Mindflex - Using "brainwaves" to move a foam ball?? I don't think so.

Miscellaneous and Personal Stuff I Must Share
1

Mindflex (Wikipedia article) (Mattel promo)

Anyone played around with this? Do you really think it uses “brainwaves” to move the ball?

I don’t think so.

I’d never heard of this game before, but someone had one this last weekend, so I was curious. At first it is very uncanny. The ball levitates and moves, and , yes, you do seem to, sort of, control it with your mind. Well, the ball levitates by being blown by a little fan, and the fan, in a little holder, moves by a motor on a track. But your mind does seem to control the fan and motor.

All this is supposedly done by your brainwaves being picked up by some funky head gear you strap on, comprised of a clip on your earlobe, a forehead sensor, and two little black boxes pressed against either side of your head.

Naturally I was skeptical, so I did some 'spermentin. Turned out I could levitate and move the ball just as amazingly, more so even, by clipping the earlobe thingy to my finger and holding the forehead sensor in my other hand, while the two little black boxes made no contact with my body, let alone my head. In fact, no part of the system was closer to my head than arm’s length.

My current theory is that the forehead sensor is basically a thermometer and the earlobe clip is a simple electrode measuring galvanic skin response. The black boxes are wireless transmitters transmitting the readings to the console, which uses that crude data to vary the fan and track motor.

Nothing to do with “brainwaves”, but still a sort of simple biofeedback system that is kind of neat in itself.

What surprises me is that all the tekkie web sites that review this product, including a video of an “examination” done by three so-called “neuroscientists”, seem to swallow hook, line and sinker that this really is operating with some kind of EEG technology. At the risk of hubris, it took me less than 5 minutes to debunk that.

2

Moved Cafe Society --> MPSIMS.

3

It is perfectly possible to build a device like this that reads and reacts to brain waves although this one may not be that sophisticated. Check out this hacking teardown of the device. It does have a chip in it that provides some EEG capability (the NeuroSky TGAT1-L64 D498Q-010 0924). It is also more complicated in general than it would be if it just depended on random fluctuations.

My take on it is that the Mindflex does have some EEG capability and is reading your brain waves to control the movement, it just isn’t that sophisticated and can pick up false readings from other things because it is a toy rather than a lab instrument.

4

I have personally hacked one one of the headsets and used the data output of the EEG for another circuit. It’s pretty touchy and interference-prone, but as best I can tell, it does measure actual brainwaves.

5

Looked at the teardown. Thanks. But I still have doubts. As my little experiment suggested, there doesn’t seem to be any sensors in the headset. The only sensors seem to be the little on?e on the forehead and the earlobe clip. The teardown didn’t get into those. I’m by no means an expert on EEGs, but I just don’t see how those dinky things could be EEG sensors.

The NeuroSky chip is intriguing. But couldn’t that chip just be processing galvanic skin response input rather than “brainwaves”? Both are just small electrical currents. The processing task should be the same with either input.

Balance, are you confident your hack was really picking up brainwaves? From your brief description, it sounds like it was just redeploying the output. Does/did redeploying the output help elucidate the source of the input? Couldn’t the input still have been galvanic skin response rather than EEG readings of neural activity in the brain?

Scouts honor I am not just being a naysayer. I am skeptical, but genuinely interested. Thanks.

6

I am not familiar with that one specifically - but there is no reason to believe it doesn’t use brain waves. I have two other gadgets that read brain waves and I have no doubt they work (ZEO alarm and Jedi Mind Force). Just cause you can trick it with other input doesn’t mean it isn’t reading brain waves.

When designing a device to read brain waves - you are using some sort of algorithm that processes a signal (same with GSR) - the idea is make a classification prediction based on whatever training data is presented. It is not to prevent any other method from triggering the sensor. In fact doing so will almost always degrade the accuracy of the device.

I often see people claiming a device is inaccurate cause it can be tricked (pedometers, sleep monitors, polygraphs, whatever). This isn’t how even fancy devices work. Unless it is designed to prevent fraud or something - designing a device to reject false positives (that can easily be avoided) is counterproductive.

That being said - I wouldn’t expect it to be that great of a device. The Jedi mind force was meh - only so so, but it wasn’t expensive either.

Also - the forehead isn’t the best location for a GSR device.

7

No, I took the data on several different types of brainwaves from the serial output to perform some calculations, which meant looking at the actual data itself. I saw the variations in the measurements that I would expect as my testers engaged in various forms of mental activity–meditating, reading, performing mental calculations, and even sleeping. (I was building a gadget that would only “unlock” if my players could reach a state of sufficient calm while a fight was going on around them.)

I’m not an EEG expert, and my impression is that the measurements are fairly crude, but as best I can tell, it is at least trying to take legitimate measurements.

For what it’s worth, here’s a demo of a video game using the same basic tech, although it’s with a better quality headset directly from NeuroSky: Throw Trucks With Your Mind - Developer Demo - YouTube

8

There is a “game” at the Montreal Science Centre that sounds similar - You and a competitor wear a headset that measures alpha and theta brainwaves, and the idea is to push a ball towards your competitor.

9

Consider me schooled by all the above. Thanks. I will now transition from being wholly skeptical to intrigued and puzzled.

The thing I don’t understand, and one of the main reasons for my doubts, was the sensor situation. In the Mindflex Duel, the only things making direct contact with the skin are the earlobe clip and the single forehead sensor. It doesn’t seem possible to pick up brainwaves from the earlobe (or can you?). That only leaves the forehead sensor, and for an EEG to work, doesn’t there have to be at least two points of contact, the voltage difference between them being what is measured? I suppose there may be two closely positioned electrodes in the forehead sensor. But then that sensor is covered by the elastic fabric that makes up the headstrap. It just seems that that material would cause too much resistance to measure what are pretty small currents. If it really is measuring current, why didn’t they make the forehead contact points metal? As is, the only metal contacts are, again, on the earlobe clip.

You all seem know a thing or two about these kind of gadgets. If you wouldn’t mind, could you please explain how you think it works, sensor-wise?

Thanks.

PS. Balance, what are you doing, unning a Ninja school?! Intrigued all the more…

10

The ear-clip serves as an electrical ground for the circuit and as a reference point for signal analysis. It’s subject to the same environmental noise as the forehead sensor, but is not in a position to pick up significant neural activity.

If you look at the forehead sensor, you’ll notice that the fabric covering it is different from the surrounding material. It’s actually conductive, and the fact that it’s soft and flexible makes it easy to get it smoothly seated against the skin and relatively comfortable. It picks up minute variations in voltage, amplifies them, and feeds the result into NeuroSky’s fancy proprietary signal processing algorithm. The signal processor subtracts any signal from the earpiece (which would be environmental noise) and certain signals that NeuroSky has identified as blink signals and other muscle-related noise. Then it feeds what’s left into a set of band-pass filters that separate it into different categories of brain waves based on frequency–delta, theta, and high- and low-alpha, beta, and gamma waves.

The serial output has channels for each of those types of brain wave, plus “attention” and “meditation” channels (which are based on some analysis of the various wave types) and a signal quality channel.

There’s a study (PDF) comparing NeuroSky results with a research-grade EEG system. Their take seems to be essentially that the headset doesn’t provide the greatest quality data, but it really is reading brainwaves, and it’s easy to use.

There was this artifact that was angering some spirits…

11

Peyton Manning moves a ball with his mind. I see him do it every Sunday.

Now if Mattel has invented a method for moving objects with your mind that doesn’t involve a physical interface, I’ll be impressed.