The Brain-Computer Interface: a firsthand account

March 2, 2012

For people who have lost the ability to communicate as a result of severe or total paralysis, including individuals with brain stem stroke and amyotrophic lateral sclerosis (ALS), a breakthrough in technology may enable them to dramatically improve the quality of their lives.

Over the past 20 years, a team at Wadsworth Center has developed a Brain-Computer Interface (BCI) device that may provide communication functions to individuals who have lost the ability to breathe, talk or even move their eyes. A gel-filled cap, which records brain waves from the scalp and then decodes them, allows the user to communicate by making selections on a computer screen. The device may provide Internet access and enable the user to express basic needs, such as a request for food or medication, or allow the operation of environmental controls.

Staff at Wadsworth and Helen Hayes Hospital, which are both New York State Department of Health facilities, are now transitioning the BCI from the laboratory to the home environment via a pilot study that will place the device in selected users’ homes. Through an Internet link and periodic visits, researchers will monitor and assess the extent and success of the BCI usage and its impact on quality of life.

I have had ALS for fifteen years. I am a quadriplegic, but I can use my facial muscles and move my head a little. I spent my professional career making new technologies user friendly. So when Debra Zeitlin, director of the Center for Rehabilitation Technology at Helen Hayes Hospital, and Theresa Vaughan of the Wadsworth Center, asked me about six years ago to participate in the brain-computer research, it was a natural fit.

To get started, my nurse, Mila, puts the electroencephalogram (EEG) cap on. She measures the distance from the cap to the bridge of my nose to ensure the electrodes are at the right locations. Then she is ready to insert the conductive gel. This is sometimes a challenging task because I have unruly, thick, curly red hair. The electrodes in the cap pick up the electrical activity in my brain from the surface of my scalp with the aid of the gel.

Mila quickly goes through the diagnostics until she gets to the impedance check. Impedance is a measure of resistance to the current. To get a good signal from the electrodes, the impedance must be below 20 ohms. We typically have some of the eight electrodes with impedance that is unacceptably high. She parts my hair with an orange stick and inserts more gel. When the impedance of each electrode is in the acceptable range, we are ready to begin.

First, we do a calibration, which consists of a string of letters I am to copy. My job is to count each time the desired letter flashes in the matrix of letters, numbers and other symbols. The researchers are looking for a component of my brain waves called p300. It has been called the “aha response.” Since the letters I am to select are known, the researchers can use the calibration to determine problems in the BCI and also adjust certain parameters unique to me.

Next, Mila starts the Main Menu and I can operate the BCI on my own. I have a choice of email, Wordpad and the RSS Reader, which has the RSS feeds from the New York Times, BBC, Reuters, ALS Association and many others. Today I will use my favorite application: email. I often use email to send my suggestions about the features and the user interface of the BCI. The email application has an inbox from which I can select an email to reply to, and also the capability to compose a new email. As the symbols flash, I count each time “New” flashes to compose a new email. My address book consists of names I can select by using the Alt symbol followed by the number corresponding to the name I want. I move from the “To” field to the “Cc” field to the “Subject” field by selecting Tab.

In the subject field as well as the body of the email, a box appears after I type a letter with suggestions for the word I am typing. For example, if I select “I”, the choices “In, I, Is, It, In the” appear in a vertical, numbered list. This is called word prediction. The next word is also predicted. If I select “Is”, “your, the, a, that, not” will be displayed in the word prediction box. I select the word I want by selecting Alt followed by the number of the word. This is a good example of combining two technologies. The BCI, without word prediction, would be much slower. If my word is not displayed in the word prediction of box, I keep selecting letters from the matrix. If I select the wrong letter, I use Backspace to delete it. Sometimes my attention wanders while counting the flashes, but an error snaps me back. I finish my email and select Send twice to send my email. I have to select Send twice to avoid the inadvertent sending of the email.

-Catherine G. Wolf