BRAIN POTENTIALS DURING SLEEP

Electrical potentials of the human brain, led from the intact scalp, afford an objective measure of the activity of this organ which may be car-related with other physiological or psychological states. The striking change in neural function between wakefulness and sleep has already been examined with the electrical technique by several workers. Adrian and Yamagiwa (1935) as well as' Loomis, Harvey and Hobart (1936) have found the lO/sec. rhythm of the resting wakeful state gone or rare in sleep; when it is replaced by random potentials, or more regular ones at 3-5/set. (Gibbs, Davis and Lennox, 1935), or brief trains at 14/set. (Loomis et al.) which wax and wane during a period of a second. The present experiments were likewise directed to a study of the depth of sleep, and to a lesser extent of hypnosis, with the aid of cerebral potentials. METHODS. 1. Electrodes. Copper or silver disc electrodes 1 or 2 cm. in diameter were placed on occiput and forehead. Electrode jelly insured good contact, with the total input resistance regularly under 5000 ohms. Stray potentials from heart, respiratory and eye muscles, etc., were largely eliminated or were controlled. 2. Recording system. A four-stage, resistance capacity-coupled amplifier fed a loud speaker, a cathode ray oscillograph, and a crystograph (Offner and Gerard, 1936) or some pair of these. The amplifier reproduced 3 microvolt well above the base line and sine waves at l/set. showed a reduction in amplitude of only 10 per cent. At times a variable low-pass filter was inserted in the output to attenuate the higher frequencies, chiefly above lOO/sec. Records were taken continuously through the night or at desired intervals. 3. Measurement of sleep depth. Since the primary objective of this work was to establish a correlation between depth (or intensity) of sleep and electrical patterns, it was necessary to test the former independently. Sleep intensity is usually measured by the minimal intensity or duration of a stimulus necessary to elicit some predetermined response. Obviously sleep intensity might be modified by each test, so that valid depth curves could be obtained only when 1 or 2 tests were made a night and many data considered statistically. We have ordinarily used 3 or 4 stimuli in one night, which probably have not seriously altered the curve; but even con-692