A Finnish study followed young people for eight years, tracking their physical activity and screen time at different points and then examining how these were associated with their cognitive performance in adolescence. Sounds impressive, and it is. And things get particularly interesting when more screen time appears to be positively associated with cognition. But… we may want to take a closer look.

Petri Jalanko and his colleagues used data from the Finnish PANIC study. The children were around eight years old at the start and were examined again eight years later. For this analysis, the researchers were able to use data from 260 young people. Physical activity and sedentary behaviour were not only assessed using questionnaires, but also with an Actiheart, a device that records movement and heart rate. This allowed the researchers to compare self-reported behaviour with objectively measured physical activity.

At sixteen, the participants completed a series of cognitive tests measuring attention, reaction time, working memory and associative learning, among other things. One important detail straight away: the researchers only administered this cognitive test battery at the end of the study. They did, however, measure non-verbal reasoning at the start, allowing them to control for it in their analyses.

You might expect young people who move more to perform better cognitively, and those who spend more time on screens to perform worse.

But that is not what the researchers found. When they looked at physical activity and sedentary behavior, they found no link between the two and cognitive performance in the sample. However, things changed when they considered subjective data. Specifically, young participants who took part in physical activities without supervision didn’t do well on a working-memory task. But we should be careful here. This finding is questionable since it didn’t hold up well after adjusting for multiple comparisons.

For screen time, the researchers found something even more surprising. Young people who reported spending more time on screens over the years responded faster on two working-memory tasks and scored better, on average, on the overall measure of cognition. These connections remained significant even after adjusting for multiple tests.

So, do screens suddenly make children smarter? No. This is where the word but in the introduction, and seems in the title, become important.

To start, we are looking at correlations. Yes, those correlations are back. Based on this study, the researchers cannot say that more screen time leads to better cognitive performance. It may be that people with strong cognitive skills spend their time in front of the screen differently. The researchers tried to reduce this problem by accounting for non-verbal reasoning at the start of the study, but this does not eliminate the possibility of reverse causality. The authors explicitly acknowledge this themselves. However, there is something even more basic in all this: What is screen time?

Screen time, according to this research, means watching television programs or movies, working on the computer, playing video games, using the mobile phone, and playing mobile games. This covers quite varied activities. One hour of gaming is not equivalent to an hour of viewing videos. There can be huge variations even within one activity.

This is important because maybe it is not the screen that counts, but what the individual is actually doing in front of the screen. The researchers themselves suggest that both screen time and other sedentary behaviour can include activities that are cognitively quite demanding. Think, for example, of reading and writing, doing homework, playing video games or playing board games.

And that may also help explain the strange finding for physical activity. The researchers offer one possible explanation: young people who spend a lot of time on unsupervised physical activity may spend less time reading or doing homework.

But this is just that, a possible explanation, not something the study actually demonstrates.

There is another interesting warning hidden inside this study as the results change depending on how the researchers measure physical activity. Most of the significant correlations appear in the self-reported data. But when the researchers look at the activity actually recorded by the device, most of it disappears. This could be because the questionnaires are inaccurate. But it could also be the case because the devices miss something important too. Such a device can tell us how much someone moves, but not what they are doing while moving. So again context may matter more than the simple number of minutes spent moving or sitting still.

There was one exception among girls in this study. Those who engaged in more objectively measured light physical activity performed better on a working-memory task. The researchers did not find the same association for boys. They are cautious about explaining this difference, but suggest that light physical activity among girls may, for example, include more time spent socialising.

So, what can we actually take from all of this?

Above all, this study shows how risky it can be to treat concepts such as screen time, sedentary behaviour and physical activity as if they were single, clearly defined activities. They’re not.  An hour spent in front of a screen tells us remarkably little, by itself, about what someone actually did during that hour. And a movement sensor tells us how much someone moved, but not necessarily what that activity meant.

To be clear: this remains a genuinely strong study: eight years of follow-up, several measurement points, both questionnaires and objective measures of physical activity, and controls for several potential confounding variables. But, as good scientists do, the researchers themselves also point to the important limitations I mentioned above.

And because of those limitations, this study ultimately turns things less upside down than the results might initially suggest.