6 Learning and Neural Plasticity
Learn how activity-dependent synaptic changes, firing timing, neural adaptation, and stabilizing mechanisms shape learning-related changes in neural circuits.
Learning changes neural circuits
Learning changes how neural circuits respond to experience. These changes can occur at synapses, across networks, and in the responses of individual neurons to ongoing input. Synaptic plasticity is an important mechanism for learning, but no single mechanism explains every form of learning or memory.
Synapses strengthen and weaken
A synapse's strength describes the influence that activity in one neuron has on another. is a lasting increase in synaptic strength, whereas is a lasting decrease. Both are activity-dependent and can be specific to the synapses that were active. By allowing connections to strengthen or weaken, circuits can refine their connections rather than simply accumulating stronger signals.
How strengthening can occur
At many excitatory synapses, LTP is triggered when presynaptic activity and postsynaptic depolarization coincide. Depolarization relieves a voltage-dependent block of NMDA-type glutamate receptors. When these receptors open, calcium ions enter the postsynaptic neuron and activate signaling pathways that can increase the number or effectiveness of AMPA-type glutamate receptors at the synapse. The result is a larger response to later input. Some forms of LTP also involve changes on the presynaptic side.
How weakening can occur
LTD can arise through different mechanisms. In some hippocampal synapses, a more modest rise in postsynaptic calcium favors signaling that removes or reduces AMPA receptors, weakening transmission. The exact mechanisms vary across brain regions and synapse types.
A particular input that repeatedly helps activate a postsynaptic neuron may strengthen its connection. If the input is weak or poorly correlated with the neuron's activity, the connection may weaken. This illustrates how the history and pattern of activity help determine how a synapse changes.
Timing and context shape plasticity
In , the relative timing of presynaptic and postsynaptic firing can influence whether a synapse strengthens or weakens. In a commonly observed pattern, presynaptic firing shortly before postsynaptic firing favors strengthening, while the reverse order can favor weakening. The precise timing rules depend on the circuit and conditions.
Early plasticity may rely on modifying existing proteins and receptors. Longer-lasting change can involve altered gene expression, protein production, and structural changes to synapses. These processes help stabilize learning-related changes, but synaptic plasticity is a cellular correlate of learning, not by itself a complete explanation of how a complex memory is represented.
Neuromodulators, including dopamine, can influence plasticity by signaling whether an outcome was rewarding, unexpected, or otherwise important. Consequently, the same neural activity can have different effects depending on the context in which it occurs.
Neural responses adapt to input
describes changes in responsiveness following sustained or repeated stimulation. For example, a neuron's response to an unchanging stimulus may diminish, helping the nervous system emphasize new or changing information. Adaptation depends on recent sensory context and can involve multiple stages of processing. It is related to plasticity but is not identical to learning.
Plasticity needs stability
Plasticity must be balanced. If strengthening continued without constraint, circuit activity could become excessive; if connections weakened too much, activity could fade. helps stabilize overall activity.
One example is , which adjusts many of a neuron's excitatory synapses up or down in response to prolonged changes in activity. This differs from Hebbian plasticity, which tends to modify particular active connections. A neuron's prior activity can also change how easily it undergoes further plasticity; this regulation of the conditions for future change is called .
Together, synapse-specific learning rules and broader stabilizing mechanisms allow circuits to remain adaptable without becoming unstable.