Browse all publications by topic
Browse all publications by year
- J. Mayhew, Y. Zheng,
D. Johnston, P. Coffey, J. Berwick, J. Porril, J. Stone, M. Stetter,
I. Schießl, and K. Obermayer. Weak models in the analysis of optical
imaging data: haemoglobin changes following neural activity.
.
In Soc. Neurosci. Abstr. 25, page 1638, 1999.
A major difficulty in the analysis of optical imaging data sets is
the presence of low frequency oscillations and vascular artefacts that can
mask the activation signal. Recent developments in signal source separation
algorithms have been successfully applied to optical imaging data to produce
maps of functional architecture. The research reported here describes an
extension to this work which utilises weak temporal models to constrain the
spatial analysis. Weak temporal models can be used to improve both linear and
non-linear analysis methods by using ratios of variance to constrain the
amplitude of the time series. The research reported here will: Describe the
use of Extended De-correlation (ref to ingo) and Independent Component
Analysis (Bell and Szne..) which have been extended to incorporate weak
models and their application to the analysis of spatial spectroscopic data
sets showing the hemodynamic changes in response to whisker stimulation in
rat. Present a randomisation method by which the ?statistical reliability? of
the results can be assessed. Present data which suggest that the maps showing
regions of the increase in oxygenated hemoglobin are very similar to the maps
of the decrease in de-oxy hemoglobin, and can be highly localised to the
region of the stimulated barrel. However, we find no evidence for the
presence of the ?early deoxy peak? reported by Malonek and Grinvald (1996)
and by Mayhew (1998) using multi wavelength slit spectrographic data.
Simulations using comparable artificial data show that were such changes
present they could be recovered.
|
