Cortical astrocytes are star-shaped brain cells that tile the outer layers of your cerebral cortex, the folded sheet of tissue responsible for thought, sensation, movement, and language. Once dismissed as mere structural filler between neurons, these cells are now recognized as active participants in nearly every aspect of brain function, from shaping synaptic communication and regulating blood flow to clearing metabolic waste during sleep. In humans, they are remarkably large and complex compared to those in other mammals, hinting that astrocytes may be part of what makes our brains distinctive.
Where Cortical Astrocytes Come From
During embryonic brain development, a population of stem cells called radial glia serve as the scaffolding for the growing cortex. These elongated cells stretch from the brain’s inner ventricles to the outer surface, and migrating neurons climb along them like vines on a trellis. Radial glia also divide to produce new neurons and, eventually, new glial cells. Toward the end of development, the radial glia themselves transform into astrocytes, retracting their long processes and settling into their final positions throughout the cortical layers.1Genes & Development. Reinduction of ErbB2 in astrocytes promotes radial glial progenitor identity in adult cerebral cortex In human fetuses, this transformation proceeds through the later stages of gestation and into the period around birth, coinciding with the final waves of neuronal migration.2PubMed. Cortical radial glial cells in human fetuses: depth-correlated transformation into astrocytes Growth factors like TGF-β1 help drive this transition, nudging radial glia toward an astrocyte fate rather than continuing to produce neurons.3PubMed Central. TGF-β1 promotes cerebral cortex radial glia-astrocyte differentiation in vivo
This developmental origin matters because it means astrocytes are not just passive bystanders that arrive after the cortex is built. They share a lineage with the neurons they support, and their placement across the cortical layers is determined by the same developmental programs that organize neuronal circuits.
Not All the Same Across Cortical Layers
The cerebral cortex is organized into roughly six layers, each with different types of neurons handling different jobs. Astrocytes follow a similar logic. Rather than being a uniform population, they vary in shape, orientation, and gene expression depending on which layer they occupy. In the mouse somatosensory cortex, astrocytes in upper layers have different territorial volumes, branching patterns, and cell orientations compared to those in deeper layers.4Nature Communications. Layer-specific morphological and molecular differences in neocortical astrocytes and their dependence on neuronal layers
At the molecular level, this layered organization shows up clearly in gene expression. A transcriptomic mapping study found that certain genes mark upper-layer astrocytes while others mark deep-layer ones. For instance, the gene Chrdl1 is expressed in astrocytes of layers 2 through 4, while Interleukin-33 concentrates in layers 5 and 6.5PubMed Central. Astrocyte layers in the mammalian cerebral cortex revealed by a single-cell in situ transcriptomic map These molecular differences likely shape how astrocytes in each layer interact with their neighboring neurons, tailoring their function to the local circuitry. An astrocyte sitting among sensory input neurons in layer 4 faces different demands than one nestled among the output neurons of layer 5.
How Human Cortical Astrocytes Differ from Those in Other Species
If you compared a human cortical astrocyte to a mouse cortical astrocyte under a microscope, the size difference alone would be striking. Human protoplasmic astrocytes, the most common type in the cortex’s gray matter, have roughly three times the diameter and about ten times as many primary branches as their rodent counterparts.6PubMed. Astrocytic complexity distinguishes the human brain Each one covers a much larger territory, meaning a single human astrocyte contacts far more synapses than a mouse astrocyte does.
Beyond size, humans and other primates have astrocyte subtypes that simply do not exist in rodents in any meaningful form. Interlaminar astrocytes, found in layer 1 of the cortex, send long fibers that can stretch for hundreds of micrometers into deeper layers. In humans, these processes average about 593 micrometers in length and sprout nearly 28 branches on average, compared to roughly 133 micrometers and about 6 branches in mice.7Frontiers in Neuroanatomy. Human and mouse cortical astrocytes: a comparative view from development to morphological and functional characterization Humans also possess polarized astrocytes in layers 5 and 6 that project long processes, another feature unique to primates.8PubMed. Astrocytic complexity distinguishes the human brain What these primate-specific subtypes actually do remains an open question, but their existence suggests that astrocyte complexity scaled up alongside the expansion of the human cortex.
