I am just finishing a book review for the Lancet Neurology (a British medical journal) on the book, Who Needs Emotions? The Brain Meets the Robot (Oxford University Press, 2005). Since they own the copyright I can't published the review until it appears in print, but here is a glimpse of what I thought the book. Let me know if you have read it and if you agree.

The editors (Jean-Marc Fellous of Duke University and Michael A. Arbib of USC) assembled leading researchers in both fields to bridge the gap between the latest findings in the neurobiology of emotions and state of the art in computer science. Rather than building on the hype surrounding “thinking machines” the book provides a superb scientific analysis of the current state of emotions research in animals, humans and man-made systems. AI researchers are interested leveraging emotions to build systems that can perform unanticipated tasks in unpredictable environments.

The book is divided into four parts, opening with an entertaining conversation between two fictional characters Edison, a theoretical neurobiologist, and Russell, an established roboticist, discussing the important role definitions play in understanding and analyzing emotions. The second part contains several chapters that analyzes the neural mechanisms of emotions in both animals and humans. This is followed by a general discussion of computational architectures of emotions is explored in detail. You'll have to read the book or the review to read the last part, "Beware of the Passionate Robot".

While technical in parts, this book is an important contribution to the emerging field of emotional neurotechnology. It is a stimulating book that is well edited and researched. I highly recommend Who Needs Emotions? for researchers and graduate students across neuroscience and computer science.