Title: Artificial Neural Networks for Analysis of Cartilage Repair
using Hydrogel Scaffolds

Speaker: Mansoor Haider, Department of Mathematics

Abstract:

Articular cartilage is the hydrated biological soft tissue that lines
surfaces of bones in joints such as the knee, shoulder and hip. 
Cartilage is populated with cells (chondrocytes) that maintain the
extracellular matrix by regulating their metabolic activity in
response to the local extracellular environment. In osteoarthritis
cartilage loses its structural integrity and can, ultimately, result
in complete tissue degradation with painful bone-on-bone contact
necessitating joint replacement.  Osteoarthritic cartilage can
exhibit ~Sholes~T called osteochondral defects that, in theory, could
be ~Sfilled~T with biocompatible materials that facilitate
restoration of the tissue~Rs structural integrity.   In this
talk, the use of artificial neural network models (ANNs) for
analyzing structure-function relationships for one such class of
materials will be discussed.  Specifically, elastin-like polypeptides
(ELPs) are injectable in situ polymerizing biomaterials that can be
genetically engineered to exhibit a fluid-to-gel phase transition at
physiological temperature and, thus, show promise in filling
osteochondral defects. However, the optimal design and use of such
material scaffolds for cartilage repair is not yet known. 
Realization of functional gel-tissue constructs depends on many
diverse factors including ELP hydrogel mechanical properties,
biocompatibility between the hydrogel and cells, as well as nutrient
diffusion, cell proliferation and cell metabolic response in the
evolving gel-tissue construct.  The use of both unsupervised and
supervised ANN learning algorithms will be discussed in the context
of ELP data sets obtained from experimental collaborators at Duke
University.