Data Sculptures

neuroscience
brain research
physical networks

With over 100 billion nodes, the human brain is perhaps the most complex network known to science. Neuroscientists call the whole brain network, mapped at the level of individual neurons, the »connectome.« The connectome plays a vital but still unexplored role in brain functioning and consciousness; understanding its structure is therefore crucial.

The 3D rendition of the »Connectome«, the neuronal network of a mouse brain, is the BarabásiLab’s first attempt to visualize brain structure. Its configuration is based on data collected over the course of a multiyear project at the Allen Institute, a bioscience research center in Seattle, USA. The mouse connectome was laid out using the same algorithm that generated »The Flavor Network« and the other data sculptures in the exhibition. Its arrangement exposes the exceptional complexity of the brain’s wiring patterns, and opens up new avenues towards understanding brain function.

physical networks
thermal gradient
network temperature
infinity network

The physicality of a network strongly limits its structure. Since the links of a physical network cannot cross each other, they must deviate from straight lines to reach their destination. In the brain, neurons have to bend and twist to avoid each other. Since 2006 the BarabásiLab has been involved in a systematic effort to understand the properties of physical networks. To quantify the deviations in links, the Lab developed the concept of network temperature. A zero-temperature network has only straight links, while the links of warmer networks swerve and curve to get around each other. Through temperature controls, the Lab created a series of hot networks from a wide variety of materials for the »Heat« series. The starting point was a two-dimensional grid whose temperature was raised in the middle and kept cooler on the outside. A continuation of the series in three dimensions starts from a cube with fixed nodes; the temperature of all links is raised evenly, helping contrast the regularity of the lattice with the randomness of the links. In addition to the data sculptures, »Infinity Network« was recently created to show that most real networks are not finite, but grow to infinity over time.

3D visualization
physical networks
inflatables

One challenge in 3D visualization is representing large dense networks, whose inner structure is obstructed by its large number of links. If the links between the nodes are thick enough, they must wind around each other to reach their destination rather than following straight lines. When such curvatures become extreme, dense physical networks can change their state of aggregation and enter a new phase of existence.

Albert-László Barabási refers to these dense networks as »Gurgas«, a term rooted in the Latin word »gorge.« Thanks to their compact shapes and lack of loose links, the »Gurgas« became the lab’s testing ground for 3D-printing materials, such as bronze and plastic. The most recent material experiment is the work »Inflatable Gurga«. The process of inflation embodies the transition from a data sculpture to a »Gurga«, with a structure shaped by the extreme expansion of its links. The project is inspired by the genre of inflatable pneumatic sculptures in art, and embodies the strategy of inverse appropriation, an artistic practice of the BarabásiLab that involves the transfer of the formal and visual language of art into science.

3D printing
random graph
scale-free networks

After developing a mathematical toolset for data sculptures, in 2018 the BarabásiLab once again returned to the »Network Canon« project, which had begun in 2000 and aimed to illustrate the canonical network models. Using 3D printing, it was now possible to render the structural differences between the two most studied network models, the Erdős-Rényi and Barabási-Albert models. The former is based on work done in 1960 by the Hungarian mathematicians Pál Erdős and Alfréd Rényi on randomly connected networks. The latter is based on the theory developed by Albert-László Barabási and Réka Albert in 1999, introducing so-called scale-free networks, which are held together by hubs with an exceptional number of links. This discovery contradicted the prevailing theory of random networks, which assumed that the nodes of real networks are randomly linked to each other, resulting in a relatively uniform, »democratic« network architecture. The two models from 2018 presented in the exhibition, constituted the first successful attempt to implement the random and scale-free network as data sculptures.

The data sculpture »Human Disease Network«, also on display, is the first prototype of a multicolor 3D printing. Because of its spatial composition, the final 3D color version unveils relationships between disease classes that are largely invisible in the 2D map.