Predicting alcohol-related memory problems in older adults: A machine learning study with multi-domain features

Abstract

Memory problems are common among older adults with a history of alcohol use disorder (AUD). Employing a machine learning framework, the current study investigates the use of multi-domain features to classify individuals with and without alcohol-induced memory problems. A group of 94 individuals (ages 50-81 years) with alcohol-induced memory problems (Memory group) were compared with a matched Control group who did not have memory problems. The Random Forests model identified specific features from each domain that contributed to the classification of Memory vs. Control group (AUC=88.29%). Specifically, individuals from the Memory group manifested a predominant pattern of hyperconnectivity across the default mode network regions except some connections involving anterior cingulate cortex which were predominantly hypoconnected. Other significant contributing features were (i) polygenic risk scores for AUD, (ii) alcohol consumption and related health consequences during the past 5 years, such as health problems, past negative experiences, withdrawal symptoms, and the largest number of drinks in a day during the past 12 months, and (iii) elevated neuroticism and increased harm avoidance, and fewer positive "uplift" life events. At the neural systems level, hyperconnectivity across the default mode network regions, including the connections across the hippocampal hub regions, in individuals with memory problems may indicate dysregulation in neural information processing. Overall, the study outlines the importance of utilizing multidomain features, consisting of resting-state brain connectivity collected ~18 years ago, together with personality, life experiences, polygenic risk, and alcohol consumption and related consequences, to predict alcohol-related memory problems that arise in later life.

The collaborative study on the genetics of alcoholism: Brain function

“…For example, we reported that ML models combining EEG and AUD associated genetic variants outperformed other models based on a single type of data, suggesting that each contributed unique and significant information 136 . Using a Random Forests method, we found that EEG hyperconnectivity across the default mode network regions, PGS for AUD, alcohol consumption and related health consequences, elevated neuroticism, increased harm avoidance, and fewer positive life events could all be used to classify individuals who would develop alcohol induced memory problems 20 years later 31 . More recently, COGA has used ML algorithms to predict the difference in AUD recovery status, identifying several discriminative features, including PGS related to alcohol use, personality and psychopathology, psychosocial factors and electrophysiological indicators including lower default mode network and fusiform connectivity and higher insula connectivity 137 (Figure S3).…”

“…Functional Connectivity (eLORETA lagged connectivity) Phenotypic findings : Dysregulation in network communication in Default Mode Network (DMN) seen in those who developed alcohol‐related memory problems 20 year later. 31 Genetic findings : PGS for AUD, along with other multimodal features (including functional connectivity in default mode network), are associated with alcohol‐related memory problems. 31 …”

“… 31 Genetic findings : PGS for AUD, along with other multimodal features (including functional connectivity in default mode network), are associated with alcohol‐related memory problems. 31 …”

“…Phenotypic findings : Dysregulation in network communication in Default Mode Network (DMN) seen in those who developed alcohol‐related memory problems 20 year later. 31 …”

“… 136 Using a Random Forests method, we found that EEG hyperconnectivity across the default mode network regions, PGS for AUD, alcohol consumption and related health consequences, elevated neuroticism, increased harm avoidance, and fewer positive life events could all be used to classify individuals who would develop alcohol induced memory problems 20 years later. 31 More recently, COGA has used ML algorithms to predict the difference in AUD recovery status, identifying several discriminative features, including PGS related to alcohol use, personality and psychopathology, psychosocial factors and electrophysiological indicators including lower default mode network and fusiform connectivity and higher insula connectivity 137 (Figure S3 ). Taken together, these ML findings highlight the strength of the multidomain data and analytical expertise in COGA and provide examples of how this information can be synthesized across modalities to identify risk factors and improve prediction of the development of AUD, and patterns of persistence, resistance and recovery from AUD, across the lifespan.…”