Dendritic action potentials connect distributed dendrodendritic microcircuits

Abstract

Lateral inhibition of cells surrounding an excited area is a key property of sensory systems, sharpening the preferential tuning of individual cells in the presence of closely related input signals. In the olfactory pathway, a dendrodendritic synaptic microcircuit between mitral and granule cells in the olfactory bulb has been proposed to mediate this type of interaction through granule cell inhibition of surrounding mitral cells. However, it is becoming evident that odor inputs result in broad activation of the olfactory bulb with interactions that go beyond neighboring cells. Using a realistic modeling approach we show how backpropagating action potentials in the long lateral dendrites of mitral cells, together with granule cell actions on mitral cells within narrow columns forming glomerular units, can provide a mechanism to activate strong local inhibition between arbitrarily distant mitral cells. The simulations predict a new role for the dendrodendritic synapses in the multicolumnar organization of the granule cells. This new paradigm gives insight into the functional significance of the patterns of connectivity revealed by recent viral tracing studies. Together they suggest a functional wiring of the olfactory bulb that could greatly expand the computational roles of the mitral-granule cell network.

Fig. 1

Basic properties of the two kind of cells that we used in our model. (a) (left) Schematic representation of the mitral cell morphology; colored squares indicate recording locations for which membrane potential is shown on the right; open circles indicate locations of the inhibitory synapses from granule cells; (right) membrane potential at different locations (lines color correspond to locations indicated on the left part) during a weak (top) or strong (bottom) odor activation. In both cases, the first action potential is shown on an expanded scale. (b) a Schematic representation of a granule cell morphology used in the simulations; b simulation of the somatic membrane potential during a suprathreshold current step injection under control conditions (light line) and after the pharmacological block of the K_A current (thick line); c, d, e membrane potential of the soma (black lines) and spines (colored lines) of a granule cell during a train of mitral cell action potentials (bottom plots) activating the red synapse; different synaptic locations were used in c, d, and e; the dashed lines represent the threshold for GABA release

Fig. 2

Non-topographical connectivity of the granule–mitral cells network requires a new paradigm for an efficient lateral inhibition mechanism; traces are mitral cell model (M1-2) somatic recordings; dendrodendritic synapses between mitral and granule cell (GC1-2) indicated with curved arrows and open (excitatory) and closed (inhibitory) circles. (a) Original model of spatial center-surround organization of lateral inhibition in the olfactory bulb. (b) Increasing the distance between mitral cells (belonging to different glomeruli) weakens the inhibition of the distant cell: the inhibitory postsynaptic signal (IPSP) spreads decrementally in the lateral M2 dendrite to reach the soma. (c) Inhibition can be independent of distance if it is imposed locally by granule cells activated by a backpropagating action potential (bAP). (d) Schematic representation of the larger network used to test action potential-mediated lateral inhibition in a more realistic arrangement for a glomerular unit; groups of 20 GCs were clustered around each mitral cell soma, forming random dendroden-dritic contacts with overlaying lateral dendrites. Traces are mitral cell somatic recordings from 10 simulations; for clarity, only one granule cell and its connectivity (open and closed circles) are explicitly drawn

Fig. 3

Schematic representation of the reduced network used in most simulations, and its effect for different odor inputs (n=3…11). M1, M2, and M3 indicate Mitral cells, GC1, GC2, and GC3 indicate granule cells. Uniform all-to-all connectivity was assumed in this case. Each histogram represents the number of somatic APs elicited in a given mitral cell by each odor input without (top histograms) or in the presence of granule cells in the network (bottom histograms). The effect of lateral inhibition in abolishing M2 output for flanking odors is schematically represented by the black bars; Weak input was modeled with a 60% reduction of the strong one. Locations of dendrodendritic synaptic contacts are indicated with small open and closed circles. In all cases, synapses on the granule cells were positioned as in Fig. 1(b d); for clarity, each GC is represented with only the branch including the synapses

Fig. 4

Effects of lateral inhibition and comparison with experiments. (a) (left) Single unit recordings of mitral cells activity during presentation of different odors (4-, 6-, and 8-CHO); (right) somatic potential of a mitral cell from our model (M2) during presentation of different odors (5, 7, and 9); (b) (left) experiments, effects of bicuculline application; (right) model, somatic membrane potential of M2 during odor presentation after block of dendrodendritic interactions. Experimental traces taken and redrawn from Yokoi et al. (1995)

Fig. 5

Detailed dynamics of the interaction between mitral and granule cells during an odor presentation, using the network in Fig. 3. (a) Time course of the input excitatory conductance, corresponding to odor 9, activated in the tuft of each mitral cell (M1, M2, M3); for the purpose of this figure, the total peak excitatory input on M3 and all the peak inhibitory conductances were 17 and 18 nS, respectively. (b) Somatic membrane potential of the three mitral cells during odor presentation. (c) Membrane potential of one of the secondary dendrites of mitral cell M3 at different distances from the soma; the dotted line indicates the threshold for the activation of the mitral-to-granule excitatory synapse. (d) Membrane potential of the granule cells at the location of the dendrodendritic synapse with mitral cell M3 (top plot) and at the soma (bottom plot) during odor presentation. (e) Time course of the local dendritic inhibitory conductances activated on the M3 secondary dendrite during odor presentation

Fig. 6

Schematic representation of a network using mitral cells without lateral dendrites, to test the effect of different odor inputs as in Fig. 3. M1, M2, and M3 indicate mitral cells, GC1, GC2, and GC3 indicate granule cells. Uniform all-to-all connectivity was arranged through mitral cells soma and granule cell dendrites; Top histograms represents the number of somatic APs elicited in a given mitral cell by each odor input without granule cells in the network. The response of M1 in the presence of the granule cells network is reported in the bottom histogram. Weak input was modeled with a 60% reduction of the strong one. Locations of dendrodendritic synaptic contacts are indicated with small open and closed circles. In all cases, synapses on the granule cells were positioned as in Fig. 1(b d), with a peak inhibitory conductance of 9 nS; traces were obtained from a simulation with a total synaptic input on M3 of 17 nS, as in Fig. 5

Fig. 7

Model of predicted gating effects during odor presentation. Arrow thickness represents the strength of the input to each mitral cell tuft, and locations of dendrodendritic synaptic contacts are indicated with small open and closed circles. (a) Traces are shown at selected locations (colored squares) and shown in different colors; (red) mitral cell soma, (blue) granule cell soma, (cyan) granule cell synapse with M2 lateral dendrite, (orange) M1 lateral dendrite at 200 and 400 μm from the soma; note that the odor does not activate M2. (b) As in (a) but with an odor that also activates M2. Note the EPSPs in GC2, activated by M2 (cyan). The spike train of M1 (orange) decrements because local synaptic activity in GC2 (cyan). Experimental traces (exp) from Lowe (2002)

Fig. 8

Gating of action potential backpropagation with a larger GC network. (a) Typical traces from a simulation during presentation of odor-9; (top) membrane potential of M2 at the soma and lateral dendrite at 200 and 400 μm from the soma; (bottom) overlapped dendritic recordings from five GCs. (b) Somatic and dendritic recordings during presentation of odors a and b as in Fig. 6; (red) mitral cells soma; (orange) M1 lateral dendrite at 400 μm from the soma; compared with the analogous traces in Fig. 6