Some neuroscientists think a healthy brain works like a system poised on a knife edge, neither too quiet nor too wild.1PubMed. Metastability, criticality and phase transitions in brain and its models That balance isn’t fixed at birth. The brain keeps retuning itself, which may tie the idea to how and when we learn, including why some kinds of learning come easily in childhood and harder later. It remains a hypothesis, and researchers still argue over whether the edge is truly required for thought.2Trends Open. A critical assessment of the brain criticality hypothesis
What It Means for a Brain to Sit on the Edge
Take a model network of neurons and turn a dial that controls how strong inhibition is. When inhibition is weak, a small burst of activity quickly amplifies and spreads.3PubMed Central. Addressing skepticism of the critical brain hypothesis When inhibition is strong, the same burst is snuffed out almost at once.4PubMed Central. Addressing skepticism of the critical brain hypothesis At an in-between setting, activity fades only very slowly, which is the signature of a tipping point, or criticality.5PubMed Central. Addressing skepticism of the critical brain hypothesis
The brain’s fingerprint of this state is the neuronal avalanche. Recordings show cascades of firing that are mostly small but occasionally huge.6PubMed Central. Statistical Analyses Support Power Law Distributions Found in Neuronal Avalanches Their sizes follow a regular pattern with no typical size, limited only by how much brain is recorded.7PubMed Central. Statistical Analyses Support Power Law Distributions Found in Neuronal Avalanches That pattern is what physics predicts for a system at a tipping point.8PubMed Central. Hierarchical networks, power laws, and neuronal avalanches Such signals appear in animals from leeches to humans, and across sleep and waking.9PubMed Central. Self-organized criticality as a fundamental property of neural systems
In practice, researchers count how many avalanches of each size occur in a recording and plot the result. Small events are common and large ones rare, in a smooth, steady relationship. A too-quiet brain would show mostly tiny events and a runaway brain mostly giant ones, so the smooth mix is what stands out.
Why the Edge Could Help You Think
Brain tissue that produces avalanches responds to the greatest range of input strengths, from faint to strong.10PubMed Central. Neuronal avalanches imply maximum dynamic range in cortical networks at criticality In living animals, the cortex’s range of response peaked when its state sat right at the tipping point and fell on either side.11PubMed Central. Maximizing Sensory Dynamic Range by Tuning the Cortical State to Criticality A quiet network misses faint signals, while an overexcited one saturates on everything.
Your senses constantly face inputs of wildly different strengths, such as a whisper in a quiet room and a siren on the street. Sitting at the tipping point would let the same tissue handle both without being retuned each time. This is a claim about how tissue responds, and the tests were done in tissue slices and animals, not in people judging sounds.
Balanced networks at the edge can also be both highly sensitive and reliable in how they respond.12PubMed. Critical Avalanches in Excitation-Inhibition Balanced Networks Reconcile Response Reliability with Sensitivity for Optimal Neural Representation Sensitivity means a small input produces a detectable change in activity. Reliability means the same input gives a similar response each time rather than random variation. The two usually trade off, since a system tuned to catch faint signals tends to be jumpy and a steady one tends to be dull. Edge-balanced networks are proposed to ease that trade-off. For everyday thinking, studies link edge-like signals to working memory and reasoning, but this is still being tested.
That is why the idea attracts chip designers too. Computer networks reach peak information richness close to the tipping point, so engineers are exploring edge-tuned networks for low-power computer chips, working with as few as 100 simulated neurons.13PubMed Central. A mean-field approach to criticality in spiking neural networks for reservoir computing
How the Brain Keeps Its Balance
Edge-like avalanches and brain rhythms both emerge when excitation and inhibition are balanced.14PubMed Central. Critical-state dynamics of avalanches and oscillations jointly emerge from balanced excitation/inhibition in neuronal networks That balance is what sets up the tipping point in cortical tissue.15PubMed Central. Neuronal avalanches imply maximum dynamic range in cortical networks at criticality
Excitatory neurons push their neighbors toward firing, while inhibitory neurons push their neighbors toward staying quiet. Neither is the “good” side: too much excitation leads toward runaway activity, and too much inhibition toward silence. Brain rhythms, the waves seen in recordings, arise from the back-and-forth between the two, which is why rhythms and avalanches can share a cause. The hypothesis cares about the ratio more than the absolute amount of either.
What keeps the ratio right? A leading explanation is plasticity, the brain’s ability to strengthen or weaken connections.16Frontiers in Physics. Self-Organization Toward Criticality by Synaptic Plasticity In models, it builds steady background activity that takes on edge-like features, and it pulls the network back to the edge after disturbances.17PubMed Central. Synaptic Plasticity Enables Adaptive Self-Tuning Critical Networks Rewiring rules modeled on the brain’s self-stabilizing processes make the tipping point a state the network settles into.18PubMed Central. Analytical investigation of self-organized criticality in neural networks Some proposed rules, though, need extra mechanisms to stop the network drifting away.19Frontiers in Physics. Self-Organization Toward Criticality by Synaptic Plasticity In this view, neurons sense their own activity and adjust their connections, so the edge is maintained rather than fixed, which may explain how balance survives constant learning.
Chemical messengers may add extra dials. Dopamine, noradrenaline and acetylcholine shift alertness and mood, and they are proposed to nudge how close the brain sits to the edge.
Windows of Plasticity and How They Close
The brain’s ability to rewire peaks in early-life windows called critical periods. That is why children pick up skills such as native-language sounds more easily than adults. One fast-firing type of inhibitory cell is central to when these windows open and close.
