| Previous work in this laboratory has shown that for many neurons in the inferior colliculus, responses to dynamic stimuli cannot be predicted on the basis of static tuning to an important binaural cue to sound source location, the interaural phase disparity (IPD). Instead, the response of the neuron at a particular value of IPD may be dramatically enhanced or suppressed— conditioned—by the temporal context in which that value occurs. I will show that profound sensitivity to stimulus context is the rule, rather than the exception in the central auditory system, and that adaptive gain control is a general feature of central auditory processing. Conditioning effects in the gerbil auditory midbrain generalize to dynamic interaural level differences, which alter the balance of excitation and inhibition without engaging binaural pathways related to IPD processing. Monaural frequency transitions, which do not simulate spatial motion, also condition the responses of IC neurons. Using sets of glide and step stimuli converging on a common target, the response to an identical stimulus was shown to be a function of the preceding frequency, even when that frequency was outside the excitatory frequency response area of the cell. Many neurons responded significantly differently to targets following origins that elicited statistically indistinguishable responses. Selective adaptation among the neuron's variously tuned afferents is proposed to help engender stimulus-specific conditioning of this sort. Encoding a given stimulus parameter in firing rate appears to be a necessary and sufficient condition for exercising adaptive gain control. Conditioning also generalizes across species, structure, and anesthetic state. In awake rhesus monkeys, the cortical representation of IPD is also profoundly context-dependent—changes in firing rate associated with equivalent stimuli occurring in different contexts are comparable to changes in discharge rate that establish cortical tuning to the cue itself. This process is subserved by multiple adaptive mechanisms operating on timescales ranging from tens of milliseconds to seconds, and reflects an adaptive coding strategy that prioritizes the representation of stimulus changes over actual stimulus values. I discuss the mechanisms that could support sensitivity to recent stimulus history, and how they might contribute to processing dynamic acoustic signals. |
