Biological and Neural Computation
The UH Biological and Neural Computation Research Group represents the fusion between the Neural Computation Group and the Biocomputation Research Group. It was formed in 2005 in recognition of the large amount of overlap in term of projects and staff between the two original groups. The Biological and Neural Computation Group forms part of the Centre for Computer Science and Informatics Research (CCSIR) which is based within the Science and Technology Research Institute (STRI). The Biological and Neural Computation Group also has strong links with the Psychology Department. Links with Biology and Psychology have allowed a number of inter-discplinary PhD projects to be explored, which is a strong theme running throughout the Group.
Biologically-inspired computing, the concept of using computers to model nature, and simultaneously the study of nature to improve the usage of computers, is an area that is growing rapidly.
Our expertise in modelling and pattern recognition, combined with our studies of the molecular biology of neural development, evolution and physiology, places us in a very strong position to contribute to the development of computational methods in molecular and neural biology.
Our collaborative link with the Division of Biology at the California Institute of Technology (CalTech), which started in computational neuroscience in 1996 with Professor Eric Davidson, has rapidly developed into several large scale, multinational projects.
Since 2001 this research has expanded to include collaboration with:
- Dr Alistair Rust at the Institute for Systems Biology in Seattle
- Professor Lee Kroos at Michigan State University
- Dr Irena Abnizova at the Medical Research Council (MRC) Biostatistics Unit
- Dr Nicholas LeNovere at EMBL-EBI (European Bioinformatics Institute) in Cambridge (UK)
- Dr Steve Martin at the MRC NIMR (National Institute for Medical Research) in London.
Since those beginnings over 10 years ago, we have formed collaborations with several high profile institutions in Europe, the USA and Japan.
Neural computation
Research in neural computation, led by Professor Rod Adams, focuses on evolutionary systems, which has led to close collaborations with both psychologists and biologists.Research in neural computation can be clustered into three related themes:
- Examination of the fundamental issues in neurobiology centred around the practical study of biologically realistic models of high capacity associative memory.
- Using neural networks and advanced computational tools for pattern recognition to solve practical problems, typically motivated by external clients and collaborators.
- Modelling of cognitive phenomena as part of a long-standing collaboration which Dr Neil Davey has developed with the School of Psychology.
Current projects and collaborations
The following are some of our current research projects carried out at the University of Hertfordshire:
Computational models
We use computational models at different levels of complexity and collaborate with experimentalists in the UK, Europe and the USA to model neural coding in the cerebellum, the part of the brain that is involved in movement control and motor learning.
The research is funded by a collaborative systems biology grant from BBSRC (Biotechnology and Biological Sciences Research Council) and ANR (National Agency for Research) (France) to Dr Volker Steuber.
Mammalian cortex
The mammalian cortex acts as a massive associative memory – it contains a vast number of neurons.
We are using some of the known properties of the cortex as an inspiration for building artificial neural networks made of tens of thousands of neurons that function as large-scale associative memories.
Face images
We are collaborating with Queen Elizabeth II Hospital to classify face images so that we can identify the emotional state of people from their faces.
Detection of errors in optical data
We are collaborating with Aston University in an investigation of how rapid detection of errors in data transmitted over optical fibre can be carried out using contemporary machine learning techniques.
Neural controllers
We are developing neural controllers for simulated robots that can detect and react to simulated pheromones.
Other related research within the faculty and elsewhere at UH can be found by following the links below:
If you have any comments or any contributions to make regards the NCGroup webpages feel free to drop us a line.