Electrical synapses are direct physical channels between neurons that let ionic current flow from one cell to the next without any chemical intermediary. Where chemical synapses rely on neurotransmitter release, diffusion across a gap, and receptor binding, an electrical synapse is essentially a tiny protein tunnel bridging two cells, fast enough to transmit signals in a fraction of a millisecond. For decades they were considered a primitive leftover, useful mainly for speed, but research over the past twenty years has overturned that view. Electrical synapses turn out to be tunable, computationally sophisticated, and deeply intertwined with the chemical synapses most neuroscience courses emphasize.
How They Were Found
The first electrical synapse in a nervous system was identified in 1957 by Edwin Furshpan and David Potter, who were studying the giant escape circuit of crayfish.1Frontiers in Physiology. The giant escape neurons of crayfish: Past discoveries and present opportunities At the time, chemical transmission was the accepted model for how neurons communicated, and many researchers were skeptical that direct electrical coupling played a meaningful role. The crayfish lateral giant neuron circuit proved otherwise: signals crossed the synapse far too quickly for chemical transmission, and the connection let current pass in one direction much more easily than the other. That directional bias, called rectification, hinted early on that electrical synapses were not just passive conduits.
The Physical Structure
An electrical synapse is formed by a structure called a gap junction. Each of the two connected neurons contributes a half-channel, called a connexon (in vertebrates) or innexon (in invertebrates). When the half-channels dock across the narrow extracellular space, they create a continuous pore roughly one to two nanometers wide. Small ions, metabolites, and signaling molecules can pass through, directly linking the interiors of the two cells.
In vertebrates, connexons are built from proteins called connexins. The most important connexin for neuronal electrical synapses in the brain is connexin-36, or Cx36. A recent cryo-electron microscopy study resolved the structure of Cx36 at high resolution, showing how the anti-malarial drug mefloquine wedges into pore-lining pockets formed by transmembrane helices, partially blocking the channel with six drug molecules arranged in a ring.2bioRxiv. Structural basis of connexin-36 gap junction channel inhibition That structural detail matters for pharmacology, as we will see later.
Invertebrates use a different family of proteins, called innexins. Despite the different gene families, innexins and the vertebrate pannexins (which are distinct from connexins but related to innexins) belong to a single superfamily, established through statistical and sequence analyses.3PubMed Central. Gap junctional proteins of animals: the innexin/pannexin superfamily The upshot is that electrical synapses are an ancient form of cell-to-cell communication, conserved across essentially all animal lineages.
More Than Speed and Synchrony
The traditional textbook story is that electrical synapses exist to make things fast and to synchronize groups of neurons so they fire together. Both of those things are true, but they are far from the whole story. Research has shown that rectifying electrical synapses can act as coincidence detectors, and that regulated changes in electrical synapse strength can actually promote oscillatory or even asynchronous neural activity.4PubMed Central. Electrical synapses: beyond speed and synchrony to computation In other words, electrical synapses do not simply lock neurons into lockstep. Depending on their properties, they can sharpen timing, selectively filter signals by frequency, or create complex patterns of coordinated and uncoordinated firing.
One striking example of computational sophistication is coincidence detection. Rectifying electrical synapses, those that pass current more easily in one direction, can distinguish inputs arriving at the same moment from inputs that are slightly staggered. When multiple presynaptic neurons fire simultaneously, the synaptic currents add up and produce a strong response in the postsynaptic cell. But if some inputs are delayed by even a hundred microseconds, the early ones partially block the contribution of the late ones through a shunting mechanism. This was demonstrated in the lateral giant neurons of crayfish, the same circuit where electrical synapses were first discovered.5PubMed. Neuronal coincidence detection by voltage-sensitive electrical synapses Coincidence detection at that temporal precision is something chemical synapses simply cannot achieve.
Measuring how much an electrical synapse actually influences a neuron’s behavior turns out to depend on context. One study developed a metric based on how much an electrical synapse shifts the timing of a postsynaptic neuron’s spike when the neuron is near its firing threshold. By that measure, electrical synapses were roughly an order of magnitude more powerful than excitatory chemical connections: the average shift in spike timing was about 16.5 milliseconds for electrical synapses, compared to about 1.4 milliseconds for a chemical synapse.6PubMed Central. A new measure for the strength of electrical synapses That finding challenges the instinct to treat electrical synapses as weak background connections.
How Rectification Works
Not all electrical synapses pass current equally well in both directions. Some are rectifying, meaning current flows preferentially from one neuron to the other. This makes the electrical synapse behave a bit like a diode in an electronic circuit and was the very property Furshpan and Potter noticed in the 1950s.
