Definition
Synaptic plasticity: Synaptic plasticity is the ability of synapses, the junctions between neurons, to become stronger or weaker in response to activity. It includes long-term potentiation, long-term depression, short-term changes and homeostatic scaling, and it is widely regarded as the main cellular basis of learning and memory.
Also known as: neural plasticity, LTP, LTD, synaptic modification
Synaptic plasticity is the ability of synapses, the connections between neurons, to become stronger or weaker depending on their activity. It includes long-term potentiation (LTP), long-term depression (LTD), short-term changes and homeostatic scaling. Synaptic plasticity is widely considered the cellular basis of learning and memory, and the mechanism that could let living neurons learn in biocomputers.
Key takeaways
- Synaptic plasticity changes how strongly one neuron influences another, reshaping the flow of signals through a network.
- Long-term potentiation was first described by Bliss and Lømo in 1973; long-term depression is its counterpart.
- Timing matters: spike-timing-dependent plasticity strengthens or weakens synapses based on millisecond spike order.
- Homeostatic plasticity keeps overall activity in a workable range so networks do not saturate or go silent.
- For organoid intelligence, plasticity is the biological counterpart of weight updates in machine learning.
How does synaptic plasticity work?
When an action potential reaches a synapse, it releases neurotransmitter that produces a response in the receiving neuron. Synaptic strength is how large that response is. Plasticity changes this strength, through mechanisms on both sides of the synapse.
What happens on the receiving side?
At many excitatory synapses in the brain, a key player is the NMDA receptor, a glutamate receptor that only lets calcium through when the receiving neuron is already depolarized. This makes it a coincidence detector: it responds strongly when input arrives while the receiving cell is active. The resulting calcium signal triggers biochemical cascades. Large, fast calcium rises tend to produce LTP, often by adding more AMPA receptors to the synapse. Smaller, prolonged rises tend to produce LTD, often by removing them (Citri & Malenka, 2008).
What happens on the sending side?
Plasticity can also change how much neurotransmitter is released per spike. Short-term facilitation and depression, lasting milliseconds to minutes, depend largely on presynaptic calcium and vesicle supply. Longer-lasting changes can involve growth or loss of synaptic contacts.
What are the main types of synaptic plasticity?
| Type | Timescale | What changes | Role |
|---|---|---|---|
| Short-term facilitation and depression | Milliseconds to minutes | Neurotransmitter release probability | Temporal filtering of signals |
| Long-term potentiation (LTP) | Hours to much longer | More or more effective receptors, structural growth | Strengthens useful connections |
| Long-term depression (LTD) | Hours to much longer | Fewer or less effective receptors | Weakens or prunes connections |
| Spike-timing-dependent plasticity (STDP) | Induced by millisecond timing | LTP or LTD depending on spike order | Learns causal relationships |
| Homeostatic synaptic scaling | Hours to days | All synapses on a neuron scaled up or down | Keeps firing rates stable |
Turrigiano (2008, Cell) described synaptic scaling as a way neurons “self-tune”: if activity is chronically too high, synapses shrink proportionally; if too low, they grow. This stabilizing process works alongside the faster, input-specific changes that store information.
Where did the idea come from?
The idea that learning depends on changing connections predates the experiments. In 1949, Donald Hebb proposed that a connection strengthens when one cell repeatedly helps fire another, a principle now called Hebbian learning. The first direct physiological evidence came from Bliss and Lømo in The Journal of Physiology (1973). They showed that brief high-frequency stimulation of a pathway into the hippocampus of anaesthetized rabbits caused a lasting increase in responses.
Later work refined the rules. Bi and Poo (1998, The Journal of Neuroscience) showed in cultured hippocampal neurons that the order of pre- and postsynaptic spikes within a window of tens of milliseconds decides whether a synapse strengthens or weakens.
Why does synaptic plasticity matter for biocomputing?
In conventional machine learning, a network learns by updating numerical weights. In biocomputing with living neurons, there are no weights to set directly. The only way to change what a culture computes is to give it experiences that trigger plasticity. That raises practical questions:
- Which stimulation patterns reliably induce lasting change? High-frequency trains, paired stimulation and timing-based protocols all have different effects.
- How do we give feedback? Closed-loop stimulation delivers inputs that depend on the network’s own output, which can shape learning.
- How long do changes last? Cultures show ongoing spontaneous activity, including network bursts, that can overwrite or wash out learned changes.
- How do we measure it? Plasticity is inferred from changes in evoked responses recorded on a microelectrode array.
