A two-layer biophysical model of cholinergic neuromodulation in olfactory bulb

Abstract

Cholinergic inputs from the basal forebrain regulate multiple olfactory bulb (OB) functions, including odor discrimination, perceptual learning, and short-term memory. Previous studies have shown that nicotinic cholinergic receptor activation sharpens mitral cell chemoreceptive fields, likely via intraglomerular circuitry. Muscarinic cholinergic activation is less well understood, though muscarinic receptors are implicated in olfactory learning and in the regulation of synchronized oscillatory dynamics in hippocampus and cortex. To understand the mechanisms underlying cholinergic neuromodulation in OB, we developed a biophysical model of the OB neuronal network including both glomerular layer and external plexiform layer (EPL) computations and incorporating both nicotinic and muscarinic neuromodulatory effects. Our simulations show how nicotinic activation within glomerular circuits sharpens mitral cell chemoreceptive fields, even in the absence of EPL circuitry, but does not facilitate intrinsic oscillations or spike synchronization. In contrast, muscarinic receptor activation increases mitral cell spike synchronization and field oscillatory power by potentiating granule cell excitability and lateral inhibitory interactions within the EPL, but it has little effect on mitral cell firing rates and hence does not sharpen olfactory representations under a rate metric. These results are consistent with the theory that EPL interactions regulate the timing, rather than the existence, of mitral cell action potentials and perform their computations with respect to a spike timing-based metric. This general model suggests that the roles of nicotinic and muscarinic receptors in olfactory bulb are both distinct and complementary to one another, together regulating the effects of ascending cholinergic inputs on olfactory bulb transformations.

Figure 1.

A, Schematic representation of dendrodendritic synaptic connectivity among MCs, PGCs, and GCs. The first computational layer of the olfactory bulb is the glomerular layer (GL), in which each glomerulus comprises a single input from a convergent olfactory sensory neuron population (OSN input) along with dendrodendritic interactions between an MC and a PGC. Two glomeruli are depicted. Mitral cells innervating different glomeruli interact with one another (indirectly) in the second computational layer, the EPL. Here, MC lateral dendrites excite GC dendrites and GCs inhibit MCs, producing a functionally lateral inhibitory network among MCs. See Materials and Methods for details. B, Example of the sigmoidal, noninstantaneous time course of OSN input as delivered to MCs and PGCs in the network model. The dotted vertical line indicates the nominal time of odor onset.

Figure 2.

Firing properties of the model MC. A, Voltage responses of the MC model to four different levels of current injection. B, Expanded time view of A to highlight STOs between spike clusters. C, STO and intraburst spike frequencies as a function of injected current. D, The number of spikes per cluster as a function of injected current. E, Transiently opening a shunting inhibitory conductance resets the STO phase. F, A larger IPSC resets the STO phase and also generates a rebound spike.

Figure 3.

Firing properties of the model GC. A, Voltage responses of the GC model to three different levels of current injection. B, Voltage responses of the GC model to a large (220 pA) current injection in the presence (top) and absence (bottom) of the I_CAN current. The ramping depolarization evoked by this large injected current was suppressed by blockade of the I_CAN current. C, The effect of muscarinic neuromodulation (carbachol, CCh) was simulated by blocking the I_M and I_KCa currents. This suppressed the AHP that was normally evoked by depolarizing steps and revealed a latent ADP. The prepulse resting membrane potential was −64 mV. D, The ADP produced under muscarinic neuromodulation depended on the prepulse resting membrane potential [−63 mV (top), −61 mV (bottom)] and could reach spike threshold.

Figure 4.

Firing properties of the model PGC. Two different levels of depolarizing current injection (top and middle) produce calcium T current-dependent LTSs. Release from a hyperpolarizing current injection (bottom) can produce an anode break LTS.

Figure 5.

Simulated MC and PGC responses in the glomerular-layer model. A, Current-clamp responses of seven representative MCs under control conditions. B, As in A, but with nAChR neuromodulation active. C, Current-clamp responses of seven PGCs associated with the same glomeruli as the MCs in A, under control conditions. D, As in B, but with nicotinic cholinergic neuromodulation active. E, PGC-mediated GABA_A conductances measured from the same representative MCs under control conditions. F, As in E, but with nicotinic cholinergic neuromodulation active. Black horizontal bars indicate the duration of odor input (same for below), and the numbers on the top of the voltage traces in C and D indicate the baseline voltage levels during odor presentation (i.e., excluding spikes).

