Congrats to Antonia Kaczkurkin, whose paper “Neurostructural Heterogeneity in Youths With Internalizing Symptoms” is now out in Biological Psychiatry. Using semi-supervised machine learning and multi-modal imaging, she parses heterogeneity in youth mood and anxiety symptoms.We leveraged a large sample from the Philadelphia Neurodevelopmental Cohort who completed structural brain imaging and a recently-developed semi-supervised machine learning technique (HYDRA, from collaborators at CBICA) to find two data-driven subtypes. Despite relatively similar levels of symptoms, one subtype was marked by globally reduced gray matter volume and cortical thickness, while the other subtype had spared (or greater than typical) gray matter volume and cortical thickness.These subtypes also differed in multiple independent data types that were not used in clustering: the subtype with structural deficits also showed cognitive deficits, whereas the subtype with preserved brain structure also had intact cognition. These differences generalized to other imaging measures; the subtype with prominent structural deficits also reduced amplitude of functional signals in frontal cortex, and reduced white matter fractional anisotropy in tracts that connect the frontal lobe. This paper was also identified in a press release from the American College of Neuropsychopharmacology, where Toni presented this work.
Finally, Antonia Kaczkurkin (the first author) just started her lab, and is recruiting students at the Department of Psychology at Vanderbilt University (http://bit.ly/2KVthdu).