An early demonstration came from Shahaf and Marom (2001, The Journal of Neuroscience). They stimulated cultured cortical networks and stopped stimulation when a chosen response appeared. Over repeated cycles, the networks produced the desired response more readily, which the authors interpreted as learning driven by activity-dependent change. More recently, Kagan et al. (2022) reported cultured neurons improving at a simplified game when given structured feedback, as covered in our article on how neurons learn in a dish.
Smirnova et al. (2023, Frontiers in Science) identified learning and memory in brain organoids as a central goal for the field, and noted that organoids would need to show plasticity that supports training. Our complete guide to organoid intelligence covers the wider roadmap.
Synaptic plasticity vs backpropagation: how do they compare?
Backpropagation computes an exact error gradient for every weight and updates all of them together. Biological plasticity is mostly local: a synapse changes based on activity at that synapse, plus broad modulatory signals such as dopamine. This difference shapes how living networks can be trained.
| Property | Synaptic plasticity | Backpropagation |
|---|---|---|
| Information used | Local activity, neuromodulators | Global error signal |
| Update control | Indirect, through stimulation | Direct, numerical |
| Continuous learning | Yes, always on | Usually separate training phase |
| Precision | Noisy, variable | Exact to floating-point precision |
| Energy | Very low per synapse | High at scale |
What are the limitations?
- Immaturity. Organoid synapses develop over months and may not show adult forms of plasticity.
- Variability. Plasticity rules differ by cell type, synapse type and developmental stage.
- Measurement. Extracellular electrodes see network outputs, not individual synaptic strengths.
- Stability. Learned changes can decay, and homeostatic processes may counteract them.
Building protocols that induce, track and verify plasticity over weeks is one of the problems we want to make routine at Organoid Programming.
Conclusion
Synaptic plasticity is how neurons turn experience into lasting change. Its many forms, from fast facilitation to long-term potentiation and homeostatic scaling, give living networks a built-in capacity to learn. Learning to steer synaptic plasticity reliably is one of the defining challenges for organoid biocomputing.
Frequently asked questions
What is synaptic plasticity in simple terms?
Synaptic plasticity is the way connections between neurons change strength based on how they are used. A synapse that is repeatedly involved in making a neuron fire can get stronger, while one that is out of step can get weaker. These changes alter how signals flow through a network, which is how the brain stores memories and learns new skills.
What is the difference between LTP and LTD?
Long-term potentiation (LTP) is a lasting increase in synaptic strength, so the same input produces a larger response afterward. Long-term depression (LTD) is a lasting decrease. Both can be triggered by specific patterns of activity, such as high-frequency stimulation for LTP or low-frequency stimulation for LTD. Together they let networks both reinforce useful connections and prune unhelpful ones.
Who discovered long-term potentiation?
Long-term potentiation was described by Tim Bliss and Terje Lømo in a 1973 paper in The Journal of Physiology. They showed that brief high-frequency stimulation of a pathway into the hippocampus of anaesthetized rabbits produced a long-lasting increase in synaptic responses. This finding became a cornerstone of research on the cellular basis of memory.
Does synaptic plasticity happen in brain organoids?
Brain organoids contain neurons that form functional synapses, and studies of cultured neurons and organoids have reported activity-dependent changes in network responses. However, organoid synapses are immature compared with an adult brain, and measuring plasticity precisely in 3D tissue on electrode arrays is difficult. Characterizing and controlling plasticity in organoids remains an open research goal.
Why is synaptic plasticity important for biocomputing?
Plasticity is what could let a living neural network learn rather than just respond. If stimulation and feedback can reliably reshape synapses, a culture can adapt its behavior to a task over time. This makes plasticity the biological equivalent of weight updates in machine learning, and understanding its rules is key to designing training protocols for organoids.
Sources
- Bliss & Lømo (1973). Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. The Journal of Physiology.
- Citri & Malenka (2008). Synaptic plasticity: multiple forms, functions, and mechanisms. Neuropsychopharmacology.
- Turrigiano (2008). The self-tuning neuron: synaptic scaling of excitatory synapses. Cell.
- Bi & Poo (1998). Synaptic modifications in cultured hippocampal neurons: dependence on spike timing, synaptic strength, and postsynaptic cell type. The Journal of Neuroscience.
- Shahaf & Marom (2001). Learning in networks of cortical neurons. The Journal of Neuroscience.
- Hebb (1949). The Organization of Behavior: A Neuropsychological Theory. Wiley. Reissued by Psychology Press.
- Kagan et al. (2022). In vitro neurons learn and exhibit sentience when embodied in a simulated game-world. Neuron.
- Smirnova et al. (2023). Organoid intelligence (OI): the new frontier in biocomputing and intelligence-in-a-dish. Frontiers in Science.
