Ion-channel regulation of response decorrelation in a multi-scale model of the dentate gyrus

Abstract

The dentate gyrus (DG) is uniquely endowed with multiple forms of biological heterogeneities owing to the expression of adult neurogenesis and sparse connectivity, and has been functionally implicated in response decorrelation and pattern separation. Although channel decorrelation could be achieved through synergistic interactions between these heterogeneities, the impact of individual ion channels on channel decorrelation has not been explored. Here, to systematically assess the cascading impact of molecular-scale (ion channel) perturbations on network-scale outcomes (decorrelation), we first quantified the impact of eliminating individual ion channels on single-cell physiology of heterogeneous populations of granule cells (GCs) and basket cells (BCs). Employing virtual knockout simulations involving both populations, we found that the mapping between ion channels and nine distinct physiological measurements was many-to-many. Next, to assess the impact of ion channel elimination on channel decorrelation, we employed a conductance-based multi-scale network model of the DG. This network was endowed with four distinct forms of heterogeneities (intrinsic, synaptic, structural and afferent), with afferent inputs from the entorhinal cortices driven by virtual arena traversal. We show that individual ion channels expressed in GCs govern DG network excitability, and critically regulate the ability of the network to perform channel decorrelation. The impact of eliminating individual ion channels on network excitability and channel decorrelation was differential and variable, with local heterogeneities playing a pivotal role in determining the strength of such impact. Specifically, in the presence of structurally immature neurons in the DG network, the impact of ion channel elimination on channel decorrelation was considerably lower when compared with a network exclusively constructed with structurally mature neurons. Finally, we show that for any given ion channel knockout, the average percentage change in output correlation was invariant to the specific values of input correlation, across different network configurations endowed with disparate structural and afferent heterogeneities. Our analyses emphasizes that the mapping between components and function is many-to-many across scales, and assign critical roles for biological heterogeneities in conferring multi-scale functional robustness in the face of physiological and pathological perturbations.\n\nSIGNIFICANCE STATEMENTThere are precise sets of computation spanning different scales of analyses that drive behavioral states and responses of an animal. Perturbations to components that drive these computations at one scale could result in a cascading effect that alters physiological properties across several scales. Multi-scale computational models that account for biological heterogeneities at each scale are ideal tools to approach this complex problem, where systematic analyses of such cascades could be rigorously accomplished. Here, we systematically assessed the impact of eliminating individual ion channels, first on neuronal intrinsic properties, and consequently on network excitability and response decorrelation. Our results unveil important roles for biological heterogeneities in conferring multi-scale functional robustness in the face of physiological and pathological perturbations, achieved through many-to-many mappings between constitutive components and physiological outcomes.