Patients with glioma are frequently operated in awake conditions, with the aim to intraoperatively map different brain functions. Surprisingly, in a picture naming task, Broca’s area is only responsive to electrical stimulation in 50 % of patients. Currently, it is not known if this is due to a specific pattern of structural and functional connectivity of Broca’s area that would be found in a similar rate in healthy people or if this is an effect of tumor-induced changes in connectivity of the region. Preliminary observations indeed suggest a wide variability of Broca’s area connectivity in healthy subjects (see Margulies & Petrides, 2013, J Neurosci ; Jakobsen et al., 2016, Eur J Neurosci).
In this project, we propose to investigate the patterns of structural and functional connectivity of Broca’s area in a sample of 500 subjects from the Human Connectome Project. For each subject, Broca’s area will be automatically segmented and divided in pars triangularis (PT) and pars opercularis (PO). These ROIs will be used as seed regions to determine functional connectivity (from resting-state MRI data) and structural connectivity (from diffusion-weighted MRI data). Then, clustering approaches will be used to parse distinct classes of connectivity. These results will help to interpret connectivity imaging in patients and will provide the basis for future work correlating different connectivity classes with observed responses to electrical stimulation.
Profile of the candidate.
Expertise in Python
Basic level in machine learning
Strong interest in cognitive neuroscience and neuroimaging
Co-supervisors : Daniel Margulies and Demian Wasserman