Keeping Synapses Running Cleanly
Neurons communicate at synapses by releasing chemical messengers, and glutamate is the most common excitatory one in the cortex. After glutamate does its job, it needs to be cleared from the synaptic space quickly. If it lingers, it can overstimulate neurons and damage them, a process called excitotoxicity. Cortical astrocytes handle the bulk of this cleanup through specialized transporter proteins known as EAAT1 and EAAT2, which pull glutamate out of the synapse and into the astrocyte.9PubMed Central. The Regulation of Astrocytic Glutamate Transporters in Health and Neurodegenerative Diseases Without these transporters working properly, extracellular glutamate accumulates, and the consequences range from subtle synaptic dysfunction to outright neurotoxicity.10PubMed Central. The role of astrocytic glutamate transporters GLT-1 and GLAST in neurological disorders: Potential targets for neurotherapeutics
But astrocytes do more than mop up after neurons. There is strong evidence that they also release their own signaling molecules, sometimes called gliotransmitters, that can tune synaptic activity. One of the best-studied is D-serine, a small amino acid that acts as a co-activator of NMDA receptors, a key type of receptor involved in learning and memory.11Frontiers in Cellular Neuroscience. D-serine as a gliotransmitter and its roles in brain development and disease In mouse somatosensory cortex, astrocyte-released D-serine has been shown to be necessary for a specific form of synaptic plasticity at connections between neurons in layer 2/3, directly shaping how those circuits change with experience during development.12PubMed Central. Developmental Spike Timing-Dependent Long-Term Depression Requires Astrocyte d-Serine at L2/3-L2/3 Synapses of the Mouse Somatosensory Cortex Astrocytes also store and release glutamate itself through vesicle-based mechanisms, providing another channel through which they can influence nearby synapses.13Journal of Neuroscience. Storage and Uptake of d-Serine into Astrocytic Synaptic-Like Vesicles Specify Gliotransmission
Calcium Signals as the Astrocyte Language
Neurons communicate with rapid electrical impulses. Astrocytes do not fire action potentials, so they rely on a different signaling system: waves of calcium ions flowing within the cell. When nearby neurons are active or when neuromodulators arrive, astrocytes respond with increases in internal calcium that can spread along their branches. The fine processes of an astrocyte, the thin tentacles that reach toward synapses, display fast calcium events roughly 50 to 100 times more frequently than the cell body does.14PubMed Central. Calcium Signals in Astrocyte Microdomains, a Decade of Great Advances Many of these tiny, local calcium flickers arise spontaneously and involve mitochondria within the astrocyte’s branches, independent of the classical signaling pathways researchers initially expected.15Neuron. Astrocytes Retain and Propagate Ca2+ in Microdomains Mediated by Mitochondrial Permeability Transition
Three-dimensional imaging of entire astrocytes in awake mice has revealed that calcium activity is scattered across the cell, largely compartmentalized between different regions, and much higher during wakefulness than in brain slices removed from the animal.16Science. Three-dimensional Ca(2+) imaging advances understanding of astrocyte biology Astrocytes can even respond locally to minimal firing from a single nearby axon, suggesting they track neural activity with surprising spatial precision. These calcium signals are what trigger many downstream astrocyte functions, from releasing gliotransmitters to adjusting local blood flow.
Coordinating Blood Flow With Neural Demand
Active neurons need more oxygen and glucose, and the brain solves this by rapidly increasing blood flow to busy regions. Cortical astrocytes are physically positioned to orchestrate this: their cell bodies sit among neurons, while specialized structures called endfeet wrap around nearby blood vessels. When neurotransmitters from active neurons trigger calcium increases in an astrocyte, the endfoot releases chemical signals onto the vessel wall. Some of these signals, like prostaglandin E2 and potassium ions, cause vessels to dilate and increase flow. Others, like 20-HETE, cause constriction.17PubMed Central. Astrocyte regulation of blood flow in the brain The balance between dilation and constriction depends on local metabolic conditions, giving astrocytes a kind of thermostat role for cortical blood supply.
This process is relevant to brain imaging as well. Functional MRI measures changes in blood flow to infer which brain regions are active. Because astrocytes are major intermediaries between neural activity and vascular response, the signal researchers see on a brain scan is partly an astrocyte-mediated signal, not a pure readout of neuronal firing.
