Book Review: Neural Networks

The book reviewed here is Neural Networks by Phil Picton. The book is relatively short running at 195 pages. Picton begins with a helpful introduction to neural networks. The author then moves onto a discussion of ADALINE (Adaptive Linear Neuron) which uses a supervised learning method . In the third chapter, perceptrons are examined – software approximations to the behaviour of neurons and there is a clear explanation of single layered and multi-layered perceptrons as well as the rules that are used to create effective learning. The author then goes onto discuss a number of different types of neural network including associative networks, Hopfield networks, Kohonen networks (self-organising) as well as some practical applications although this is in engineering. The authors include exercises at the back of each chapter with worked answers. There is lots of maths throughout the book (mainly matrix algebra).

Why is this relevant to psychiatry. Well, neural networks provide a biologically plausible framework for how neurons can act together to learn information about the world around them. Thus we can hypothesise about different brain regions based on their neuroanatomical connections and structure. Relating this to the neurophysiology is a little more tricky as sophisticated methods are required for detecting activity in vivo. This usually occurs during neurosurgical operations and involves single cell electrode recording. Using non-invasive imaging technologies is fraught with difficulties as there signals are dampened by the cranium, the intervening connective tissues as well as by interference  from neighbouring neurons or collective groups. Projects such as the Blue Brain may illustrate ways forward however. The other method that neural networks can have an impact on practice is on diagnosis where neural networks have demonstrated utility in various branches of medicine. Neural networks are an interesting means of looking at brain function in a structured manner and this book would be of particular use to people intending to undertake research in this area or who have had some practical exposure to neural networks.

References

Phil Picton. Neural Networks. Second Edition. Grassroots Series. 2nd Edition. 2000.

Disclaimer

The comments made here represent the opinions of the author and do not represent the profession or any body/organisation. The comments made here are not meant as a source of medical advice and those seeking medical advice are advised to consult with their own doctor. The author is not responsible for the contents of any external sites that are linked to in this blog.