The brain depends on a continuously changing balance between excitation and inhibition. That balance is not a simple on/off switch, and it is not accurately described as ‘glutamate turns neurons on while GABA turns them off’. In real neural circuits, the effect of a signal depends on receptor type, ion gradients, timing, cell type, location within the neuron and the state of the wider network.
This matters in epilepsy because seizures emerge when networks enter patterns of abnormal synchrony and excitability. Changes in excitatory and inhibitory signalling are central to that process, but epilepsy is not caused by one universal ‘chemical imbalance’.
Excitation and inhibition describe effects, not simply chemicals
An excitatory postsynaptic effect makes a neuron more likely to reach action-potential threshold. An inhibitory effect makes firing less likely, delays it, or constrains when and where it can occur. These effects depend on which ion channels open and on the electrochemical gradients across the membrane.
Glutamate is the principal fast excitatory neurotransmitter in the mammalian brain, while GABA is the principal inhibitory neurotransmitter. But the effect of either system depends on receptor subtype and cellular context, so the shorthand should not be treated as an absolute rule.
How glutamate excites neurons
Glutamate acts through several receptor families. The main ionotropic receptors are AMPA, NMDA and kainate receptors. AMPA receptors mediate much of the brain’s fast excitatory synaptic transmission by allowing positively charged ions, especially sodium, to enter the postsynaptic neuron.
NMDA receptors are unusual because their opening depends both on glutamate binding and on membrane voltage. At resting potentials, magnesium blocks the channel. Depolarisation relieves that block, allowing sodium and calcium to enter. This gives NMDA receptors an important role in coincidence detection, synaptic plasticity and activity-dependent changes in circuits.
Metabotropic glutamate receptors work more slowly through intracellular signalling pathways. Depending on subtype and location, they can increase or decrease excitability and transmitter release.
How GABA inhibits neurons
GABA acts mainly through GABAA and GABAB receptors. GABAA receptors are ligand-gated ion channels. When they open, chloride conductance usually increases, tending to stabilise the membrane near the chloride equilibrium potential and reduce the effect of excitatory inputs.
GABAB receptors are G-protein-coupled receptors. Their effects are slower and include opening potassium channels, reducing calcium entry and suppressing neurotransmitter release. Together, these mechanisms can inhibit firing over longer timescales than fast GABAA-mediated synaptic inhibition.
Inhibition is not always just hyperpolarisation
One of the most important refinements in modern neuroscience is that inhibition does not always mean making the membrane much more negative. GABAA signalling can inhibit by increasing membrane conductance and ‘shunting’ incoming excitation, reducing the impact of excitatory currents even when the membrane voltage changes only slightly.
This is one reason simple diagrams showing inhibition as a downward voltage deflection can be misleading. Inhibitory control can operate through timing, conductance and local dendritic effects as well as through classic hyperpolarisation.
Chloride gradients can change what GABA does
The effect of GABAA receptors depends strongly on the chloride gradient across the membrane. Mature neurons usually maintain low intracellular chloride through transporters such as KCC2. This makes GABAA signalling predominantly inhibitory.
During intense neuronal activity, however, chloride can accumulate inside neurons. Changes in chloride transporters, bicarbonate movement and activity-dependent ionic shifts can reduce the inhibitory effect of GABA. In some seizure states, GABAergic signalling can become less inhibitory and, under certain conditions, depolarising.
This does not mean that GABA ‘becomes excitatory’ in every seizure. The effect depends on cell type, location, timing and ionic conditions. Modern epilepsy research therefore treats GABAergic signalling as dynamic rather than uniformly protective.
Interneurons are more than brakes
GABAergic interneurons are highly diverse. Different interneuron classes target different parts of principal neurons: some synapse near the soma or axon initial segment, while others target dendrites. This lets inhibition control not just whether a neuron fires, but also when it fires and which inputs are allowed to influence it.
Interneurons also help organise normal brain rhythms. They participate in oscillations important for attention, memory, sleep and sensory processing. In other words, inhibition is not simply suppression; it is a major organiser of timing within neuronal networks.
Why synchrony matters
A seizure is not produced simply because one neuron becomes highly excitable. Large populations of neurons must become abnormally coordinated. Excitatory and inhibitory circuits both contribute to that synchronisation.
In some experimental seizure models, inhibitory interneurons fire intensely at seizure onset. This may initially restrain principal neurons, but it can also synchronise large groups of cells and contribute to activity-dependent chloride loading. The relationship between inhibition and seizure spread is therefore more complex than ‘too little GABA’.
