To tap or not to tap
Theoretical and cognitive neurosciences to explain smartphone behaviour
Jean-Pascal Pfister and Arko Ghosh
Recorded 29 May 2017 in Lausanne, Vaud, Switzerland
Event: IC Colloquia - EPFL IC School Colloquia
Discovering the fundamental principles that govern human behavior is a major and exciting challenge. The rise of smartphones has created unprecedented opportunities for cognitive neuroscience to study unconstrained human behavior reversing the traditional focus on highly artificial laboratory experiments. For theoretical neuroscience, this wealth of digital data represents a stimulating opportunity to develop a new class of computational models that focus on natural behavior. In this talk, we combine the cognitive and theoretical approaches to explain the timing of touchscreen interactions observed in the real world. We illustrate this approach with two specific examples.
First, we develop an optimality framework that captures the relationship between the statistics of touchscreen events during information gathering (such as checking the news or the weather forecast) and during information sharing (such as writing a text message). In a second study, by combining ideas from priority-based decision process developed in communication sciences with the unconstrained nature of smartphone behaviour we developed a theoretical framework to study the scale invariant dynamics of touchscreen interactions. By applying this framework, we found that smartphone use is of a higher priority than other tasks for 76% of users.
Both of these examples offer new perspectives in computational neuroscience by rising the question of scale invariance in existing paradigms such as in reinforcement learning. The discovery of statistical patterns in touchscreen interactions also yields a clear line of questions for cognitive neurosciences on whether these patterns are rigid or plastic, and how cognitive processes are differently engaged in information gathering and sharing. We conclude that the areas of communication sciences, theoretical neurosciences and cognitive neurosciences must merge to comprehensively understand and benefit from the surge of high-resolution behavioural data made available through modern digital interactions.
Watched 257 times.Watch