THANKS. Listen to this article using the player above. ✖
Do you want to listen to this article for free?
Fill out the form below to unlock access to all audio items.
A study published in The Journal of Nutrition Used automatic learning techniques to determine what health and lifestyle factors best predict cognitive performance throughout the lifespan of adults. Research has evaluated how variables such as age, blood pressure, body mass index (BMI), diet and physical activity are linked to the performance of a task that tests attentional control and response speed.
The task, known as the Flanker test, obliges individuals to identify the direction of a central arrow while ignoring the distracting flanked arrows. This is a commonly used measurement of attention and inhibitory control.
Age and cardiovascular measures have become dominant predictors
The data was collected from 374 adults aged 19 to 82, who were evaluated on a wide range of measures, including BMI, eating habits, blood pressure and physical activity levels. The participants also finished the Flanker test, which measured precision and response time.
Among the automatic learning models evaluated, those which most precisely predicted the performance highlighted age as the strongest predictor. Diastolic blood pressure, BMI and systolic blood pressure followed the influence. While food quality, measured by adhesion to the health of healthy food, was less predictive than age measures or cardiovascular, it was associated with an improvement in the performance of tasks.
Lifestyle models can alleviate certain risk factors
The study noted that if a high BMI and high blood pressure were generally linked to less good results, higher physical activity and better food membership could partially compensate for these effects. Physical activity in particular appeared as a moderate predictor, and his interaction with other lifestyle factors has suggested a complex relationship with cognitive performance.
Automatic learning provides analytical depth
Traditional statistical methods often examine the unique variables in isolation. On the other hand, automatic learning allows a simultaneous evaluation of several variables potentially in interaction. This approach revealed nuanced associations between the lifestyle factors that could otherwise remain obscured.
The researchers tested several algorithms to identify the best able to predict cognitive performance according to the data collected. The models were validated using standard automatic learning practices to assess robustness.
Research was supported by the personalized nutritional initiative and the National Center for SuperComputing requests at the University of Illinois Urbana-Champaign.
Reference: Verma S, Holthaus Ta, Martell S, et al. Predict cognitive results thanks to nutrition and health markers using supervised automatic learning. J nutr. 2025. Doi: 10.1016 / J.TJNUT.2025.05.003
This content includes text generated with the help of AI. IA policy of technological networks can be found here.