Controversies in Computerized Cognitive Training
A few months ago, a scientific colleague – Dr. Phil Harvey, of the University of Miami – came to me with a fun idea. He had been giving talks at various universities about cognitive training, and always got a lot of questions about the back-and-forth that seems to dominate our field (Cognitive training works! Cognitive training doesn’t work!). Phil wanted to write a review article that addressed the main scientific issues around cognitive training, and publish it in a leading journal – so we could move forward in the science of cognitive training, and address new questions instead of rehashing the same debates over and over again.
Phil was kind enough to invite me to contribute a section on the use of cognitive training to improve cognitive function in healthy older adults, and he recruited two internationally recognized experts in cognitive training as applied to schizophrenia – Drs. Til Wykes and Susan McGurk – as well (Phil himself is an expert in cognition in both aging and schizophrenia, so it made sense that he brought us all together).
Our paper “Controversies in Computerized Cognitive Training” came out this week in the journal Biological Psychiatry: Cognitive Neuroscience and Neuroimaging – you can read the whole paper here at no cost.
Our central point about cognitive training in normal aging is pretty simple – when you look at the largest and best conducted clinical trials, and when you look at the meta–analyses (which average together the results from many studies) – the results are pretty clear – the right kind of cognitive training improves cognitive performance and improves real-world performance. Reviews of cognitive training in normal aging need to engage with this consensus, and sadly some have not done so.
And our central point about cognitive training in schizophrenia is only a bit more complicated. In that field, we observe that most clinical trials with computerized cognitive training programs have also used bridging groups – in-person discussions led by a trained therapist who helps connect the cognitive training exercises to the real-world goals of people doing the training. These studies have frequently shown strong benefits in cognitive performance and real-world activities. Reviews of cognitive training in schizophrenia need to consider cognitive training as it as often delivered – with therapist guidance – and once again, some reviews have missed the mark.
One great part about publishing this article is that the journal published two additional commentaries from yet more experts to further dig into the ideas we discussed in our paper. It’s pretty fun when your ideas stimulate this kind of discussion!
In the first commentary, Dr. Robert Bilder (a noted neuropsychologist) points out that our definition of benefit from cognitive training depends on our definition of generalization and transfer – we want to see people get better not just at the cognitive exercises, but also on neuropsychological tests and measures of real-world performance that are not directly trained. He notes that whether or not a trial is deemed successful depends on this definition, and he makes some quantitative suggestions about how research in cognitive training can better evaluate generalization and transfer.
I agree with Dr. Bilder – and would add to his argument about how to think about generalization and transfer: There’s a pernicious notion in the field of cognitive training that we should expect cognitive training programs show benefits in “untrained areas” – for example, that a visual speed processing training program is only valuable if it shows benefits in an executive function measure of planning. That’s a bit like saying running is only good for your health if it also makes you a better weight-lifter (that’s obviously not true!). What makes a cognitive training program beneficial is that it shows benefits on untrained standard measures, or, better still, on standard measures of real world activities, or in actual real-world activities. I believe that the transfer debate often misses the mark – because the brain will only change from training in areas where the training is relevant. That’s one reason why our programs emphasize speed and attention, which are relevant to, and building blocks for, most higher cognitive skills. We’re better off using the pattern of generalization (measured in the ways that Dr. Bilder suggests) to give us insight into the way the brain works.
In the second commentary, Drs. Richard Keefe and Luca Pani point out that the debates in the field of cognitive training need to stop being a bar fight and start being a boxing match (in their welcome and colorful metaphor!) – with an unbiased referee to establish what cognitive training programs are safe and effective for what people. They are exactly right! The FDA should (and does) regulate claims about efficacy regarding specific cognitive training programs in clinical indications (like Alzheimer’s or schizophrenia, although no cognitive training programs have yet received this clearance from the FDA). But we also need a group to take the lead in performing this service for data regarding cognitive enhancement in normal adults – an area that the FDA does not want to regulate. I’d like to see an academic or non-profit group take this on, and rate evidence-based cognitive training programs for normal adults. I know Posit Science would come out well: in a recent review published by a network of academic researchers, we were the only cognitive training program rated as having multiple gold-standard clinical trials showing efficacy.
It’s exciting for me to see this kind of productive exchange of ideas in the field of cognitive training – they really help push the field further, in a scientific and evidence-based way. I know everyone at Posit Science is ready to start the next set of research studies to ask better questions and advance our field, and it’s great to see so many others doing the same.