\documentclass[11pt]{article}
\setlength{\oddsidemargin}{0.0truein}
\setlength{\evensidemargin}{0.0truein}
\setlength{\textwidth}{6.5truein}
\setlength{\topmargin}{0.0truein}
\setlength{\textheight}{9.0truein} \setlength{\headsep}{0.0truein}
\setlength{\headheight}{0.0truein} \setlength{\topskip}{10.0pt}
\setlength{\parskip}{5mm}
\usepackage{url}
\usepackage{amsmath}
\usepackage{amssymb}
\begin{document}

\begin{center}
  \textbf{\textsc{STANFORD UNIVERSITY}}\\[5pt]
  \textbf{\textsc{DEPARTMENT OF STATISTICS}}\\[5pt]
  \Large{\textbf\textsc{{DEPARTMENTAL SEMINAR}}}
\end{center}
\begin{center}
  4:15 p.m., Thursday, August 3, 2006\\
  Sequoia Hall Room 200\\
  (Cookies at 3:45 in 1st Floor Lounge)
\end{center}


\begin{center}
  \textsl{Karl SjÄostrand} \\
Technical University of Denmark\\
University of California, San Francisco\\
\end{center}
\begin{center}
  \textbf{ Modern Statistical Methodology in Medical Image Analysis}
\end{center}


\noindent
Modern Statistical Methodology in Medical Image Analysis
Statistical analysis of medical image data may be aimed at predicting disease progression, separating
\documentclass[11pt]{article}
\setlength{\oddsidemargin}{0.0truein}
\setlength{\evensidemargin}{0.0truein}
\setlength{\textwidth}{6.5truein}
\setlength{\topmargin}{0.0truein}
\setlength{\textheight}{9.0truein} \setlength{\headsep}{0.0truein}
\setlength{\headheight}{0.0truein} \setlength{\topskip}{10.0pt}
\setlength{\parskip}{5mm}
\usepackage{url}
\usepackage{amsmath}
\usepackage{amssymb}
\begin{document}

\begin{center}
  \textbf{\textsc{STANFORD UNIVERSITY}}\\[5pt]
  \textbf{\textsc{DEPARTMENT OF STATISTICS}}\\[5pt]
  \Large{\textbf\textsc{{DEPARTMENTAL SEMINAR}}}
\end{center}
\begin{center}
  4:15 p.m., Thursday, August 3, 2006\\
  Sequoia Hall Room 200\\
  (Cookies at 3:45 in 1st Floor Lounge)
\end{center}


\begin{center}
  \textsl{Karl SjÄostrand} \\
Technical University of Denmark\\
University of California, San Francisco\\
\end{center}
\begin{center}
  \textbf{ Modern Statistical Methodology in Medical Image Analysis}
\end{center}


\noindent
Modern Statistical Methodology in Medical Image Analysis
Statistical analysis of medical image data may be aimed at predicting disease progression, separating
the normal from the abnormal or aiding in the comprehension of a natural or pathological process.
This talk will present a couple of important applications within this ¯eld, with an emphasis on
methodology developed at Stanford University.\\
Clinically, the material is focused on brain morphometry { the measuring of shape and shape
changes within the brain. A study is presented where the shape of the corpus callosum brain
structure has been extracted from magnetic resonance images of a large number of patients. The
shape data can be decomposed into a small set of interpretable template shapes using sparse
principal component analysis. These templates may then be related to patient outcome data
through the LASSO regression framework. The aim of the current analysis is to learn more about
the relationship between localized brain tissue loss and the decline in e.g memory performance
and motor skills. The more ambitious long-term goal is to predict and detect early stages of
neurodegenerative diseases directly from image data. Examples of other applications include sparse
modeling of functional magnetic resonance data (fMRI) and the statistical analysis of whole-brain
deformation data, where the number of variables is in the order of millions.\\
Another task with several important applications in image analysis is the ordering of a set of
observations according to the probability of each observation, given the entire data set. This can
be implemented using a technique for one-class classi¯cation known as the support vector domain
description. An e±cient algorithm is presented and applied to medical shape data.
\end{document}