High-resolution fMRI mapping of ocular dominance layers in cat lateral geniculate nucleus

Abstract

In this work, we exploited the superior capability of high-resolution functional magnetic resonance imaging (fMRI) for functional mapping of ocular dominance layer (ODL) in the cat lateral geniculate nucleus (LGN). The stimulus-evoked neuronal activities in the LGN ODLs associated with contralateral- and ipsilateral-eye visual inputs were successfully differentiated and mapped using both blood-oxygenation-level dependent (BOLD)-weighted and cerebral blood volume (CBV)-weighted fMRI methods. The morphology of mapped LGN ODLs was in remarkable consistency with histology findings in terms of ODL shape, orientation, thickness and eye-dominance. Compared with the BOLD signal, the CBV signal provides higher reproducibility and better spatial resolvability for function mapping of LGN because of improved contrast-to-noise ratio and point-spread function. The capability of fMRI for non-invasively imaging the functional sub-units of ODL in a small LGN overcomes the limitation of conventional neural-recording approach, and it opens a new opportunity for studying critical roles of LGN in brain function and dysfunction at the fine scale of ocular dominance layer.

Figure 1

Typical functional maps of LGN ODLs based on the BOLD-weighted fMRI measurement from two adjacent image slices shown in (A) and (B). Green color indicates that the fMRI signal is dominated by the right-eye stimulation; red color indicates that the fMRI signal is dominated by the left-eye stimulation. Three major monocular ocular dominance layers in LGN (i.e., layers C_M, A1 and A) can be reliably differentiated and identified, as depicted by white solid lines delineating the territory boundaries among the ODLs. The corresponding LGN ODLs in the two slices show consistent eye dominance, similar ODL orientation, and a smooth transformation in the LGN shape. The fMRI map in (B) is more anterior compared to the map in (A).

Figure 2

Comparison of functional LGN ODL maps created from the BOLD and CBV signals. Both BOLD- and CBV-weighted fMRI maps are capable of differentiating the ODL structures in the LGN. The position of each corresponding ODL is almost identical between the BOLD and CBV maps. The shape and orientation of each corresponding ODL are also similar. However, the thickness of layers A and C_M is systematically wider in the BOLD map than that in the CBV map, while that of layer A1 is remarkably narrower.

Figure 3

Functional LGN ODL maps including both BOLD- and CBV-weighted fMRI maps from all cats. White profiles in each fMRI map are the LGN body boundaries drawn based on the corresponding anatomical image. For each image slice, its position relative to the most posterior end of LGN body (indicated by the number on each map) was identified by comparing anatomical landmarks in the anatomical image of the selected slice with those in a set of 3D high-resolution T₁-weighted anatomical images that covered the entire LGN body (Supplemental Figure S1). On the left of each fMRI map, the outlines (black color) of each LGN ODL at the corresponding image position were drawn based on histological images for comparison (Supplemental Figure S2, adapted from the reference of (Sanderson, 1971)). It is evident that the mapped LGN ODLs and whole LGN body by fMRI are in excellent agreement with the anatomic LGN architectures.

Figure 4

Illustration of the effect of the point spread function in BOLD- and CBV-weighted fMRI signals on mapping the LGN ODL structures. The traces of BOLD and CBV signal are along the brain region marked by the white dash lines in the fMRI maps. The red traces indicate profiles of CBV (left panel, Figure 4A) and BOLD (right panel, Figure 4B) signals induced by ipsilateral-eye stimulation. The green traces are profiles of the BOLD and CBV signals induced by contralateral-eye stimulation. The blue traces are the profiles after subtracting the contralateral-eye activity from the ipsilateral-eye activity for the BOLD signal and vice versa for the CBV signal. These traces reveal well-defined CBV signal profiles across the adjacent LGN ODLs compared to BOLD signal profiles.

Figure 5

Distinctive effect of the bleeding signals in BOLD (top panel) and CBV (bottom panel) time courses. The red traces indicate the BOLD or CBV signal profiles measured from the left-eye LGN ODLs; and the green traces indicate the BOLD or CBV signal profiles measured from the right-eye LGN ODLs during left- or right-eye stimulation. Bleeding signal in the BOLD time courses is substantial, yet it is negligible in the CBV time courses due to a better CBV spread point function.

Figure 6

Examining the reproducibility of LGN ODL mapping. Data are presented in the original imaging spatial resolution (390×390 μm²). The BOLD (left panel) and CBV (right panel) data acquired within the same fMRI session were split into two halves, respectively. (A) The LGN ODL maps created from all the BOLD or CBV datasets. (B) The LGN ODL maps created from the first halves of fMRI datasets. (C) The LGN ODL maps created from the second halves of fMRI datasets. There are consistencies among the split fMRI maps.

Figure 7

Correlation of the first and second halves of fMRI data in the signal amplitude differences between the left- and right-eye stimulated LGN ODL activities. The sign of the amplitude differences can be positive or negative, and it depends on which eye being stimulated as demonstrated in Figure 5. The left panels are plots from one cat and the right panels are plots from all cats. The upper panels are plots for the BOLD data and the lower panels are plots for the CBV data.