Learning and memory are acquired in the brain by long-term synaptic potentiation, a process consisting of consecutive, long-lasting increases in synaptic strength. Usually, this process results in runaway excitation of synapses and leads to imbalance and instability in neural network dynamics. After years of research on the acquisition of learning and memory, scientists are still searching for a definitive answer to how our brains store and preserve it.
To understand how memory is preserved despite excessive excitation, scientists including Yogesh S. Virkar, Woodrow L. Shew, Juan G. Restrepo, and Edward Ott have proposed a mechanism through a dual model that uses glial cells to preserve learning and memory in the brain. According to this study, published in May 2016, glial cells may play a pivotal role in the stabilization of these neural dynamics and the preservation of memory.
Spike Timing Dependent Plasticity (STDP)
STDP is a mechanism of learning in which neural firing is dependent on the timing of firing (or spikes) a neuron receives from other neurons, often resulting in long-term potentiation of synapses. STDP commonly leads to excessive, unstable growth in the strength of these synapses. Additional mechanisms are required to prevent and stabilize this runaway excitation in order to preserve what is learned in the brain. Much energy is required for the stabilization of these neural synapses—so, where does it come from?
Neurons Vs. Glial Cells
While neurons undergo electrical impulses and action potentials, glial cells usually serve to regulate neural systems and synaptic changes. Early studies indicate that glial cells form bridges between neural synapses and vascular structures of the brain. More recent studies indicate that glia supply metabolic resources directly to synapses, influencing neural dynamics.
Scientists developed a model consisting of two interacting layered networks: a neural network and a glial network. This model serves as a mechanism for the stabilization of neural dynamics involved in learning and memory. The neural network consists of neurons connected through synapses that function by undergoing Spike Timing Dependent Plasticity (STDP), while the glial network consists of glial cells connected through gap junctions that diffuse metabolic resources. These resources are then transported to and consumed by neural synapses. Through this system, glial cells supply the energy required to stabilize synaptic processes. This model provides a mechanism that prevents the tendency of runaway excitation among synapses during learning.
Researchers conducted three numerical experiments in order to mathematically present how this model can function effectively.
- The first quantitatively evaluates the stability of the neural network and shows that it is balanced by excitation and inhibition, ultimately stabilized from the transported resources.
- The second numerical experiment demonstrates that turning off the diffusion of resources from glial cells results in runaway synaptic growth in neurons. When diffusion was turned off, instability levels rose rapidly.
- In the third experiment, scientists divided neurons into two groups of 500. They stimulated each group of neurons and analyzed synapses in both groups as well as synaptic interactions between the two groups during learning. They found that synaptic strength increases among interconnected synapses from the first group to the second, following STDP, but decreases from the second to the first. The learning phase resulted in a decrease in resources that were replenished once the learning stimulus was removed, thus allowing neurons to store what was learned.
This proposed model of two interacting networks explains how neural systems maintain and remember what is learned in the brain. The glial network serves as a necessary homeostatic mechanism in learning and memory, playing an important role in supplying and replenishing metabolic resources to neurons and maintaining stability in neural dynamics.
The model that scientists have proposed in this study provides a possible solution to the runaway excitation that occurs during learning in the brain as part of STDP. As evaluated in numerical experiments, memory may occur because of the important role glial cells play in stabilizing neural dynamics, allowing the brain to preserve what it learns.