Matched Signal Detection on Graphs: Theory and Application to Brain Network Classification using PIB-PET data

Speaker: Quanzheng Li , Harvard Medical School, MGH

Date: Wednesday, October 09, 2013

Time: 4:00 PM to 5:00 PM Note: all times are in the Eastern Time Zone

Public: Yes

Location: 32-D507

Event Type:

Room Description:

Host: Polina Golland, CSAIL

Contact: Polina Golland, polina@csail.mit.edu

Relevant URL:

Speaker URL: None

Speaker Photo:
None

Reminders to: seminars@csail.mit.edu

Reminder Subject: TALK: Matched Signal Detection on Graphs: Theory and Application to Brain Network Classification using PIB-PET data

We developed a matched signal detection (MSD) theory using general
likelihood ratio test to detect whether a signal is embedded in an
intrinsic structure described by a weighted graph. We considered different
levels of complexity in the signal models. In the simplest scenario, we
assumed that the signal is deterministic with noise in a subspace spanned
by a subset of eigenvectors of the graph Laplacian. The conventional
matched subspace detection can be easily extended to this case. n
Furthermore, we studied signals with certain level of smoothness. The test r
turns out to be a weighted energy detector, when the noise variance is
negligible. More generally, we presumed that the signal follows a prior
distribution, which could be learnt from training data. The test statistic
is then the difference of signal variations on associated graph
structures, if an Ising model is adopted. Effectiveness of the MSD on
graph is evaluated both by simulation and real data. We apply it to the
network classification problem of AD. The preliminary results demonstrate
that our approach is able to exploit the sub-manifold structure of the
data, and therefore achieve a better detection/prediction accuracy than
the traditional principle component analysis (PCA), linear discriminant
analysis (LDA) and support vector machine (SVM) in the application of
brain network classification using PIB PET data.

Research Areas:

Impact Areas:

See other events that are part of the Biomedical Imaging and Analysis 2013/2014.

Created by Polina Golland at Monday, September 30, 2013 at 9:00 PM.