Phasic and tonic inhibition
GABAA-mediated inhibition occurs in at least two broad forms. Phasic inhibition is produced by brief bursts of GABA released into synapses, generating discrete inhibitory postsynaptic currents. Tonic inhibition comes from more persistent activation of high-affinity extrasynaptic GABAA receptors by low ambient concentrations of GABA.
Tonic inhibition can have a major effect on the overall excitability of a neuron even though it does not appear as individual synaptic events. Changes in tonic and phasic inhibition have both been implicated in different epilepsy mechanisms.
Astrocytes help regulate the balance
Excitation and inhibition are not controlled by neurons alone. Astrocytes help clear extracellular glutamate and GABA through neurotransmitter transporters, regulate potassium in the extracellular space and support energy metabolism.
If glutamate uptake, potassium buffering or other homeostatic functions are impaired, neuronal networks can become more excitable. This is another reason the excitation–inhibition balance cannot be reduced to transmitter release from neurons alone.
Ion channels shape excitability before neurotransmitters even arrive
Intrinsic membrane properties also determine whether a neuron responds strongly or weakly to synaptic input. Sodium, potassium, calcium and HCN channels influence threshold, burst firing, adaptation and rhythmic behaviour.
A neuron with altered ion-channel function may become hyperexcitable even if glutamate and GABA release are unchanged. Conversely, strong inhibitory input may fail to control firing if intrinsic membrane properties are substantially altered.
What ‘E/I balance’ really means
The phrase ‘excitation/inhibition balance’, often shortened to E/I balance, is useful but can be misunderstood. It does not mean the brain maintains one fixed numerical ratio of excitation to inhibition. The appropriate balance changes moment by moment depending on behaviour, brain region and network state.
During normal sensory processing, for example, excitation and inhibition can rise together. Inhibition may sharpen responses, limit spread and improve timing rather than simply reducing overall activity. Healthy brain function depends on coordinated interaction, not on keeping excitation permanently low.
How epilepsy can disturb this system
Epilepsy can alter E/I regulation through many routes. These include changes in receptor expression, ion-channel mutations, loss or dysfunction of interneurons, synaptic reorganisation, altered chloride transport, inflammation, gliosis, abnormal extracellular potassium handling, impaired neurotransmitter uptake and structural rewiring after brain injury.
Different epilepsies can involve very different combinations of these mechanisms. Even within the same person, the mechanisms promoting seizure initiation may differ from those supporting seizure spread or termination.
Why simply increasing inhibition is not always enough
Many antiseizure medicines enhance inhibitory signalling or reduce excitatory transmission, but the relationship is not straightforward. Some treatments act mainly on sodium channels, calcium channels, synaptic vesicle proteins or other targets rather than directly changing GABA or glutamate.
And because GABAergic effects depend on ionic conditions, ‘more inhibition’ is not always equivalent to better seizure control in every circuit or seizure type. The effectiveness of a treatment depends on the underlying mechanism and the network in which it acts.
Excitation, inhibition and absence seizures
Generalised absence seizures are a useful example of why the old imbalance model is incomplete. They involve abnormal oscillation within thalamocortical networks, where both excitatory and inhibitory connections contribute to rhythmic activity. Excess inhibition in one part of the circuit can, paradoxically, help sustain abnormal oscillations rather than simply suppressing them.
What researchers can actually measure
Researchers study excitation and inhibition using intracellular recordings, patch-clamp electrophysiology, receptor pharmacology, optogenetics, calcium imaging, microelectrode recordings and computational models. Human evidence also comes from resected surgical tissue, intracranial recordings, spectroscopy and genetic studies.
No single measurement provides a complete ‘E/I balance score’ for a person with epilepsy. Measurements made in one cell type, one brain region or one experimental preparation cannot automatically be generalised to the whole brain.
What this page does — and does not — explain
This page explains how excitatory and inhibitory signalling regulate neuronal circuits. It does not explain the electrical signal of an individual neuron, the full network sequence of a seizure, or the long-term biological process by which a brain becomes epileptic. Those are treated separately in the Science & Discovery section to avoid repeating the same material under different titles.
Sources
Jasper’s Basic Mechanisms of the Epilepsies, 5th ed. — GABAA-receptor signalling and ionic plasticity in seizure generation and spread
Jasper’s Basic Mechanisms of the Epilepsies, 5th ed. — GABAA receptors, seizures and epilepsy
Bryson, Reid & Petrou — GABAA receptor neurotransmission and epilepsy
Sarlo & Holton — Brain concentrations of glutamate and GABA in human epilepsy: review
NCBI Bookshelf — Basic mechanisms underlying seizures and epilepsy
Evidence checked: September 2026.