Until I can figure out how to swap my starred for my shared tags in Google Reader, here is a link to my neuroscience/BCI/robotics related items. Note that I will remove papers from here when Super Paper Friday is posted, but you can easily find them by just searching for the "papers" tag.
Couple quick grabs here.
First, Medgadget mentioned a front end to PubMed that looks interesting called Ologeez (as in plural for the suffix "-ology" - those Stanford kids are clever). They basically apply the hive mind/social bookmarking/Digg/aggregator schemes that are all the rage right now to rate, recommend, and tag papers. I'll be messing with that more in a month or so, but it looks like a very beefed up version of CiteULike, which is kinda "Meh". Anyone have experience with Ologeez?
On the topic of PubMed, the is an iPhone App called PubGet which allows quick, on-the-go searching. It's actually just a mobile version of the PubGet site, but I didn't know PubGet existed and this brought my attention to it. It offers direct display of the linked PDFs, and is basically everything PubMed should be in a much more compact interface, but remember that you will have to be on your campus network to actually view full articles (or connected via VPN). I did a few quick searches and it was surprisingly good. I haven't tested whether it works at Brown (or any institutions not signed up through the site, which appears to total 7), but when I am back on campus I will test it out and report back.
Next, Microsoft's Live Mesh is now open to the public, at least as of right now there are spots, and it looks very very cool IMO. They did an incredibly crappy job of advertising it, but basically it is the same thing as Apple's MobileMe, but with a respectable history of actually working (neener neener neener). It takes the features of FolderShare, which I absolutely love, adds 5GB of online storage, public doc sharing, and remote access (in IE only for remote access) across any number of PCs, Macs, and mobile devices. I have an account, but haven't had time to set it up, so again, I'll report back when I do. Chances are that this will not remain free once released, but I would expect it to be bundled with every MS app purchase, and with so much competition in this particular market, I expect it to be cheap on its own (I'm thinking $40/year, if they don't have some type of tiered service where another $10 gets you 25GB and music streaming, or some such tactic).
Next up, two quick Windows apps. I have been on a hunt for the perfect launcher to help me declutter. If you want an OS X rip off with more functionality than Apple allows, RocketDock looks like the way to go. Low memory footprint, smooth, customizable. BUT, inefficient for my needs. I have really been liking SideSlide, which creates a little work area with different hide-able panes. Oh yeah, and they're both free.
If you deal with constant file operations, like renaming, selective copying, etc then you know that Explorer is limited. Xplorer2 is like Windows Explorer on steroids (free and pay-for versions). Think a million and one keyboard shortcuts, multipaned interface, saved layouts, and you've begun to scratch the surface.
Vista, for some ungodly reason, finds that copying files it tough. Part of this has to do with a bug that neither antivirus makers or Microsoft want to acknowledge, but you can avoid it by using TeraCopy. It basically copies. Yeah, you'd think a modern OS could accomplish that. Also, it will integrate into the shell, so you can just use it instead of the built in copying technique, which apparently requires a delicate combination of black magic, luck and celestial alignment for proper operation. Of course, it does do a good deal more, like batch jobs, schedules, and resumable/selectable jobs, too.
TrueCrypt is now on version 6. Major change? You can now boot into a completely separate, fake account which appears to be the primary account, hiding all evidence of the real partition. Good for when you are planning on leaving or entering good ole xenophobic USA. Get ready for SfN folks!
And lastly, a quick tip. If you have virtual stacks and stacks of audiobooks, but can't get them to show up in iTunes as, well, Audiobooks, the solution is simple. iTunes only recognizes .m4b files as audiobooks. That's the way Steve wants it, and you should want it like that, too. Really? No, not really. I'm being sarcastic. To convert those lovely gems, I like Easy CD-DA Audio Extractor, which converts every type of audio to every other type. I haven't looked around lately, but might, since the copy protection/licensing scheme keeps getting worse and worse with each version, bogging down the whole damn program. But, if you have a preferred program, go ahead and post it. Anyhow, re-encode those novels at 64kbps/mono and save yourself mucho space. You might be tempted to merge the files then, but don't. You'll have to put up with several files for each book due to a long-standing bug that Apple hasn't fully corrected that causes larger books to crash iPods and iTunes. You can use something like ReNamer to rename the files something convenient in the meantime, or TagScanner - my favorite IDv2 tagger - to tag and rename them. TagScanner is nice because it will add every type of tag to avoid situation where a program or device will first try to read IDv1 tags first or doesn't support IDv2 tags
I want to swap my starred and shared items in Google reader (that way it's still one click to mark a story for the tech site and DNI), but I see no possible way to do it. I was able to move all the items in Shared to a temp tag using the privacy settings, but I can't figure out a way to get the starred items to do the same thing or unstar and mark Shared.
