This Article  • References  • Full text PDF  • Send to a friend  • Save in My Folder  • Save to citation manager  • Permissions  Citing Articles  • Citing articles on HighWire  • Contact me when this article is cited  Related Content  • Similar articles in JAMA  Social Bookmarking   What's this?

Artiss L. Powell, MD; Jill Brooks, PhD
University of Medicine and Dentistry of New Jersey New Brunswick

Daniel A. Zahner, MS; Zhizhong M. Zhang, PhD; Yasemin M. Akay, BA; Evangelia Micheli-Tzanakou, PhD
Rutgers University Piscataway, NJ

JAMA. 1994;272(18):1405-1406.

	Since this article does not have an abstract, we have provided the first 150 words of the full text PDF and any section headings.

To the Editor.

—The article by Dr Stern and colleagues¹ reports an important relationship between risk for AD and education, a relationship that came under scrutiny following the report of Zhang et al.² Subsequent reports have suggested a protective role of education, possibly through increased neuronal synapse density.³

For several decades, computational neuroscientists, cognitive psychologists, and computer scientists have incorporated anatomical and physiological data into neural network computer simulations to replicate⁴ and, more recently, to study^5,6 brain function.

In general, neural network development can be separated into three phases (design, training, and testing) with simple models usually constructed with at least three layers: input (I), hidden units (HUs), and output (0). The I layer is equivalent to a sensory system such as vision, while HUs correlate with association cortices. Hidden units are sandwiched between the I and O layers with the I output serving as the . . . [Full Text PDF of this Article]

CiteULike    Connotea    Del.icio.us    Digg    Reddit    Technorati    Twitter     What's this?