Using machine learning methods to predict cognitive age from psychophysiological tests

TL;DR

The paper tackles predicting cognitive age from psychophysiological tests collected via a remote online platform. It compares a range of regression algorithms and emphasizes preprocessing to handle outliers and multicollinearity, finding ensemble methods (AdaBoost, Bagging) to provide the most accurate predictions (R^2 ≈ 0.66, MAE ~5). SHAP analysis identifies Stroop task variability as a key predictor and demonstrates nonlinear feature interactions. The approach supports scalable, privacy-conscious remote screening for cognitive aging on mobile platforms, particularly with small sample sizes.

Abstract

This study introduces a novel method for predicting cognitive age using psychophysiological tests. To determine cognitive age, subjects were asked to complete a series of psychological tests measuring various cognitive functions, including reaction time and cognitive conflict, short-term memory, verbal functions, and color and spatial perception. Based on the tests completed, the average completion time, proportion of correct answers, average absolute delta of the color campimetry test, number of guessed words in the Münsterberg matrix, and other parameters were calculated for each subject. The obtained characteristics of the subjects were preprocessed and used to train a machine learning algorithm implementing a regression task for predicting a person's cognitive age. These findings contribute to the field of remote screening using mobile devices for human health for diagnosing and monitoring cognitive aging.

Figures (13)

• Figure 1: Distribution of the target variable "age".
• Figure 2: Boxplot plot of data variables.
• Figure 3: Triangular correlation matrix of data variables.
• Figure 4: Selected 12 variables with the best VIF value in the range from 0 to 10.
• Figure 5: Comparison of models on test data by MAE and $R^2$.