Subcortical Translational Neuroimaging

Much of the points below came from a article by Klein-Fugge et al. (2022)

Overall goal

identify specific relatoinship between dimensions of mental health and connectivity in precise subcortical networks

predictions in translational neuroimaging studies typically rely on a large number of brain regions, networks, or edges. impressive prediction accuracies come at the cost of reduced anatomical specificity.
another problem, unrelated to improving anatomical specificity, is relating baseline neural measures to mental well-being. disorders are ill-defined and span a broad range of impairments that are not consistently present in all patients diagnosed with the same disorder, and symptoms are partly overlapping between disorders.
^^ one of the reasons why a classifier trained to distinguish a depressed from a non-depresse person reveals a broad network of regions instead of mapping onto well-defined and anatomically interpretable brain circuits.
new approach: focus on specific rather than broad symptom categories - may be possible to relate them to specific brain circuits

authors examined variance in mental health dimensions along a continuum spanning the subclinical range..
they analysed 33 markers including measures from the NIH toolbox “emotion” Ususcales: psychological well-being, social relationships etc.)
To capture latent dimensions that produce mental well-being, they performed a factor analysis that resulted in four main factors: social and life satisfaction, negative emotions, sleep, anger and rejection
visualize factor loadings onto behavioral scores.

does FC between specific amygdala nuclei and ROIs carry information about mental well-being as captured by four latent mental health dimensions?
can you predict mental health dimensions in held-out data using a weighting of FC values derived from an independent dataset?