Countering the idea of inevitable age-related cognitive decline

Everyone is aware of all the reports that suggest that our cognitive powers go into decline with increased age. As a result, as some of us get into our senior years, we can’t help but identify as symptoms of that mental decline every time we do not remember some thing that we think we should, or when it takes awhile to do something that we think we used to do faster. We may joke about having ‘senior moments’ but those jokes are accompanied by a twinge of anxiety as to whether they are precursors of more serious problems to come.

So it was something of a boost to read (via Machines Like Us) about a new study by Michael Ramscar, Peter Hendrix, Cyrus Shaoul, Petar Milin, and Harald Baayen (Topics in Cognitive Science 6 (2014) 5–42) that says that the slowness associated with age arises not because there is a deterioration in our mental processing capabilities, but because age brings with it increased knowledge and a greater appreciation of fine distinctions and that taking into account all this extra knowledge requires longer processing times. They did the research using computers that simulated learning so that they could isolate the effect of learning by eliminating as a variable the decline in processing power.

This extended excerpt (citations omitted) to the paper titled The Myth of Cognitive Decline: Non-Linear Dynamics of Lifelong Learning pretty much says it all.

In what follows, we consider the question of whether one might reasonably expect that performance on any measure of cognitive performance could or should be expected to be age- or, more specifically, experience-invariant. We shall suggest that, since the answer to this question is no, many of the assumptions scientists currently make about “cognitive decline” are seriously flawed and, for the most part, formally invalid. We will show that the patterns of response change that are typically taken as evidence for (and measures of) cognitive decline arise out of basic principles of learning and emerge naturally in learning models as they acquire more knowledge. These models, which are supported by a wealth of psychological and neuroscientific evidence, also correctly identify greater variation in the cognitive performance of older adults, and successfully predict that older adults will exhibit greater sensitivity to the fine-grained properties of test items than younger adults. Given that the models run (and can be rerun) on computers, the possibility that any differences in their performance are due to aging hardware can be eliminated; instead, their patterns of performance reflect the information-processing costs that must inevitably be incurred as knowledge is acquired. Once the cost of processing this extra information is controlled for in studies of human performance, findings that are usually taken to suggest declining cognitive capacities can be seen instead to support little more than the unsurprising idea that choosing between or recalling items becomes more difficult as their numbers increase. (p.6)

What about all the neurobiological evidence for cognitive decline? Our answer is that except in the case of neurological diseases where there is evidence of pathology, there is no neurobiological evidence for any declines in the processing capacities of healthy older adults. (p. 34)

They also look at the social consequences of the belief of cognitive decline with age.

The tithonean account of aging echoes loudly in the literature of the psychological and brain-sciences, which portrays adulthood as a protracted episode in mental decline, in which memories dim, thoughts slow, and problem-solving abilities diminish, and where researchers seem to compete to set the advent of cognitive decrepitude at an ever younger age. Thus, although studies indicate that older adults are, on average, happier than younger adults, in the light of the foregoing, even this small crumb of comfort might be seen as further evidence of their declining mental prowess.

Because it is believed that cognitive abilities wither over the course of adulthood, population aging is thought to pose a serious threat to the world’s economic well-being: As the proportion of cognitively impaired adults in the population increases, it is feared they will impose an ever-larger burden on the ever-smaller proportion of society still in full command of its cognitive faculties. Given this uncertain scenario, understanding the way our minds age could be considered the most significant matter that the psychological and brain sciences address. (p.6)

It is abundantly clear, for example, that learning is influenced by social as well as environmental factors, and that self-perception can exert a strong influence on what is actually learned from the environment. Because of this, the ideas about “cognitive decline” we have critiqued here are likely to be exerting a strong, negative influence on the lives of many millions of older adults. We hope this can change. Formal models of learning and information processing offer practical as well as scientific insights, and a better, more widespread understanding of these ideas can help people manage their memories more effectively in the future. At the outset, we noted that population aging is seen as a problem because of the fear that older adults will be a burden on society; what is more likely is that the myth of cognitive decline is leading to an absurd waste of human potential and human capital. It thus seems likely that an informed understanding of the cognitive costs and benefits of aging will benefit all society, not just its older members. (p.35)

So all of us oldsters, take heart! We may not be losing our marbles that quickly after all, and may be actually getting better!

