We develop a theoretical framework that shows how mesencephalic dopamine systems could distribute to their targets a signal that represents information about future expectations. In particular, we show how activity in the cerebral cortex can make predictions about future receipt of reward and how fluctuations in the activity levels of neurons in diffuse dopamine systems above and below baseline levels would represent errors in these predictions that are delivered to cortical and subcortical targets. We present a model for how such errors could be constructed in a real brain that is consistent with physiological results for a subset of dopaminergic neurons located in the ventral tegmental area and surrounding dopaminergic neurons. The theory also makes testable predictions about human choice behavior on a simple decision-making task. Furthermore, we show that, through a simple influence on synaptic plasticity, fluctuations in dopamine release can act to change the predictions in an appropriate manner.

1. A network model of thalamocortical (TC) and thalamic reticular (RE) neurons was developed based on electrophysiological measurements in ferret thalamic slices. Single-compartment TC and RE cells included voltage- and calcium-sensitive currents described by Hodgkin-Huxley type of kinetics. Synaptic currents were modeled by kinetic models of alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA), gamma-aminobutyric acid-A (GABAA) and GABAB receptors. 2. The model reproduced successfully the characteristics of spindle and slow bicuculline-induced oscillations observed in vitro. The characteristics of these two types of oscillations depended on both the intrinsic properties of TC and RE cells and their pattern of interconnectivity. 3. The oscillations were organized by the reciprocal recruitment between TC and RE cells, due to their manual connectivity and bursting properties. TC cells elicited AMPA-mediated excitatory postsynaptic potentials (EPSPs) in RE cells, whereas RE cells elicited a mixture of GABAA and GABAB inhibitory postsynaptic potentials (IPSPs) in TC cells. Because of the presence of a T current, sufficiently strong EPSPs could elicit a burst in RE cells, and TC cells could generate a rebound burst following GABAergic IPSPs. Under these conditions, interaction between the TC and RE cells produced sustained oscillations. 4. In the absence of spontaneous oscillation in any cell, the TC-RE network remained quiescent. Spindle oscillations with a frequency of 9-11 Hz could be initiated by stimulation of either TC or RE neurons. A few spontaneously oscillating TC neurons recruited the entire network model into a "waxing-and waning" oscillation. These "initiator" cells could be an extremely small proportion of TC cells. 5. In intracellular recordings, TC cells display a reduced ability for burst firing after a sequence of bursts. The "waning" phase of spindles was reproduced in the network model by assuming an activity-dependent upregulation of Ih operating via a calcium-binding protein in TC cells, as shown previously in a two-cell model. 6. Following the global suppression of GABAA inhibition, the disinhibited RE cells produced prolonged burst discharges that elicited strong GABAB-mediated currents in TC cells. The enhancement of slow IPSPs in TC cells was also due to cooperativity in the activation of GABAB-mediated current. These slow IPSPs recruited TC and RE cells into slower waxing-and-waning oscillations (3-4 HZ) that were even more highly synchronized. 7. Local axonal arborization of the TC to RE and RE to TC projections allowed oscillations to propagate through the network. An oscillation starting at a single focus induced a propagating wavefront as more cells were recruited progressively. The waning of the oscillation also propagated due to upregulation of Ih in TC cells, leading to waves of spindle activity as observed in experiments. 8. The spatiotemporal properties of propagating waves in the model were highly dependent on the intrinsic properties of TC cells. The spatial pattern of ...

Synaptic events are often formalized in neural models as stereotyped, time-varying conductance waveforms. The most commonly used of such waveforms is the a-function (Rall 1967): where gsyn is the synaptic conductance and to is the time of transmitter release. This function peaks at a value of l l e at t = to + 7 , and decays exponentially with a time constant of 7. When multiple events occur in succession at a single synapse, the total conductance at any time is a sum of such waveforms calculated over the individual event times.There are several drawbacks to this method. First, the relationship to actual synaptic conductances is based only on an approximate correspondence of the time-course of the waveform to physiological recordings of the postsynaptic response, rather than plausible biophysical mechanisms. Second, summation of multiple waveforms can be cumbersome, since each event time must be stored in a queue for the duration of the waveform and necessitates calculation of an additional exponential during this period (but see Srinivasan and Chiel 1993). Third, there is no natural provision for saturation of the conductance.An alternative to the use of stereotyped waveforms is to compute synaptic conductances directly using a kinetic model (Perkel et al. 1981). This approach allows a more realistic biophysical representation and is consistent with the formalism used to describe the conductances of other ion channels. However, solution of the associated differential equations generally requires computationally expensive numerical integration.In this paper we show that reasonable biophysical assumptions about synaptic transmission allow the equations for a simple kinetic synapse model to be solved analytically. This yields a mechanism that preserves the advantages of kinetic models while being as fast to compute as a single a-function. Moreover, this mechanism accounts implicitly for sat-

Recent work has identified a neuron with widespread projections to odour processing regions of the honeybee brain whose activity represents the reward value of gustatory stimuli. We have constructed a model of bee foraging in uncertain environments based on this type of neuron and a predictive form of hebbian synaptic plasticity. The model uses visual input from a simulated three-dimensional world and accounts for a wide range of experiments on bee learning during foraging, including risk aversion. The predictive model shows how neuromodulatory influences can be used to bias actions and control synaptic plasticity in a way that goes beyond standard correlational mechanisms. Although several behavioural models of conditioning in bees have been proposed, this model is based on the neural substrate and was tested in a simulation of bee flight.