On the basis of the total information obtained from the collection of these data, four basic behavioral-electrophysiological patterns were defined: (a) awake, activated; (b) drowsy, slow wave; (c) asleep, spindle burst; and (d) asleep, activated, and each minute of recording was classified according to its dominant pattern. Analysis of these data disclosed several quantitative measures which were highly consistent both for a given animal and among the animals observed. These were: (1) The percentage occurrence of each of the patterns during a 23 h period. The group means and standard deviations were: awake, 28 ± 4.2%; drowsy, 14.3 ± 1.2%; spindle burst sleep, 42.2 ± 4.1%; and activated sleep, 15.5 ± 2.2%. The percent of total sleep time for activated sleep was 27.5%. (2) Pattern percentage in relation to time: a general periodicity in the occurrence of sleeping and waking patterns indicated reciprocal peaks related to specific time periods. During the peak occurrence of the sleep patterns, a sleep EEG cycle of 25.8 ± 7.1 min was determined.
Other interesting electrophysiological characteristics of the observed patterns were noted, including transitional phenomena associated with their respective interactions. It is suggested that the stable percentage data provide the basis for a statistical approach to the experimental study of sleep mechanisms.