not yet a clinical preventive. Blood-derived brain immune cells are not yet a map of ageing or disease. But across development, degeneration, pain, and cognition, the brain increasingly appears not as an isolated command center, but as an organ continuously remade by its connections.