A computational theory of the subjective experience of flow

Abstract

Flow is a subjective state characterized by immersion and engagement in one’s current activity. The benefits of flow for productivity and health are well-documented, but a rigorous description of the flow-generating process remains elusive. Here we develop and empirically test a theory of flow’s computational substrates: the informational theory of flow. Our theory draws on the concept of mutual information, a fundamental quantity in information theory that quantifies the strength of association between two variables. We propose that the mutual information between desired end states and means of attaining them — $$I({M;E})$$ I ( M ; E ) — gives rise to flow. We support our theory across five experiments (four preregistered) by showing, across multiple activities, that increasing $$I({M;E})$$ I ( M ; E ) increases flow and has important downstream benefits, including enhanced attention and enjoyment. We rule out alternative constructs including alternative metrics of associative strength, psychological constructs previously shown to predict flow, and various forms of instrumental value.

The Structure of Immersive and Engaging Activities: Insights From a Computational Model of Flow

“…To the extent that 𝐼(𝑀; 𝐸) really does underlie flow, it also serves as an answer to the basic question of how representations of task structure and flow interact, and as a tool that can help people harness flow in their daily lives. It remains to be seen whether the informational theory of flow stands up to empirical scrutiny, though initial tests give reason for optimism (Melnikoff et al, 2022). Regardless, testing the predictions presented in this chapter would yield valuable insights, and, we believe, should be a major item on the agenda for future research.…”

“…Critically, mutual information is one of many ways of quantifying associative strength. We have shown that it outperforms other metrics as a predictor of flow (Melnikoff et al, 2022), but it may not fare as well as a predictor of other self-regulatory phenomena.…”

“…We recently obtained direct empirical support for the informational theory of flow across five experiments (Melnikoff et al, 2022). Most experiments used a computerbased task in which 𝑝 (𝑚) and 𝑝 | (𝑒|𝑚) can be precisely manipulated.…”

“…Research inspired by goal-systems theory has produced two insights of particular relevance to the present analysis. One is a computational model of flow whose formal structure permits quantitatively precise predictions about immersion and engagement (Melnikoff, Carlson, & Stillman, 2022). The other is a taxonomy of ways in which people represent task structure (Kruglanski, 2022;Kruglanski et al, 2015;Kruglanski et al, 2002).…”

Pushing the Bounds of Bounded Optimality and Rationality

“…This view comports with accounts that operationalize effort and information as inherently rewarding, to explain seemingly irrational behavior (Inzlicht, Shenhav, & Olivola, 2018). Furthermore, accounts on flow (Csikszentmihalyi, 1990;Melnikoff, Carlson, & Stillman, 2022;Wilson, Shenhav, Straccia, & Cohen, 2019), boredom (Geana, Wilson, Daw, & Cohen, 2016), curiosity (Schmidhuber, 1991), and fatigue (Agrawal et al, 2021) suggest mechanisms for investing cognitive resources not only to accommodate current bounds, but to optimally change those bounds. In line with normative theories of learning (Dubey & Griffiths, 2020;Kidd & Hayden, 2015), human infants and macaques will allocate attention to stimuli that are intermediately surprising (Cubit, Canale, Handsman, Kidd, & Bennetto, 2021;Wu et al, 2021), and adults will self-organize their curricula to maximize learning and reward (e.g., Ten, Kaushik, Oudeyer, & Gottlieb, 2021).…”