While I was gone...

Here are a few stories worth noting from the last few weeks along with some random ramblings.

Now you see it, now you don't (remember)!
First, I'm watching a really interesting piece from 60 minutes on how flawed eyewitness testimony can be. What's most interesting is the way that psychologists can alter the report of an eyewitness in a predictable manner. To sum up one technique, present two faces that are similar to the perpetrator, present the previously chosen face and another that is very different, present the reaffirmed face and the perpetrator. These little psychological tricks provide important clues to the inner workings of memory and an important bridge between experimental and applied psychology. Whether they will be equally useful for neuroscience, we'll have to wait and see. Think of these types of stories as similar to the barrage of visual illusions you would see in an intro to neuroscience course (like these). Part 1 sets up the case, Part 2 is where it gets interesting (those pressed for time can skip to part 2).

Eat it ASIMO!
Toyota has been trumpeting the creation of a noninvasive BCI (EEG) for controlling a wheelchair. Basically, it sounds like a 3 state decoder (forward, left, right), with an emergency stop signal (the user puffs out their cheek). Usually this isn't too exciting (like that Honda media circus the previous month), but the 125ms lag is actually pretty good for EEG. Worth looking into more for those even slightly interested in EEG based systems.

Mendeley-he-hoo!
v0.9 (and now v0.9.1 which fixes a problem with PDF linking), is now available. PDF reader with annotation tools, more control over syncing collections, and a bunch of other features. Go get em, tiger!

Honda: Our in-house research can't read

Apparently, Honda has decided that they are the first to use a BCI (NIRS and EEG combo in this case) to control a robot (video here-and, oh, is it worth it). Not the first to use this technique, but the first to do it, ever. Okay, let's pretend that the Fetz 1969 experiments drove a 'device' instead of a robot (I'm not even sure how far back a device driven by surface electrode input might have been), but this is massively misleading. Sure this is just a stupid company statement, but Google "first robot controlled by a BCI" in two days and this crap will be the #1 result. *facepalm* I'm sure I can't really have an impact on such a large corporation's ridiculous press statement, so instead enjoy this video of ASIMO falling down stairs:



Thanks Katie for the heads up.

EEG consumer BCIs spanked by the Schwartzinator

Forbes has a fun little story on the state of consumer BCIs that hits on the major consumer players, but also throws real BCIs researchers John Wolpaw and Andy Schwartz into the fray. Each echos the officially sanctioned position of their respective scalp electrode versus implanted microwire/fancier probe electrode camp.

I prefer to call this the Head-Computer Interface versus Brain-Computer Interface debate. I try to be even keeled on the subject. EEG is obviously a worthwhile endeavor because there's none of that messy surgery involved, and thus appeals to a very large audience (at least in the near term). Hey, if you get a signal that can be controlled by the user reliably and with an acceptable level of precision, you have something useful. BUT, EEG is a degraded, garbled, sloppy signal prone to nearly limitless interference sources. Anything that can bork an implanted BCI can bork an EEG-based system, but EEG throws the doors open to the world because you are essentially wearing an antenna (or several).

One major problem with EEG is the fallacy that somehow destroyed information can be recovered by some form of fancy filtering. This is simply not so. Think of the electrode as a point of convergence for all electromagnetic signals of a measurable intensity. Even after narrowing frequency bands and implementing funky probabilistic decoders, any squiggle can be the convergence of several squiggles of indeterminable sources. In other words, at a specific time, a signal of amplitude +5 can be two signals of +4 and +1, -2 and +7, or 5 signals of +1-4-2+3+7. And don't even start with harmonics.

The second problem is population size. I know we like to think that motor cortex responds to only movement. It makes life easier. What life? Life in La-La-Land. MI responds to visual stimuli, movement preparation, auditory stimuli, imagined movements, movement related words (heard, internally rehearsed or spoken), reward, attention, cutaneous and proprioceptive feedback, and a bunch of other factors I'm not even mentioning. Until the impact of these influences is understood and quantified, there will not be any BCI that translates the neural activity for "move my arm to point x,y,z". EEG will never have the fidelity to isolate the differences feedback has at the single neuron level, so the nature of the recorded signal will never allow the 'direct' mapping of neural activity to output. That is, the activity that once gave rise to movement can never be harnessed with an acceptable degree of control to recreate the movement. Yes, I am using the word never. Never. There, I said it again.

Like I said, this isn't an attack on EEG, just reality. Different uses for different technologies. I can drive a car on a road. That doesn't mean I can drive a road, or that a road has no use.

