Susan L. Wearne, Ph.D.

  • Associate Professor of Biomathematics
  • Laboratory of Biomathematics
  • Fishberg Department of Neuroscience
  • Mount Sinai School of Medicine, Box 1065
  • One Gustave L. Levy Place
  • New York, NY 10029, USA
  • Phone: 1-212-659-5572
  • Fax: 1-212-996-9785
  • Email: susan.wearne@mssm.edu

Modeling Contributions of Morphology to Neuronal Function, and Dysfunction in Neurodegeneration and Disease

Despite the extraordinary diversity of structure in neurons of the central nervous system, we still understand very little of its role in determining neural function. Our research aims to provide a mechanistic understanding of the cognitive decline that accompanies morphologic dystrophy in aging and neurodegenerative disorders. To do this, we establish quantitative I/O relations through mathematical modeling at multiple levels, from molecular through cellular, to network levels and ultimately, cognitive function. Importantly, through mathematical modeling we can construct quantitatively precise, mechanistic links between successive levels of function, allowing model validation by experimental or clinical data, and falsifiable predictions to be made.

Cellular and Subcellular Levels

We use compartment models based on high-resolution 3D reconstructions (see CNIC website for details), to understand how spine shape, spine density and global dendritic structure determine cell level functions, such as electrical and biochemical signaling.

Network Level

Working memory can be modeled as an attractor network, whose ability to retain a short-term memory is quantified by network stability and robustness. Our current work studies how cellular-level changes, including morphologic perturbations, affect the stability and robustness of Hopfield-style attractor models of working memory. At the network level, we compare the effects of parameters describing connectivity with those intrinsic to single cells, on network output. In this way, mechanistic links between levels of function can be built up incrementally.

Predicting Homeostatic and Compensatory Mechanisms for Restoring Normal Function following Morphologic Dystrophy or Other Disease-related Perturbation

We use these models to develop, and test experimentally, new computational methods for predicting compensatory strategies that can reverse the effects of pathologic changes accompanying normal aging and neurodegenerative disorders. These compensatory mechanisms can either restore a given cellular function, perturbed by disease, or maintain that function at a constant level in a homeostatic manner, despite ongoing morphologic changes that occur in normal development and plasticity. Neuroanatomic and electrophysiological data from our experimental collaborators are used to constrain the models, and validate their predictions.

Physical and Biological Bases of Fractional Order Dynamical Systems

A major research interest is applying the mathematics of fractional calculus to understanding electrodiffusion and diffusive biochemical signaling in complex biological systems. This work is in collaboration with Dr. Bruce I. Henry of the Department of Applied Mathematics, University of New South Wales, Australia. This new mathematical framework more realistically represents the biophysics of irregularly shaped, non-homogeneous biological systems than standard diffusion based models. As an example, neither the intracellular media in which biochemical signaling takes place, nor the spiny dendritic cables that support electrical signaling in neurons, are homogeneous or smooth. Our recent modeling work shows that these structural irregularities significantly alter the molecular diffusion and electrical signaling predicted from standard diffusion models. Currently we are applying these techniques to understand how electrical and chemical signaling in neurons depends on spine shape, spine density and spatial distribution. Accurate 3D reconstructions of spine morphology, crucial for such modeling, are developed by our image analysis team in the CNIC. (e.g. Figure 2, Rayburst estimation of spine head diameter).

Public Dissemination of Modeling Tools

All modeling and analysis software developed to perform these studies is distributed publicly from our labs CNIC website, and in the online public model repository ModelDB, and announced with a posting to the NEURON Users Forum online. Reconstructed neuron morphologies used for compartment modeling are also archived both on our, CNIC repository and on the public morphologic repository NeuroMorpho.org, a centrally curated inventory of digitally reconstructed neurons.