Export
Mehraveh Salehi
· 2 previous affiliations
- Yale UniversityCurrent
- Current
- PhD student · 2014–2019
- Summary Analytics2019–2020
- Chief Experience Officer · 2019–2020
Research Overview
Mehraveh Salehi focuses on functional neuroimaging, human brain mapping, and the development of individualized functional parcellation methods. Her work explores how functional networks in the brain vary across individuals and dynamically reconfigure in response to different cognitive states. The most recent research in her trajectory investigates the flexibility of these networks and their capacity to predict task performance.
Research Trajectory
Salehi’s research trajectory is defined by a sustained effort to refine how functional networks in the human brain are mapped using functional Magnetic Resonance Imaging (fMRI). Her early work addressed a fundamental limitation in conventional brain mapping: the reliance on group-level analyses that collapse data across populations. Because these traditional methods obscure significant inter-individual variability, Salehi focused on developing computational techniques to generate individualized functional atlases.
In 2017, as first author, she introduced A Submodular Approach to Create Individualized Parcellations of the Human Brain. This work proposed a novel method for defining nodes in brain networks by grouping similar voxels while accounting for the unique connectivity patterns of individual subjects. By leveraging submodularity, the approach provided a mathematically grounded framework for individualized parcellation, which is particularly critical for patient and developmental studies where a universal functional atlas falls short.
Building on this computational foundation, Salehi expanded her methodology in 2018 with An exemplar-based approach to individualized parcellation reveals the need for sex specific functional networks. Also a first-author contribution, this research utilized an exemplar-based method to achieve a joint group- and individual-level parcellation. Crucially, this technique maintained network correspondences across a population while preserving individual variations. The application of this method demonstrated the importance of individualized networks for predictive modeling and provided evidence for the necessity of sex-specific functional networks.
By 2020, Salehi’s focus evolved from examining variability between individuals to investigating dynamic variability within a single individual. Her first-author paper, There is no single functional atlas even for a single individual: Functional parcel definitions change with task, challenged the assumption of static functional subunits in the brain. The research demonstrated that functional parcel boundaries actively reconfigure depending on a subject's cognitive state. Furthermore, these state-dependent reconfigurations were shown to be robust across sessions and capable of predicting both the task condition and the subject's performance.
This intra-individual dynamic approach was further detailed in Individualized functional networks reconfigure with cognitive state. In this work, Salehi and her co-authors analyzed how nodes form flexible networks that shift as the brain transitions between different task demands. The study categorized nodes into distinct classes based on their network assignment flexibility, revealing that these state-evoked reconfigurations are not only functionally relevant but also sufficiently robust to serve as predictive markers for cognitive states.
Current Directions
The most recent papers in the evidence, published in 2020, center on the dynamic and state-dependent nature of functional brain organization. Recent work investigates how individualized functional networks are not fixed anatomical maps, but rather flexible systems that reconfigure in response to shifting cognitive demands. This direction emphasizes the quantification of network flexibility, exploring how the shifting boundaries of functional parcels can be utilized to predict specific task conditions and overall cognitive performance.
Profile updated
Loading new papers…
Loading research neighbors…
See something wrong? Report a mistake