- Levels of plasticity
- Plasticity occurs at multiple distinct levels, and these should not be conflated:
**Synaptic plasticity**: changes in the strength or efficacy of existing synapses (LTP, LTD, receptor trafficking). Well characterised in animal models and brain slice preparations.
**Structural plasticity**: physical changes to neural structures — dendritic spines, synapse formation or elimination, axonal remodelling. Evidence from animal models is strong; measuring subtle structural changes in living humans with neuroimaging is considerably more difficult and methods remain debated.
**Functional/network plasticity**: changes in how existing circuits participate in a task, particularly relevant after injury or during rehabilitation.
**Developmental plasticity**: experience-dependent changes during sensitive or critical periods when circuits are especially malleable.
**Myelin plasticity**: experience can alter oligodendrocytes and the myelination of axons, influencing conduction speed and circuit timing — an important but often overlooked form of plasticity.
Evidence for each type differs substantially in strength, species generalisability, and applicability to humans.
- Hebbian plasticity ("fire together, wire together")
- The principle, proposed by Donald Hebb (1949), that repeated co-activation of a presynaptic and a postsynaptic neuron can lead to lasting strengthening of the connection between them. This is a general theoretical principle, not a description of a single mechanism. NMDA-receptor-dependent mechanisms provide one important biological route through which activity-dependent synaptic strengthening can occur — by acting as "coincidence detectors" requiring simultaneous presynaptic glutamate release and postsynaptic depolarisation — but Hebbian-like plasticity can be implemented through multiple molecular pathways. The rule has been influential in computational models of learning, though modern artificial neural networks use variants that depart substantially from the original biological proposal.
- Synaptic plasticity
- The ability of synapses to strengthen or weaken over time in response to activity — changes in neurotransmitter release, receptor number, receptor sensitivity, and spine morphology. Synaptic plasticity underlies most current accounts of learning and memory at the cellular level, though the relationship between synaptic changes and the systems-level process of memory formation involves multiple additional mechanisms and levels of organisation.
- Long-term potentiation (LTP)
- A persistent increase in synaptic strength following high-frequency stimulation of a neural pathway. First demonstrated in hippocampal tissue of anaesthetised rabbits by Bliss and Lømo (1973). LTP depends on NMDA receptors acting as coincidence detectors and subsequent AMPA receptor insertion into the postsynaptic membrane. LTP is a major experimental model of synaptic plasticity and is strongly associated with learning and memory mechanisms. However, memory is not reducible to LTP — there are multiple forms of plasticity and multiple levels of memory organisation. LTP has been difficult to demonstrate during actual learning in vivo (outside slices or artificial stimulation), and the relationship between synaptic LTP and systems-level memory consolidation is still being worked out.
- Long-term depression (LTD)
- A persistent decrease in synaptic strength following low-frequency or asynchronous pre- and postsynaptic activity. LTD is mediated by NMDA receptors and mGluR receptors and leads to AMPA receptor internalisation. LTD is not simply a "forgetting" mechanism — it contributes to circuit refinement and some forms of learning, and may help prevent excessive synaptic strengthening and keep networks adaptable. Plasticity requires both strengthening and weakening: without a mechanism to reduce synaptic weights, circuits would become saturated and lose the ability to encode new information.
- Sensitive and critical periods
- Developmental windows during which specific circuits are especially sensitive to environmental input. In strict usage, a **critical period** is a window during which a particular experience is necessary for normal development and after which the circuit's plasticity is substantially reduced. The classic example is the visual critical period in kittens: monocular deprivation during the first weeks of life (Hubel & Wiesel, 1963) causes permanent loss of binocular vision and cortical columns for the deprived eye — demonstrating experience-dependent plasticity specifically in the visual cortex during a restricted developmental window.
**Sensitive periods** are broader windows during which experience has heightened influence but the constraints are less absolute. Language development involves multiple overlapping sensitive periods for different components (phonology, syntax, vocabulary), with considerable individual variability — not a single, clean critical period analogous to visual deprivation. Critical periods are defined by elevated synaptic plasticity regulated by the balance of excitation and inhibition (parvalbumin interneurons and perineuronal nets play key roles).
- Cortical reorganisation and cross-modal plasticity
- Topographic cortical maps can reorganise in response to altered sensory input or motor demands. Following limb amputation, the somatosensory cortex region representing the missing limb shows altered responses from adjacent body representations. This reorganisation has been associated with phantom limb phenomena, though the relationship between cortical reorganisation and phantom limb pain is considerably more complex than early accounts suggested — peripheral, spinal, and higher-level mechanisms all contribute.
In sensory deprivation, **cross-modal plasticity** can occur: cortical regions associated with one sensory modality become responsive to input from another. In profoundly deaf individuals, regions of auditory cortex may show altered responses to visual input; in blind Braille readers, regions of visual cortex are engaged during tactile reading. "Recruited" is a useful shorthand, but the underlying changes are in connectivity and responsiveness rather than a clean switch of function. Remapping can be rapid (hours after nerve block) or gradual (years of practice), suggesting multiple mechanisms at different timescales.
- Adult neurogenesis
- The generation of new neurons in the adult brain. In rodents and other mammals, neurogenesis is well established in two regions: the subgranular zone of the hippocampal dentate gyrus and the subventricular zone (with new neurons migrating to the olfactory bulb). In rodent hippocampus, neurogenesis is increased by physical exercise, environmental enrichment, and antidepressants; decreased by stress, alcohol, and ageing.
The human situation is considerably more contested. Some studies report evidence for adult hippocampal neurogenesis in humans; others, using different methodological approaches, have found little or none at functionally significant levels in adulthood. Whether adult human hippocampal neurogenesis occurs at a scale that influences cognition or mood remains an active controversy. The SVZ-to-olfactory-bulb pathway seen in rodents does not appear to operate in the same way in adult humans. This is a case where a famous neuroscience finding has become substantially more complicated as methodological standards improved.
- Use-dependent plasticity and experience
- Changes in neural circuitry driven by repeated use of specific circuits. Maguire et al. (2000) found that London taxi drivers — who spend years learning the city's complex street network — showed larger posterior hippocampal volume than controls, with the size difference correlating with years of experience. This is a striking cross-sectional association between experience and brain structure in adulthood, though cross-sectional differences do not by themselves establish that the experience caused the anatomical difference.
Draganski et al. (2004) provided longitudinal evidence: grey matter changes were reported in human adults following juggling skill training, with increases during training and partial reversal after practice stopped. Methodological debates continue about what structural MRI measures actually represent biologically, but the overall picture from cross-sectional and longitudinal evidence supports experience-dependent structural differences in adulthood. Distinguishing **cross-sectional** (experienced people differ from controls) from **longitudinal** (brain measures change during training) evidence is important for evaluating these claims.
- Maladaptive plasticity
- Plasticity is not synonymous with improvement. The same mechanisms that support learning and recovery can also maintain dysfunction. Examples include:
- **Chronic pain**: synaptic strengthening in pain pathways can maintain pain long after the original injury has healed (central sensitisation)
- **Addiction**: repeated drug exposure produces lasting changes in reward circuits that drive compulsive drug-seeking
- **Maladaptive fear**: fear conditioning can become entrenched, as in PTSD, when normal extinction mechanisms fail
- **Some forms of tinnitus**: thalamocortical reorganisation following auditory input loss may contribute to persistent phantom sounds
Understanding maladaptive plasticity is as important as understanding beneficial plasticity — and treatments for these conditions often work by targeting the same molecular and circuit mechanisms that underlie adaptive learning.