Figure 6.

Activation of nicotinic cholinergic receptors sharpens MC tuning curves in the glomerular-layer model. A, Histograms of activity in all 25 glomeruli under control conditions, artificially arranged according to their levels of afferent activation for ease of visualization. Top, Steady-state OSN input levels; middle, odor-evoked spike rates in MCs; bottom, odor-coding spike rates in MCs (see Materials and Methods). B, As in A, but with nicotinic cholinergic neuromodulation active. C, Raster plot of MC spiking activity immediately before and after the onset of odor presentation, under control conditions. D, As in B, but with nicotinic cholinergic neuromodulation active. E, Simulated LFP (top) during odor presentation, with autocorrelation (middle) and power spectrum (bottom), under control conditions. F, As in E, but with nicotinic cholinergic neuromodulation active.

Figure 7.

In the full model, including both GL and EPL circuitry, muscarinic activation improves MC spike synchronization. A, Current-clamp responses of two pairs of MCs under control conditions. B, As in A, but with mAChR neuromodulation active. C, Current-clamp responses of two pairs of GCs under control conditions. D, As in B, but with muscarinic cholinergic neuromodulation active. E, Time series of total GABA_A-mediated synaptic chloride conductance received from all presynaptic GCs by three representative MCs under control conditions. F, As in E, but with muscarinic cholinergic neuromodulation active. Black vertical arrows indicate the onset of odor stimulation (same for Fig. 10).

Figure 8.

Activation of muscarinic cholinergic receptors in the full model potentiates GC excitability with little impact on MC firing rates. A, Raster plot of MC spiking activity immediately before and after the onset of odor presentation, under control conditions. B, As in A, but with mAChR neuromodulation active. C, Raster plot of GC spiking activity under control conditions. D, As in C, but with muscarinic cholinergic neuromodulation active. E, Histograms of activity in all 25 glomeruli under control conditions, arranged according to their levels of afferent activation for ease of visualization. Top, Steady-state OSN input levels; middle, odor-evoked spike rates in MCs; bottom, odor-coding spike rates in MCs (see Materials and Methods). F, As in E, but with muscarinic cholinergic neuromodulation active.

Figure 9.

Activation of muscarinic cholinergic receptors in the full model enhances sLFP oscillatory power and imposes more stringent phase-locking between MC spikes and sLFP oscillations. A, Simulated LFP (top) during odor presentation, with autocorrelation (middle) and power spectrum (bottom), under control conditions. B, As in A, but with mAChR neuromodulation active. C, Distribution of MC spike phases with respect to sLFP oscillations under control conditions. D, As in C, but with muscarinic cholinergic neuromodulation active. E, Raster plot of MC spike phases with respect to sLFP oscillations, under control conditions. F, As in E, but with muscarinic cholinergic neuromodulation active. Note that muscarinic activation narrows the distribution of spike phases.

Figure 10.

Coactivation of nicotinic and muscarinic cholinergic receptors in the full model both sharpens the MC representation and increases MC spike synchrony. A, Raster plot of MC spiking activity immediately before and after the onset of odor presentation, under nicotinic neuromodulation alone (nAChR). B, As in A, but with muscarinic neuromodulation active in addition to nicotinic neuromodulation (nAChR + mAChR). C, Histograms of activity in all 25 glomeruli under nicotinic neuromodulation alone, arranged according to their levels of afferent activation for ease of visualization. Top, Steady-state OSN input levels; middle, odor-evoked spike rates in MCs; bottom, odor-coding spike rates in MCs (see Materials and Methods). D, As in C, but with both nicotinic and muscarinic neuromodulation active. E, Current-clamp responses of two pairs of MCs under nicotinic neuromodulation alone. F, As in E, but with both nicotinic and muscarinic neuromodulation active.

Figure 11.

Coactivation of nicotinic and muscarinic receptors in the full model enhances coherent sLFP oscillations. A, Simulated LFP (top) during odor presentation, with autocorrelation (middle) and power spectrum (bottom), under nicotinic neuromodulation alone. B, As in A, but with both nicotinic and muscarinic neuromodulation active. C, Distribution of MC spike phases with respect to sLFP oscillations under nicotinic neuromodulation alone. D, As in C, but with both nicotinic and muscarinic neuromodulation active. E, Raster plot of MC spike phases with respect to sLFP oscillations, under nicotinic neuromodulation alone. F, As in E, but with both nicotinic and muscarinic neuromodulation active. Note that muscarinic activation narrows the distribution of spike phases. To compare responses to those under muscarinic neuromodulation alone, compare with Figure 9, B, D, and F.

