When they were tested 12 h later, their performance varied as a function of whether they had slept in the intervening time period or not. Fenn et al. ran a large number of conditions to tease apart sleep from the amount of time between training and test, the time of day of training, or the time of day of testing