Background noise improves gap detection in tonically inhibited inferior colliculus neurons.

Single units in the inferior colliculus (IC) in the C57Bl/6 inbred mouse strain were tested for their temporal processing ability as measured by their minimum gap threshold (MGT), the shortest silent interval in an ongoing white-noise stimulus which a unit could encode. After ascertaining the MGT in quiet, units were re-tested in various levels of background noise. The focus of this report is on two types of tonically responding units found in the IC. Tonically inhibited (TI) units encoded gaps poorly in quiet and low levels of background noise as compared with tonically excited (TE) units. In quiet, the MGTs of TI units were about an order of magnitude longer than the MGTs typical of TE units. Paradoxically, gap encoding was improved in high levels of background noise for TI units. This result is unexpected from the traditional viewpoint that noise necessarily degrades signal processing and is inconsistent with psychophysical observations of diminished speech and gap detection processing in noisy environments. We believe the improved feature detection described here is produced by the adaptation of inhibitory input. Continuous background noise would diminish the inhibitory efficacy of the gap stimulus by increasing the latency to the onset of inhibition and decreasing its duration. This would allow more spontaneous activity to "bleed through" the silent gap, thus signaling its presence. Improved feature detection in background noise resulting from inhibitory adaptation would seem an efficient neural mechanism and one that might be generally useful in other signal detection tasks.