Gain modulation in visual cortex by noisy synaptic inputs

Gain modulation of neuronal responses is widely observed in the cerebral cortex of both anesthetized and behaving animals. This dissertation examines mechanisms that may underlie different forms of gain modulation observed in visual cortex. We examine mechanisms that arise at many levels, beginning at the single-cell level (modulating the effective reversal potential of synaptic inputs to a cell) and ending at the network level (nonlinear interactions arising through recurrent connectivity).;In Chapter 2, we examine how the balance between excitation and inhibition in a neuron's input may affect how that neuron integrates its inputs. We have found that if excitatory and inhibitory inputs are balanced at a specific potential (we refer to this potential as the "balanced potential"), these do not alter the firing-rate responses, provided that they do not introduce additional noise. If these inputs do increase noise, their effect is close to input-gain modulation, an effect best described as scaling of the input. If the inputs are noisy but instead balanced at a more hyperpolarized potential, their effect is more like response-gain modulation, a divisive scaling of the neuron's firing rate.;In Chapter 3, we compare the effects of noisy synaptic inputs in a configuration that divisively scales firing-rate curves with the effects of subtractive mechanisms of inhibition and explore their possible role in producing multiplicative gain modulation effects observed in vivo, such as surround suppression and attention. The divisive mechanism of inhibition is more successful in producing both response-gain and inputgain modulation effects although the effects of subtractive inhibition can appear nonlinear.;Finally in Chapter 4, we examine how network effects can produce gain modulation phenomena. Specifically, we construct a recurrent network model of attention in visual cortex, and demonstrate how multiple forms of gain modulation can arise through nonlinear interactions that arise within a recurrent framework of connectivity. Gain modulation phenomena observed in vivo may arise from all or only a subset of the mechanisms examined in this study. As discussed in the final chapter, whole-cell recording techniques that measure in vivo conductance changes may allow distinguishing between these mechanisms.