When I went to my first SfN back in 2007, I was told that I would be the equivalent of a sad, sad child who just wanted a drink of water but instead pulled up a chair to a fire hydrant. And this is true: there is just way, way too much science. It is overwhelming, humbling, inspiring, to see so much work in one place.
And yet…how much of it is truly good work? The last couple of years I have found myself seriously disenchanted. Many of the posters seem slapped together at the last minute, and, yes, even the dreaded “array of 8.5×11″ paper” posters continue to make appearances in this, the post-Kinko’s age in which you can buy a 3-D printer for less than $500. Beyond issues of design is the issue of scientific quality. For those who don’t know, all posters submitted to SfN are accepted. This may horrify people like my brother who work in fields where conference submissions are strictly vetted and require the submission of a real paper, but I assure you, it’s true.
Don’t get me wrong: I am all for the democratization of science in the age of Google, where a search can find you the paper you didn’t know existed in less than a second. But in a conference without peer review, mediocrity reigns on the poster floor. Many students (myself included) submit something every year, regardless of whether they have anything particularly new or interesting to say. Who wouldn’t want to take a few days on the NIH’s dime? And while SfN’s website provides a means to search the posters and find topics of interest, I have found that abstracts and titles for posters almost always overpromise and underdeliver. This is in part because abstracts are due a number of months before the conference, so many people have not even analyzed the data they plan to present. I call this the Field of Dreams strategy: if you submit it, the results will come.
Some, however, did show they were up to the challenge, and many of these found new and interesting ways to spice up their posters. I saw several posters utilizing iPads to great effect. One poster presenter, whose title I sadly did not write down, even went so far as to cut a hole in their poster and slide an iPad behind, so that a time-lapse video could be shown on a loop. Another set up a projector next to their poster to show data in multiple dimensions. These are not just gimmicks: neuroscience data has become increasingly dynamic over the last few years, and flat paper is a truly horrible way to illustrate dynamic datasets. As one of my colleagues pointed out, the cost of printing a poster is such that if you added up the costs of all presenters who use a given poster board during the conference, you could purchase a sweet flat screen TV to put up on each board. Yes, yes, electricity, global warming, etc–but it would be pretty awesome.
While the poster sessions continue to simultaneously under- and overwhelm me, the talks I went to, mostly large lectures, were stellar. Ann Graybiel of MIT reminded us that the cortex ain’t the only belle at the ball–the basal ganglia has fundamental importance in our selections of actions. Dora Angelaki of Baylor showed fascinating data arguing that multisensory integration of information from the visual and vestibular systems makes us better at moving through a complex visuophysical environment. Specifically, she showed that how much our brain “listens” to our vestibular system depends on the reliability of our visual input, and vice versa. The fact that the brain can do these calculations on the fly and adapt to the results is simply amazing.
The best symposium I went to, though, was the most argumentatively titled: “Beyond the Connectome: Why the Wiring Diagram is Not Enough.” There is a bit of a war going on in neuroscience these days with regards to the connectome. Scientists like Olaf Sporns and the other leaders of the Human Connectome Project, which is extremely well funded by the NIH, are using neuroimaging techniques to make a “wiring diagram” of the human brain. Their methods include diffusion tensor imaging, which uses probabilistic methods to determine the directions and connections of neural fibers, and resting state fMRI, where subjects lie still in an fMRI scanner and…rest. These data, the scientists argue, reveal correlations that show which areas are connected to which other areas.
Meanwhile, other scientists, including Partha Mitra of Cold Spring Harbor, are working to build an actual map of actual tissue by using classical and novel neuroanatomical track-tracing methods to map the connections in rodent and, in the future, macaque and human tissue. They argue that direct measurements are essential, and indirect neuroimaging methods should be used to supplement them only when necessary.
Yet the speakers in the Beyond the Connectome symposium do not study brain-wide connectivity: they study local modulation of functional circuitry, and they argued that a static map of the brain is not so helpful in understanding how it works. One scientist pointed out, for example, that not all neuromodulation is local: neuromodulators, for example, can diffuse over an area of tissue, causing changes in firing and network function that would be invisible in all but the most ridiculously detailed of wiring diagrams. Similarly, the physics of neuronal networks can be such that repeated firing can itself change the dynamics of a network without changing the wiring–a type of subtle plasticity that can only be seen via dynamic recordings of activity over time. As modelers have known for a long time, getting a bunch of neurons together suddenly leads to properties far more complicated than those of the neurons themselves. Their point: a connectome is all well and good, but save some money for us.