A Subway Map Metaphor for Visualizing Nanoscale Neuronal Connectivity
IEEE InfoVis 2014
||Jeff W. Lichtman
We present NeuroLines, a novel visualization technique designed for scalable detailed analysis of neuronal connectivity at the nanoscale level. The topology of 3D brain tissue data is abstracted into a multi-scale, relative distance-preserving subway map visualization that allows domain scientists to conduct an interactive analysis of neurons and their connectivity. Nanoscale connectomics aims at reverse-engineering the wiring of the brain. Reconstructing and analyzing the detailed connectivity of neurons and neurites (axons, dendrites) will be crucial to understanding the brain and its development and diseases. However, the enormous scale and complexity of nanoscale neuronal connectivity pose big challenges to existing visualization techniques in terms of scalability. NeuroLines offers a scalable visualization framework that can interactively render thousands of neurites, and that supports the detailed analysis of neuronal structures and their connectivity. We describe and analyze the design of NeuroLines based on two real-world use-cases of our collaborators in developmental neuroscience, and investigate its scalability to large-scale neuronal connectivity data.
NeuroLines neurite visualization.
We abstract the original 3D structure and topology of neurites segmented in nanoscale brain tissue data into a 2D subway map visualization that preserves topology and relative distances. Left: Volume rendering of a dendrite (red) and connected axons (blue). Right: NeuroLines abstraction of the same data, represented as subway lines to more clearly show branches, clusters of synapses in proximity, individual synapses, and the actual connections (shown on demand).
Data processing workflow.
After slicing and imaging the tissue block, registration, segmentation, and synapse labeling are performed. The segmented neurites are then skeletonized, which forms the basis for subsequent interactive visualization and analysis in NeuroLines.
Neurolines design prototypes.
a) 3D volume rendering of a cylindrical region of interest of the segmented data; b) First visualiza- tion approach in 3D for directly displaying neurite skeletons; c) Ab- stract 2D graph visualization showing the connectivity between neu- rites, without spatial information; d) First 3D subway map prototype.
We visualize neurites using a multi-scale approach with three tiers of linked views (a,b,c), which provide overview as well as details. The navigation bar (a) shows individual neurites as lines color-coded according to selected neurite attributes and sorting criteria. The neurite overview (b) allows inspecting all neurites at a medium level of abstraction. The workspace view (c) allows the inspection of neurites at a detailed level of abstraction. The neurite analysis (d) shows statistics for a selected neurite. Pinning (e) a neurite in the workspace allows keeping a specific neurite in focus while exploring others. The synapse analysis (f) shows synapse details and its neighborhood in the original EM volume. (g) NeuroLines is also coupled with a 3D volume visualization for additional visualization and exploration in 3D.
We employ different abstraction levels to visualize neurites at different levels of detail. (a) Medium-level abstraction showing an overview with collapsed branches. (b) Detailed view (low-level abstraction) with individual synapses shown as diamonds (spinal) or circles (non-spinal). (c) Synapses overlapping in screen space are automatically clustered. Clusters show the number of contained synapses and can be fanned out to show them in sequence.
The current sorting of neurites is depicted in the navigation bar by color-coding the sorting criteria according to neurite attributes. (a) Multi-critera sort- ing order from left to right: neuron id, function (excitatory/inhibitory), type (axon/dendrite), no. synapses, assigned neurite color. (b,c) Multi-critera sorting and color map (left to right: neuron id, type, function). (d) Tool tip for a single neurite showing its attributes. (e) Neurites can also be color-coded with a single attribute (no. synapses) regardless of the sorting criteria. (f) The navigation bar also displays a slidable focus window, which is linked to the neurite overview for navigation.
This view combines a medium-level abstraction of a neurite with detailed statistics over all of the neurite’s synapses, e.g., percentage of spinal vs. non-spinal synapses.
Spatial 2D and 3D views.
The synapse analysis view shows detailed synapse information and a 2D inset view that shows the synapse neighborhood in the original electron microscopy volume. The 2D inset view allows users to toggle the display of segmentation information, and is linked with a 3D volume visualization for further exploration.