Search

Abstract Machine learning (ML) techniques have gained popularity in the neuroimaging field due to their potential for classifying neuropsychiatric disorders. However, the diagnostic predictive power of the existing algorithms has been limited by small sample sizes, lack of representativeness, data leakage, and/or overfitting. Here, we overcome these limitations with the largest multi-site sample size to date (N = 5365) to provide a generalizable ML classification benchmark of major depressive disorder (MDD) using shallow linear and non-linear models. Leveraging brain measures from standardized ENIGMA analysis pipelines in FreeSurfer, we were able to classify MDD versus healthy controls (HC) with a balanced accuracy of around 62%. But after harmonizing the data, e.g., using ComBat, the balanced accuracy dropped to approximately 52%. Accuracy results close to random chance levels were also observed in stratified groups according to age of onset, antidepressant use, number of episodes and sex. Future studies incorporating higher dimensional brain imaging/phenotype features, and/or using more advanced machine and deep learning methods may yield more encouraging prospects.

Abstract Evidence from previous studies suggests that elevated body mass index (BMI) and genetic risk for obesity is associated with reduced brain volume, particularly in areas of reward-related cognition, e.g. the medial prefrontal cortex (AC-MPFC), the orbitofrontal cortex (OFC), the striatum and the thalamus. However, only few studies examined the interplay between these factors in a joint approach. Moreover, previous findings are based on cross-sectional data. We investigated the longitudinal relationship between increased BMI, brain structural magnetic resonance imaging (MRI) parameters and genetic risk scores in a cohort of n = 502 community-dwelling participants from the Study of Health in Pomerania (SHIP) with a mean follow-up-time of 4.9 years. We found that (1) increased BMI values at baseline were associated with decreased brain parameters at follow-up. These effects were particularly pronounced for the OFC and AC-MPFC. (2) The genetic predisposition for BMI had no effect on brain parameters at baseline or follow-up. (3) The interaction between the genetic score for BMI and brain parameters had no effect on BMI at baseline. Finding a significant impact of overweight, but not genetic predisposition for obesity on altered brain structure suggests that metabolic mechanisms may underlie the relationship between obesity and altered brain structure.