Imaging the Serotonin System 2013–2020
The structure of the serotonin system: a PET imaging study
The human brain atlas of the serotonin (5-HT) system does not conform to commonly used parcellations of neocortex, since the spatial distribution of homogeneous 5-HT receptors and transporter is not aligned with such brain regions. This discrepancy indicates that a neocortical parcellation specific to the 5-HT system is needed.
We first outlined issues with an existing parcellation of the 5-HT system, and presented an alternative parcellation derived from brain MR- and high-resolution PET images of five different 5-HT targets from 210 healthy controls. We then explored how well this new 5-HT parcellation can explain mRNA levels of all 5-HT genes. The parcellation derived here represents a characterization of the 5-HT system that is more stable and explains the underlying 5-HT molecular imaging data better than other atlases, and may hence be more sensitive to capture region-specific changes modulated by 5-HT.
The clustering is available on our serotonin atlas website. Publication: "The Structure of the serotonin system: A PET imaging study" (see Publications).
Multimodal brain imaging of the serotonin system Danish Council for Independent Research | Medical Sciences
Over the last decades, immense progress has been made in mapping human brain structure and function, but much less in mapping the brain's neurochemical signaling systems. Serotonin (5-HT) is an important modulator of e.g. mood, appetite and diurnal regulation, and is involved in frequent neuropsychiatric disorders.
With this computationally advanced project, we utilized unique existing molecular, functional and structural brain imaging data from healthy individuals to generate a high-resolution atlas of the human brain 5-HT system, and to examine the functional connectivity between the 5-HT "control center" and corresponding brain regions. The 5-HT functional connectivity was examined in winter-depressed individuals, in their symptomatic and asymptomatic phases. The outcome was open-access tools and a high-resolution normative database atlas, along with novel insight into brain 5-HT functional connectivity patterns.
Publications: "A High-Resolution In Vivo Atlas of the Human Brain's Serotonin System" and "Cerebellar heterogeneity and its impact on PET data quantification of 5-HT receptor radioligands". Link to the atlas: www.nru.dk/FS5ht-atlas/.
Exploring a multimodal serotonin atlas with machine learning methods Lundbeck Foundation
Neuroimaging is an interdisciplinary field covering biology, chemistry, computer science, mathematics and medicine, and has contributed immensely to our understanding of the nervous system. In this project, we explored how machine learning techniques can be applied to robustly identify variations in the human brain as determined by molecular and structural brain imaging obtained with functional PET and structural MRI.
Specifically, we used a unique multi-modal dataset of the serotonin neurotransmitter system collected at the Center for Integrated Molecular Brain Imaging (Cimbi): approx. 200 high-resolution brain PET scans of healthy volunteers targeting four different serotonin receptors (5-HT1A, 5-HT1B, 5-HT2A, 5-HT4) as well as the serotonin transporter (SERT), paired with corresponding MR scans. This high-dimensional data was examined with machine learning tools such as clustering methods and multivariate regression, and used to build a population-based atlas of the serotonin system for use as a reference and for constructing biomarkers of brain disease.
In a univariate two-tailed t-test of patients having smaller cortical thickness than controls, correction for multiple comparisons was needed; a cluster-wise correction (as done with FreeSurfer, voxel-wise threshold 0.01, cluster-wise threshold 0.01) left only two surviving regions (insula and temporal cortex). By contrast, a multivariate feature selection reliably identified additional regions found in the intersection of feature selections performed over 5 independent subsets – demonstrating how multivariate analysis can add information beyond a univariate analysis.
Publication: "Relevant feature set estimation with a knock-out strategy and random forests".