Simplification Strategies for Biologically Constrained Hippocampal CA1 Microcircuit Models: How Far Reduction Preserves Firing and Driven Theta but Not Emergent Gamma

Biologically detailed models of the hippocampal CA1 microcircuit reproduce cellular and network dynamics with high fidelity, but their multi-compartmental morphologies and Hodgkin-Huxley (HH) channel kinetics make them computationally expensive. This study asks a deliberately narrow question: how far can such a circuit be simplified before its dynamics degrade, and which properties break first? Three reduction strategies (morphological reduction to single-compartment neurons, abstraction of HH channels to a leaky integrate-and-fire (LIF) formulation, and connectivity sparsification) are evaluated at four scales: single neurons, a 10-neuron microcircuit, a 100-neuron scaled network, and a hybrid network combining LIF pyramidal cells with HH interneurons. To keep comparisons fair, all models are integrated with the same fixed-step scheme, and rhythmicity is quantified using relative band power and peak prominence above a fitted 1/f background rather than raw in-band power. Firing rates and a 6 Hz theta component are preserved across all three simplifications, and single-neuron computational cost falls roughly twentyfold from HH to LIF. However, the theta component is imposed by the external drive rather than generated by the network, and none of the configurations (detailed, simplified, scaled, or hybrid) produced a genuine emergent gamma rhythm; the apparent gamma peaks reflect firing-rate harmonics or band-edge artifacts. Reduced CA1 models therefore faithfully preserve firing statistics and driven theta at a fraction of the cost, but emergent gamma is not captured at this scale and cannot serve as a marker of model fidelity here.