Jessica Vermeer
17 June

Scientists of Eindhoven University of Technology (TUE), together with American and Italian researchers from Stanford University and Italian Institute of Technology, respectively, have developed artificial synapses that can communicate with living cells. These synapses could possibly ‘connect’ prostheses to the brain. The results have been published in Nature Materials.

The brain consists of nerve cells that send electrochemical signals to each other. These cells talk to each other through synapses and a narrow cleft, which serves as the medium for transport. When a signal passes the cleft, the connection between the synapses becomes stronger and the transmission expends less energy. As the path strengthens, the brain learns. All in all, the receiving synapse both processes and memorizes the signals, which makes for an ultra-efficient learning system.

The artificial synapse made of organic materials. The electrical probes (metal pieces) measure the conductivity. The microfluidic system (tubes above) feed the living cells and restore the synapses to their original state. Credit: Yoeri van de Burgt

TUE researcher Yoeri van de Burgt developed an artificial synapse made of organic materials in 2017, when he was a postdoc at Stanford. He has now succeeded in making this synapse communicate with living cells that resemble nerve cells. He explains: “Just like a real brain, our system appears to have a learning and a memory function. This brings us one step closer to an adaptive connection with the brain, which makes advanced prostheses and regenerative medicine possible.”

At TUE, Van de Burgt and PhD student Setareh Kazemzadeh mainly worked on the transport of liquids between the two synapses. According to Van de Burgt, where most research groups are only able to measure electrical signals, his group can truly mimic the process in the synapse. “That makes our approach more efficient but also more relevant.”


Machine Learning Conference: limited number of visitors

The second edition of the conference brings together industrial and scientific specialists who are applying ML techniques in high tech domains, or are interested in doing so. Keynote by Nicolas Lehment (NXP). Other sessions by ASML, TUE, Quby, Ghent University, Grai Matter Labs, Edgise and Verhaert AI Lab. There are only 13 seats left, so register in time!

Tests showed that the neurotransmitter dopamine did indeed cause a permanent change in the second electrode, thus altering the conductive state of the system. The next step for Van de Burgt is to apply his system to real cases in medicine.