1213

MouseStream: A Software Suite for Mapping and Analyzing Mouse Cortical Functional Architecture In Vivo Using Magnetic Resonance Microscopy
Jia Guo1, Xinyang Feng1, Hannah Sigmon2, Frank Provenzano2, and Scott A. Small3

1Department of Biomedical Engineering, Columbia University, New York, NY, United States, 2Departments of Neurology, Columbia University College of Physicians and Surgeons, New York, NY, United States, 3Departments of Neurology, Radiology or Psychiatry, Columbia University College of Physicians and Surgeons, New York, NY, United States

Synopsis

The functional architecture of the cortex has never been mapped in vivo with the fidelity necessary to distinguish “functional unit”. Mapping the functional architecture of the cortex, reflecting the observed regional and layer differences in synaptic density and its correlate energy metabolism, has been very challenging. Here we set out to address this issue by in vivo high-resolution cerebral blood volume (CBV) mapping of the mouse cortex. Tailored software, MouseStream, was developed to reconstruct the cortical CBV data through the normalized curved cortical coordinate (NCCC). NCCC allows projection of functional architectures mapped using CBV onto the cortical surface across the whole cortex and from different cortical depths.

INTRODUCTION

The functional architecture of the cortex reflects cellular compositions and metabolic states that vary across the cortical surface and among its different layers. The cortical functional architecture has never been mapped in vivo with the fidelity necessary to distinguish “functional unit”. Here we address this issue by in vivo high-resolution relative cerebral blood volume (rCBV) mapping1 and generated microscopic three-dimensional functional maps of the cortex in the living mouse.

Mapping the cortical functional architecture, reflecting regional and layer differences in synaptic density and its correlate energy metabolism, has been very challenging. Tailored software, MouseStream, was developed to address this challenge through the normalized-curved-cortical coordinate (NCCC).

METHODS

A Bruker BioSpec 9.4T system was used with a 23mm 1H circularly-polarized transmit/receive capable mouse head volume coil. Mice were anesthetized using the medical air and isoflurane. T2-weighted images were acquired before and 36-min after intraperitoneal injections of the contrast agent Gadodiamide at the dosage of 10 mmol/kg. T2-weighted images were acquired with RARE (TR/TE: 2,500/45 ms; RARE factor, 16; in-plane resolution, 50 µm; slice thickness, 250 µm), from 10 wildtype (C576BJ) mice in total.

As previously described1, rCBV was mapped according to changes in the transverse relaxation time (∆R2) induced by the contrast agent. rCBV images were preprocessed with skull stripping, large vessel filtering, isotropic up-sampling and co-registered into a common template space2.

Co-registered rCBV volumes in Euclidean coordinate have to be first transformed to the NCCC, which aligns better with the cortical anatomy. The proposed pipeline is illustrated in Figure-2. We used a similar technique as mapping mouse cortical thickness by solving the Laplace’s equation3,4. This was the first time applying this technique in this context to process in vivo functional MRI data of the mouse brain.

The cortical functional architecture patterns we observed vary by cortical depths and regions. A whole brain depth & region of interest (DROI) analysis was performed to illustrate those features. MouseStream also generates the cortical thickness map as one output.

To visualize the whole cortex in one view, flat maps can be generated for both cortical rCBV and the thickness.

RESULTS

Figure-1 shows a typical group-wise in vivo rCBV map compared with the ex vivo COX staining.

The proposed pipeline is illustrated in Figure-2.

rCBV surface projection results generated by MouseStream at 20 depths of the cortex are displayed in Figure-3.

Figure-4 shows cortical DROI results of rCBV calculated from 20 depths and 26 regions.

Figure-5 shows the flap maps of cortical labels, rCBV and thickness.

DISCUSSION

MouseStream utilizes the high contrast, high-resolution rCBV maps in vivo and is capable to map the three-dimensional functional architecture of the whole cortex noninvasively, across the cortical surface and among its different depths.

