Applying Machine Learning to Analyze Psychological Data from Human Brain Scans for Understanding Age-Related Memory Deterioration

Abstract

In this project, we employed Machine Learning (ML) techniques alongside the analysis of psychological survey data to investigate the decrease in memory as a function of age. Our objective was to uncover predictors and mechanisms underlying these changes by relying on a person’s memory through their storytelling. By thoroughly analyzing an existing psychological dataset from various age groups, we utilized machine learning algorithms such as Logistic Regression, Decision Trees, and k-Nearest Neighbors to detect subtle alterations in memory performance across a person’s lifespan. By integrating ML methodologies with psychological data, we aimed to identify distinct patterns associated with memory encoding, storage, and retrieval. Based on the observations, “recalled” and “retold” stories were often very similar, while “imagined” stories had distinct characteristics, which made it easier to distinguish them. Studying memory loss with machine learning opens avenues to explore how machine learning and neurological disorders tie in with memory loss. Ultimately, our findings contributed to a deeper understanding of human memory and aging by highlighting the specific patterns and mechanisms involved in memory changes over time.