Learning and memory, which require a strengthening of synaptic connections known as long-term potentiation (LTP), have traditionally been thought to be synapse specific, with one synapse unable to influence LTP induction in even its closest neighbors. But computational modeling has suggested that this kind of influence could allow individual neurons to store more information.
To test whether brains exploit this theoretical advantage, the authors stimulated individual dendritic spines with the neurotransmitter glutamate. Next, a weaker stimulus was applied to nearby spines. This weak stimulus is too low to trigger LTP on its own, but it caused robust potentiation when following the stronger stimulus.
“It made it easier for [synapses] to learn in the future if their neighbors had learned something in the past,” Harvey says. But too far in the past, or too distant a neighbor, didn't help: the subthreshold stimulus had to occur within 10 min and 10 μm of the first. The two synapses also had to be on the same branch, suggesting that the bolstering signal probably travels intracellularly from spine to spine. So far, the group has no leads on this roaming internal signal.
The authors suggest that such “clustered plasticity” may link memories that are laid down in close succession on the same dendritic branch. Whether this neighbor effect increases storage capacity remains to be seen.