Brain-computer interface

Don't you think these people have it bad enough without subjecting them to voice recognition sofware? Those automated agents all the banks and telephone companies are using these days drive me up a tree. Much worse than the old touch tone menus. Typical interaction:

Automated teller: "Please say your last name."
Me: "Solberg"
AT: "I'm sorry. I couldn't understand you. Please say your last name again."
Me: "Sooool Beeerg."
AT: "I'm sorry. I still couldn't understand you. Please spell your last name."
Me:"S-O-L-B-E-R-G"
AT: "You entered F-O-O-D-D-R-P. If this is correct, say 'yes'. If this is incorrect, say 'no.'"
Me: "No. Speak to an agent."
AT: "I'm sorry. I couldn't understand what you just said. Please choose from the following options. To access your account information, say 'account information.' To change your pin, say 'change my pin'. Blah blah blah."
Me: "Operator."
AT: "I'm sorry. I couldn't understand what you just said. Please choose from the following options. To access your account information, say 'account information.' To change your pin, say 'change my pin'. Blah blah blah."

If I had the choice between controlling them with a chip in my brain or trying to talk my way through one of those things, you better believe they'd be drilling a hole in my skull right now.

No way voice interfaces are superior. Think of the frustration in shared-space environments from having to listen to your coworkers command their computers to browse the web and check their email.
We have to start with crude brain interfaces if we're going to get to the good ones.

Jenny,

I suspect your frustration with the brain-control interface would be similar. If you had as much training on voice-control as the participant population in this brain-control task, my guess is that you'd be legions better than the brain-control participants.

James,

Point taken about needing to start somewhere, but I still wonder if this is a solution in search of a problem. If you can talk, you don't need brain control. The plan is not to implement this in an office environment, but rather to help disabled people control their computers.

Supposing this was implemented in an office environment, I still wonder if it would ever be more effective than the human body itself. What's wrong with good old-fashioned keyboards? After all, if computer voice recognition is difficult, wouldn't computer thought recognition be even more fraught with the type of problems Jenny describes above?

I have to wonder what kind of training is needed to control the device but that's not the point I want to talk about, It was more about the pertinence of voice vs brain based control.

would it be possible that the voice is only a middleman between our thought and their expression and that some people are able to do much deeper thought than they are able to express them (this might explain the effect usually diagnosed as language development delay) ?

Of course, the example given here is pretty simple or crude but I think there's lot of potentially rewarding development to be made in that field.

Alain

I'm not sure where that video was made, but there's an almost identical project at Georgia State University, where the aim was to get paraplegics to be able to use a computer. So voice wasn't an option.

I still wonder if this is a solution in search of a problem

What about RSI?

What of those who can't speak?

Hello,
It is always great to see people interested in the research field I love. I was the project lead on the BCI research at Georgia State University. The approach that we used, and that most BCI researchers use, was not similar to the approach pursued by John Donoghue and colleagues. His group, along with Phil Kennedy's group, has made great progress getting BCIs to work with patients. However, this creates the implication that surgery is necessary to use a BCI.
Most BCIs rely on noninvasive measures of brainwaves, or EEGs. These can be recorded from sensors placed outside the head. You do not need to cause any pain, shave hair, etc. We tried several different BCI approaches with a variety of patients. The patients were not simply paraplegics (who can talk); these were patients with little or no voluntary muscle control who could not even control breathing. However, BCI technology is advancing to the point that is it more useful than other assistive technologies (such as a sip and puff switch) and hence is becoming helpful to a wider population.
To address other questions: Most BCIs require little or no training. The effects of long term BCI use have not been well explored. The bandwidth of BCIs is nowhere near keyboards; that's not the point. They are used primarily to enable communication for patients with severe disabilities. They will not replace conventional interfaces for conventional users in conventional situations any time soon, and all senior researchers in the field agree.
More information is at the website above.
Brendan Allison PhD