Dream interpretation

The notion that dreams contain hidden meanings has been popular for centuries. During the second century AD, Artemidorus Daldianus produced a five-volume work entitled “The Interpretation of Dreams”. In the 1890s, Sigmund Freud associated specific meanings to characters, objects, animals, and scenarios that frequently appeared in dreams. More recently, sleep scientists have developed increasingly sophisticated ways of coding dreams.
There are more than 150 dream rating and content analysis scales today, and the best validated and most widely used one remains the Hall and Van de Castle’s. This coding system sees a dream as:
a) a cast of characters (for example, a depressed man, a friendly woman, a cute dog);
b) a plot in which characters interact with each other (for example, the depressed man patting the dog);
c) a process reflecting different affective states (for example, the depressed man now feels safe and relaxed).

Problem

Dreams are coded according to the Hall and Van de Castle’s scale manually and as such the number of dreams that can be coded is very limited.

Solution

We designed a Natural Language Processing tool that automatically scored 24K dream reports according to the Hall and Van de Castle’s scale. From each dream report, the tool extracted nouns to identify people, animals, and fictional characters, and verbs to classify interactions in terms of friendly interactions or acts of aggression. The results suggest that it is possible to build future technologies that bridge the current yawning gap between real life and dreaming, ultimately making our ‘sleeping mind’ quantifiable.