Signal processing by single neurons: Biophysical mechanisms and implications for nociception

Neurons transform analogue input into digital representations (i.e. spike trains) that can be communicated to other neurons. In the process of incorporating information from different sources, representations are modified. These modifications constitute computations. This is particularly relevant in the pain system where, according to the gate control theory, sensory input from multiple modalities as well as cognitive information influence the relationship between noxious input and pain perception. Much of this control is exerted in the spinal dorsal horn, including lamina I, but the biophysical basis for those computations remains unknown. This thesis aims to elucidate basic computational processes involved in signal processing by single neurons, especially as they relate to nociceptive processing by neurons in lamina I. Intrinsic cellular properties of lamina I neurons were characterized using a spinal slice preparation and whole cell patch clamp recordings. The neuron population is divisible into four classes on the basis of spiking pattern. Neurons from different classes encode information in fundamentally different ways. Subsequent investigation revealed that tonic neurons act as integrators because of voltage-dependent inward currents that prolong subthrehsold depolarizations and allow repetitive spiking. Single spike neurons, on the other hand, act as coincidence detectors because of their predominant outward current. Encoding mode (integration vs. coincidence detection) determines the information transmitted by each cell type. Inhibition is also known to play an important role in nociceptive processing. It has been debated whether shunting inhibition is able to divisively modulate firing rate (i.e. implement gain control). Results presented here demonstrate that divisive modulation by shunting inhibition depends on background synaptic noise and dendritic saturation. Further analysis reveals that effects of noise can be understood on the basis of a noise-induced switch in spike generating mechanism. Implications of these computational processes for nociception are discussed.