cognitive fun!

Learn your mind. Play it too.
The 3D Dual n-back Speed Run
cognitivefun | 13 years ago Reply Link me
The 3D Dual n-back Speed Run
This is an experimental test that involves many concepts rolled into one. The main purpose is to test learning efficiency during immediate feedback and dynamic difficulty adjustment. The design is based on the idea that a cognitive training task, which is constantly adjusted to be at or slightly above the level of peak performance, would lead to a faster learning effect. This idea is not at all new, and the results should be far from surprising. Nevertheless, this test attempts to blend a demanding working memory task with the appearance of an approachable game. The game setup takes from Damien Clark's Missile Game 3D, which is itself a very interesting spatial processing task involving immediate and dynamic feedback.

As a game, this test brings about several confounds which could make results pretty much incompatible with the other, more controlled setups. I will list a few below and discuss in further detail at a later time.

1. Control scheme. Arrow keys for movement and space for auditory responses. The space is simple enough, but the arrow keys clearly are not the best input method. Four keys, usually controlled by three fingers, are used for eight directions. Four directions require one finger, and four require two. The access time of the up arrow and down arrow are unequal, and dexterity heavily influences response time. It is possible to forego this control scheme completely, and reduce visual response input to a single button (such that the ship moves to the correct edge when pressed), but the arrow keys are used to give the user a greater control over movement, for the sake of providing a more immersive and entertaining experience. Whether this is truly the case, I do not know.

2. Sound scheme. Sounds are sampled from the percussion channel from the Jurgen GM GS midi SoundFont. Originally, sequences of notes were used, but different users would respond differently to "musical stimuli," based on their background and musical training. I am guessing it is very hard to ensure equal novelty or equal storage and retrieval speed for auditory stimuli, so I have simply gone with selecting 8 sounds that are easily discernible from each other. Again, this may not be the case.
However, sounds without obvious phonologies were deliberately used for this test. That is, whereas in the previous dual n-back tasks, the auditory stimuli could be remembered via the phonological loop, users will have a much harder time doing the same with these sounds; they should carry minimal linguistic information.

3. Perspective elements. Lines, circles, and shapes are used in the "tunnel" to enhance the motion parallax and depth perception. These serve absolutely no other purpose than to help create a more immersive environment. They may very well be nothing but distractors, and this point can easily be negotiated in later versions.

4. Timing of stimulus presentation. It is easily confusing that sounds are played the moment the images cross the green crosshair. For those interested, this is essentially the "binding problem" done incorrectly: it is easy to think that the sound is somehow associated with the images, but they are, in fact, two independent stimuli! I have explored other timing schemes, but unfortunately, given the setup, there are only three options: sound before crosshair, sound at crosshair, and sound after crosshair. Since depth is "infinite," it is unclear where the visual stimuli begin and end their journey in the tunnel. Hence, the only definite location seems to be the green crosshair. This point is also negotiable.

5. Visual feedback elements. Both auditory and visual stimuli show feedback when incorrect responses are given. It would seem reasonable to reply to sound responses using sounds, but using auditory feedback would cause a much greater interference than visual. The visual feedback for auditory responses may very well hamper performance on visual responses. To what degree, I don't know. The colors and shapes for the feedback have also been chosen, arbitrarily if you will, to be distinctive, quickly recognizable, but quickly disappear from the scene. I am certain there are better ways to do this.

6. On the choice of 3D. Playing a test in 3D gives a more immersive feeling, which means that perceiving an imminent collision towards the screen might be more stressful than watching a block fall through the floor. This is a guess, but a test done in a "fall down" fashion clearly involves more eye movement and peripheral vision. The other benefit from having a 3D arena is that more information can be fit in the same space, by being far away and smaller. As you fly through the current visual stimuli, you will be able to see the next stimuli coming into view. I don't know if this "increased" information makes the task easier.

So that's a quick list off the top of my head. There are definitely more, but the point is clear. This "test" is less of a test and more of a game, with the possible side effect of drilling your working memory and maybe your dexterity, and the definite side effect of training the domain specific task of playing tunnel games. But always keep in mind that this test is loaded with variables that may make the results largely meaningless. Play it at your own risk -- and for your own enjoyment!
cognitivefun | 13 years ago Reply
? | 6 years ago Reply
? | 6 years ago Reply
w h a t d i d y o u s a y
wildfire | 6 years ago Reply
hi friend!
Natalia8 | 6 years ago Reply
how old are you?
Natalia8 | 6 years ago Reply
? | 6 years ago Reply

Login to save scores

© 2008-2021 (yes!) | about | widgets | blog | cognitive neuroscience for everyone