FOR 2093

Research program

The goal of the research group is the technological simulation of neuronal information processing in the trisynaptic circuit of the hippocampus based on memristive devices.

The main focus of the research group is the transfer of the neuronal, synaptic plasticity and information storage into an analogue circuit technology with the help of nanoelectronic, memristive devices. The challenge will be to transfer synaptic plasticity mechanisms as a basis of learning- and memory processes in biological systems to memristive technological circuits. This includes the simulation of neurobiological cellular mechanisms as they occur in neurons and synapses via memristive devices. The aim is the structural and technological reconstruction of cellular learning- and memory forms that enable the simulation of implicit and explicit learning processes. A detailed characterisation of static and dynamic characteristics of memristive devices and a simulation of cellular learning mechanisms by using memristive systems enables the emulation of learning- and memory processes in neurobiologically inspired network circuits. Furthermore, we aim to adapt neurobiological circuit processes based on the special structure of the hippocampal trisynaptic circuit that closely adhere to the hippocampus of the mammalian brain. The hippocampus represents the central brain area for memory, learning and separation and completion of patterns. The aim of the research group is the simulation of functional sequences within the tri-synaptic, hippocampal circuit in a memristive circuit. This includes clarifying to what extent it is possible to transfer neurological concepts of cognitive maps and place cells to analogue circuit architectures. The research group aims to reach this ambitious goal through a network of scientists with complementary expert knowledge. 

Distinctive characteristics of the research group

  • Forming-free, memristive quantum-mechanical tunnel contacts, memristive nanocomposite devices and lateral devices with resistive circuit mechanisms that are not based on the formation of filamentary structures for charge transfer but on homogeneous interface effects (the devices should be highly resistive and low in dissipation)
  • Silicon-based floating-gate transistors with memristive characteristics (MemFlash)
  • Use of in-situ interface analyses for correlation of the electrical, ionic and structural characteristics of the devices
  • Kinetic Monte-Carlo simulations to describe memristive circuit effects in nanoionic devices
  • Efficient extraction of the essential memristive characteristics and neuronal circuits via hardware and software emulators
  • Development of electronical neurotransmitters and electronical simulation of acquisition rates
  • Transmission of the basic mechanisms of the tri-synaptic, hippocampal functionality into neuromorphic circuits with memristive devices
  • Long-term potentiation-dependent learning on network level in fault-tolerant auto- and hetero-associative circuits
  • Transmission of principles of place cells and cognitive maps into electronical circuits
 
The successful realization of neuronal information processing via memristive, nano-electronical devices will establish various possibilities within neuromorphic circuit technology. Self-adapting systems with a parallel architecture with at the same time extremely low energy dissipation are conceivable. Possible applications of memristive, neuromorphic circuits are, for instance, visual pattern recognition, auditive real-time signal processing, autonomous robots, intelligent machines and Green IT. This circuit architecture could be the corner stone for a paradigm shift in information technology. Nanoelectronics, system theory and neurosciences are the keystones of the interdisciplinary research group. The research group has designated expert knowledge of materials science, interface analytics, production of devices, circuit technology, system theory and neurology.