During a window, wiring is especially responsive to experience, so what a child sees, hears, or does shapes circuits strongly. Infants are open to the sounds around them, and later learners often find unfamiliar sounds harder to hear and produce. Different abilities have windows at different times, so there is no single moment when the whole brain closes. Most of the detailed biology comes from animal experiments, especially on vision.
Those inhibitory cells mature gradually in early life, and their maturation helps determine when a window opens. As they mature, their influence on neighboring neurons grows and rewiring gets harder. Then mesh-like nets build up around them and the windows close.20eLife. Perineuronal nets control visual input via thalamic recruitment of cortical PV interneurons The nets reduce those cells’ plasticity and set their connections.21PubMed Central. Perineuronal Nets Regulate the Inhibitory Perisomatic Input onto Parvalbumin Interneurons and γ Activity in the Prefrontal Cortex In adult mice, the nets also boost the cells’ influence on visual responses and brain rhythms.22eLife. Perineuronal nets control visual input via thalamic recruitment of cortical PV interneurons
A closed window lowers plasticity but doesn’t end it. Adults still learn and form memories daily, and the stability protects useful skills from being overwritten. Models hint at a link to the edge: learning works best in an edge-like state, and slower, gentler rewiring helped models learn complex rules. That connection to windows of plasticity is a proposal, not settled fact.
Can Adult Windows Reopen
In adult mice, the antidepressant fluoxetine restored visual-cortex plasticity for as long as treatment continued.23PubMed Central. Fluoxetine‐induced plasticity in the visual cortex outlasts the duration of the naturally occurring critical period Researchers tested whether the visual cortex could change in response to altered visual input, something normally possible mainly in youth. With the drug, it could. Because the effect lasted only during treatment, the drug maintained a state rather than flipping a permanent switch. The reopened window also outlasted the natural one, so closure doesn’t look irreversible.24PubMed Central. Fluoxetine‐induced plasticity in the visual cortex outlasts the duration of the naturally occurring critical period Enriched surroundings have shown similar plasticity effects in mice.25PubMed Central. Fluoxetine‐induced plasticity in the visual cortex outlasts the duration of the naturally occurring critical period In a small study, healthy men did better on a perfect-pitch test after 2 weeks of valproate than after placebo.26PubMed Central. Valproate reopens critical-period learning of absolute pitch
These are prescription drugs with real side effects, and the pitch study was a small proof of concept. Don’t take them hoping to boost learning. Practice and rich experiences remain the safe route.
Reopening isn’t purely good, either. Closed windows keep established skills stable, and too much plasticity could unsettle what’s already learned. Researchers are looking for targeted, temporary reopening.
What Pushes the Brain Off Balance
Edge signals in human brain recordings fade the longer people stay awake.27PubMed Central. Fading signatures of critical brain dynamics during sustained wakefulness in humans Activity tilts toward excitation, so large cascades dominate.28PubMed Central. Fading signatures of critical brain dynamics during sustained wakefulness in humans The measure here is the average number of new firings each firing triggers, and the edge value is 1. Over 36 hours without sleep, it rose from about 0.98 to 1.08, past the edge value of 1.29PubMed Central. Inhibitory Decay and Supercritical Brain Dynamics During Sleep Deprivation In a system at the edge, a tiny tilt above 1 favors bigger cascades.
This fits the fog after an all-nighter: overexcited networks respond less selectively, which may relate to poorer focus and slower reactions. Sleep restores the edge pattern,30PubMed Central. Fading signatures of critical brain dynamics during sustained wakefulness in humans and one proposed job of sleep is resetting the brain for the next day awake.31PubMed Central. Fading signatures of critical brain dynamics during sustained wakefulness in humans That is one proposed function among several, not a settled explanation. Practically, it supports protecting a regular sleep schedule rather than relying on occasional catch-up.
The other ways off the edge sit at opposite ends. During seizures, the brain is driven past the edge into inflexible, high-activity chaos,32PubMed Central. Microscale Neuronal Activity Collectively Drives Chaotic and Inflexible Dynamics at the Macroscale in Seizures and it loses its range of response and flexibility.33bioRxiv. Single-cell Networks Reorganise to Facilitate Whole-brain Supercritical Dynamics During Epileptic Seizures Anesthesia with propofol or xenon pushes the brain the other way, into fading, less varied activity.34PubMed Central. Critical dynamics in spontaneous EEG predict anesthetic-induced loss of consciousness and perturbational complexity Edge-related brain-wave measures can even help tell consciousness from anesthetic unconsciousness.35PubMed Central. Why Brain Criticality Is Clinically Relevant: A Scoping Review
How Solid Is the Evidence
Some researchers argue the edge may not be necessary for thinking.36Trends Open. A critical assessment of the brain criticality hypothesis The avalanche-size patterns used as evidence can appear in irregular networks far from any tipping point,37PubMed. Power-law statistics and universal scaling in the absence of criticality so those patterns alone cannot prove the brain is at the edge.38PubMed. Power-law statistics and universal scaling in the absence of criticality Stronger tests would combine several signatures, such as slowly fading activity and the response-range findings.
Seeing the edge is also different from needing it. Skeptics argue the edge could be a side effect of balanced circuits rather than a goal. Testing necessity would mean pushing brains off the edge and showing thinking suffers, which sleep loss and anesthesia hint at but don’t cleanly prove. Disease studies hint at links too, such as weaker long-range brain-wave patterns in adults with moyamoya disease.39PubMed Central. Faded Critical Dynamics in Adult Moyamoya Disease Revealed by EEG and fMRI
When reading headlines, be wary of “the brain is proven critical” claims, and ask whether a study was in a model, a mouse, or people. Treat brain-boosting products that cite criticality with skepticism. The core picture of balanced excitation and inhibition is well supported.