EDITORIAL

This is the first issue of Network: Computation in Neural Systems to be published by Taylor & Francis, after the companies acquisition of the Journal from Institute of Physics Publishing at the beginning of 2005. It is also the last issue of Network with which my name is linked as Editor-in-Chief. After involvement with the Journal from its inception, firstly as a Board Member and then for the past seven years as Editor, I felt that the new chapter in Network's history, which is just starting with the arrival of a new publisher, should be written by a new Editor. As a result, sometime ago I decided that this would be an appropriate time for me to step down.

Network was started in 1989 by IOPP, the publishing arm of the UK Institute of Physics, in response to the large numbers of physics-based neural networks papers being submitted to IOPP journals. Network was designed to stand at the intersection of neuroscience and physics and mathematics-based science. It was intended to be interdisciplinary and in the early days we went as far as to review papers initially for interdisciplinary content and then involve both experimental and computational people in the review of each paper.

As time went on, it became clear that essentially the Journal was serving two communities: people working in artificial neural networks and in computational neuroscience. When I became Editor-in-Chief in 1998, my view was that both of these areas had grown to such a level that each on their own would merit a specialist publication. Although I have worked in both fields, my ongoing interest has always been to understand how the nervous system develops and functions. It was therefore relatively straightforward for me to decide in 1999 to focus the Journal on computational neuroscience, i.e., mathematical and computer-based modelling and analytical work concerned with advancing our understanding of the nervous system at all levels (Willshaw 1999).

The field of computational neuroscience has existed for many years, under a variety of names; I always cite the classic paper on the propagation of the nerve impulse by Hodgkin and Huxley (Hodgkin & Huxley 1952) as the first landmark of computational neuroscience. When I started out as a PhD student in 1967, computational neuroscience was a minority interest. Since then, and particularly over the past 10 years, the subject has become recognised as an essential part of neuroscience. It is always tempting to advance reasons as to why one's own field has developed; quantitative evidence for the development of computational neuroscience can be obtained from the number of academic and postdoctoral positions in the area, the visibility at leading conferences, the number of papers published and the number of relevant funding schemes.

Here are a few reflections on how the journal has fared during my tenure as Editor-in-Chief.

Scope of Network

Network's scope is very broad, potentially embracing computational neuroscience at all levels, from molecular to cellular to network to behavioural levels, in vertebrate and invertebrate systems alike. Over the past years there has been a change in the types of papers published. The change in scope announced in 1999 meant that artificial neural networks papers were no longer considered. Many of the neurobiological papers published have tended to be at the network level (perhaps reflecting the Journal's roots in physics and its title). Papers on single cell modelling and on modelling of the development of the nervous system now appear; in addition, we have started to take neuroscience methods papers.

Network's target audience

Predominantly the audience of the Journal has been the community of mathematically literate scientists who are interested in using mathematical/computational methods to model or analyse the nervous system. A minority of our audience are experimental neuroscientists who are interested in computational approaches.

Engaging experimental neuroscientists

Computational neuroscience is relevant to neuroscience as a whole and so we have always endeavoured to make the Journal attractive to all neuroscientists. One challenge has been to attract experimental scientists as Board Members, readers, authors and reviewers. We have used a variety of means. In 2001, 2002 and 2003, Network was represented at the annual meeting of the Society for Neuroscience. Over the past few years, individual issues (or parts of issues) have been devoted to special topics (Willshaw 2001), usually arising from a workshop or discussion meeting. This has given neuroscientists of all types the opportunity to sample a specific field of computational neuroscience research by reading high quality papers that had been subject to our normal high standards of refereeing. Our first attempt of this type of publishing was an entire issue devoted to Natural Scene Statistics (Reinagel & Laughlin 2001). I received many positive comments about this issue, setting the trend for other more recent special issues.

