Buzsaki Lab, NYU Neuroscience Institute, New York University, New York, NY, USA

Most memory models assume that information is stored in the synaptic strength between neurons. To date, it has been difficult to record this synaptic coupling while monitoring the coding properties of neurons and the behavior of animals engaged in learning. One solution to this problem is to leverage correlations between neurons at fine time scales to infer causal relationships due to underlying anatomy. We developed a statistical algorithm to infer such synaptic connectivity between excitatory neurons and neighboring inhibitory cells, and tested the validity of an in vivo synapse detector by artificially stimulating pre-synaptic neurons and observing a postsynaptic response. We then sought to quantify changes in synaptic coupling strength across multiple time scales. At short time scales (10-1000 ms), we found that synapses often behave as bandpass filters, whose cutoff frequency varies orders of magnitudes across the recorded population. In other experiments, we found that optogenetic stimulation in CA1, but not CA3 or the dentate gyrus, caused a reorganization of hippocampal place fields and feedback inhibition. Syn-aptic coupling also changed on a moment-to-moment basis depending on an animal’s goals.

These findings show that a potentially powerful synapse can be rendered ineffective due to ongoing activity within the network. Therefore, memory-related plasticity may depend not only on the strength of the EPSP, but also on probabilistic changes in the receptivity of the postsynaptic neuron upon the arrival of presynaptic input.

Lecture Poster

Series Overview >