Spatial information transfer in recurrent place-cell networks depends on excitation-inhibition balance, neural-circuit heterogeneities, and trial-to-trial variability

Abstract

A key challenge in understanding spatial navigation and memory is explaining how hippocampal networks sustain robust spatial information transfer despite pronounced trial-to-trial variability and pervasive neural-circuit heterogeneities. Although hippocampal heterogeneities and physiological variability are well-characterized, circuit-scale understanding of stable information transfer in recurrent place-cell networks remains limited. Here, we first show that even recurrent networks composed of intrinsically identical neurons and receiving identical place-field inputs express pronounced neuron-to-neuron variability in spatial tuning profiles, place-field widths, subthreshold ramp amplitudes, and spatial information transfer. Introduction of intrinsic within-type heterogeneities to excitatory and inhibitory neurons further increased diversity in firing properties, but strikingly improved robustness of spatial information transfer under high trial-to-trial variability. Although strengthening inhibition expectedly narrowed place fields and reduced firing across all networks, the impact of inhibition on information-transfer profiles was stronger in heterogeneous networks manifesting high degree of trial-to-trial variability. Across networks, increasing trial-to-trial variability reduced information transfer and shifted the spatial location of peak information from the high-slope regions of the tuning curve to its peak-firing location. Finally, incorporating afferent heterogeneities allowed neurons to be tuned to distinct place-field centers, reducing peak information values while amplifying neuron-to-neuron diversity in information-transfer profiles. Together, we demonstrate that excitation-inhibition balance, trial-to-trial variability, within-type heterogeneities, and afferent input diversity jointly regulate spatial information transfer in recurrent place-cell circuits. We also highlight intrinsic heterogeneities as substrates for enhanced robustness of spatial coding against perturbations. Importantly, the convergence of multiple disparate mechanisms in yielding similar information-transfer profiles underscores degeneracy as a fundamental organizing principle in neural-circuit physiology.