Untitled Document

Integrate-and-Fire (IAF) Model of Cost-Benefit Decision-Making in Pleurobranchaea

This IAF model is the first step in modeling the neural network interactions of cost-benefit decision-making. It reproduces major aspects of feeding network function, based on the known structure. It replicates the ability of a command paracerebral neuron (PCN) to drive the feeding network, suppression of the network by food-avoidance learning through tonic activity of the presynaptic interneurons I2 and I1, and suppression by escape swim through corollary inputs from the swim CPG to I1.

In this network a feeding stimulus excites protractors P, retractors R and the I2 cell simultaneously.
Depolarization of the PCN excites the P population, causing potent positive feedback.
The R->I2 loop is slowly recruited via C, resulting in onset of the burst phase with inhibition of P<->PCN via the potent I2 <--> I1 --> PCN pathway.
The I2<-> I1 connection acts as a pacemaker element, mediating the transition from protraction to retraction.
The synaptic strength of the R->I2 connection requires delicate adjustment in order to avoid an indefinitely prolonged retraction burst.
Similar regulation of I2 activity simulates effects of food-avoidance learning (see further).

Three behavioral consequences of PCN induction of the above IAF Model

	Induced Feeding Rhythm This network version permits pattern generation by activation of the PCN, as in the real animal, due to the putative connector neuron (C) that has been introduced in the model.
	Escape Swimming Suppression of Feeding Input from the swimming network disturbs the feeding rhythm in the old network, but does not suppress it entirely. Arrows indicate the period of current injection into I1.
	Satiation/Learning-Induced Suppression of Feeding Satiation and food avoidance learning regulate activity of the feeding network by increasing the networks threshold of excitability. There are a number of ways of effecting this change. It can either function by changing the threshold of excitability in certain network neurons, which is most likely the case in satiation, or it can change synaptic strengths, as seen in learning. This picture demostrates disruption of feeding rhythm when threshold of excitation of I2 is increased. I2 was chosen, because it participates in I1<->I2 loop that plays the important role of driving the network, as indicated above, and thus is likely to be a site for regulation.

We plan to replicate cost-benefit decisions by incorporating the bilateral neuron populations of orienting/avoidance turning network and chemo- and nociceptive sensory lines. We expect also to add virtual neuromodulatory elements that regulate network arousal state and appetite in an effort to obtain thresholds for network activation consistent with those of the utility theory model and the actions of 5-HT in the real animal.

Back

Home