Particle.news

Download on the App Store

MIT Proposes Astrocyte-Based Model to Account for Brain’s Massive Memory Capacity

The hypothesis invites experimental tests of astrocyte-neuron interactions to verify their impact on memory storage capacity.

MIT researchers have a new hypothesis for how brain cells called astrocytes might contribute to memory storage in the brain. Their model, based on dense associative memory networks, would help explain the brain’s massive storage capacity.
Image

Overview

  • MIT scientists propose that astrocytes play an active computational role in memory storage after developing a neuron-astrocyte associative network model.
  • The model conceptualizes each astrocyte process as an independent computational unit at tripartite synapses to massively boost storage capacity.
  • Simulations indicate this framework can store many more patterns than traditional Hopfield or dense associative memory networks.
  • Previous studies demonstrate that disrupting astrocyte-neuron connections in the hippocampus impairs memory formation and recall.
  • Researchers hope experimentalists will manipulate astrocyte process connections to validate the model and anticipate results could guide artificial intelligence design.