Presynaptic input synchrony at scale

We present, for the first time, insights into functional synchronization of individual neurons under a complete anatomical neural connectivity map at a massive scale. While functional synchrony has been extensively studied as a mechanism of information processing in the brain, previous work has been limited to small populations of neurons in vitro, analyses across entire brain regions, or simulations in silico. Here, we leverage the large scale connectomics dataset MICrONS to analyze synchrony through the proxy of Pearson correlation across neurons at a larger scale in vivo with direct anatomic connectivity. We investigate functional synchronization between direct neighbors as a form of Hebbian synchrony, and input synchrony of neurons with a common postsynaptic neighbor.

While synaptic connections significantly increase Hebbian synchrony, we find that distance substantially impacts such synchronization. With increasing distance, among both neighbors and non-neighbors, synchronization significantly decreases, reaching almost the same level. For neighbors, such a trend alludes to reasons for maintaining long-range connections other than immediate synchronization. Interestingly, for non-neighbors, this lends further evidence to the role of local, non-synaptic, diffuse signaling mechanisms. Additionally, we find that the number of synaptic connections affects synchronization in a very sharp manner.

For the first time, we show that excitatory postsynaptic neurons exhibit input synchronization in vivo, while inhibitory postsynaptic neurons do not. Furthermore, we find that inhibitory postsynaptic neurons are more involved in recurrent connections with their presynaptic neighbors, and hypothesize that these two phenomena taken together could be related to winner-take-all dynamics within excitatory-inhibitory networks. We also observe that presynaptic neurons that have a greater difference in travel distance to the postsynaptic neighbor are less correlated, but that this can be offset by accounting for the relative distance traveled. This suggests that presynaptic neurons time their inputs to a postsynaptic neuron to drive postsynaptic activity more efficiently.

Recommended citation: Kong, L.*, Durresi, H.*, Mi, L., & Shavit, N. N. (2025, March). Presynaptic input synchrony at scale [Poster presentation]. Computational and Systems Neuroscience (COSYNE 2025), Montreal, QC, Canada.
Download Paper