Name: Metric: Scale:
This matrix diagram visualizes large-scale white matter connectivity in the human brain. The connectivity model was derived from magnetic resonance imaging (MRI) and tractography as follows. First, A high-resolution T1 MRI was parcellated into gray matter regions using Freesurfer, then streamline fiber tractography was performed on a diffusion MRI. Next, fibers beginning and terminating near gray matter areas were mapped to pairs of regions from the Desikan-Killany Atlas, and a variety of metrics of connectivity strength were computed from combinations of fiber length and diffusion indices. These metrics have been found to be especially sensitive to white matter integrity . This visualization includes the mean and variance of these connectivity metrics across a population of 80 normal subjects.
Each colored cell represents the value of a connectivity metric between a pair of gray matter areas, where darkness of a cell is mapped according to the scale and the metric value. Scales specify how the connectivity metrics map to cell color, and the scale can be selected in a list above. Linear scales preserve relative differences between metric values but can wash out interesting details if there are outliers. Using a logarithmic or exponential scale can give more emphasis to lower or high values, respectively. By default, the minimum and maximum metric values are mapped to the lightest and darkest colors, respectively. However, the slider above can be used to adjust the range of this mapping. Scales that begin "inv" will reverse the orientation of the mapping (making smaller values darker and bigger values lighter). Finally, a binary scale can be used to show the gross structure of the network by thresholding the data.
The names of the gray matter regions are shown along the border, and hovering over a cell will show its corresponding metric value and connecting region names. The cells may be reordered lexicographically by region name or numerically by the column sum, min or max of the connectivity metric.
Built with d3.js and derived from this example.