[reprinted by permission from APS]
In a recent Phys. Rev. E article, David Nelson, together with Ariel Amir (SEAS; formerly Harvard Physics Junior Fellow) and Naomichi Hatano (UTokyo), analyzed the eigenvalues and eigenvectors of certain asymmetric tridiagonal matrices. The authors found a rich and complex behavior in the distribution of eigenvalues and the localization properties of eigenvectors. This has implications that can help understand features of the dynamics of neural networks, as well as neural development in organisms.