Temporal processing and activity-dependent cellular learning rules

One of the major questions in neuroscience research is how information processing emerges from the propagation and regulation of activity in networks of neurons. In this dissertation I examine the functional characteristics of a particular type of information processing as well as two of the mechanisms that contribute to shaping network computation.; Temporal processing on the millisecond scale is an important aspect of several tasks done by the nervous system. However, little is known about the specificity of the underlying mechanisms. We conducted auditory psychophysics studies in which subjects learned to improve their discrimination of a specific unfilled interval. Learning generalized to the trained time span marked by filled interval (duration) stimuli, and untrained tone frequencies. However, there was no generalization to untrained intervals. Thus, our data indicates that temporal processing relies on centralized, but interval-specific timers.; These timing mechanisms must ultimately rely on the ability of networks of neurons to perform temporal computations. Spike-timing dependent plasticity (STDP) is a process by which the order of, and interval between pre and postsynaptic activity determines whether the strength of the synapse is potentiated or depressed. The established model of synaptic plasticity depends on the coincidence detection properties of the postsynaptic NMDA receptor. In computational models we show that plasticity based only on this receptor can not produce order-specific STDP, and suggest that a second coincidence detector must be present as well.; Finally, while STDP produces synapse-specific changes, intrinsic and synaptic properties can be modified homeostatically due to longer-term shifts in overall network activity. We demonstrate that both intrinsic excitability and functional inhibition exhibit bi-directional homeostatic plasticity in organotypic hippocampal slices, and that intrinsic plasticity is engaged before that of inhibition. These results suggest that different forms of plasticity may be engaged independently in order to minimize changes in the relative balance of excitation and inhibition while preserving information stored in synaptic strengths.; Overall, the work I present here details the expression of both synapse-specific and global network learning rules. Additionally, I show psychophysical data that can be used to constrain models of how this plasticity may produce or affect temporal processing.