A core circuit module for cost/benefit decision

Abstract

A simple circuit for cost-benefit decision derived from behavioral and neural studies of the predatory sea-slug Pleurobranchaea may closely resemble that upon which the more complex valuation and decision processes of the social vertebrates are built. The neuronal natures of the pathways in the connectionist model comprise classic central pattern generators, bipolar switch mechanisms, and neuromodulatory state regulation. Marked potential exists for exploring more complex neuroeconomic behavior by appending appropriate circuitry in simulo.

Keywords: Pleurobranchaea; approach/avoidance; central pattern generator; decision making; neuroeconomics; neuronal switch; simulation.

Figure 1

A Pleurobranchaea with chemotactile oral veil and rhinophores indicated.

Figure 2

Conservation of appetitive state in the isolated CNS. Spontaneous feeding nerve burst frequency correlated with donors’ sensory feeding thresholds to the appetitive stimulant betaine (trimethylglycine). (A) Spontaneous burst frequency recorded from buccal motor nerve R3 of isolated CNSs was less from high-threshold donors than from low-threshold donors. (B) R3 burst frequency was an approximately linear function of donor feeding thresholds on a log–log plot (n = 25; R² = 0.54 and 0.59 for proboscis extension and biting, respectively). Line fits were by least squares. Three high-threshold donor CNSs did not show burst patterns in R3 and were excluded from the figure. From Hirayama and Gillette (2012).

Figure 3

Modeling homeostatic decision. (A) Excitation state of the homeostatic feeding network switches avoidance to approach. (B) Sensory inputs for resource quality, sensory signatures, and nociception access sensory networks for Incentive and Deterrence, which promote excitation of feeding and avoidance turns, respectively. Excitation in the homeostatic network suppresses avoidance and promotes orienting turns (approach). Active avoidance and satiation inhibit the homeostatic network, while homeostatic network activity suppresses Deterrence. Modulatory feedback pathways from the Feeding and Avoidance networks potentiate learning of sensory signatures, mediating reward. Modified from Hirayama and Gillette (2012).

Figure 4

Simulation suggests that the turn network is default organized for avoidance responses to unilateral sensory inputs (left), and that inputs from the feeding network could reverse responses to orienting (right) via a hypothetical dyad of switch neurons (circled with a broken orange line). Double-headed connections indicate reciprocal excitatory connections.

Figure 5

Dependence of approach/avoidance decision on the excitation state of the feeding motor network. Fictive avoidance (A) was switched to orienting (B) following penetration of a feeding command neuron (PCp; Gillette et al., 1982) whose firing induced rhythmic bursting in a feeding nerve (R3). Hyperpolarization of PCN (C) suppressed fictive feeding and restored avoidance. From Hirayama and Gillette (2012).