A review such as this can do no more than provide an indication of the issues involved in using autonomic activity as a means of providing a "window on the brain." Several points arise. One of the most important is that of careful and appropriate use of techniques available. One well-known textbook of experimental psychology published some time ago advocated the use of two dimes applied to the palm of the hand for the measurement of electrodermal activity. It was this sort of recommendation that led to the use of psychophysiological measurement falling into disrepute. As indicated in the second section, it is important to understand fully the peripheral mechanisms involved before measurement of electrodermal activity can be usefully carried out. Appropriate use of silver/silver chloride electrodes and physiologically appropriate levels of saline in the electrolyte medium can lead to accurate and repeatable measurement where artefact is not carelessly introduced. Equally important is the context in which studies are carried out. The psychological invasiveness of the technique is important to recognize, and it is here that measurement of autonomic activity probably scores over other methods that are available insofar as very little restriction of the subject is required and the number of transducers that must be applied is minimal. The measurement of autonomic activity within the totality of the experimental context is all important. As an example Dawson and Schell investigated the SCR to words which had previously been associated with shock. When these words were presented to the ear to which attention was not directed in a dichotic listening paradigm, an SCR could be elicited although the subject was unaware of the presentation of the stimulus. The importance of the Dawson and Schell study was the care that they took to make sure that the subject really was unaware of the critical stimulus and had not momentarily switched attention from the attended ear. More important, their experiment, in contrast to some which had gone before, used a balanced design in which the critical stimuli were presented on different occasions to each ear. As a result of this it was found that critical stimuli, which were presented to the left ear, right hemisphere, gave rise to SCRs, even when the subject was not aware of their presentation, whereas stimuli presented to the right ear, left hemisphere elicited no response.(ABSTRACT TRUNCATED AT 400 WORDS)