Besides providing insight into the cortical organization, MouseStream allows a more precise spatial-temporal mapping of how the brain changes its function and structure in response to experience. And more importantly, being able to map function and structure simultaneously opens up the opportunity to better understand the functional-structural relationships in neuroplasticity.

CONCLUSION

We have successfully mapped the in vivo functional architecture of the mouse cortex across its surface and among its different depths for the first time, using rCBV as the image contrast. MouseStream provided an automated processing, analysis and visualization pipeline for mapping mouse cortical functional architecture in vivo. Surface-based vertex-wise data comparisons can be easily applied.

Besides clarifying the functional organization of the cortex, we believe that the more important application of the reported MRI tools is to map functional changes over time. When investigating brain function and how and where it changes with experience, the MRI tools add an important complement to electrophysiology and confocal microscopy.

Acknowledgements

No acknowledgement found.

References

1. Moreno, H., Hua, F., Brown, T. & Small, S. Longitudinal mapping of mouse cerebral blood volume with MRI. NMR Biomed 19, 535-543 (2006).

2. Avants, B.B., et al. A reproducible evaluation of ANTs similarity metric performance in brain image registration. Neuroimage 54, 2033-2044 (2011).

3. Jones, S.E., Buchbinder, B.R. & Aharon, I. Three-dimensional mapping of cortical thickness using Laplace's Equation. Human Brain Mapping 11, 12-32 (2000).

4. Lerch, J.P., et al. Cortical thickness measured from MRI in the YAC128 mouse model of Huntington's disease. Neuroimage 41, 243-251 (2008).

Figures

Figure 1. rCBV was derived by normalizing ∆R2 to the mean ∆R2 signal present in the posterior cerebral artery. (A) The averaged rCBV map and the surface 3D-plot to show rCBV intensity as height. Individual modules with high basal metabolism were observed at layer-IV of the barrel cortex. (B) Comparing metabolic patterns in the barrel cortex between ex vivo COX staining (left) and in vivo rCBV maps (right). Qualitative comparable patterns were observed at the layer-IV in the barrel cortex. Spatially separated functional modules with locally high metabolism are observed in the barrel fields (between the red arrows).

Figure 2. Overview of the proposed pipeline. (A) The input cortical rCBV volume. After borders of the mouse cortex were defined (B), Laplace’s equation was solved between cortical surface and white matter surface by evenly subdividing the cortex into multiple equipotential intermediate surfaces (color-coded) as shown in (C). (D) Streamlines (light blue) starting from vertices (dark blue) on the meshed cortical surface (gray mesh) were computed by searching for the steepest descent path through the equipotential field. (E) rCBV values at different cortical depths can then be projected along the streamlines onto the cortical mesh surface which allows further integration or comparison.

Figure 3. Mapping functional architectures in vivo with rCBV maps at different depths across the cortex. We displayed rCBV cortical surface projection results at 20 different depth ranges from 0-5% (closest to the cortical surface) to 95-100% (closest to the white matter surface). When viewing rCBV data projected onto this normalized curved cortical surface, functional architectures such as the barrel fields and limb fields in the somatosensory cortex, motor cortex, auditory cortex and the visual cortex become visually distinct (with labels displayed in the zoom-in views).

Figure 4. Generating the (depth & region of interest) DROI rCBV results. (A) Cortical rCBV can be generated at a certain depth range and/or a given ROI using MouseStream. (B) DROI rCBV results generated from 20 depth ranges (y-axis) and 26 cortical regions (x-axis). Each DROI result of a given region is normalized to its highest rCBV across the depths within the same region. Red ribbon shows regions which have highest rCBV in the shallower layers, blue ribbon shows regions which have highest rCBV in the deep layers while other colors showing cases that are in between.

Figure 5. Flat maps of the cortical ROI labels, the cortical rCBV (averaged rCBV along each streamline) and the cortical thickness.

Proc. Intl. Soc. Mag. Reson. Med. 26 (2018)
1213