I apply complex network concepts to different domains. My strengths and intersts are in computational and analytical modeling as well as data mining and statistical analysis.
Modeling brain developmentHow does the human brain develop through the time? This work is in collaboration with Prof.Olaf Sporns. Structural network of Major Depressive DisorderIn this work, we investigate the structural connectivity of the depression network. We focus on 18 ROIs that are considered key mediators in the putative depression network. Our objective is to identify how these 18 ROIs are interconnected, using Diffusion Tensor Imaging (DTI) data and tractography methods, and to reveal the special role (if any) of the SCC region in network-centric terms. We focus on the 18 ROIs in the MDD circuit model proposed by Mayberg [Mayberg 1997]. These regions include mood regulation, mood monitoring, exteroception and interoception areas. The ROIs were drawn on the 2mm standard space using the SPM pickatlas tool. We perform Diffusion Tensor Imaging (DTI) and probabilistic tractography on diffusion MRI data from 30 healthy subjects using FDT tool in FSL. The tractography results allow us to identify the connectivity from each voxel of each source ROI to all target ROIs. We use this data to create both unweighted (binary) as well as weighted versions of the depression network interconnecting these 18 ROIs. We also evaluate the robustness of the inferred networks with respect to variations in tractography parameters as well as network inference parameters. In the resulting (robust) networks, we apply network analysis methods to characterize each ROI in terms of various centrality measures (degree, closeness, betweenness and others) using both unweighted and weighted metrics. Our focus is on the SCC region, given that that ROI is the target of DBS treatment for MDD. The variability of the depression network connectivity across different subjects is also quantified and discussed. We also compare the SCC with other regions that have been recently proposed as potential DBS targets for treating MDD. This work is in collaboration with Dr. David Gutman from Emory.
Optimized vs Evolutionary Network DesignThe problem of network topology design is often studied as a “clean-slate” optimization, however in practice most service-provider and enterprise networks are designed incrementally over time. This evolutionary process is driven by changes in the underlying parameters and constraints (the “environment”) and it aims to minimize the modification cost after each change in the environment. In this project, we formulate the incremental design approach (in three variations), and compare that with the more traditional optimized design approach in which the objective is to minimize the total network cost. We evaluate the cost overhead and evolvability of incremental design under two network expansion models, comparing incremental and optimized networks in terms of cost, topological similarity, performance and reliability.PublicationsSaeideh Bakhshi, David Gutman, Constantine Dovrolis, SubCallosal Cingulate Gyrus Is The Most Central Region In The Network Of Major Depressive Disorder(MDD) , Complenet 2012, Melbourne, Fl Saeideh Bakhshi, Constantine Dovrolis, The price of evolution in incremental network design (the case of ring networks) , Bionetics 2011, York, UK Saeideh Bakhshi, David Gutman, Constantine Dovrolis, The Structural Network of Major Depressive Disorder , INCF Neuroinformatics Congress meeting, Boston, MA, 2011 poster abstract Saeideh Bakhshi, Constantine Dovrolis, Incremental versus Optimized Network Design, technical report, 2011 Saeedeh Bakhshi, Marzieh Bakhshi,Hamid Sarbazi-Azad, Efficient VLSI Layoutof Grid Pyramid Networks,1st International conference on Contemporary Computing, India, 2008. Saeedeh Bakhshi, Marzieh Bakhshi, Hamid Sarbazi-Azad, One-to-one and One-to-many node- disjoint Routing Algorithms for RTCC networks, 1st International conference on Contemporary Computing, India, 2008. Saeedeh Bakhshi, Hamid Sarbazi-Azad, Efficient VLSI layout of Edge Product Networks, 4th IEEE International symposium on Electronic, Design, Test and Application, 2008 Saeedeh Bakhshi, Hamid Sarbazi-Azad, One-to-one and One-to-many Node-disjoint Routing Algorithms for WK-Recursive Meshes, ninth International Symposium on Parallel Architectures, Algorithms, and Networks (I-SPAN), 2008. Saeedeh Bakhshi, Marzieh Bakhshi, Hamid Sarbazi-Azad, Efficient Parallel Routing Algorithms for Cartesian and Composition Networks, 13th International CSI Computer Science conference, 2008. Saeedeh Bakhshi, Hamid Sarbazi-Azad, Efficient VLSI Layout of WK-Recursive and WK-Pyramid Interconnection Networks, 13th International CSI Computer Science conference, 2008. CoursesToolsI use a wide variety of tools for my research. Some of them are:Useful meta-research linksPersonalI am from beautiful city of Shiraz, located in southwest of Iran.Contact InformationKACB 3337, 266 Ferst DriveGeorgia Institute of Technology Atlanta, Georgia 30318 Email: "First Name" AT gatech dot edu |