Anatomical definition of hMT+ using quantitative R1 mapping

2309 

Abstract Submission 

Woon Ju Park¹, Kelly Chang¹, John Pyles¹, Ione Fine¹

¹University of Washington, Seattle, WA

Woon Ju Park  
University of Washington
Seattle, WA

Kelly Chang  
University of Washington
Seattle, WA

John Pyles, PhD  
University of Washington
Seattle, WA

Ione Fine  
University of Washington
Seattle, WA

Human middle temporal complex (hMT+) is traditionally identified functionally, as the cortical area that responds more strongly to visual motion versus static stimuli. In recent years, quantitative R1 mapping has emerged as a valuable tool for measuring the local myeloarchitecture of the human brain in vivo. hMT+ has higher myelination than neighboring cortical areas (Bridge et al., 2014), and previous studies using a group-averaged approach have found that group-averaged functional hMT+, as defined by retinotopy, overlaps with a region of high myelination (Sereno et al., 2013). Here, we examined whether the location of hMT+ based on quantitative mapping is consistent with the functional location, within individuals.

Structural and functional MRI images were collected in 8 neurotypical observers (ages 33–72, 4 females). Quantitative structural mapping was performed by acquiring PDw, MTw, and T1w images using a multi-echo 3D FLASH pulse sequence. R1 maps were derived using the hMRI toolbox (Tabelow et al., 2019). Vertex-wise residuals from a linear regression, with local curvature as a predictor for R1 values sampled at 50% cortical depth, were used as the final estimates of the R1 map. Functional characterization of hMT+ was based on standard T2* EPI scans. Participants viewed alternating blocks of moving and static dots within an 8 deg radius circular aperture while fixating on a cross at the center of the screen. The BOLD activity for moving and static blocks were contrasted to find a region selectively responsive to visual motion. The Dice coefficient was used to quantify the overlap between anatomical and functional definitions of hMT+ for each hemisphere of each individual. Bootstrapping combined with a Receiver Operating Characteristic (ROC) analysis was used to examine whether the spatial overlap between the two definitions of hMT+ is greater within than between individuals.

In all participants, quantitative R1 mapping identified a highly myelinated area in the lateral occipital cortex, which aligned with regions identified as MT/MST in the HCP-MMP1 atlas. There was a strong overlap between group-averaged anatomical and functional hMT+ (Dice coefficient of 0.81 on the left and 0.58 on the right hemisphere; Figure 1). There was also considerable overlap between the anatomical and functional definitions of hMT+ within individuals (left hemisphere Dice coefficient mean = 0.6, SD = 0.16; right hemisphere mean = 0.42, SD = 0.25; Figure 2A). An ROC analysis comparing the cumulative distribution function of bootstrapped Dice coefficients within and across individuals suggested that the two definitions of hMT+ were more consistent within than between individuals (Figures 2B &C). The area under the ROC curve (AUC, an aggregate measure of whether overlap was higher within than between individuals) was 0.94 for the left hemisphere and 0.74 for the right hemisphere.

The anatomical delineation of hMT+ using quantitative myelination mapping shows substantial overlap with regions functionally responsive to visual motion within individuals. Thus, quantitative mapping provides a potential method for identifying hMT+ in populations where the functional localization of hMT+ might pose practical challenges (e.g., infants or individuals with vision loss).

Anatomical MRI ¹

Perception: Visual ²

Cortex

FUNCTIONAL MRI

STRUCTURAL MRI

Vision