University of Rochester

Home

Facilities

Research

Events & Courses

Policies

People

Directions & Contacts

News

FAQs

RCBI: Rochester Center for Brain Imaging

Bold Signal Imaging


High resolution fMRI of LGN

The lateral geniculate nucleus (LGN) is a small subcortical visual structure, contained within a cubic centimeter in humans, that serves as relay and filter for information traveling from the retina to the cortex. It is composed of six layers, three innervated by each eye, and the ventral two magnocellular differ in function from the dorsal four parvocellular layers. Keith Schneider at the Rochester Center for Brain Imaging (RCBI) has been studying the LGN and the superior colliculus (SC), an even smaller visual structure involved in attention and eye movements, in humans using fMRI. He has co-authored studies investigating the topographic organization of the LGN (Schneider, Richter et al. 2004) and SC (Schneider and Kastner 2005) and the responses of the LGN during binocular rivalry (Wunderlich, Schneider et al. 2005). The latter study revealed that our conscious perceptions are reflected throughout the visual system, even as early as the level of the LGN, a finding that was unexpected based on non-human primate studies. Schneider's current studies, in collaboration with Jianhui Zhong and Tianling Gu at RCBI, seek to push the limits of the spatial resolution of fMRI using zoomed fields of view and segmented k-space and 3D-EPI sequences with an 8-channel head coil. The goal is to achieve 1 mm3 resolution to better image the structure of the LGN while keeping a reasonable SNR, TR and TE. Below is an image acquired at RCBI showing the upper (green) and lower (purple) hemifields of the left (L) and right (R) LGN and SC activated by flickering checkerboard stimuli and using the standard EPI sequence with a 128 x 128 matrix, 1.5 x 1.5 x 1.9 mm3 voxel, TR = 2 s, TE = 42 ms, and 8-channel coil with GRAPPA acceleration factor of 2.

High res fMRI of LGN

Brain BOLD fMRI applications - intermolecular DQC fMRI

We are developing a technique called iDQC imaging, which allows sensitive detection of microscopic susceptibility changes with minimal contamination from macroscopic field inhomogeneity and therefore suitable for fMRI studies in brain areas susceptible to inhomogeneity problems. Furthermore, the sensitivity of iDQC to susceptibility of small-length scales selected by dc will provide enhanced fMRI responses specific to capillary levels, and therefore true neuronal activities.

intermolecular DQC fMRI

fig. Results from experiments with presentations of hemifield visual stimulus that last 30 sec for the left side, then 30 sec for the right side, and repeat 24 times for iDQC, and 8 times for the conventional SQC BOLD. TR is 5 sec for iDQC and 3s for SQC. (Left): GE-SQC activation pattern superimposed on the raw EPI images in a single subject. Hot and cool colors represent activations to separate hemifield stimulations. (Middle) The same activation pattern (interpolated to higher resolution) superimposed on the MR angiograph show locations of major blood vessels. (Right): SE-iDQC activation map of the same slice using the imaging pulse sequence based on Fig 5.(b). Note the less susceptibility-related distortion in iDQC image due to its spin-echo nature in readout. Robust activation is however detected. These observations are consistent among 5 subjects studied so far. Locations of activations from both types of images share some common voxels (note the cross hair on both) but fewer voxels are activated in iDQC, which may reflect its lower SNR and/or spatial resolution, but could also be due to selection of iDQC to different vessel scales.

The DQC gradients are applied on the x- and z- direction of the magnetic field (with angle of 0 degree meaning the gradients are on the z gradient, and 90 degree along the x direction). The signal of the iDQC acquisition has a dependence of (3.sqr(cos(theta))-1) on the angle (theta), as shown in Figure 2 (a) acquired in a phantom study. Also instead of a pure T2 signal decay on TE, the iDQC signal has an increasing-decay pattern, as shown in Figure 2(b) in a phantom study. The iDQC signal reaches the maximum when the TE equals to the T2 of the sample, which is about 150ms in this case.

iDQC signal dependence graph

iDQC signal dependence on the DQC gradient angle (a) and on TE (b)


Mapping of the visual cortex

A retinotopic map projected on a 3D reconstruction of the occipital lobe showing the wave of activation driven by an expanding ring stimulus along the calcarine sulcus.

retinotopic map of the occipital lobe

Working memory study in Deaf

Activation maps illustrating MTMST and left DLPC activation in an encoding task

Activation maps illustrating MTMST