BCI for the kids

Yup, it's finally happened. The Consumer Electronics Show (CES) is going on in Las Vegas, NV and I'll post a research/cool gadget wrap up later. But, I thought it was worth noting that there are not one, but TWO BCI devices being marketed towards children. Both use EEG, so I'm sure they won't offer a frustrating and imprecise experience (note the sarcasm).

The most notable is the Mattel MindFlex. The user wears a headband with what looks like electrodes over the temples (+ ground to the earlobes), and navigates a ball around an obstacle course. The folks over at Switched (part of the Engadget universe) tried it out and found it almost completely unworkable. More coverage here, here, and a video here.

The second is marketed as The Force Trainer, a Star Wars toy, and all it does is train you to raise and lower a ball in a tube. Wheee! Ball goes up! Wheee! Ball goes down! What will your child will do after he or she 'masters' this game or decides that it's boring? I don't know, but whatever it is, I'd bet you could find out within 5 minutes of giving them this. These Gizmodo reader comments pretty much sum it up:
simi510: does it work or is a gimmick? IMO both. It works as a gimmick.
hofmann: My kids would honestly play with that for 5 minutes and then beat each other with it, and they actually like Star Wars.

What IS interesting about these toys is that there will be a generation of kids that have this idea of using neural signals to control external objects. If you read this blog, you probably grasp the basic concepts behind BCI systems pretty 'natively', but ask your average Joe on the street, and chances are you would have to explain the most basic ideas repeatedly before he/she could wrap their head around it.

The second interesting bit to come out of this is that no one knows the long term impact of 'training' yourself to repeatedly activate particular rhythms. I'm not being alarmist - oh noz! Johnny is dead from over training his brain! - but if you had a kid play with one of these for days, weeks, months, what changes might occur? We know that repeatedly engaging in 'mind altering experiences' (meditation, drug use, chronic stress, etc) can alter behavior significantly. What happens when a kid eats, sleeps and breathes The Force Trainer is unknown. Would it make them more docile and calm, or unable to focus on critical thinking? I guess we'll find out!

Rockin tidbits

First, as I type, I'm watching the premier episode of Prototype This, and I gotta say I really like it so far. Check it out on Discovery Channel. As a bonus, the first episode features the Emotiv EEG BCI system. The task is to design a car that is driven by biofeedback (they started with the idea of shutting off a car in response to road rage, but it seems to be drifting towards the more fun idea of controlling the car's navigation). They also visit BioPac - which makes a number of biofeedback-related systems and sensors. Perhaps they should get one of these to add to the fun (sample video below).

The most important thing about this episode is that it highlights a concept that everyday people don't think about much, namely that any controllable signal can be translated into a computer input, and thus tied to anything a computer can control. This is a great concept to drive home for BCI enthusiasts and general hacker types. It also hits the real life issues introduced by using niche/'vertical market' devices. We know those problems all too well (and by we, I mean JDS, or as we call him "The BrainGate").



And a plane, which is the awesome-er-est:



Next, just a reminder that there are two available sites for news stories which are constantly updated .

For neuro/bci items:
http://www.google.com/reader/shared/user/12240293210094451962/state/com.google/starred

For tech items:
http://www.google.com/reader/shared/12240293210094451962

Consciousness obliterated

Birbaumer and Kubler have a paper in Clinical Neurophysiology on the idea that completely locked-in patients essentially slip off into unconsciousness once they become completely locked-in.

Objective: To investigate the relationship between physical impairment and brain–computer interface (BCI) performance.
Method: We present a meta-analysis of 29 patients with amyotrophic lateral sclerosis and six patients with other severe neurological diseases in different stages of physical impairment who were trained with a BCI. In most cases voluntary regulation of slow cortical potentials has been used as input signal for BCI-control. More recently sensorimotor rhythms and the P300 event-related brain potential were recorded.
Results: A strong correlation has been found between physical impairment and BCI performance, indicating that performance worsens as impairment increases. Seven patients were in the complete locked-in state (CLIS) with no communication possible. After removal of these patients from the analysis, the relationship between physical impairment and BCI performance disappeared. The lack of a relation between physical impairment and BCI performance was confirmed when adding BCI data of patients from other BCI research groups.
Conclusions: Basic communication (yes/no) was not restored in any of the CLIS patients with a BCI. Whether locked-in patients can transfer learned brain control to the CLIS remains an open empirical question.
Significance: Voluntary brain regulation for communication is possible in all stages of paralysis except the CLIS.

