|
A fundamental problem facing sensory systems is how to encode accurately the vast range of stimulus intensities present in the environment. In the case of the auditory system, one challenge is to encode sound intensities extending over 12 orders of magnitude (120 decibels of Sound Pressure Level, dB SPL - a range that extends from the sound of a pin-drop to that of a jet engine). Human listeners can discriminate changes in sound intensity of approximately 1 dB at any sound level over this range. Although the auditory system has likely not evolved to take account of the specific examples provided, natural sounds encountered cover a similar range. In this context, an important principle concerns the brain ' s ability to adapt to the local (in time) sound environment. Brain mechanisms responsible for such adaptive coding include spike-frequency adaptation, one definition of which might be that it is the reduction over time of the number of action potentials generated to a constant stimulus. At this juncture, however, the question arises as to what constitutes a constant stimulus. In the case of the auditory system, for example, this might be a sound of constant intensity, or a distribution of sounds with constant mean intensity or variance. Although spike-frequency adaptation invariably contributes to the coding of such sound features - and is evident at the earliest stages of auditory processing - the sheer complexity of the acoustic scene means that adaptive mechanisms are likely based on the interplay of cell- and circuit-based mechanisms. Using spatial hearing as an example, I will discuss the contribution of brain circuits to adaptive coding, demonstrating that although spike-frequency adaptation is often a critical contributing factor to adaptive coding, the extent to which it occurs in any one brain nucleus can be limited by specific requirements of the task in hand. I will then discuss the influence of cortical feedback on adaptive coding, demonstrating that although neural responses in the auditory midbrain readily adapt in the absence of cortical input, this feedback input is necessary if the population is to adapt in a manner that is informative about the stimulus features. |
|