Neuroracer game promises far transfer in supposedly well controlled study

[Ram]

http://www.theatlantic.com/health/archive/2013/09/how-to-rebuild-an-attention-span/279326/

The conclusion that brain training had no transfereffect, and the publicity it generated around the world, only fueled rising skepticism about outsize claims of those promoting everything from supplements to quick-fix gimmicks in the multimillion dollar enterprise aimed at the burgeoning population of aging adults concerned about fading memory. In an overall review of those claims, the Stanford Center on Longevitypreviously warned that an exploitative hard sell for products from computer training to supplements ranged “from reasonable though untested to blatantly false.” There were few provable links between discrete activities and broader benefits, the review concluded.  In other words, taking your vitamins or doing a variety of brain training exercises was unlikely to help you retrieve the name of your boss’s partner, painfully tripping along the tip of your tongue.

In the current report in Nature, however, 11 researchers in Gazzaley’s laboratory presented persuasive evidence of transfer. They ventured one step further, detailing the underlying neural mechanisms at play.   “… (A)ge-related deficits in neural signatures of cognitive control, as measured with electroencephalography were remediated by multitasking training (ie. enhanced midline frontal theta power and frontal-posterior theta coherence),” they write. Use of the specially-designed game, in other words, markedly improved performance of older adults not merely on the game itself but also led to robust increases in activity in those parts of the brain’s prefrontal cortex associated with greater cognitive control.

In an email several days later, he summed up the significance of this breakthrough: “What we have here is a link between neural plasticity and behavioral plasticity, pointing to a neural basis of transfer effects…Transfer has become the holy grail for training studies, but it is not a magic trick. Our data suggest that there must be a common neural mechanism of cognitive control that underlies working memory, sustained attention and multitasking and we put pressure on it with our game.”

This was a point fleshed out by the lead author of the study, Joaquin Anguera. “Many groups have found transfer to other abilities, other groups have shown neural changes following training, but we are the first to show both – and that they are correlated with each other,” he said. “That’s something no other cognitive training study has previously shown, ever – hey, with video game training!” The next step, he said, would be refining NeuroRacer, fleshing out the role of the prefrontal cortex, and devising interventions that even more effectively reverse the “costs” of distraction for older adults. If they can succeed in these goals, the implications for helping people navigate the disruptiveness of modern life would be immense.

[Gwern Branwen]

On Wed, Sep 4, 2013 at 5:42 PM, Ram <raman...@gmail.com> wrote:
> In the current report in Nature, however, 11 researchers in Gazzaley’s
> laboratory presented persuasive evidence of transfer. They ventured one step
> further, detailing the underlying neural mechanisms at play. “…
> (A)ge-related deficits in neural signatures of cognitive control, as
> measured with electroencephalography were remediated by multitasking
> training (ie. enhanced midline frontal theta power and frontal-posterior
> theta coherence),” they write. Use of the specially-designed game, in other
> words, markedly improved performance of older adults not merely on the game
> itself but also led to robust increases in activity in those parts of the
> brain’s prefrontal cortex associated with greater cognitive control.

Not sure why this is being hyped, seems like near transfer to me.

Anyway, here is the paper:
https://dl.dropboxusercontent.com/u/182368464/2013-anguera.pdf

I skimmed it; I noticed that the active and passive control groups are
very different in some of the scores, as one would expect at this
point.

The university press release notes the common conflict of interest
here http://www.ucsf.edu/news/2013/09/108616/training-older-brain-3-d-video-game-enhances-cognitive-control
:

> Gazzaley is co-founder and chief science advisor of Akili Interactive Labs, which is developing cognitive video game software as diagnostic and therapeutic tools, and has a patent pending on a game-based cognitive intervention he developed from the research presented in the paper.

--
gwern
http://www.gwern.net

[polar]

very nice, thanks ram and gwern 

Dne středa, 4. září 2013 23:42:39 UTC+2 Ram napsal(a):

[mushin]

Is it possible to download this game? Has it 'leaked out'? It looks fun. 

[Palguay]

It seems like the objective is to keep the car at the center and either respond or not respond to signs , revised space fortress seems more difficult as it has a lot more rules while keeping the plane inside the hexagon .

This seems interesting thought and combining this with other tasks while moving the signs outside the field of view might be a fun task. 

I was looking at implementing a 3d game for brainturk.com and it looks like a variant of this might be a good way to experiment.

On Sat, Sep 7, 2013 at 11:50 PM, mushin <daniel....@gmail.com> wrote:

Is it possible to download this game? Has it 'leaked out'? It looks fun. 

