Mahta Mousavi
Technische Universitat Berlin
mahta.mousavi@gmail.com
Bio
Dr. Mahta Mousavi is an Einstein international postdoctoral fellow at the Technische Universitaet Berlin working with Professor Stefan Haufe on developing and benchmarking brain functional connectivity measures for electroencephalography (EEG) data. Prior to joining TU Berlin she was a postdoctoral scholar at UC San Diego where she also earned her doctoral degree in 2019 in Electrical Engineering and Cognitive Science with a specialization in machine learning and data science advised by Professor Virginia de Sa. During her career Dr. Mousavi won several awards and grants including the Mary Anne Fox dissertation year fellowship Institute for Neural Computation predoctoral fellowship German Academic Exchange Service (DAAD) research grant and UC San Diego Chancellors Research Excellence Scholarship. Her main research interests lie at the intersection of machine learning and neuroscience where she develops neurophysiologically interpretable machine learning methods that can serve as diagnostic measures or engineering solutions for patients in need.
Evaluation of EEG-based functional connectivity measures through simulations of whole-brain dynamics
Evaluation of EEG-based functional connectivity measures through simulations of whole-brain dynamics
Brain functional connectivity is one of the widely used measures to investigate the neural correlates of psychiatric and neurodevelopmental disorders. As interactions between functionally specialized brain regions are crucial for normal brain function characterizing and further contrasting communication among brain regions in healthy and patient populations can provide biomarkers for the diagnosis of disorders such as autism spectrum disorder (ASD) and attention deficit hyperactivity disorder (ADHD). Functional connectivity refers to the statistical dependencies of brain activity among anatomically distinct regions of the brain. To measure functional connectivity brain activity can be recorded through various imaging modalities among which electroencephalography (EEG) is popular due to its high temporal resolution. Nevertheless the EEG signal suffers from low spatial resolution which can affect the accuracy of the estimated brain functional connectivity measures. Therefore these measures must be carefully verified before being used for making inferences on real data. To address this we are developing a simulation framework that makes realistic assumptions about the dynamics of the underlying brain sources through modeling neural mass models that contribute to EEG signal generation. Our goal is to investigate and verify the robustness of popular EEG-based functional connectivity measures with simulated ground truth data from our model-based framework. Furthermore our simulation framework provides the necessary tools for other researchers to evaluate their newly developed connectivity measures prior to applying them to real data and will be disseminated as an open-source toolbox.