A rat trained to control a robot is an interesting ‘Pavlovian’ experiment by itself. But, independently-cultured rat brain cells aiding a robot navigate through an obstacle seems straight out of science fiction.
Having cultured brain cells on a glass plate and kept in sterilized conditions, a group of researchers at Indian Institute of Science (IISc) demonstrated that this tissue culture can read signals from an infra-red enabled robot, process the problem of obstacles, and give an appropriate, accurate solution.
The results of the 2.5-year-long experiment were described in an award-winning paper presented recently at the 28th International Conference on VLSI Design and Embedded Systems.
Scientists — from the Center for Nanoscience and Engineering (CeNSE) and Electrical Communications Engineering at IISc — took the rat brain cells (hippocampus of just-born rat pups) and cultured it on a specialized glass plate that is covered with multiple electrodes that can detect the most minute spiking in voltages generated by the cells. The cultured cells start to grow dendrites — the branched structure of a neuron that is capable of transmitting to and fro electrical simulation — and rudimentary communication between them is established.
The cells form a network that shows spontaneous electrical activity through tiny voltage spikes.
Interpretation of these spikes is done through an electronics platform that can detect as well as send electrical impulses (of just 500mV amplitude) to the cultured tissue through the embedded electrodes.
Much like the brain, a coding system is developed to decode the spikes. Furthermore, “training” is imparted through instructions coded as electrical spikes.
One of the experiments involve an infra-red robot that ‘senses’ obstacles. These impulses are fed through the computer to the cells, which process the information, and the resulting voltage spikes — for commands of front, back, left and right — are translated into codes for the robot. Run over 10 minutes, with obstacles moved around in random, the robot was able to navigate successfully nearly 98 per cent of the time.
Though still nascent, the “Neuro-electronic hybrid systems” experiment allows researchers to develop electronic systems that use the learning and processing abilities inherent in brain cells. Jude Baby George from CeNSE believes the system has the potential to build a computing system with “wetware” — a combination of hardware, software and neural functions of organic matter — capable of adapting and solving real-world problems.
