On the homepage to my website, I refer to the power of cognitive training and illustrate that power by reference to research with chicken sexing and landmine detection. I cite a study in which novices were taught to identify the sex of very young chickens in a few days. This is a skill that normally takes a couple of years to develop. I also cite a study in which approximately 20 hours of targeted cognitive training improved detection rates for low-metal landmines from 20% to 95%. In each of these cases, a modest amount of practice produced enormous gains in cognitive performance.
Am I promoting brain training?
There is, however, a crucial difference. Both chicken sexing and landmine detection are operational tasks and that cognitive training, while undertaken in a simulated environment, provided intensive practice on particular elements of those operational tasks; elements assessed by a work analysis as posing the greatest challenge to normal progression to expertise.
Thus, the chicken sexing and landmine detection research should be viewed as explorations of cognitive training rather than of brain training. There is no suggestion in this research that there is a generalized benefit to the brain and there is no claim here that either training program will improve capability on other tasks.
On the other hand, beyond their obvious practical application to the targeted work tasks, these studies demonstrate the importance of assessing transfer in the evaluation of training effectiveness.
Skill transfer refers to the idea that something learned in a training program can enhance performance on an operational task. That, of course, is the whole idea behind training. If all you ever do in training is learn how to perform better in training, there is no point.
A common (now decades-old) problem in training research is that conclusions about training effectiveness are often based on performance during training.
For example, distributed practice has generally been viewed as more effective than massed practice. That belief is based on experiments in which performance during distributed trials was much better than during massed trials. However, those rather large effects disappeared in tests of transfer to a common condition, whether it was a distributed- or a massed-trials condition. Despite the fact that the benefits of distributed practice are touted in many textbooks, the evidence that distribution of practice offers any benefit is mostly negative.
Similarly, motion systems are regarded as essential for flight training simulators. Experimental data clearly show they help trainee pilots fly the simulator. Nevertheless, as with distributed-trials training, the transfer evidence is mostly negative.
Brain training revisited
In his book, Smarter: The New Science of Building Brainpower, Dan Hurley shares his experience with the Lumosity program. As you may recall from my April post, Lumosity is one of the commercial purveyors of brain exercises. Over a period of a few months, Hurley improved on these exercises from the 43rd percentile (about average) to the 93rd percentile (well into the elite range). He does not, however, offer any worthwhile evidence of transfer; that is, of any cognitive improvement within his professional or everyday activities.
This powerful and obvious improvement on the trained tasks is what makes brain training so appealing. Those who undertake brain training have concrete evidence of a benefit. The assumption that this benefit will transfer to daily activities is apparently so compelling that people typically do not step back to critically assess whether or not they are more cognitively adept in their everyday life.
As I mentioned in my earlier post, even Buschkuehl and Jaeggi, two of the proponents of brain training, observe that there is no evidence that these training effects extend beyond laboratory tasks to standardized measures or even academic achievement or into daily life in general.
What to make of this
The lesson I leave you with in this post is that transfer of cognitive skills to everyday life, even powerful transfer, is possible, but it takes far more specific training; training that is carefully tuned for the targeted tasks. All of this requires a systematic work analysis and then careful design and development of the training regime.
I have a video of a presentation on distributed training and a paper on distributed training on the Training Systems page of my website. I also have a paper on the evaluation of motion systems for flight simulators on the same page.