Neuroscience · Neurobiology

Neuroplasticity

Reference entry · last updated September 12, 2026

Neuroplasticity is the capacity of the nervous system to modify its structural organization and functional connectivity in response to experience, sensory perturbation, developmental stages, or injury [1, 2].

1. First principles and physical constraints

Neural tissue balances computational stability against adaptability. If synapses modify too easily, existing memories degrade through catastrophic interference. If synapses resist alteration, the organism cannot acquire new associative representations. Donald Hebb formulated the biophysical basis of associative adjustment in 1949: when axon A repeatedly takes part in firing cell B, metabolic or structural changes occur in one or both cells that increase A's efficiency in firing B [3].

In idealized mathematical formulations, Hebbian weight modification updates synaptic strength \(w_{ij}\) as a function of pre-synaptic activation \(x_j\) and post-synaptic activation \(y_i\):

Because pure Hebbian correlation leads to unbounded runaway excitation, physical neural circuits enforce homeostatic synaptic scaling. Cells adjust total surface receptor densities to maintain stable mean firing rates within bounded dynamic ranges.

2. Synaptic plasticity mechanisms

Synaptic efficacy alters across timescales ranging from milliseconds to lifetimes. Functional modulation centers on two primary phenomena:

• Long-Term Potentiation (LTP): High-frequency stimulation induces persistent increases in synaptic response amplitude. Bliss and Lømo demonstrated LTP in the rabbit perforant path and dentate gyrus in 1973 [4]. Post-synaptic depolarization expels magnesium ions (\(\text{Mg}^{2+}\)) blocking NMDA receptor channels, permitting calcium (\(\text{Ca}^{2+}\)) influx. This biochemical cascade recruits AMPA receptors into the post-synaptic density.
• Long-Term Depression (LTD): Low-frequency stimulation produces prolonged decreases in synaptic efficacy, driven by modest, sustained calcium elevations that trigger receptor internalization.
• Spike-Timing-Dependent Plasticity (STDP): The sign and magnitude of synaptic modification depend on the microsecond-to-millisecond interval between pre- and post-synaptic action potentials. Pre-before-post timing induces potentiation; post-before-pre timing triggers depression.

3. Structural remodeling and adult neurogenesis

Plasticity extends beyond chemical sensitivity to physical morphology. Synaptic changes stabilize through physical alteration of dendritic spines:

• Dendritic Spine Remodeling: Actin filament rearrangement expands or retracts dendritic spines within minutes following stimulation. Persistent potentiation converts small thin spines into stable mushroom-shaped spines.
• Axonal Sprouting: Damaged or denervated cortical pathways form novel collateral axon branches that establish functional synapses on adjacent receptive fields.
• Adult Neurogenesis: Neural progenitor cells generate functional neurons throughout adulthood in restricted niches: the subgranular zone (SGZ) of the dentate gyrus in the hippocampus and the subventricular zone (SVZ) lining lateral ventricles [2]. These new cells migrate and integrate into existing synaptic circuits.

4. Systems-level remapping and critical periods

Cortical representations do not remain fixed across an organism's lifespan. In sensory cortices, map boundaries shift when afferent input distributions change:

• Cortical Remapping: Merzenich et al. demonstrated that transecting median nerves or digit amputations in adult primates causes neighboring cortical representations to expand into deactivated areas within weeks [5]. Cross-modal plasticity occurs in blind individuals, whose primary visual cortices activate during tactile Braille reading.
• Critical Periods: Heightened plastic states occur during early postnatal development, bounded by perineuronal net consolidation and maturation of inhibitory GABAergic interneuron circuits. While adult brains retain plasticity, alterations require stronger behavioral salience and focused attention.

5. Computational analogies and artificial networks

In artificial neural networks, learning operates via parameter optimization rather than biological morphogenesis. Numerical optimization adjusts continuous weight matrices using stochastic gradient descent and backpropagation:

Artificial systems lack the autonomous physical re-wiring, energetic constraints, and homeostatic normalization present in biological tissue. Biological neuroplasticity operates locally and asynchronously through chemical diffusion and membrane physics, whereas artificial network plasticity relies on centralized matrix computation across explicit training passes.

The direct correspondence between biological learning rules and backpropagation updates is an active theoretical topic rather than an established biological equivalence.