Topical Reviews

Another significant way in which we have approached this challenge is in publishing Topical Reviews that would also be attractive to the experimental neuroscientist. These are usually long reviews of a whole field of research, and they have received wide acclaim. I am particularly proud of the reviews on the basal ganglia (Wickens 1997) and on molecular computation in the olfactory system (Mori et al. 1998), which I advised on before becoming Editor-in-Chief; and those on temporal aspects of visual coding (Victor 1999) and on competition in the developing nervous system (van Ooyen 2001). This issue contains a review of another very topical aspect of the developing nervous system, by our new Editor-in-Chief, Geoff Goodhill (Goodhill & Xu 2005). The publication of Topical Reviews has become one of the unique selling points of Network. A frequent topic of discussion at our Board Meetings has been the possibilities of publishing other types of review material such as shorter, more focused reviews or acccounts of meetings and workshops.

Interaction with electronic publishing

There are various ways in which Network, still essentially a paper-based journal, has profited from recent advances in web-based publishing.

Electronic access

IOPP has always been at the forefront of electronic publishing and readers and authors have enjoyed excellent electronic access to the journal.

Web-based publishing

In the early days of the Journal, all communication with the Editor was via email. This provided a personal link between the editorial office and author/referee/Board Member but had its limitations, like any manual system. In 2002 it was replaced by a proprietory web-based submission and reviewing system that was designed specifically for IOPP journals. This has been a success and now Taylor & Francis are installing a new system, from Manuscript Central, whose products are already used very successfully by many leading journals. This system will be tailormade to the needs of Network and it will be a great asset for everyone concerned with the Journal.

Publishing the code

Finally, here is something that is a wish for future computational neuroscience papers. Most computational neuroscience papers publish the results of applying an algorithm, in the form of a computer program, to a particular neurobiological problem. In the same way that authors of experimental papers are making their published experimental data available to others in electronic form, I have always thought it should be beneficial for all concerned if the web-accessible versions of computational papers contained the code of the models. It should then be possible for readers to generate their own results with the code provided, to test out and extend the results in the paper. This is one feature that I had always wanted Network to have but was never able to achieve.

In conclusion

I would like to thank all those people who, during my time as Editor-in-Chief, have helped or advised me in all the various tasks and have contributed to the pleasure that I have enjoyed as Editor. Thank you to the staff at IOPP and, more recently and more briefly, at Taylor & Francis. I owe many thanks to the members of the Board who have given me much good advice, carried out urgent last-minute refereeing, and have helped to develop the Journal in a variety of ways.

I have two special votes of thanks to make. During my entire time as Editor, Fiona Jamieson has worked with me on the Journal, initially as Managing Editor when the Journal was run from Edinburgh. Fiona set a standard for commitment to the Journal and for involving authors, referees, Board Members and publishers that has not been surpassed. Over all this time, I have come to rely on her sound advice, editorial correctness and encouragement. Thank you, Fiona, for all this.

Finally, I would like to thank the authors and the referees for jointly continuing to make Network the high-quality computational neuroscience journal that it is. The papers speak for themselves but the refereeing that goes into the submitted papers tends to go unnoticed. I have always been very impressed by the amount of thought that referees invest in reviewing our papers. I hope that the Journal will continue to benefit from your expertise.

My future

In the post-editorial part of my career, I intend to return to research and to my number one interest – developmental computational neuroscience, particularly the interaction of neural activity-dependent and activity-independent mechanisms for the formation of ordered nerve connections. As you will read in Geoff's Topical Review, this is a topic that, under the impetus of many new experimental findings, offers much scope for computational work and I intend to take up where I left off some time ago!

The future of Network

I have restricted myself to reviewing what has happened rather than to predict what will happen as clearly this is in the hands of the new Editor-in-Chief, Geoff Goodhill. I am very happy that he has agreed to take on this post. Geoff has an impressive pedigree in several different fields of computational neuroscience. As well as being interested in what we communicate, he is also very interested in how we communicate; arguably the two most desirable qualities for an Editor-in-Chief! I do hope that you will support him in this venture and I wish him all good luck for the future.