Time

Prog #

Jozien Goense¹, Nikos Logothetis¹, ²

¹Max-Planck Institute for Biological Cybernetics, Tuebingen, Germany; ²University of Manchester, Manchester, UK

High-resolution fMRI can aid in determining to what extent the BOLD signal arises from capillaries or larger vessels. In high-resolution functional activation maps both positive and negative BOLD signals associated with vessels were observed; this was seen for both GE- and SE-BOLD. Because of its higher specificity, the SE-BOLD signal was used to investigate the origin of these vessel signals. The location of the SE-BOLD signal from veins changed when the direction of the gradients was changed. This is in contrast to the peak SE-BOLD occurring in layer IV which arises from capillaries, and was insensitive to gradient reversal.

Seong-Gi Kim¹, Chan-Hong Moon¹, Mitsuhiro Fukuda¹

¹University of Pittsburgh, Pittsburgh, Pennsylvania, USA

To understand the vascular source of functional signals which are specific to neuronal activity, it is critical to examine spatiotemporal characteristics at high resolution. We examined BOLD and CBV signals in a well-established cat orientation column model.  Orientation-specific BOLD signal is minimal, while orientation-specific CBV response is significant, but increases slowly compared to non-specific BOLD and CBV signals. This indicates that the orientation-specific signal originates from microvessels including capillaries.

Junjie V. Liu¹, Mriganka Sur¹, Christopher I. Moore¹, Jitendra Sharma¹, ²

¹Massachusetts Institute of Technology, Cambridge, Massachusetts, USA; ²Martinos Imaging Center, Charlestown, Massachusetts, USA

We present the first measurements of high-resolution (200 um) orientation maps in visual cortex of the ferret, using multi-slice gradient-echo EPI (without contrast agent). A novel coil pocket glued on top of skull near visual cortex is used. Orientation maps are measured in 3-4 slices parallel to cortical surface near V1, and are comparable to maps measured by optical imaging. The laminar profiles of BOLD signals show that the orientation-specific responses are relatively uniform across cortical layers, whereas the non-specific visually-driven responses are stronger near superficial layers.

Xu Chen¹, Fuqiang Zhao², Seong-Gi Kim², Jean Tkach¹, Stephen Strother³, ⁴

¹Case Western Reserve University, Cleveland, USA; ²University of Pittsburgh, Pittsburgh, Pennsylvania, USA; ³Baycrest, Toronto, Canada; ⁴University of Toronto, Canada

The spatial scale of  neural representation structures plays an important role in neuro-imaging studies and is in itself a booming area of  research.  In this abstract, we demonstrate a framework to estimate the spatial scale of columnar neuronal activity in a cat's primary visual cortex based on the analysis of ultra-high-resolution CBV-weighted functional MRI (fMRI) data. With matched filter theorem as its rationale, this framework takes advantage of the distribution of the reproducibility between spatial patterns from re-sampled fMRI datasets and has the potential to be extended to estimate the spatial scale of other important neuronal activities.

Amir Shmuel¹, ², Guenter Raddatz¹, Denis Chaimow¹, Nikos Logothetis¹, Kamil Ugurbil, ¹², Essa Yacoub³

¹Max Planck Institute for Biological Cybernetics, Tuebingen, Germany; ²University of Minnesota, Minneapolis, Minnesota, USA; ³University of Minnesota, Minneapolis, USA

Recent studies have demonstrated that classification algorithms applied to human fMRI data can decode information segregated in cortical columns, although the voxel-size was large relative to the width of columns. The mechanism by which low-resolution imaging decodes information represented at higher resolution is not clear. We show that using GE-fMRI signals, the mechanism underlying the decoding signals involves contributions from both gray matter and macroscopic blood vessels. We hypothesize that draining regions biased towards columns with preference to one eye underlie the specificity of vessels. Decoding at high-resolution is superior to low-resolution when applied to data from small cortical volumes.

Govind Nair¹, Yoji Tanaka¹, Tsukasa Nagaoka¹, Machelle T. Pardue², Darin Olson², Peter Thule², Timothy Q. Duong¹

¹Emory University, Atlanta, Georgia, USA; ²VA hospital, Atlanta, Georgia, USA

The thin retina is organized into multiple cell layers and two vascular layers. The retinal and choroidal vasculatures are located on either side of the retina with an avascular photoreceptor layer in the middle. Importantly, the two vascular layers independently regulated and respond differently to physiologic stimulations. The retina’s remarkable anatomical and functional layout lends itself to be an excellent model for testing layer-specific MRI. We recently reported multiple anatomical layers and unique layer-specific BOLD fMRI responses in the retina. In this study, we extend previous findings to include layer-specific blood-volume MRI and layer-specific physiologically evoked BV changes.

Wendy Ringe¹, Kaundinya Gopinath¹, Sergey Cheshkov¹, Subhendra Sarkar¹, Richard Briggs¹, Robert Haley¹

¹UT Southwestern Medical Center, Dallas, Texas, USA

Converging evidence indicates that thalamic nuclei are specifically involved in category-dependent processing which may not be resolvable with conventional fMRI voxel sizes. High resolution fMRI demonstrated material-specific activation in each of four material types: Words, Objects, Faces, Nature Scenes. Each activated bilateral medial dorsal nucleus, with additional activation of: Words - left inferior pulvinar, anterior, ventral lateral, and ventral posterior nuclei; Nameable Objects - left ventral lateral, ventral posterior and right inferior pulvinar and ventral lateral nuclei; Faces - left inferior pulvinar and ventral lateral and right anterior nuclei;  Nature Scenes - bilateral pulvinar and right ventral posterior nuclei.

Nanyin Zhang¹, Xiao-Hong Zhu¹, Jae-keun Park¹, Yi Zhang¹, Ye Li¹, Wei Chen¹

¹University of Minnesota, Minneapolis, Minnesota, USA

Lateral geniculate nucleus (LGN) is the most important component in the thalamus in relaying visual information flow from the retina to primary visual cortex (V1). In this study, we demonstrate the feasibility of mapping the LGN laminar structures in cat using BOLD-based and CBV-MION-based functional magnetic resonance imaging. Both BOLD and CBV data successfully differentiate individual laminae is LGN. The maps generated using BOLD and MION signals are highly reproducible for intra-subject comparison. In addition, the results suggest CBV signal has significant narrower point spread function compared to BOLD signal.

Xiaomu Song¹, ², Limin Li¹, ², Daniil Aksenov¹, Alice M. Wyrwicz¹, ²

¹Evanston Northwestern Healthcare Research Institute, Evanston, Illinois, USA; ²Feinberg School of Medicine, Northwestern University, Chicago, Illinois, USA

This abstract introduces a novel method to detect and visualize distinct functional cortical units in rabbits' brain using kernel principal component analysis of high field fMRI data. In the abstract, the research background and motivation are first introduced, followed by the two sections describing the imaging and data analysis method. After presenting the experimental results showing activation maps and temporal profiles, the final conclusions are made.

Christopher Paul Pawela¹, Dan L. Shefchik¹, Younghoon R. Cho¹, Rupeng Li¹, Seth R. Jones¹, Ji-Geng Yan¹, Safwan S. Jaradeh¹, Hani S. Matloub¹, James S. Hyde¹

¹Medical College of Wisconsin, Milwaukee, USA

High resolution Gradient Echo BOLD images were acquired of the rat somatosensory cortex at 9.4 T. We were able to resolve the representations of the individual digits and the data showed a high degree of layer specificity.