They examined 29 ALS patients (+ 6 others) who had been trained on a P300 based BCI system prior to the onset of completely locked-in syndrome. Once CLIS set in, they were no longer able to communicate using the system. Literature related to two other non-invasive techniques were used in addition to the P300 based system - low cortical potentials and sensory-motor rhythms. While performance across the board was "usable" (66% success is hardly usable, but I digress), once completely locked-in, patients' performance dropped to chance (except for one individual).

I have a better hypothesis. Use of an EEG based system becomes impossible once completely locked in. EEG relies on large scale, detectable changes in millions of neurons. Locked-in syndrome likely results in large scale changes to cortical topography, due to the lack of feedback - both reinforcing and resultant - resulting in slow wave potential differences unique to this, what might be called 'normalized', state. Nevertheless, an interesting and somewhat calming idea - no pain or negativity, you just drift off into nothingness.

The logical follow-ups are to take this up and down a level. As anti-fMRI as I am (not too strongly, but I think the studies waaay overstate their significance usually), throw them in the magnet and condition different modalities on different answers. Given a multiple choice question, choice A = imagine seeing fireworks, choice B = imagine hearing a train whistle, choice C = imagine throwing a baseball, etc. Just in testing the ability to before these imagined tasks should be enough to assess their ability to respond to a command. Then to provide some very rudimentary, slow communication, use the above association technique. And by down a level, I of course mean single unit recording.

A final thought. Since language is so tightly tied to the motor system, it is possible that communication becomes obfuscated by some severe, hybrid aphasia. If that is the case, the question becomes whether consciousness relies on language. And if that is the case, how does that change the way we see animals, robots and babies (or baby animal robots!). Just something to think about the next time you hippie liberal academics are smoking up at the drum circle. :P

Alright, more substance

A few quick links on actual brain, robotics, AI, BCI type stuff... (I know, how novel.)

Neat-o-rama has a post on brain controlled cell phones. Okay, it's more like an application from Neurosky that happens to run on a cell phone, but to the media it's all the same. Only problem is, and funny that the story should break today, there appears to be a part of the brain that shuts off when you sent a text message. Yeah, I'm buying it. Let's look at the study... awww it's some test center without any real review process. Is this some type of complete hoax? They say that the "ratio communis" shuts off. WTF is the ratio communis? AFAIK, there's no such area of the brain. It translates from Latin to the "system common", or "commonsense". Feh. When they have to make up brain regions to explain poorly controlled results, I tend to not pay much attention.

DIY biohacking. Nothing like gene therapy from a back alley geneticist!

And ascribing human intellect to the performer of an observed strategy over at Mind Hacks. Nice experiment, worth the read.

Damn you Hack A Day!

Looks like Hack A Day has a link to a complete teardown of the OCZ NIA. I'm not much of an electronics guy, but it looks like they almost have all the components pegged. Each little bit is labeled and it looks like there is a programming head (that mystery connector I noted), though no word on whether anyone has successfully written to the PIC.

Interesting to note, that the device is USB 1.1 only (well USB 2 'full speed'), meaning 12Mbps max transferred. Sampling rate is 3.9kHz, 24-bit. Like I mentioned, there is an amp that measures the difference in voltage between the outer electrodes, presumably for eye movement detection. Lots of other neat stuff, so have a look.

Another interesting NIA thing I found was a guy who has been 'training' with this and does a decent job of moving forward and shooting, though not much else. He, too appears to have grounding issues.

OCZ NIA gets nekkid

Some pics to whet your appetite (much better pics to follow)...

Meh on the construction. Look ma! No metal. Good thing they're not trying to conduct electri... oh wait...
The amp (sans warranty!)


3 Prongs for the headband piece. Probably the center channel grounded via USB and the side channels recorded differentially.

Note the mystery connector.

Neural Impulse Actuator shipping

OCZ's Neural Impulse Actuator, is supposedly shipping. This is the headband of dry electrodes that's being marketed for gamers, with claims that it uses "brain signals" to control your standard PC games. I'm on the waiting list, so we'll see...

Here's a link of the first review of it (directly to the video/testimonial page).

Free wheelin'

Here are the articles related to wheelchairs that have been clogging the tubes...



First, the laser guided docking wheelchair.

Next, the NOA wheelchair. Discussed more here.


The Pegasus wheelchair concept can extend to a near standing position, which I have been told is not comfortable for people

A semi-old New Scientist article on EEG controlled wheelchairs. (Medgadget has a video.)

