Abstract
The middle latency response (MLR) using wide band-pass filters shows marked changes in amplitude, latency and configuration in sleep. The components with latencies greater than 20 ms show the greatest variability. There is a significant increase in Pa latency in stages 2 and stages 3/4, and in some cases a disappearance of the Nb component with the development of a broad positivity of latency intermediate to Pa and Pb which dominates the response. The responses in REM are of similar latency and configuration as in wakefulness but of reduced amplitude. The 40 Hz response is markedly reduced in amplitude in all sleep stages reflecting a decrease in the contribution of the middle latency components to this composite response. This appears to arise through a loss of 40 Hz periodicity in slow wave sleep and an increase in the slow 10 Hz component. In REM sleep, there is an overall reduction in amplitude. Much of the reported variability of the MLR in the literature arises from the widely differing band-pass filters used and the inadequate control for level of arousal. Both these factors have been shown to produce significant changes in the response.