Who would have thought of asking violinists to help research decision-making in groups – a broad subject related to economics, politics, human sciences and more?
This unusual technique was used by a group of Israeli and American researchers who recruited 16 musicians playing the noble stringed instruments to study the behavior of a human network. The team – headed by Dr. Moti Fridman of the Kofkin Faculty of Engineering at Bar-Ilan University (BIU) in Ramat Gan (near Tel Aviv), Prof. Nir Davidson of the Weizmann Institute of Science in Rehovot, and Elad Shniderman from Stony Brook University in New York – created a musical ensemble that acted like a network. Their findings have just been published in the prestigious journal Nature Communications.
The team wanted to find out what sets human networks apart from other networks, such as those comprising animals, computers and other objects. The results combine science and aesthetics and also evoke thoughts about the spread of the coronavirus.
Synchronization, in which a complex system operates as one body, is an important phenomenon that takes place in a vast range of scales – from subatomic particles to galaxies. In biology, fish, birds and even cells synchronize so they can survive. Group synchronization is essential to human beings and critical to our physical and mental health.
The research has led to a new model for simulating human networks, which is important for several applications. The dynamics of human networks are essential for understanding decision-making in groups. Since the experiment is the first to measure the dynamics of complex networks, it can help understand how and when a group of people in a social network exposed to false information arrives at wrong conclusions. Thus it can prevent what is known as “fake news” from spreading uncontrollably. In addition, the research is related to epidemic control and understanding how many connections one can preserve and still prevent an epidemic from spreading.
The results are also related to any network where each node in the network has decision-making ability, such as autonomous cars, or introducing AI into our highly-connected world. This model can predict with high accuracy the dynamic of such systems, beyond what was possible before.
Examples of synchronization can be seen in drivers on the road or in a crowd of people clapping hands together. Thanks to today’s modern telecommunications, aor a group of people to make a decision, they don’t have to meet. Instead, there is a complex network of connections that helps them make decisions. The phenomenon of synchronization among humans in a complex network, yet it hasn’t been studied until now in a measurable and accurate way.
The ensemble was composed of 16 violinists making soundless music while wearing headphones; each repeatedly played an identical short musical phrase that could be heard through the headphones his/her own performance along with the performance of two or more violinists. Visual information was also neutralized by separating the musicians with partitions. All they were asked to do was to synchronize with each other – or in other words, to play along with what they heard in the headphones.
The experimental setup created by the researchers allowed them to control the connectivity of the network, such as how many of the members of the ensemble each musician was connected to and the intensity in which each musician heard the other players. What the musicians heard in the headphones was one or two violinists or more playing with them in real time while an increasing delay was imposed on the system.
“Here we have a different phenomenon than a regular musical piece. There is no global clock, but many people within a certain network of communication responding independently. In fact, it’s an aesthetic object that reveals the behavior of people in a group personally, or as an ensemble,” explained Fridman.
“By introducing a delay between the coupled violinists so that each violinist heard what his/her neighbors played a few seconds ago, we prevent the network from reaching a synchronized state,” said Fridman. This is called a ‘frustrated situation’ and is well studied in different types of networks. According to current network theory models, in a frustrated state, each node will try to compromise between all its inputs.
“Humans behave differently,” the Bar-Ilan scientist explained. “In a state of frustration, they don’t look for a ‘middle’ but ignore one of the inputs. This is a critical phenomenon that is changing the dynamics of the network. It has not been addressed to date because the measurements weren’t clean and couldn’t be shown.”
The research, which actually began as a scientific-artistic project for BIU’s Fetter Museum of Nanoscience and Art, offers two innovations: the first is methodological – a platform that measures human network dynamics accurately and cleanly. The second is evidence that a human network has two unique characteristics – the flexibility to change pace and the ability to filter and even ignore inputs that create frustration. These capabilities fundamentally change the dynamics of human networks relative to other networks and necessitate the use of a new model to predict human behavior.
“f you take humans and study how they clap their hands together, you have no control over who hears what. While working on this project we discovered that human networks behave differently than any other network we’ve ever measured. Human networks are able to change their inner structure in order to reach a better solution than what’s possible in existing models. This concept is the core of our scientific and aesthetic discovery,” concluded Fridman.