Rectification can arise in different ways. One route is through structural asymmetry: when the two half-channels are made of different protein subunits, they respond differently to the voltage difference across the junction. In the fruit fly’s giant fiber escape circuit, two splice variants of the Shaking-B innexin, called Shaking-B(Neural+16) and Shaking-B(Lethal), form asymmetric junctions that are gated by voltage in a one-directional way, producing classical rectification.7Current Biology. Molecular Mechanism of Rectification at Identified Electrical Synapses in the Drosophila Giant Fiber System Work on Drosophila innexins further showed that the amino-terminal tail of the protein is critical for voltage gating, but it only confers sensitivity when matched with a compatible innexin body, not on its own.8PubMed. Role of amino terminus in voltage gating and junctional rectification of Shaking B innexins
A complementary picture comes from the roundworm C. elegans, where the innexins UNC-7 and UNC-9 form rectifying junctions. Neutralizing charged residues in the UNC-7b amino terminus abolished rectification, indicating that electrostatic interactions at that end of the protein are what create the directional bias.9PubMed Central. Molecular basis of junctional current rectification at an electrical synapse So across both vertebrate and invertebrate systems, the amino-terminal region of gap junction proteins emerges as a key structural element controlling whether current flows both ways or mostly one way.
Electrical Synapses Are Plastic
Perhaps the biggest shift in thinking about electrical synapses over the past two decades is the recognition that they are extensively plastic. Their strength can be turned up or down across at least three distinct time scales.10PubMed Central. Design principles of electrical synaptic plasticity On the fastest time scale, milliseconds, changes in a neuron’s membrane conductance caused by chemical synaptic input or its own spiking activity can shunt current away from the gap junction, effectively decoupling the two cells without any change to the junction itself. This same mechanism can also alter coupling symmetry, making a previously bidirectional synapse behave as a rectifying one.
On a minutes time scale, phosphorylation of connexin proteins changes the conductance of the gap junction channel itself, with an order-of-magnitude dynamic range. On the slowest time scale, hours to days, the number of connexin molecules expressed in the membrane changes. This can happen over the course of the circadian cycle, during development, or in response to injury.11PubMed Central. The ever-changing electrical synapse The retina and the inferior olive, a brainstem nucleus involved in motor learning, have been particularly revealing systems for studying this plasticity. Combined, these mechanisms allow electrical coupling to be tuned at the level of a whole network, a local cluster of cells, or a single synapse.12PubMed Central. Characteristics and plasticity of electrical synaptic transmission
A vivid example of developmental plasticity comes from the thalamic reticular nucleus (TRN), a structure involved in controlling attention and sleep rhythms. In mice lacking the Cx36 gene, electrical coupling in TRN neurons was reduced but not eliminated, implying that other connexins can partially compensate. The residual junctions, however, had different biophysical properties: they were voltage-sensitive and often asymmetric, meaning the backup gap junctions were rectifying where the original Cx36 synapses were not.13PubMed Central. Developmental functions of electrical synapses This kind of compensatory remodeling underscores how dynamic and responsive electrical synapse networks are.
Mixed Synapses and the Chemical-Electrical Partnership
One of the most important insights of recent years is that electrical and chemical transmission are not alternative strategies. They are often partners at the same contact point. Many terminals in the mammalian brain contain both gap junctions and the machinery for chemical neurotransmitter release, forming what are called mixed synapses.14PubMed Central. On the occurrence and enigmatic functions of mixed (chemical plus electrical) synapses in the mammalian CNS Rather than thinking of synaptic transmission as “either chemical or electrical,” the evidence now supports seeing it as “both chemical and electrical,” with interactions between the two forms being important for normal brain development and function.15PubMed Central. Electrical synapses and their functional interactions with chemical synapses
A concrete example was recently described in the thalamic reticular nucleus, where a pair of neurons was found to be coupled both electrically and chemically. A single action potential in one cell produced an electrical spikelet in the other, followed by a longer-lasting inhibitory chemical postsynaptic potential.16PubMed. A mixed electrical and chemical synapse in the thalamic reticular nucleus The fast excitatory electrical component and the slower inhibitory chemical component would create a complex, time-varying effect on the partner neuron. Such mixed connections were rare in that study, turning up in just one pair out of hundreds of recordings, which raises the question of how many mixed synapses go undetected when experimenters are only looking for one modality at a time.
Electrical Synapses in Vision
The retina has long been a showcase for electrical synapses. Among the best-studied examples are AII amacrine cells, interneurons that sit at a critical juncture in the rod and cone pathways. These cells form gap junctions with one another and with ON cone bipolar cells, using Cx36. In the primate retina, the density of Cx36-positive puncta, sites of presumptive gap junctions, is concentrated in the ON sublamina, where it is more than three times higher than in the OFF sublamina.17Scientific Reports. AII amacrine cells in the primate fovea contribute to photopic vision This asymmetric distribution allows AII cells to participate in both dim-light and bright-light visual processing, routing signals through different pathways depending on conditions.
