Marieke van Vugt: Computational Cognitive Modeling of Mind-Wandering through Machine Learning

Mind-wandering occurs when our mind drifts away from what we are doing to other thoughts and experiences. Sometimes mind-wandering can be beneficial, but it can also lead to accidents or even worse - it is a risk factor for psychiatric problems such as depression.

To better understand when and how mind-wandering is helpful, and when it is harmful, Prof van Vugt employs computational modeling of cognitive processes and detailed analyses of EEG activity. Mind-wandering is modeled as a process in competition with an attention process, which both interact with a simulated cognitive task. Fitting this model with large-scale data about the thought processes of depressed participants, Prof van Vugt showed that performance on a sustained attention task suffered as a result of the depression. This prediction was subsequently verified in empirical data. On the other hand, she seeks to track mind-wandering processes in real time in EEG data. She demonstrated that a task-general SVM can be built that predicts mind-wandering in single trials in two different psychological tasks.

The next research steps will be bringing those two approaches together to see how, in brain activity, rumination is different from general mind-wandering, and what their respective temporal dynamics are. This will increase our understanding of how, when and why we mind-wander and when mind-wandering goes awry in depressive rumination.