Lots of recent neuronews out of Boston so we’ve put together a Boston Neuroscience Twitter list, including local players, with a few national Tweeps of interest.
Summary: MIT’s Picower Institute for Learning and Memory prepares for its upcoming conference on, “New Insights on Early Life Stress and Mental Health” and two neuroscience imaging papers of note came out of Boston this week.
- One paper used the much invoked map metaphor to talk about a study out that looked at, “relationships of adjacency and crossing between cerebral fiber pathways.” Using technology developed at Mass General Hospital called, “diffusion magnetic resonance imaging,” the researchers were able to identify a grid-like pattern. (Link here to the Harvard Psych Neuroimaging lab.) The paper got a lot of attention, but The Boston Globe described it as, “not surprising.” More here from Harvard:
The researchers imaged cubes of the brain a couple of millimeters across in living humans and half a millimeter across in dead animals from four other primate species. They used a technique called diffusion spectrum magnetic resonance imaging, which traces the path of axons by analyzing the flow of water through the brain.
- MGH radiologists also contributed to another neuroscience imaging paper – like the first, published in Science – this one compared their imaging results to, “a human brain atlas based solely on genetically informative data….” Using a data-driven, fuzzy clustering technique with magnetic resonance imaging data from 406 twins, they parceled cortical surface area into genetic subdivisions, creating a human brain atlas based solely on genetic information.
- Now comes a paper from out west: UMass-Amherst: The Computational Power of Interactive Recurrent Neural Networks
Jérémie Cabessa of the Umass Computer Science Department, writes in the journal Neural Computation. A Umass press release quotes her saying that her a computing system is “modeled on the brain.” Here’s the abstract:
In classical computation, rational- and real-weighted recurrent neural networks were shown to be respectively equivalent to and strictly more powerful than the standard Turing machine model. Here, we study the computational power of recurrent neural networks in a more biologically oriented computational framework, capturing the aspects of sequential interactivity and persistence of memory. In this context, we prove that so-called interactive rational- and real-weighted neural networks show the same computational powers as interactive Turing machines and interactive Turing machines with advice, respectively. A mathematical characterization of each of these computational powers is also provided. It follows from these results that interactive real-weighted neural networks can perform uncountably many more translations of information than interactive Turing machines, making them capable of super-Turing capabilities.
For more on neuroscience in Boston we have listed the latest seminars below:
- Tufts The Department of Neuroscience coordinates neuroscience educational and research programs within the Sackler School of Biomedical Sciences and the Tufts University School of Medicine
- The Boston Action Club “presents a forum of interactive meetings on movement neuroscience for researchers in the Boston area. The series of talks is interdisciplinary bringing together presenters and audience from such disparate fields as Biology, Mechanical and Electrical Engineering, Physical Therapy and Rehabilitation Science, Psychology and Cognitive Science, Physics, and Kinesiology. The presentations are held in a highly interactive style, followed by more interactions in social gatherings.”
- The Northeastern Undergraduate Researchers of Neuroscience (NEURONS)
- Picower Institute
- McLean Hospital
On Yahoo; Neurotalks