Reverse Engineering Brain Networks: Testing the Brain like One would Test a Neural Network

Typical steps in a neural network modeling study are the definition of a particular cognitive phenomenon, the creation and definition of the network model by specifying the neural architecture, activation functions and the learning rules. One then sets out to test
how well the network model matches up with the phenomenon, and to the extent that it does and is parsimonious, and has neural plausibility, it is a good model.

It is then not difficult to imagine how we can do the same thing with the brain by treating it as a neural model that's already built, and we're just trying to discover its architecture, its activation functions and its learning rules. Thus, we run the brain through simulations, observe the input and resulting output, and hypothesize the parameters that led to the observation.

We can then reverse engineer these parameters into the model (which is what we do anyway), and again, test how good the model is.

Neurotree

This is a cool website that tracks who was taught by who, and through it, you could trace how ideas get formulated and how they evolve...or not!!!

My Neurotree

HBM Conference 2007 Chicago

Went to Human Brain Mapping Conference last week. Stayed at Sheraton Hotel. We ate at several places during that time, so hungry...because always using brain, so all the sugar used up. But it was fun. First night we arrived, there were fireworks and the Blues Fest was on in Grant Park. Anyway, no photos because too busy working. Check out what was cool at the conference [click here].

Here are the eatery reviews:
Happy Chef Chinatown
Always my standard place to eat...we had dim sum there, although usually for dinner, you can do a set meal for very cheap (works out usually to be ~$8), and it comes with free lobster or crab.

Oysy
Japanese sushi place. Not bad...but probably Sai Cafe is better but more expensive. Anyway, its one of those modern sushi places. There's two outlets, one on Grand, and one one Michigan. Average cost ~$20-$30, depending on what you order.

Dao Thai Restaurant
Good food and cheap prices (~$6). Has some pretty authentic Thai dishes, and the restaurant has nice Thai looking decor where you sit on floors with wells in them so that you are the same height as the table...get it?

Gino's East
Took a while to get our pizzas, but that's because they were making it fresh, as any good pizza should be. The pizza is pretty good. Nice and thick. We had a mix of supremo and vege. The sauce was not too sweet or rich, so its good. ~$15. Giordano's near the Water Tower still has my vote.

John Hancock, the Signature Room on 95th Floor
What would a trip to downtown Chicago be without a brunch at the Signature Room? $20. Best view with food combination. Menu changes. Brunch is from 11.30am to 2.30pm. This is the same view you get as when you go to the observatory, for which you must pay ~$15. So go do the math...

Andy's Jazz Pub
In my opinion, this is the heart of Chicago, and the most worthwhile place to hang out for good music. Don't come here to have deep conversation. Come here to listen to the depths of music. Cover $5-$20. Drinks are decent, not too wide ranged, but enough, serves dinner too. Historically, several notable musicians have jammed here. I like the way you can sit right in front of the band, and also the jam nights...informal, naughty, crazy, eclectic, and it all gets into your ears. Trumpets, guitars, drums, piano, bass, the whole shebang.

House of Blues
Significant venue where many musicians have performed including Louis Armstrong. Its a theatre set up inside. Cover is more expensive, concert based. You'll have to check online to get tickets in advance. Serves food. I think its worthwhile to visit once.

Red Fish
Cajun style. Food is pretty good. ~$20. They have a live band. Plays rock.

Some Tapas bar somewhere...

Cool Studies at HBM 2007 Chicago

Day 1
Cultural Neuroscience
Thinking of culture using a top-down versus bottom-up framework. Trey Hedden (MIT) and Angela Gutchess (Harvard University)...notable speaker presentations.

Visual Field Maps, Plasticity, Reading
Discovery of retinotopic visual representation in visual cortical areas other than V1. Apparently, V2, V3, even V4 and MT have some retinotopy. Speaker session by Brian Wandell (Stanford University).

Day 2
Brain Noise
Didn't attend this one, but it seems that people are looking into neural noise as a predictor of subsequent brain activity and behavior. McIntosh was one of the speakers.

Manipulative Neuroscience
Awesome talk by Mitsuo Kawato (ATR, Kyoto). He is the brainchild of DB, humanoid robot that is able to mimic human movements by visual observation, eg drumming, juggling, dancing. The talk covered latest research about controlling robot movements through brain-computer interface as well as visual and tactile feedback.

Perceptual Decision Making
Great talks relevant to the visual discrimination project. Generally, I got ideas about how to proceed with the project in terms of possible analyses, and also the fact that others have done this before. The main question is, how does the brain make perceptual discriminations of visual information? What are the mechanisms and neural correlates? Most notable speaker for me here was Paul Sajda.

Day 3
Dual Brain Systems
Control vs Representation systems in the brain. Typically showing that the control network resides in frontal, parietal regions, and representations in the primary and secondary unimodal areas. Check out www.walterschneider.net.

Repetition and the Brain
Another notable symposium of talks. Kalanit Grill-Spector hosted this one. The topic is self-explanatory, but there were some main novel directions. There is distinction between repetition suppression for immediately repeated stimuli vs stimuli repeated over interspersed trials (Grill-Spector). There is an interesting finding that for interspersed trial repetition of object naming tasks, pre-op patients for removal of lateral anterior temporal poles showed normal repetition suppression of repeated objects was observed in the ventral visual areas. But after operation with temporal poles removed, suppression disappeared even in lower perceptual areas suggesting that suppression has a top-down source in this case (Rik Henson). Another contention was Grill-Spector's testing of the fatigue vs facilitation models of adaptation effects. She found evidence for fatigue rather than facilitation, but note that her design involved immediate repetition.

Day 4
Representation and Processes
Didn't attend all, but most notable for me was John-Dylan Haynes' talk on reading hidden intentions in the human brain. They used classifier algorithms on clusters of voxels in the whole brain to identify brain areas that would reliably discriminate between stimuli. This could be applied to the visual discrimination project.

HBM: Ideas: Visual Discrimination Project: General Questions

1. Is there poorer behavioral perceptual discrimination with age?
2. If so, what are the neural correlates? Is it a perceptual representation problem (ventral visual dedifferentiation)? Or is it a selection/decision-making/control process problem (noise, non-selectivity in frontal cortex)?
3. Is this the same across all types of stimuli (eg faces, patterns, random shapes)?
4. Is there a constant in terms of brain activation pattern across all individuals which is necessary for discrimination?
5. Which parts of the brain are predictive of whether the individual is able to discriminate visual stimuli (classifier algorithms on whole brain)?
6. What leads to individual differences in performance? And if all individuals are equated at some level of performance, do the individual differences disappear?