And lastly, extreme wheelchair sports (I feel like I posted this already, but it was still in my reader.)

Another consumer BCI

The OCZ Neural Impulse Actuator, made by a mostly gaming RAM and peripheral company, is nearing completion and is set to become the first official consumer BCI. This is no Emotiv system, with only a few measly dry electrodes in a headband, but who freakin' knows? I have already placed a review request using my press creds, but will post the info here as well. The below German video is the best, I think. Watch the user's forehead twitch to control the system. There are a few others on YouTube.



The breakdown of signals will probably be something like 80% muscle EMG, 17% eye EOG, and 3% neural EEG. By that I mean the usefulness, since EMG will complete wash out any EEG and probably all EOG, depending on many factors. What will be interesting is trying to actually push the EEG portion. If/when I get ahold of one of these (I'll buy one if I have to), you can bet I'll be trying to hack the crap out of it (eventually. in the short term I'd like to see the waveforms, if the bundled software doesn't show them to you). The post here will be meant to compliment the Digital Trends post, which will focus on usability, with a more critical look at the science behind it.

Like always, there's that divide I talk about between what is really using neural signals to perform a useful action, and what is using some mesh of signals to perform a similar action in a more convenient package. The scientist side of me says, "Bah! What trash!" But the gadget geek side says, "Awesome! I want one!" Unfortunately, the gadget geek side always wins in issues concerning purchases.

Via Engadget
Official OCZ press release

Site note

Hey kids,

Sorry. Prospective student interview weekends (last week and this week), so posts will be very sparse until that's over.

And jsut so you know, yes, I have requested a review unit from Emotiv. There will be a review here, and at Digital Trends when I get it, and I'll post a little unboxing action as well. As much as I'm expecting to dump on the 'brain' component, I'm pretty excited about playing with a the first commercially available BCI unit aimed at consumers, and I think that if you like the topic of this blog, it would make a nice addition (start?) to your collection. I think it will be an 'important' product. Also, Emotiv seems to have really packed on the non-brain features,so even if the neural/EEG control is flakey, it will have a number of other. We'll see.

In the meantime, I will be destroying the dreams of prospective students. Bwahahaha! Okay, that's just mean.

BCI news items

First, pointed out by the lab, Nicolelis is in the news again. Monkey brains in North Carolina controlling walking robots in Japan. Nice. Again, dissociation of neural activity and decoder performed (monkey stopped walking, but could keep the robot walking while remaining still).
No mention of a paper with some objective measures, but a pretty decent write-up available on the NYTimes site. There's even a nice schematic and video.

Interview by PhysOrg with EEG mogul Fatourechi, related to a paper published in JNE. Take a gander here.

The Neurosky is falling!

Into fun-filled children's toys, that is! Everyone's favorite EMG, I mean no, wait TOTALLY EEG from one dry electrode because they have either signed a pact with Satan or they aren't using brain signals, has been in talks to integrate Neurosky into several toys designed by Sega Toys.

Now, I love me some BCI in the consumer marketplace, but this is starting to get a little out of hand. It was funny when Emotiv showed off the Cap of Infinite Dorkiness(tm), but Neurosky better post some concurrent EMG recordings that show they are actually using EEG before I believe it (I'm talking wet electrode, established 'real' techniques). Until then, I'll say it again, you are only getting covert muscle movements, not EEG. There's a reason EEG studies use 128 electrodes with skin abrasion and conductive gel. Oh wait, no, it is kinda fun to torture the volunteering undergrads, but no, no, it does have a real purpose.

Bah! Anyways, here's the media blitz. Here here here.

B2B v2: Brain To Bot

The EU has been hearing the call of BCI, and upping the funding for non-invasive systems left and right. One project to make the news this last week is the Brain2Robot initiative at the Frauenhofer Institute and Charite hospital in Berlin. Not much too new to report here, just a new funding source for another collaboration, but worth noting, of course. The new arm to come out of the collaboration should be shown off... oh... today in Dusseldorf.

EEG

For those of you lacking in the basics of EEG, there's a nice summary over at Brain in a Vat. EEG is the darling technology (along with EOG and EMG) of non-invasive BCIs and those trendy video game controller junkers. Of course, you're getting the activity of millions of cells, so you get what might be called a computational "schmear" of intention. Check it out! (This is actually their "Method of the Month". And I thought Super Paper Friday was a dorky name.)