Predicting task performance based on electrophysiological resting state networks

点击翻译按钮获取中文摘要

Human performance differs between individuals, but also within individuals between repetitions of the same task. The proposed project investigates whether it is possible to identify intrinsic brain features responsible for performance differences in tasks and that correspondingly allow to predict performance. This requires analyzing the dynamics of brain activity and the interaction between brain regions before the actual task. The brain networks that emerge while we are not performing a task are called resting state networks (RSNs). To study the dynamics of these RSNs, participants will be measured in the magnetoencephalogram (MEG), which is a non-invasive measurement technique that allows recording electrophysiological neural activity of the whole brain with a temporal resolution commensurate with actual neural activity. Within the MEG participants will perform a somatosensory paradigm where they have to discriminate whether one or two stimuli were applied to their index finger. The project will focus on the somatosensory domain, but its results should transfer to other sensory domains. The time interval between the two stimuli will be adjusted to ensure that participants correctly discern the two stimuli in 50% of the trials. An inter-trial period of at least 10 seconds will allow investigating the fast changing network activity by employing a Hidden Markov Model. The pattern of networks and the networks active immediately preceding the stimuli are expected to be predictive for the behavioral performance of the subjects. Identification and characterization of such a task-rest nexus will allow for a better understanding of the functional relevance of resting state brain activity. Because changes in resting state activity have been identified in many neurological and psychiatric diseases, such findings will be instrumental to better diagnose certain diseases through the recording of spontaneous brain activity.