Intrinsic dynamics of randomly clustered networks generate place fields and preplay of novel environments

• DOI: 10.7554/elife.93981.3
• Publication Date: 2024-10-18T13:45:44Z
• Authors: Jordan Breffle; Hannah Germaine; Justin D Shin; Shantanu P Jadhav; Paul Miller
• Keywords: Path integration; Biological neural network

During both sleep and awake immobility, hippocampal place cells reactivate time-compressed versions of sequences representing recently experienced trajectories in a phenomenon known as replay. Intriguingly, spontaneous can also correspond to forthcoming novel environments later, preplay. Here, we present model showing that spikes correlated with the fields underlying spatial previously future arise spontaneously neural circuits random, clustered connectivity rather than pre-configured maps. Moreover, realistic themselves circuit from minimal, landmark-based inputs. We find preplay quality depends on network’s balance cluster isolation overlap, optimal occurring small-world regimes high clustering yet short path lengths. validate results our by applying same field analyses published rat cell data. Our show recurrent generate immediate replay environments. These findings support framework whereby sensory experiences become associated preexisting “pluripotent” internal activity patterns.