Stimulus and Network Dynamics Collide in a Ratiometric Model of the Antennal Lobe Macroglomerular Complex

Abstract

Time is considered to be an important encoding dimension in olfaction, as neural populations generate odour-specific spatiotemporal responses to constant stimuli. However, during pheromone mediated anemotactic search insects must discriminate specific ratios of blend components from rapidly time varying input. The dynamics intrinsic to olfactory processing and those of naturalistic stimuli can therefore potentially collide, thereby confounding ratiometric information. In this paper we use a computational model of the macroglomerular complex of the insect antennal lobe to study the impact on ratiometric information of this potential collision between network and stimulus dynamics. We show that the model exhibits two different dynamical regimes depending upon the connectivity pattern between inhibitory interneurons (that we refer to as fixed point attractor and limit cycle attractor), which both generate ratio-specific trajectories in the projection neuron output population that are reminiscent of temporal patterning and periodic hyperpolarisation observed in olfactory antennal lobe neurons. We compare the performance of the two corresponding population codes for reporting ratiometric blend information to higher centres of the insect brain. Our key finding is that whilst the dynamically rich limit cycle attractor spatiotemporal code is faster and more efficient in transmitting blend information under certain conditions it is also more prone to interference between network and stimulus dynamics, thus degrading ratiometric information under naturalistic input conditions. Our results suggest that rich intrinsically generated network dynamics can provide a powerful means of encoding multidimensional stimuli with high accuracy and efficiency, but only when isolated from stimulus dynamics. This interference between temporal dynamics of the stimulus and temporal patterns of neural activity constitutes a real challenge that must be successfully solved by the nervous system when faced with naturalistic input.

Non-linear blend coding in the moth antennal lobe emerges from random glomerular networks

“…Individual neuron dynamics of LNs and PNs were modeled using a first-order differential equation (Chong et al, 2012) that described the evolution of the firing-rate activation variable of a neuron over time: $\tau \frac{d{a}_i}{dt}=-a_i(t)+S\left({\displaystyle \sum _{j\in P}{w}_{i,j}}{a}_j(t)+{\displaystyle \sum _{k\in L}{w}_{i,k}}{a}_k(t)+{\displaystyle \sum _{d\in R}{v}_{i,d}}{r}_d\right)$ $\text{with} S(x)=x^3/({0.5}^3+x^3) \text{for} x\ge 0,$ $\text{and} S(x)=0 \text{for} x<0,$…”

Rapid processing of chemosensor transients in a neuromorphic implementation of the insect macroglomerular complex

“…For FPA dynamics, the network encoded ratios through the spatial identity of neurons, whereas LCA networks produced a PN output continuously varying in time (bottom panels in Figure 4). While both operating regimes were able to encode the pheromone ratio reliably, we observed in previous work (Chong et al, 2012) that the ratio encoding was faster and more accurate using LCA dynamics (see Discussion). Timing simulations of the MGC design were conducted in ModelSim (Mentor Graphics, USA) and showed complex spatiotemporal dynamics in both PN and LN cell equivalent to those found in the continuous version of the model.…”

“…To perform ratiometric processing of real-time chemosensor input we first developed a biologically-constrained computational model of the early stages of ratiometric processing in insects, the MGC of the AL, in Matlab (delete first) (Mathworks, USA) using probabilistic connectivity patterns based on morphological studies of the moth brain (Akers and O'Connell, 1991; Hansson et al, 1992; Sun et al, 1997; Hansson and Anton, 2000; Carlsson and Hansson, 2002; Carlsson et al, 2002)—shown in schematic form in Figure 1A and described in detail in (Chong et al, 2012). …”

“…Probabilistic connection rules between neurons were used to ensure that results did not depend upon specific weight or connectivity configurations—see Chong et al (2012) for further continuous model details. Figure 1B shows the complex time dynamics in LN and PN population activity obtained in response to a fixed ratiometric input.…”

Data-driven inference of network connectivity for modeling the dynamics of neural codes in the insect antennal lobe

“…We model the network by three vectors $\overrightarrow{x},\overrightarrow{y},$ and $\overrightarrow{z}$, where each element in the vector represents a neuron and modeled by a firing-rate unit (Linster et al, 2005; Capurro et al, 2012; Chong et al, 2012). The three vectors correspond to the three anatomical groups RNs, PNs, and LNs, respectively:…”