Electrical coupling in the retina also contributes to fine-scale synchrony between neighboring ganglion cells, the neurons that send visual information to the brain. Coincident electrical and chemical inputs to overlapping dendritic regions of coupled ganglion cells can trigger synchronized dendritic spikes. This nonlinear integration produces precisely correlated spike output, but only for stimuli that drive the specific regions where the cells’ dendrites overlap.18PubMed Central. Nonlinear dendritic integration of electrical and chemical synaptic inputs drives fine-scale correlations The electrical synapse alone is not sufficient; it is the combination of electrical and chemical input in the right place that produces the effect. This is a good illustration of how electrical synapses act as one element in a circuit, not in isolation.
Escape Circuits and Ultrafast Responses
Some of the most dramatic roles for electrical synapses are in escape behavior, the realm where speed genuinely matters. The Mauthner cells in fish are giant reticulospinal neurons that initiate rapid escape swims in response to predator-like stimuli. These cells integrate acoustic, mechanosensory, and visual input and directly activate motor neurons. The Mauthner cell system has been a proving ground for understanding how electrical synapses contribute to ultrafast sensory-motor coupling. When zebrafish were exposed to electric field pulses, the Mauthner cells were activated even when chemical synaptic transmission was pharmacologically blocked, demonstrating that a direct, non-chemical mechanism drives the response.19PubMed Central. Direct activation of the Mauthner cell by electric field pulses drives ultrarapid escape responses In an environment where a fraction of a millisecond determines whether a fish evades a predator, the directness and speed of electrical coupling provides a survival advantage that chemical synapses cannot match.
When Electrical Synapses Contribute to Disease
The same synchronizing power that makes electrical synapses useful in healthy circuits can become dangerous when synchronization goes wrong. In epilepsy, gap junctions between neurons have been implicated in the generation and maintenance of seizures. Computer simulations have shown that neuronal firing patterns change dramatically depending on the level of gap-junctional communication, and experimental manipulations of gap junctions support their role in seizure-like synchronized firing.20PubMed. Gap junctions, synchrony and seizures The idea is that direct electrical coupling can help lock large groups of neurons into pathological rhythmic activity, amplifying what might otherwise remain a local disturbance.
Electrical synapses also play a role in ischemic brain injury, the kind of damage caused by stroke. When a region of the brain loses blood supply, neurons near the injury core die through a process involving excessive activation of glutamate receptors. Research has found that neuronal gap junctions are required for this excitotoxic death: both pharmacological blockade and genetic deletion of neuronal gap junctions dramatically reduced cell death in a mouse model of focal cerebral ischemia.21American Physiological Society (J Neurophysiol). Neuronal gap junctions are required for NMDA receptor-mediated excitotoxicity: implications in ischemic stroke The likely mechanism is that gap junctions allow toxic signals from dying cells to spread into neighboring healthy ones, expanding the damage zone. This makes gap junctions a potential therapeutic target in stroke, though no treatment based on this idea has yet reached clinical use.
The Problem with Gap Junction Blockers
One reason progress on both epilepsy and stroke applications has been slow is that the pharmacological tools available for manipulating electrical synapses are frustratingly non-specific. Carbenoxolone, a drug widely used in research to block gap junctions, also reduces chemical synaptic transmission mediated by glutamate receptors, alters the width and firing rate of action potentials, and affects inhibitory synapses. These off-target effects mean that when carbenoxolone changes a circuit’s behavior, you cannot confidently attribute the change to gap junction blockade.22PubMed Central. Direct actions of carbenoxolone on synaptic transmission and neuronal membrane properties
Mefloquine, another commonly used gap junction blocker, has a similar problem. In hippocampal brain slices, mefloquine depressed high-frequency network oscillations, but the mechanism turned out to involve increased afterhyperpolarization and reduced spike firing rather than straightforward blockade of electrical coupling.23PubMed. Nonspecific effects of the gap junction blocker mefloquine on fast hippocampal network oscillations in the adult rat in vitro The structural work on Cx36 mentioned earlier, showing exactly how mefloquine binds in the channel pore, could eventually help design more selective agents, but for now, researchers often have to rely on genetic approaches, like Cx36 knockout mice, rather than pharmacology to study electrical synapses cleanly.
Gap Junctions Beyond Neurons
Gap junctions are not exclusive to neurons. Astrocytes, the star-shaped glial cells that support neurons throughout the brain, are extensively coupled to one another through their own gap junctions, built primarily from connexin-43 and connexin-30 rather than the connexin-36 found in neuronal electrical synapses. Recordings from hippocampal astrocytes in rat brain slices showed that astrocytic electrical coupling has weak low-pass filtering properties compared to neuronal electrical synapses, and the overall characteristics of the coupling are distinct.24PubMed. Electrical coupling between hippocampal astrocytes in rat brain slices Astrocytic networks use this coupling to distribute ions and metabolites over large distances, buffering potassium levels and shuttling glucose-derived energy substrates to active neurons. While the term “electrical synapse” is typically reserved for neuron-to-neuron junctions, the broader biology of gap-junctional communication in the brain is much wider, and understanding how glial gap junctions interact with neuronal ones remains an active research question.