--
You received this message because you are subscribed to the Google Groups "Dual N-Back, Brain Training & Intelligence" group.
To unsubscribe from this group and stop receiving emails from it, send an email to brain-trainin...@googlegroups.com.
To post to this group, send email to brain-t...@googlegroups.com.
Visit this group at http://groups.google.com/group/brain-training.
For more options, visit https://groups.google.com/groups/opt_out.

[jotaro]

but wont you develop task fimiliar skills?

i mean the particular rules and recognizing them instances come alot of easier to you because you played it a bit, however how this recognition is tranferable to otehr tasks since other tasks have other cues? you only specialize on the rulez and cues of this specific game, might add i didnt read the study just from what palguy says.

[Green]

 (I think I sent a version of this post to Gwern privately twice...sorry)

 They found that the active group improved significantly more on the TOVA and something called a 'delayed recognition with distraction' task between post- and pre- testing as compared to control.

 

 As I recall, the TOVA involves looking at a screen and pressing a button every time a certain stimulus appears, but not pressing the button when a distractor stimulus appears. The TOVA is fairly significantly different from the neuroracer task.  The other task also seems significantly different.

 

I consider that far transfer. In the original paper outlining the concept of transfer, the authors talked about 'far' vs. 'near' in terms of a couple dimensions: task difference, and time were two dimensions they discussed that would seem to apply here. The active group improved at significantly different kinds of tasks, and the gains lasted six months. 

 

 

 

[Heinstar]

Neuroracer FAQ: http://gazzaleylab.ucsf.edu/neuroracer-faq.html

I'm waiting for the iOS game that's coming out.

This game seems to be about paying attention to staying on road while responding to shapes and colour.

This article states that while neuroracer seems to show benefit for older people, more difficult version maybe required for younger brains to see cognitive benefit: http://www.nature.com/news/gaming-improves-multitasking-skills-1.13674?spMailingID=17909079&spUserID=MTM0NzE2Mzg1NjkS1&spJobID=230026922&spReportId=MjMwMDI2OTIyS0 

Here's an idea I came up with to make neuroracer more difficult.

1. Instead of responding to shapes and colour, the player must memorise and match shapes(or letters like in combination N-back), sounds and colours N-back ago while driving. 

So basically combining N-back with Neuroracer instead of just responding to correct shapes/colours. 

Also, my current favourite multitasking game is "Train of Thought" from Lumosity. I don't know if it transfers but it's really fun.

[Heinstar]

More information on the upcoming game for Neuroracer: http://www.xconomy.com/boston/2014/01/09/akili-pfizer-and-the-rise-of-videogame-brain-therapies/

I'm following their twitter page for updates: https://twitter.com/AkiliLabs

Do you think this game will be as promising as Jaeggi's Dual N-back? Or is it overhyped like Gwern said?

If this game is promising, maybe we should include it in Brain Workshop? I wish there was a PC version of this game with keyboard and mouse. I don't really like playing Brain training games on tablets because it feels gimmicky.

On Sunday, September 8, 2013 4:20:26 AM UTC+10, mushin wrote:

[jotaro]

it is probably hyped considering the multi-money industry that is interested and will profit of this.
plus in history it is almost always was too hyped.

--

[Palguay]

I have uploaded an initial version of this game which is similar to Neuro racer . 

 With this we have now added our first 3d game to brainturk.com and will hopefully add/convert a few more of the old games to make it a bit more engaging. 

 This is the first version and may have a few bugs. Let me know if you think of any changes/ideas that can be added to this game or come across any bugs in the game.

 https://www.brainturk.com/cognoracer 

 you will need to register as a free user and login to play the game, This game is located in the games page Experimental section.

[MR]

Part of the reason the researchers thought the game was so effective was that the difficulty level adapts to one's performance, increasing difficulty as you become better at the task. Have you figured out how you're going to implement that into the game?

Also, when playing Space Fortress, are there missing instructions, or implicit rules one needs to discover? I ask this because I often find that at times pressing the spacebar does not fire despite having more than enough shots left. I also find that pressing a or s for the bonus doesn't always work. 

Otherwise looks really good, fun to play. 

[Palguay]

I still have not figured out how to adjust difficulty according to one's performance I wanted to get a version out there quickly so that I can get some feedback , now I can look at how to adjust difficulty :-) .

I have not looked at the space fortress in a while I will look into it and see if I can recreate the bug .

Thanks for your feedback.

For more options, visit https://groups.google.com/d/optout.