Any ideas?
(I also tried sharing the starred items and subscribing to that page, but that didn't work.)
I'm adding a list of the 10 most recent items I have been tagging for the site but haven't had time to post about. I just want to get rid of the things I marked for my general interest, and then I will open the feed to everyone.
In the meantime, enjoy these fun energy drinks from Energyfiend! I'm partial to Stewies's Mind Erase Serum.
... and for an iPhone :D
It is a wonderful little gadget, and one of those few items that make me take note because I know I will be looking back at the time before I got it, wondering how I was able to operate. If you're on the fence, just go get one. I've already hit the impassible Apple iWall - that point where I wish there were advanced options for all the settings. Everything is either too simple or, like jailbreaking, too iffy. Guess I'll have to actually simply use the damn thing.
I know I've been saying, "Updates coming soon!" but I swear they are! Oooo... I woinder if Blogger and Google Reader work on it... ~o.O~
(My background is above)
There, now I feel included. I mentioned it (going to get one, but not going to camp out). It goes on sale in 8 minutes here in the US.
Fun items I just grabbed off Gizmodo... Radiohead video shot with lasers, not cameras, and a Battlestar Galactica toaster. Toasts Cylon images into bread... Virgin Mary next? Oh, wait, someone thought of that! Does it land with the butter side down? If you believe in miracles, then no. But if you're a Mythbuster...
Looks like the interweb deities are smiling on us. As far as I can tell, I can post these without fear of losing my sweet, sweet grad student paycheck to the copyright cops.
I uploaded a new archive, though, which includes a text file that gives attribution to the authors and journal just in case the url gets passed around and people don't get it from here. No password needed.
Here it is:
http://rapidshare.com/files/128403736/Brain_map_desktop_pics_5120x1200.zip.html
Enjoy!
An awesome awesome awesome paper in PLoS Biology that has been getting media attention lately made a valiant attempt to map the brain. Yeah, the whole thing (okay, just cortex). Using Diffuse Spectrum MRI, the researchers constructed an ROI-based network of nodes, between which normalized strength measurements were calculated. No, I'm not sure how they did it exactly (I've only had time to skim the paper), nor could I find what task the people were asked to perform. They may have just asked them to lie there and taken any spontaneous activation, but that would be horribly uncontrolled.
The results point to distinct hubs:
"Based on their aggregated ranking scores across six network measures (Table 1), we identified eight anatomical subregions as members of the structural core. These are the posterior cingulate cortex, the precuneus, the cuneus, the paracentral lobule, the isthmus of the cingulate, the banks of the superior temporal sulcus, and the inferior and superior parietal cortex, all of them in both hemispheres."Check out the paper. (Free to everyone)
Brain Waves pointed out that the lectures from last month's Neural Interfaces Conference in Ohio are on the site now (in streaming form). Neat-o.
For those of you that missed it, and I assume that's 99.9% of you, Coast to Coast AM had Michael Gazzaniga a couple nights ago. I love the show, which covers all sorts of zany topics, and listening to a real scientist trying to field questions about UFOs and angels was pretty funny. The usual cast consist of 'researchers' trying to promote their books on how consciousness can influence shadow people and the Illluminati, or other such goofiness, so the audience is pretty interesting to say the least. Hey, you can find better programming at 1-5am?
There's a story making the rounds about University of Florida researcher creating an adaptive BCI that is getting misinterpreted on every site I've seen it on, including a few science-ish sites/blogs. There are just some real schlock sites out there.
Here's the real scoop. Adaptive BCIs are not new, but the implementation of this one is interesting for a number of reasons, the most notable of which is that it uses agents to evaluate the reward and output, updating each as performance improves. This requires some knowledge of the reward condition, which is a whole ball of wax we won't go into detail about because it means that reward conditions are defined and recognized prior to any attempt to learn. Though reinforcement learning does not require prior knowledge of the reward (you do something, and if you get a reward, that was what you want to do), the practical application of the system require a significant amount of knowledge about the environment.
This would be very useful for tasks such as computer control, where the environment has a limited number of components and those components can be explicitly defined and communicated. I think it is obvious by now that the first generation of truly useful BCIs will require a number of techniques working synergistically to operate in an open environment, and this 'dueling banjos' agent could very likely be one of them.
What this is NOT doing, and anyone that even glanced at the abstract should be able to tell, is alter the implant itself. It does not adjust the electrodes physically, filter the signal differently, stimulate the same or other electrodes. They are talking about a relatively standard AI learning algorithm in an interesting configuration put into action.
Don't get me wrong, it is a very interesting experiment and worth thinking about, but come on folks. At least read the abstract.
The paper.
I would link to the source sites, but they were all rubbish reporting.
Update: Okay, Medgadget did a good job, though they just posted two sentences and linked to the press release.
