|
Activity Number:
|
399
- ASA Statistics in Imaging Section Student Paper Competition
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 1, 2017 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics in Imaging
|
|
Abstract #322713
|
View Presentation
|
|
Title:
|
Evolutionary State-Space Model and Its Application to Time-Frequency Analysis of Local Field Potentials
|
|
Author(s):
|
Xu Gao* and Babak Shahbaba and Norbert Fortin and Hernando Ombao
|
|
Companies:
|
and University of California, Irvine and University of California, Irvine and KAUST and UC Irvine
|
|
Keywords:
|
Auto-regressive model ;
brain signals ;
spectral analysis ;
state-space models ;
time frequency analysis
|
|
Abstract:
|
We develop a model for high dimensional signals driven by sources whose properties evolve over epochs in experiments. We propose the evolutionary state space model(E-SSM) where signals are mixtures of sources having oscillatory activity at defined frequency bands. One unique feature in E-SSM is that sources are parametrized as second order autoregressive AR(2) processes. To account for non-stationarity, the AR(2) parameters are allowed to vary over epochs. In contrast to independent component analysis, our method captures the temporal structure. Compared to data-adaptive strategies such as filtering, E-SSM easily accommodates non-stationarity. To estimate our model, we use Kalman smoother, maximum likelihood and blocked resampling approaches. The E-SSM is applied to a multi-epoch LFP signals from an olfactory encoding task. Our method captures the evolution of power of different sources across encoding phases. The E-SSM model also identifies clusters of tetrodes that behave similarly.
|
Authors who are presenting talks have a * after their name.
