Acta Pharmacologica Sinica (2009) 30: 681-693; doi: 10.1038/aps.2009.87

Neuromodulator action has received increasing attention in theoretical neuroscience.  Yet models involving both neuronal populations dynamics at the circuit level and detailed receptor properties are only now being developed.  Here we review recent computational approaches to neuromodulation, focusing specifically on acetylcholine (ACh) and nicotine.  We discuss illustrative examples of models ranging from functional top-down to neurodynamical bottom-up.  In the top-down approach, a computational theory views ACh as encoding the uncertainty expected in an environment.  A different line of models accounts for neural population dynamics treating ACh as toggling neuronal networks between read-in of information and recall of memory.  Building on the neurodynamics idea we discuss two models of nicotine’s action with increasing degree of biological realism.  Both consider explicitly receptor-level mechanisms but with different scales of detail.  The first is a large-scale model of nicotine-dependent modulation of dopaminergic signaling that is capable of simulating nicotine self-administration.  The second is a novel approach where circuit-level neurodynamics of the ventral tegmental area (VTA) are combined with explicit models of the dynamics of specific nicotinic ACh receptor subtypes.  We show how the model is constructed based on local anatomy, electrophysiology and receptor properties and provide an illustration of its potential.  In particular, we show how the model can shed light on the specific mechanisms by which nicotine controls dopaminergic neurotransmission in the VTA.  This model serves us to conclude that detailed accounts for neuromodulator action at the basis of behavioral and cognitive models are crucial to understand how neuromodulators mediate their functional properties.

Keywords: mean-field model; nAChR kinetics; nicotine; computational model