In Search of Emotional Patterns In Dream Data: From Theoretical Frameworks To Algorithms

Due to its complexity, dreams have always been tackled from different disciplinary angles, ranging from anthropology, philosophy of mind, religious studies, psychology and, recently, cognitive science. In this talk, we will highlight how some theoretical frameworks can lead to computational examination of large collection of dreams. One such approach is based on the concept of emergent systems, omnipresent in nature. It is in human nature to search for patterns in seemingly noncoherent elements. Dreams, too, can be considered as natural phenomena (of the human mind) and can be examined for patterns. Automated recognition of such patterns can be of symbolic nature, for example, we can search for the most frequent symbols, conveyed in words or phrases. Since dreaming often times appears to be very intense nocturnal experience, the interpretation of patterns inevitably leads us to emotional aspect of the dreamed content. Random neuronal firing theory aside, what is the function of emotionally charged dreams? We will look into the second theoretical framework that deals with the notion of the self-regulating system of the psyche, where mental health activity is considered to be the key function of dreaming (also known as the compensatory function of dreaming). Would it be possible to detect prevailing moods at the population level and ultimately have an insight about the global patterns that emerge in our dreams? We will discuss the role of emotions in dreaming and highlight the current challenges in automatic detection of emotional states (‘moods’) in a large collection of dream reports.