Rewards

Rewards in neurofeedback are the alerts the brain receives to let it know that it has done something worth repeating. Rewards are set in the computer’s protocol and usually come in the form of tones, or a video playing, or points accumulating in a game.

When I trained with Joel Lubar in the early 1990s, he used 30-40% rewards, set the threshold once in the first session and left it in the same place for up to 30 sessions. Got excellent results with many of our clients.

When I trained with the Othmers (founders of EEG Spectrum) in the mid-90s, they used 70-80% rewards. I thought that was crazy, but I gave it a try. Lo and behold, some of the clients who had not done well with the lower reward rate did much better with the higher; others, who had done quite well with the Lubar approach fell off rapidly with the Othmer approach.

So, there are two rules I would suggest:

The more under-activated a person is, the more slow activity dominates, the more tendency they have to drift away, the harder I set the targets–and the shorter I make the training segment (or condition). Sometimes as low as one minute, then a break. The more over-activated the person is, the more fast activity they show, the higher the reward rate and the longer the training condition. For highly anxious or compulsive clients, I use continuous feedback combined with a very low reward level in a contiguous feedback (a chime that rings 10%), so there is never a fail condition, only a pass and extra-pass.

The trainer sets the expectations. In my first session with a client, I usually set low percents or reward–maybe 40%–but in the first minute or so of the first condition, I look at him/her and say, “this isn’t the first time you’ve played this, is it?” Of course, they tell me yes, this is their first time. I say, “Come on…NOBODY does THIS well the first time!” Sometimes with a kid I’ll say, “Do you have your dad/mom’s brain in there?” Peek in their ear (while they’re playing) and say, “Nope, that’s your brain!” After a little of that, everyone knows he or she is doing really well–even at 40% reward rates. On the other hand, if you have promised the client that this will be entertainment (the cultural promise of the last generation or so: you have a right to be entertained at all times in all things: your news will be entertaining, your job will be entertaining, the teacher has to make schoolwork “fun”, etc.), then you’d better reward them almost constantly, show them movies, have fancier and fancier games.

I like to start with all thresholds in Auto mode, the percents and training segments set according to the type of client, and then switch the inhibits to Manual and leave the rewards in Auto. My concern about leaving everything in auto is that you are training the software to follow the brain, rather than vice-versa. By setting inhibits to manual mode after about 30 seconds, you establish a baseline: This is where your brain was (in the target frequency) when the session started. [BT2 designs automatically set the target after the baseline period.] If you can improve, you’ll score more points and get more feedback. If someone starts to lose it, and they can’t keep the activity down, then I’d say it’s time to take a break, or even stop the session and shift to a different site or protocol.

In my opinion, the most important thing the trainer does in the first session or two is to guide the client to the amazing recognition: “Oh my gosh, I’m actually changing what happens on the screen by how I use my brain!” That’s something Nintendo and X-Box can’t do. I too like to use video or DVD or game rewards, but I tend to mix them into a session with what we used to call “screen control” conditions. 3 minutes of trying to keep the bar below the threshold on a bar-graph display, or 2-3 runs of PacMan to see if the client can beat his time–always asking, how did you do that? What did you try that time?