From Morphology to Computation: How Synaptic Organization Shapes Place Fields in CA1 Pyramidal Neurons

Abstract How spatial organization of synaptic inputs along the dendritic tree of neurons influence their feature selectivity is a central question in neuroscience. Here, we mapped the three-dimensional distribution of all excitatory and inhibitory synapses across the entire dendritic arbor of individual mouse CA1 pyramidal neurons in vivo and built biophysically detailed computational models to probe their functional impact on their ability to emerge as place cells. We found that excitatory, but not inhibitory, synapses are non-uniformly distributed, forming structural clusters preferentially on terminal dendrites. These excitatory synaptic clusters generate high-quality place fields more efficiently than randomized synaptic distributions, requiring fewer active synapses to achieve equivalent somatic output. Crucially, even when firing rates are matched, clustered inputs sustain significantly higher voltage-gated calcium influx and NMDA receptor activation, key substrates for synaptic plasticity. Further analysis reveals that clustering enables domain-specific computational strategies: oblique dendrites rely on cluster location, basal dendrites on cumulative synaptic strength, and the trunk on local input dispersion. Disrupting clustering collapses this compartmentalized processing into uniform summation. Our results establish synaptic clustering as a key mechanism that maximizes computational efficiency and enables sophisticated dendritic processing underlying hippocampal spatial representation. Competing Interest Statement The authors have declared no competing interest. Footnotes ↵5 Lead contact The revised manuscript transforms the study from a descriptive analysis into a robust mechanistic investigation of CA1 Pyramidal computation. A key update is the implementation of a shuffling protocol that preserves synaptic strength. This revealed that synaptic clustering is a highly resource-efficient mechanism, as shuffled configurations required 13% more activated synapses to match the somatic output of clustered inputs. We provide a new biophysical mechanism by demonstrating a dissociation between somatic spiking and dendritic integration: while somatic firing can be recovered by an increased number of dispersed inputs, dendritic calcium Ca2+ signaling and NMDA receptor activation cannot. This suggests clustering's primary role is lowering the threshold for local dendritic events. Furthermore, an unbiased Random Forest regression using new, more generalized synaptic features confirms that clustering allows the neuron to operate as a complex, multi-compartmental processor, whereas shuffled configurations collapse into simple integrators. We also refined our input spike train protocols to eliminate hypersynchrony artifacts and provide a more realistic simulation of CA3 inputs. Finally, the manuscript now explicitly highlights the structural asymmetry between clustered excitation and uniform inhibition and better contextualizes these findings within foundational anatomical and computational literature. This work represents the first direct, complete neuron-wide mapping of both excitatory and inhibitory synapses on the same cells, moving beyond the statistical extrapolations of previous studies.