In their 1983 article entitled "The function of dream sleep", Francis Crick and Graeme Mitchison proposed a revolutionary theory of the function of dreams. They believed that dreams were not the veiled expressions of subconscious fears and desires, but were simply the byproducts of a fine-tuning of memory. Due to the limits of current research methods regarding the internal processes of the mind and dreaming, it is impossible to completely validate or refute any theory regarding the function that dreams serve. Therefore in order to evaluate this theory, only the plausibility of such a theory can be examined. In order to properly evaluate any dream function theory, it is first necessary to understand the known facts regarding sleep and dreaming. In order to evaluate the function proposed by Crick and Mitchison in particular, it is also necessary to understand the basic theory and functioning of neural networks. Once a basic understanding of both dreaming and neural networks has been attained, the evidence supporting and the evidence weakening the plausibility of their theory can then be weighed against each other. Only then can a conclusion be made regarding the plausibility of the proposed function. When examined using this method, Crick and Mitchison's postulated dream theory is extremely plausible.
In order to evaluate any theory regarding dreaming it is necessary to understand the basic principles and characteristics of sleep. Immediately after falling asleep, the subject enters Stage 1, where brain activity becomes more random than when awake. After approx. 5 minutes of Stage 1, sleep passes through Stages 2, 3 and 4 (Myers 249). In these increasingly deeper sleep states, brain wave activity becomes progressively lower in frequency and higher in amplitude. After continuing through Stage 4, the subject then begins a reverse progression through the stages of sleep, returning to Stage 3 and then Stage 2 (Myers 250).