Clearing the Brain’s Waste During Sleep
The brain produces metabolic waste constantly, including amyloid-beta, the protein fragment that accumulates in Alzheimer’s disease. A waste-clearance system called the glymphatic pathway flushes cerebrospinal fluid along the spaces surrounding arteries, through brain tissue, and back out along veins. Astrocyte endfeet are central to this process because they express water channels called aquaporin-4 (AQP4) that facilitate the flow of fluid between blood vessels and brain tissue.18PubMed Central. Loss of aquaporin-4 results in glymphatic system dysfunction via brain-wide interstitial fluid stagnation When AQP4 is absent or mislocalized in mice, glymphatic exchange slows down, amyloid-beta clearance drops, and plaque formation accelerates.19PubMed Central. Loss of perivascular aquaporin-4 localization impairs glymphatic exchange and promotes amyloid β plaque formation in mice
Astrocytes also participate in the brain’s sleep physiology more directly. By releasing adenosine, a molecule that promotes sleepiness, cortical astrocytes help build up the pressure to sleep that accumulates during wakefulness.20PubMed Central. Astrocytes and the modulation of sleep In mice genetically engineered so that astrocytes could not release signaling molecules through their normal vesicle pathway, the buildup of sleep pressure was blunted: slow-wave activity during sleep was reduced, and the animals were less cognitively impaired by sleep deprivation than normal mice, because the sleep-debt signal itself was weaker.21PubMed Central. Astrocytic modulation of sleep homeostasis and cognitive consequences of sleep loss The effect worked through adenosine receptors, confirming that astrocyte-derived adenosine is a key piece of the circuit that tells the brain it needs rest.
Pruning and Remodeling Synapses
The developing brain produces far more synaptic connections than it ultimately keeps. Excess synapses need to be eliminated so that circuits can be refined, and this pruning was long attributed mainly to microglia, the brain’s resident immune cells. But cortical astrocytes actively engulf and digest synapses too. They do this using two receptor pathways, MEGF10 and MERTK, and the process depends on how active the targeted synapses are: less active connections are preferentially eaten.22PubMed Central. Astrocytes mediate synapse elimination through MEGF10 and MERTK pathways Mice lacking both of these pathways in astrocytes fail to prune their connections normally and retain too many functional synapses into later development.
This pruning role does not end after childhood. Astrocytes continue to remodel synapses in the adult brain, and the process can be hijacked under stress. Stress hormones have been shown to ramp up astrocyte-mediated synapse engulfment, contributing to behavioral changes.23Immunity. Stress hormones induce astrocyte-mediated synapse phagocytosis and behavioral deficits The APOE gene, famous for its link to Alzheimer’s risk, also influences how efficiently astrocytes clear aging synapses. The APOE4 variant, which carries the highest Alzheimer’s risk, appears to impair astrocytic phagocytic capacity, potentially allowing damaged synapses to linger and become targets for a more destructive immune cascade.24PubMed Central. Novel allele-dependent role for APOE in controlling the rate of synapse pruning by astrocytes
Responding to Brain-Wide Arousal Signals
Cortical astrocytes do not just respond to local synaptic activity. They also tune in to brain-wide neuromodulatory signals, particularly norepinephrine released from a brainstem structure called the locus coeruleus. When the locus coeruleus fires, whether in response to a startling stimulus or simply increased alertness, norepinephrine floods across large stretches of the cortex. Astrocytes throughout the cortex respond almost simultaneously with fast calcium transients, driven by activation of a type of norepinephrine receptor on their surface.25PubMed Central. α1-Adrenergic receptors mediate coordinated Ca2+ signaling of cortical astrocytes in awake, behaving mice These calcium waves are independent of the glutamate-based signaling from local neurons, representing a separate channel through which astrocytes receive information about the animal’s behavioral state.26Cerebral Cortex. Locus Coeruleus α-Adrenergic–Mediated Activation of Cortical Astrocytes In Vivo
This means cortical astrocytes simultaneously track two kinds of information: fine-grained local activity at individual synapses, and global arousal state broadcast by neuromodulatory systems. How they integrate these signals and what that integration means for cortical processing is one of the more intriguing open questions in the field.
When Astrocytes Turn Reactive
After a brain injury like a stroke, cortical astrocytes near the damaged area undergo dramatic changes. They swell, multiply, alter their gene expression, and form a dense scar around the injury site, a process called reactive astrogliosis. In the cortex surrounding a stroke, a chromatin-remodeling protein called Chd7 is switched on in proliferating reactive astrocytes within a day of the injury, and by two weeks after stroke, nearly 90% of the scar-forming astrocytes express it.27Cell Press (Cell Reports). The chromatin remodeler Chd7 regulates reactive astrogliosis and neuroinflammation after ischemic stroke The glial scar has a complicated reputation: it walls off the injury and limits the spread of toxic molecules, but it also creates a physical and chemical barrier that can prevent axons from regrowing.
