Skip to content

A Marr-ed view of connectomes

January 21, 2013

I saw an interesting post by Kresimir Josic on whole-brain simulations, and it included a link to a video of a debate between Sebastian Seung and Anthony Movshon about the importance of connectomes. The video captured my attention because of two mini-debates rolled into the big one. The first was a mini-debate about David Marr (the reason for the pun in the title of this post), and the second was about the idea of consolidating major funding to a task like building a human connectome. For background on the pro-connectome side, there is a TED talk explaining Seung’s passion. I highly recommend it simply for the amazing animations. I’m not sure what to suggest for an accessible coverage of the anti-connectome side except for this article by Ferris Jabr. At the end of this post, I’ll mention my favorite discovery in the debate: a 3-D worm!

The mini-debate that struck me most was when Seung complained that the ghost of David Marr still haunts the halls of MIT (it’s around 56:50 in the video). Seung made a striking suggestion that Marr, if he were alive today, wouldn’t make the same claims. I presume he was referring to Marr’s “three levels” and the separability of an algorithm from the hardware with which it is implemented. Many neuroscientists feel this separation is not helpful in understanding the brain. It seems that a Marr-ed view (pun intended!) of the connectome is a central part of the debate. Marr left us tragically some time ago, but Movshon suggested that Matteo Carandini is still carrying the torch, as explained in his recent paper From circuits to behavior: a bridge too far?. What’s important in the mini-debate over Marr is the relative importance of the connectome in understanding the algorithms behind various brain functions. And that brings us to the second mini-debate I mentioned.

The second mini-debate that interested me concerns whether funding should be directed toward the human connectome at the expense of other pursuits. The article by Ferris Jabr describes a fear among some people of creating a “Manhattan Project” for the human connectome but ending up with little to show for it. I was surprised that Seung did not seem to defend against Movshon’s attacks on the seemingly anti-climactic completion of the C. elegans connectome. Maybe he thought people should just read the article by Ferris Jabr which is very supportive of the outcomes of C. elegans results. What Seung did try to emphasize was that obtaining the human connectome is part of a long-term vision that does not promise immediate rewards. Movshon conceded that he and many others in the “cottage industry” of neuroscience tend to focus mainly on short-term payoff.

I’m happy to say that I did find an immediate reward by following Seung’s advice in the video to read Jabr’s article. My reward was the discovery of a 3-D interactive webpage of a C. elegans connectome. (Instructions: to see the connectome, pull the slider on the left very far down in order to reveal the neurons and connections. Then spin and zoom for a fun exploration!) I think this is awesome and demonstrates how the connectome contains information that can be highly accessible to a very broad range of people. If Henry Markram gives us public access like this, I would be willing to overlook his grandiose and misleading promises about his own connectome project. Note that Markram is perhaps the most extreme connectomist there is – maybe on the order of a David Marr. However, the information he is seeking will have tremendous potential, even if the impact he promises does not come to be.


From → Uncategorized

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: