Opponents

• James Kilner, Senior Lecturer, Institute of Neurology, University College London, UK
• Styrmir Saevarsson, Lecturer, Clinic for Neuropsychology, Bogenhausen Hospital, Germany
• Sándor Beniczky, Associate Professor, Department of Clinical Neurophysiology, Aarhus University Hospital and Danish Epilepsy Centre (Chairman of the Committee)

Supervisors

• Andreas Roepstorff, Professor, Interacting Minds Centre and Center for Functionally Integrative Neuroscience, Aarhus University
• Jørgen Feldbæk Nielsen, Professor, Hammel Neurorehabilitaiton and Research Centre, Aarhus University Hospital
• Marta Garrido, PI, Queensland Brain Institute and Centre for Advanced Imaging, University of Queensland, Australia
• Mikkel Wallentin, Associate Professor, Center for Functionally Integrative Neuroscience, Aarhus University

Summary

In this PhD project, I use electroencephalography (EEG) and magnetoencephalography (MEG) to test alternative hypotheses about the brain’s network architectures and how a network is affected by neurological disorder in patients with neglect syndrome. There is a general consensus that the brain is organised into a hierarchy of areas. The notion of a hierarchy rests on the distinction between feedforward and feedback connections, their distinct physiological characteristics and the idea that they encode different types of information. Feedforward connections link a lower area of the brain with a higher area, whereas feedback connections link a higher area with a lower area. Physiologically, an area is considered to be hierarchically lower when it can drive the activity of a higher area, whereas a higher area modulates the receptive fields of a lower area via inhibition. So why does the brain need different types of connections? One interpretation is that higher areas serve to predict the sensory information encoded in lower areas via feedback mechanisms. This means that feedforward connections only have to encode the information that cannot be explained by the brain’s predictions. This is known as predictive coding.

In the first study, I show that in the healthy brain, a network the right hemisphere is dominant when it comes to interpreting the location of sounds. In a second study, I show that neglect patients with a lesion in the right hemisphere have weaker feedforward and feedback connectivity between frontal and parietal areas in the lesioned hemisphere, compared to healthy people of the same age. Although the patients had a lesion in different parts of the brain, the dysconnection between frontal and parietal areas was consistent throughout the group of patients. This demonstrates that a lesion in a lower area can affect a higher area and vice versa, a phenomenon that speaks to the hierarchical organisation of the brain. Finally, in a third study, I show that when stimuli are unexpected or surprising, the brain relies primarily on feedforward connections. This is evidence that the brain may actively predict the world through feedback mechanisms. Feedforward mechanisms then serve to update the brain’s predictions when something unexpected occurs.

ALL ARE WELCOME