Researchers aimed to understand memory problems in older adults with a history of alcohol use disorder (AUD) using a machine learning approach. Comparisons were made between a group of 94 adults aged 50-81 with alcohol-induced memory problems (Memory group) to a matched Control group without memory issues. The Random Forests model identified specific features from multiple domains that helped classify individuals in the Memory vs. Control group. Notably, the Memory group showed hyperconnectivity in most regions of the default mode network in the brain, except for some connections involving the anterior cingulate cortex, which were hypoconnected. Other important contributing features included genetic risk scores for AUD, recent alcohol consumption and related health consequences, personality traits like neuroticism and harm avoidance, and the frequency of positive life events. The findings suggest that using various features, such as brain connectivity, personality, life experiences, and genetic risk, can help predict alcohol-related memory problems in later life.
Kamarajan C, Pandey AK, Chorlian DB, Meyers JL, Kinreich S, Pandey G, Subbie Saenz de Viteri S, Zhang J, Kuang W, Barr PB, Aliev F, Anokhin AP, Plawecki MH, Kuperman S, Almasy L, Merikangas A, Brislin SJ, Bauer L, Hesselbrock V, Chan G, Kramer J, Lai D, Hartz S, Bierut LJ, McCutcheon VV, Bucholz KK, Dick DM, Schuckit MA, Edenberg HJ, Porjesz B (2023) Predicting alcohol-related memory problems in older adults: A machine learning study with multi-domain features. bioRxiv, 2022.2012.2030.522330. DOI: 10.1101/2022.12.30.522330