Astrocytes Could Be the Brain’s Hidden Memory Engine
MIT researchers propose astrocyte tripartite synapses as computational units that could enable memory storage beyond neuron-to-neuron networks, using a dense associative memory framework. In this view, neuron–astrocyte networks might store far more memories, with capacity scaling with network size, potentially explaining why human memory has no known upper limit. The model is theoretical and awaits experimental testing, challenging the traditional neuron-first view of memory storage.