Effects of loud noise and signal probability on visual vigilance.

Abstract The combined effects of training and loud (95 dBC) white noise were investigated in a visual version of the Bakan vigilance task. Two signal probabilities, high (p=0·02) and low (p=0·), were factorially combined to give the following training/test combinations: high/high, high/low, low/high and low/low. While signal detection was sensitive to task length only, response times were sensitive to both training and noise level. Noise tended to lengthen response times for those people trained in, and then receiving, the low-probability schedule and this effect increased over time. The memory component of the task (recall of pre-signal digits) was independently sensitive to both noise and signal probability. The results imply that it is the high memory component of the task which allows a variety of strategies to be employed in order to perform the task. The measures do not allow us to differentiate performance effects due to changes in efficiency from those due to changes in strategy.