A Computational Theory of Flow

Flow is a coveted psychological state characterized by immersion and engagement in one’s current task. The benefits of flow for productivity and health are well-documented, but a rigorous description of the flow-generating process remains elusive. To achieve a deeper understanding of flow, we developed and empirically tested a theory of its computational substrates — the informational theory of flow. At the heart of our theory is the concept of mutual information, a fundamental quantity in information theory that quantifies the strength of association between two variables. Our central claim is that the mutual information between desired end states and means of attaining them — I(M;E) — gives rise to flow. We obtained support for this prediction across three experiments (two preregistered): increasing the value of I(M;E) in a computer game reliably increased the level of flow that the game evoked. Increasing I(M;E) also had important downstream benefits. It enhanced attention, as revealed by faster and less variable response times, and made the game more enjoyable. These findings could not be attributed to any of seven alternative constructs, including alternative metrics of associative strength, psychological constructs previously shown to predict flow, and various forms of instrumental value. In fact, I(M;E) was the only construct that reliably predicted flow across all three experiments, suggesting that it plays a central role in the flow-generating process. By supporting the informational theory of flow, our findings shed light on the computational substrates of task immersion and engagement and reveal a novel means of enhancing enjoyment and attention.