So as someone who likes to look on the bright side, I am going to persuade myself that my brain isn’t slowing down as I feared, it just needs more time to deal with all the good stuff I have accumulated over my lifetime.

Now where did I leave my glasses … ?


  1. badgersdaughter says

    I’m getting stupider because I’m getting smarter. Awesome. Can’t wait to tell the kids. 😀

  2. machintelligence says

    Intelligence: “Crystalized” intelligence, i.e., knowledge or experience accumulated over time, actually remains stable with age. On the other hand, “fluid” intelligence or abilities not based on experience or education tend to decline. Examples of “fluid” intelligence include ability to think and react quickly, mental flexibility or mental “multi-tasking”, and learning of new information.

    On the third hand, it is “fluid intelligence” that has been increasing at the rate of 3 IQ points per decade according to Jim Flynn (of the Flynn Effect). So perhaps we old farts are being compared to the young whipper-snappers who have been trained to use better mental tools from a young age. It may well be harder to teach an old dog new tricks.
    On the fourth hand (isn’t it great to have an arbitrary number of hands?), I seem to remember that the decline of problem solving abilities with age is inversely correlated with the starting value: the more you start with, the less you lose. This could be due to a small general decline with age. For example, if some typical scores on Raven’s Progressive Matrices were
    low 6
    average 12
    high 26
    and everyone lost five points, the new scores would be 1, 7 and 21 respectively. Low would have almost disappeared off the chart, average would be approaching low, and high would not have declined very much. The scores that I picked were arbitrary.
    Flynn has an interesting TED talk on the subject.

  3. says

    They did the research using computers that simulated learning

    I don’t think that’s very convincing – to think that computers can adequately simulate learning in a human, I’d expect we’d have to understand a whole lot more about how humans learn and remember, and we’d have to solve the hard AI problem to simulate it accurately enough that I’d believe the simulation’s results were sufficiently similar to a human’s. The problem with simulation models is that your model reflects your expectations of the outcome of your experiment, and it shouldn’t surprise anyone when it produces that outcome.

  4. moarscienceplz says

    I seem to remember that the decline of problem solving abilities with age is inversely correlated with the starting value: the more you start with, the less you lose.

    IANA expert, but I think “starting with more” might also mean you have more tools in your mental toolbox with which to solve problems, so even if you lose the ability to be effective with one method you may still have other pathways to success available to you.

  5. jamessweet says

    I’m with Marcus. I would have to learn more about their process, but I’m skeptical of the idea that their model necessarily applies. The human mind doesn’t function like a big database, it’s more nonlinear… I am not at all convinced that knowing more facts makes one think slower, in the same way that such a process might slow down a computer database.

    Nevertheless, their analysis of the social consequences is spot on.

  6. mnb0 says

    “Now where did I leave my glasses … ?”
    Don’t worry. The first time I asked that question I was younger than 10.

  7. Nick Gotts says

    The problem with simulation models is that your model reflects your expectations of the outcome of your experiment, and it shouldn’t surprise anyone when it produces that outcome. – Marcus Ranum

    On the contrary, as a fairly experienced simulator, I’m always surprised when the simulation produces anything remotely sensible!

  8. steffp says

    “Now where did I leave my glasses … ?”
    Over the last 4 decades, you put your darn spectacles in some thousand different places, without thinking about it because you’re used to putting your glasses somewhere all the time. It’s no longer an intentional process. Just getting rid of them. It’s not like parking a car (takes some physical/mental effort) or remembering your postal code. So, when looking for your specs, you apply a quasi-statistical procedure, going through the most likely places, which indeed is pretty similar to the data bank query problems of a computer system: the more possibilities, the more processing time is needed.
    And after all, specs aren’t important, aren’t they?

  9. says

    Cognitive decline is also related to things like hearing loss. As hearing loss becomes a greater concern for young people, companies like Audicus are providing a more progressive way to ensure you will always be able to hear your loved ones say “I love you”. Check out their new ad campaign here: as well as the products featured on their site:

Leave a Reply

Your email address will not be published. Required fields are marked *