The type of reactive state an astrocyte enters matters. Research has identified a harmful reactive profile, originally labeled A1, that is driven by signals from activated microglia. When microglia detect damage or infection, they secrete a trio of inflammatory molecules (IL-1α, TNF, and C1q) that push nearby astrocytes into this neurotoxic state.28PubMed Central. Neurotoxic reactive astrocytes are induced by activated microglia A1-type astrocytes lose many of their normal supportive functions and gain the ability to kill neurons and other brain cells. This microglia-to-astrocyte signaling cascade has been implicated in Alzheimer’s disease, Parkinson’s disease, and other neurodegenerative conditions.29PubMed Central. Astrocytic and microglial cells as the modulators of neuroinflammation in Alzheimer’s disease The original A1/A2 classification has since been recognized as an oversimplification, with reactive astrocytes actually falling along a spectrum of states, but the core finding that microglia can flip astrocytes into a damaging mode remains well supported.
Cortical Astrocytes and Epilepsy
Epileptic seizures arise from excessive, synchronized electrical activity in neural circuits, and astrocyte dysfunction can push circuits in that direction. One well-studied mechanism involves potassium buffering. Neurons release potassium ions into the extracellular space when they fire, and astrocytes normally absorb this excess through channels called Kir4.1. When these channels are impaired, whether by genetic mutation or reduced expression, extracellular potassium builds up alongside glutamate, making neurons abnormally excitable.30PubMed Central. Role of Astrocytic Inwardly Rectifying Potassium (Kir) 4.1 Channels in Epileptogenesis This potassium-buffering failure can contribute to the development of epilepsy over time, and Kir4.1 channels are being explored as potential therapeutic targets.
Astrocytes in Motor Learning
The motor cortex, the region that plans and executes voluntary movements, relies on astrocytes during skill acquisition. As an animal learns a new motor task, synapses in the motor cortex reorganize, and astrocytes participate in this reorganization. Manipulating astrocyte function in the mouse motor cortex disrupts both the synaptic plasticity that accompanies learning and the stable patterns of neuronal activity that encode newly learned movements.31Journal of Neuroscience. Differential Effects of Astrocyte Manipulations on Learned Motor Behavior and Neuronal Ensembles in the Motor Cortex The finding reinforces a broader point: cortical astrocytes are not passive witnesses to neural circuit changes but active enablers of them. When astrocytes are experimentally silenced or overstimulated, the neuronal ensembles that drive skilled behavior lose their coherence.
The Debate Over How Much Astrocytes Really Do
It is worth noting that the field of astrocyte biology has had its share of contentious debates. The concept of gliotransmission, the idea that astrocytes release signaling molecules in a regulated way to influence synapses, was challenged for years by researchers who argued that some of the observed effects could be experimental artifacts or that the calcium signals in astrocytes were too slow and diffuse to meaningfully regulate fast synaptic transmission. Newer tools have largely resolved this in favor of gliotransmission being real, at least for molecules like D-serine, where the evidence for astrocyte release and synaptic impact is now quite direct. But the scope and physiological importance of astrocyte-to-neuron signaling in everyday cortical function, as opposed to under experimental conditions, remains an area where researchers hold genuinely different views.
Similarly, the direction of metabolic fuel shuttling between neurons and astrocytes has been debated for decades. The classical hypothesis proposed that astrocytes take up glucose, convert it to lactate, and shuttle that lactate to neurons as fuel. However, modeling of human visual cortex data during prolonged stimulation found the measured lactate changes were actually consistent with lactate flowing the opposite direction, from neurons to astrocytes, when standard glucose transport values were used.32PubMed Central. The in vivo neuron-to-astrocyte lactate shuttle in human brain: evidence from modeling of measured lactate levels during visual stimulation The traditional astrocyte-to-neuron shuttle could only be reproduced in the model by assuming a twelve-fold increase in astrocytic glucose transport capacity, a condition not supported by existing measurements. This does not mean the traditional model is definitively wrong in all contexts, but it illustrates that basic questions about what cortical astrocytes do metabolically are still being worked out.