Figure 12.

Cholinergic effects averaged across 12 simulations with background inputs based on different randomization seeds. A, Mean MC firing rates and network oscillation frequencies. B, Mean CE index. C, Mean STS index. D, Mean spike phase-locking index. E, Mean amplitude of the power spectrum peak. F, Mean oscillation index. The phase-locking (D) and oscillation (F) indices are bounded between 0 and 1 (see Materials and Methods). Error bars denote the SD.

Figure 13.

Cholinergic neuromodulation in the full model with lower PGC→MC synaptic weight and higher GC→MC synaptic weight. The PGC→MC synaptic weight was reduced from 8 to 5 and the GC→MC synaptic weight was increased from 1 to 2. A1, Histograms of activity in all 25 glomeruli under control conditions, arranged according to their levels of afferent activation for ease of visualization. Top, Steady-state OSN input levels; middle, odor-evoked spike rates in MCs; bottom, odor-coding spike rates in MCs (see Materials and Methods). A2, Simulated LFP (top) during odor presentation, with autocorrelation (middle) and power spectrum (bottom), under control conditions. A3, Raster plot of MC spike phases with respect to sLFP oscillations, under control conditions. B1–B3, As in A1–A3, but with nicotinic cholinergic neuromodulation active. C1–C3, As in A1–A3, but with muscarinic cholinergic neuromodulation active. D1–D3, As in A1–A3, but with both nicotinic and muscarinic cholinergic neuromodulation active.

Figure 14.

Cholinergic neuromodulation in the full model with lower PGC→MC synaptic weight and higher GC→MC synaptic weight and smaller decay time constant of the GC→MC synapses. The PGC→MC synaptic weight was reduced from 8 to 5 and the GC→MC synaptic weight was increased from 1 to 2, plus the decay time constant of the GC→MC GABA_A synapses reduced from 18 ms to 3 ms. A1, Histograms of activity in all 25 glomeruli under control conditions, arranged according to their levels of afferent activation for ease of visualization. Top, Steady-state OSN input levels; middle, odor-evoked spike rates in MCs; bottom, odor-coding spike rates in MCs (see Materials and Methods). A2, Simulated LFP (top) during odor presentation, with autocorrelation (middle) and power spectrum (bottom), under control conditions. A3, Raster plot of MC spike phases with respect to sLFP oscillations, under control conditions. B1–B3, As in A1–A3, but with nicotinic cholinergic neuromodulation active. C1–C3, As in A1–A3, but with muscarinic cholinergic neuromodulation active. D1–D3, As in A1–A3, but with both nicotinic and muscarinic cholinergic neuromodulation active.

Figure 15.

Nicotinic neuromodulation of the glomerular model with stronger excitation on MCs. The maximal conductance density of the nicotinic current inserted into the MC tuft was simulated at multiple fold of the original default value (1 mS/cm²), whereas that inserted into the PGC spine was held constant at the default value (15 mS/cm²). A, Histograms of activity in all 25 glomeruli under control conditions (no nicotinic activation), artificially arranged according to their levels of afferent activation for ease of visualization. Top, Steady-state OSN input levels; middle, odor-evoked spike rates in MCs; bottom, odor-coding spike rates in MCs (see Materials and Methods). A is identical to Figure 6A and reproduced here for ease of comparison. B, As in A, but with nicotinic cholinergic neuromodulation active; the nicotinic conductance density on MCs is 1 mS/cm². B is identical to Figure 6B and reproduced here for ease of comparison. C, As in A, but with nicotinic cholinergic neuromodulation active; the nicotinic conductance density on MCs is 3 mS/cm². D, As in A, but with nicotinic cholinergic neuromodulation active; the nicotinic conductance density on MCs is 5 mS/cm². E, Average spontaneous and odor-evoked MC firing rates as a function of the maximal conductance density of the nicotinic current in MCs. F, Contrast enhancement (CE) index as a function of the maximal conductance density of the nicotinic current in MCs. The dashed horizontal line denotes the CE index in the control case.