Review of the Handbook of Functional MRI Data Analysis"

Available at Amazon.

So, you want to analyze fMRI data?  Here's a good place to start.  

Historically, researchers wanting to learn fMRI techniques have had to apprentice themselves to one or several senior researchers who have mastered the techniques.  This is still arguably the best way to familiarize oneself with everything since learning Linux/Unix and how to navigate several image processing packages can get a bit hairy in the beginning and a guiding hand is very much appreciated.  

That said, nearly everyone I know in this field also gets a textbook.  Huettel's Functional Magnetic Resonance Imaging.  This is *the* book to get if you really want to know the ins and outs of MRI physics, analysis, preprocessing etc.  But it is very dense.  Huettel's book is meant to be approached chronologically with new information building on what you learned in previous chapters making it an excellent textbook for a course on this subject.  It also becomes an invaluable reference for those who know the subject and want to quickly consult something.  

However, for someone who is being introduced to fMRI, or even a seasoned researcher who is simply more interested in the analysis methods than the physics, a new alternative came out last year.  A book by Russell Poldrack, Jeannette Mumford and Thomas Nichols (all well respected researchers using fMRI) focusses exclusively on preprocessing and analysis techniques.  

The book is really quite excellent, and eminantly readable - I read several chapters on my flight to Virginia.  It covers the regular univariate single subject and group analyses, and also discusses multivariate and machine learning approaches - very useful if you like me have previously only used one or two analysis techniques and read reports of others without necessarily knowing the theoretical drive behind them.  Very often the authors will present a list of things you the researcher will need to do in order to acomplish analysis 'x', they also provide links to various tools and more in depth resources should you wish to focus on any technique in more detail.  It's also very easy to select a topic of interest and simply read that chapter to get a good overview, while one can read the book in chronological order, it is not necessary to do so.  For instance I was very intertested to read the chapters about visualizing activation data and multivariate approaches - I read those first and then went back to the beginning

The book is an wonderful overview of current fMRI analysis techniques and I would recommend it to anyone interested in this field, it's also one of the books I intend to keep on my shelf for a long time (or at least until the next edition comes out).