A. Improvement of Spatial Resolution:

One of our main research goals is to try to improve the spatial resolution of the fMRI signal for obvious reasons. In order to improve spatial resolution and specificity of fMRI signal, three issues are considered; (1) improvement of SNR (2) elimination of large vascular contributions, and (3) fundamental limit of physiological response. We can improve SNR by using high magnetic fields, by using a small surface coil, and by using fast imaging techniques such as single-shot EPI. We have extensively targeted the issue of eliminating large vascular contribution (discussed at the Physiology section).

The major concern is a fundamental limit of hemodynamic responses induced by increased neural activity. According to data derived from optical imaging studies (Malonek and Grinvald, 1996), hemodynamic responses are widespread and diffused beyond neural active sites. It is suggested that metabolism-based imaging can give rise to a more specific signal to the neuronal active site than hemodynamic imaging. To investigate spatial specificity of fMRI, we implemented a cat orientation column model. Cat orientation columns are 500µm wide and ~ 1 mm a part. We found that late BOLD responses are widespread, and can not used to generate columnar maps. We used metabolism-based early negative BOLD responses, which occurred between 0.5 and 4.0 sec after the onset of stimulation (discussed at the Physiology section). We successfully obtained orientation column map in the cat primary visual cortex using the early negative BOLD signal. This opens a new avenue for high resolution column-level fMRI.
 
 

Click on image for viewing larger version

Composite-angle map and orientation "pinwheels" generated using "negative" fMRI signals. Panel a displays the composite-angle map obtained through pixel-by-pixel vector addition of the four single iso-orientation maps based on "negative" signal changes. The resulting orientation preference at each cortical location is color-coded according to the color key displayed left to the panel a. Note how -- tangential to the cortical surface -- the preferred orientations change smoothly, thereby forming a "map" of orientation selectivity. This continuity is interrupted at the "orientation pinwheels" where the cortical columns for different orientations are arranged in a circular manner, thereby forming two types of topological singularities according to their rotational circularity (shown in b and c). The white and black circles in panel a depict such "clockwise" and "counterclockwise" pinwheels, respectively. Scale bar for panel a: 1 mm. Scale bar for panels b and c: 200 µm. A: anterior, P: posterior, M: medial, L: lateral.

References:

S.-G. Kim, S.-P. Lee, B. Goodyear, and A. Silva, "Spatial Resolution of BOLD and Other Functional MRI Techniques", in Medical Radiology - Diagnostic Imaging and Radiation Oncology, Volume Functional MRI (eds. C. Moonen and P.A. Bandettinni), Springer-Verlag, pp 195-203, 1999.
 

D.-S. Kim, T.Q. Duong and S.-G. Kim, "High resolution imaging of Iso-Orientation Columns using fMRI", Nature Neuroscience, 3: 164-169, 2000 (PDF file) D.-S. Kim, T.Q. Duong and S.-G. Kim, "Reply to Can current fMRI techniques reveal the micro-architecture of the cortex?", Nature Neuroscience, 3: 414, 2000.  (PDF file)

B. Spatial Extent of Hemodynamic Responses:

The ultimate limit of spatial resolution of fMRI is dependent on an intrinsic hemodynamic response induced by increased neural activity.  However, spatial extent of the intrinsic hemodynamic response is controversial. Based on optical imaging and functional imaging studies (see above), a hemodynamic response is diffuse and widespread.  Since both optical imaging and functional imaging techniques are extremely sensitive to large draining veins, poor spatial specificity of hemodynamic response can result from either intrinsic spatial uncoupling between metabolism and hemodynamic response or large signal contributions from draining veins. To separate these two components, we used a large-vessel-free perfusion-based functional imaging technique, Flow-sensitive Alternating Inversion Recovery (FAIR).  We found that the hemodynamic response is well controlled to submillimeter columnar levels.
 

Perfusion-based functional map of cat visual cortex and its time course during stimulation with horizontal moving gratings.  Color pixels have statistically significant signal changes.  The white color bar = 1 mm.  Note that localized patches with semi-regular shapes are highly reproducible and occupy complementary cortical territories of two orthogonal stimuli.  Time course of activated areas shows significant CBF changes.  Black bars under the time course indicate 1-min stimulus.

References:

T.Q. Duong, D.-S. Kim, K. Ugurbil and S.-G. Kim, "Localized Cerebral Blood Flow Response at Submillimeter Columnar Resolution", Proc. Natl. Acad. Sci. USA,  98: 10904-10909,  2001. PDF file