The blog Skippy Records has an interesting post on the recent progress of robotics with incredible videos of robots demonstrating amazing capabilities. I’ve always felt that the robotics community seemed to be way ahead of the computational neuroscience community in terms of presenting deliverables. I remember attending a talk by Rodney Brooks almost a decade ago that left me astounded. It made me wonder if reverse engineering the brain is much more difficult then simply engineering a new one. Just trying to make stuff work and not worrying about anything else is a great advantage. Computational neuroscience is more of a diffusion driven field where people try out different ideas rather than build on previous work.
The increasing availability of high quality data is also both a blessing and a curse. It is a blessing because it can help constrain and validate models. It can be a curse because theorists may become less bold and creative. I’ve blogged before on my old blog that it may be possible that we already have all the biological mechanisms necessary to understand how the brain works. That is not to say that new experimental results are unimportant. Biological details are critical when it comes to understanding the genetic basis for brain function or curing diseases. However, discovering the main principles for how the brain performs computations may not require knowing all the biological mechanisms. The trend in neuroscience and biology in general is towards more and more high throughput data acquisition. The consequence may be that while data driven systems biology will thrive classical computational neuroscience could actually slow.