|
|
Brain Imaging & Modeling
The past few decades have witnessed
extraordinary progress in the development of techniques for noninvasive
structural and functional imaging of the human brain. However, despite
this progress, no existing medical imaging modality provides all of the
information required for best clinical practice or cutting edge basic
research. MRI is the premier technique for characterizing the soft
tissue anatomy of the human brain, but has significant limitations for
defining the geometry of the skull. Functional MRI, and PET
provided detailed pictures of spatial patterns of neural activation based
on associated hemodynamic changes, but cannot capture the temporal dynamics
of electrophysiological activation on its characteristic timescale.
MEG and EEG provide excellent temporal resolution of neural population
dynamics but are limited in spatial resolution by the ambiguity and ill-posed
nature of the current reconstruction problem. Electrical and magnetic
stimulation offer the possibility of direct intervention in central or
peripheral neural pathways, but depend on knowledge of anatomical and
functional organization drawn from other sources. Other methods,
including optical and magnetic resonance spectroscopy, SPECT and nuclear
medicine, histology, endoscopy, neurosurgical intervention, etc., each
provide important and unique insight into neural function and functional
organization. Although the mix and relative importance of imaging
technologies will continue to evolve, the need to integrate information
from multiple methods will remain.
The goal of research supported by the Human
Brain Project is to develop composite techniques for noninvasive,
functional brain imaging that provide spatial and temporal resolution
superior to any available imaging technique. We are developing
experimental, theoretical and computational procedures to combine anatomical
MRI, functional MRI, and MEG into an integrated structural/functional
imaging technique that exploits the strengths and minimizes the weaknesses
of each modality alone.
The second major effort addresses the development and application of
Advanced Optical Methods for tissue measurements on the
macroscopic and microscopic scale. For macroscopic measurements
the principal challenge is the highly scattering nature of biological
tissue. Tissue is relatively translucent, particularly in the
near IR, and these wavelengths provide useful spectral data, e.g. allowing
measurement of the quantity and oxygenation state of hemoglobin.
However, a photon in the near IR may scatter >10 times per cm , so that
light travels through tissue by a process that resembles diffusion.
To address the uncertainty in path and pathlength we are developing
time-resolved imaging methods that allow measurement of the time of
flight of photons launched into the medium. The first approach
employs high-speed gated intensifiers (gate time ~200 ps) coupled to
cooled CCDs. A second strategy currently being explored will employ
time-resolved photon counting technologies
developed at Los Alamos. This class of methods coupled with
sophisticated modeling techniques shows promise for noninvasive quantitative
characterization and tomographic reconstruction
of the optical properties of highly scattering media such as brain
tissue.
A second project is developing a hybrid hardware/software technology
for confocal and spectral imaging via
microscope or endoscope. A prototype imager has been employed
for in situ imaging of fast intrinsic signals associated with excitation
in the brainstem of an experimental animal, and shows promise for characterizing
patterns of cellular activity in extended neural networks. These
technologies are presently being developed with support from the DOE
Technology Transfer Initiative.
.
![](images/brain_imaging.gif)
|
|
|