Mapping the Landscape of Brain Activity using Experience Sampling

Thumbnail Image
Goodall-Halliwell, Ian W.
Psychology , Cognitive Neuroscience , Machine Learning , Mindwandering
The thoughts we have and the activity in our brain are influenced by what we are doing. Studies within the lab and in the field suggest that tasks, whether lab-coordinated or organic, consistently push our thoughts and neural activity into “states”. Drawing on this research, this study had participants perform a testing battery made up of a diverse range of cognitive tasks which had been previously scanned in an fMRI. During the testing, participants reported their thoughts during each task. Following previous work, common patterns of thought were extracted using principal component analysis. These patterns of thought were used to predict neural activity and vice versa using simple machine learning techniques. This study found that neural activity and patterns of thought could be predicted better than chance. This study also suggests that certain brain networks, notably the default mode network, were essential for the successful prediction of thought patterns. These findings may be useful to researchers and clinicians who wish to more robustly sample cognition.
External DOI