Traditional neuroimage analysis pipelines involve computationally intensive, time-consuming optimization steps, and thus, do not scale well to large cohort studies with thousands or tens of thousands of individuals. In this work we propose a fast and accurate deep learning based neuroimaging pipeline for the automated processing of structural human brain MRI scans, replicating FreeSurfer’s anatomical segmentation including surface reconstruction and cortical parcellation. To this end, we introduce an advanced deep learning architecture capable of whole-brain segmentation into 95 classes. The network architecture incorporates local and global competition via competitive dense blocks and competitive skip pathways, as well as multi-slice information aggregation that specifically tailor network performance towards accurate segmentation of both cortical and subcortical structures. Further, we perform fast cortical surface reconstruction and thickness analysis by introducing a spectral spherical embedding and by directly mapping the cortical labels from the image to the surface. This approach provides a full FreeSurfer alternative for volumetric analysis (in under 1 min) and surface-based thickness analysis (within only around 1 h runtime). For sustainability of this approach we perform extensive validation: we assert high segmentation accuracy on several unseen datasets, measure generalizability and demonstrate increased test-retest reliability, and high sensitivity to group differences in dementia.

The complex life-cycle of the human malaria parasite Plasmodium falciparum requires a high degree of tight coordination allowing the parasite to adapt to changing environments. One of the major challenges for the parasite is the human-to-mosquito transmission, which starts with the differentiation of blood stage parasites into the transmissible gametocytes, followed by the rapid conversion of the gametocytes into gametes, once they are taken up by the blood-feeding Anopheles vector. In order to pre-adapt to this change of host, the gametocytes store transcripts in stress granules that encode proteins needed for parasite development in the mosquito. Here we report on a novel stress granule component, the seven-helix protein 7-Helix-1. The protein, a homolog of the human stress response regulator LanC-like 2, accumulates in stress granules of female gametocytes and interacts with ribonucleoproteins, such as CITH, DOZI, and PABP1. Malaria parasites lacking 7-Helix-1 are significantly impaired in female gametogenesis and thus transmission to the mosquito. Lack of 7-Helix-1 further leads to a deregulation of components required for protein synthesis. Consistently, inhibitors of translation could mimic the 7-Helix-1 loss-of-function phenotype. 7-Helix-1 forms a complex with the RNA-binding protein Puf2, a translational regulator of the female-specific antigen Pfs25, as well as with pfs25-coding mRNA. In accord, gametocytes deficient of 7-Helix-1 exhibit impaired Pfs25 synthesis. Our data demonstrate that 7-Helix-1 constitutes stress granules crucial for regulating the synthesis of proteins needed for life-cycle progression of Plasmodium in the mosquito vector.

No abstract

Traditional neuroimage analysis pipelines involve computationally intensive, time-consuming optimization steps, and thus, do not scale well to large cohort studies. With FastSurfer [1] we propose a fast deep-learning based alternative for the automated processing of structural human MRI brain scans, including surface reconstruction and cortical parcellation. FastSurfer consists of an advanced deep learning architecture (FastSurferCNN) used to segment a whole brain MRI into 95 classes in under 1 min, and a surface pipeline building upon this high-quality brain segmentation. FastSurferCNN incorporates local and global competition via competitive dense blocks and competitive skip pathways, as well as multi-slice information aggregation that specifically tailor network performance towards accurate recognition of both cortical and sub-cortical structures. We demonstrate the superior performance of FastSurferCNN across five different datasets where it consistently outperforms existing deep learning approaches in terms of accuracy by a margin. Further, we perform fast cortical surface reconstruction and thickness analysis by introducing a spectral spherical embedding and by directly mapping the cortical labels from the image to the surface. Precisely, we use the eigenfunctions of the Laplace-Beltrami operator to parametrize the surface smoothly and quickly generate the final spherical map by scaling the 3D spectral embedding vector to unit length. For sustainability of the pipeline we perform extensive validation of FastSurfer: we measure generalizability to different scanners, disease states, as well as an unseen acquisition sequence, demonstrate increased test-retest reliability, and increased sensitivity to disease effects relative to traditional FreeSurfer. In total, we provide a reliable full FreeSurfer alternative for volumetric analysis (within 1 minute) and surface-based thickness analysis (within only around 1h + optionally 30 min for group registration). References1. Henschel L, Conjeti S, Estrada S, et al. FastSurfer -a fast and accurate deep learning based neuroimaging pipeline. CoRR. 2019;abs/1910.03866.

Acute exercise has beneficial effects on mood and is known to induce modulations in functional connectivity (FC) within the emotional network. However, the long-term effects of exercise on affective brain circuits remain largely unknown. Here, we investigated the effects of 6 months of regular exercise on mood, amygdala structure, and functional connectivity. This study comprised N = 18 healthy sedentary subjects assigned to an intervention group (IG; 23.9 ± 3.9 years; 3 trainings/week) and N = 10 subjects assigned to a passive control group (CG; 23.7 ± 4.2 years). At baseline and every two months, performance diagnostics, mood questionnaires, and structural and resting-state-fMRI were conducted. Amygdala-nuclei segmentation and amygdala-to-whole-brain FC analysis were performed. Linear mixed effects models and correlation analyses were conducted between FC, relVO2max, and mood scores. Data showed increases in relVO2max exclusively in the IG. Stronger anticorrelation in amygdala-precuneus FC was found, along with a stronger positive correlation in the amygdala-temporal pole FC in the IG after 4 and 6 months, while mood and amygdala volume did not reveal significant interactions. The relVO2max/amygdala-temporal pole FC correlated positively, and the amygdala-precuneus/amygdala-temporal pole FC correlated negatively. Findings suggest that exercise induced long-term modulations of the amygdala FC with the precuneus and temporal pole, shedding light on potential mechanisms by which exercise has positive influences on mood-related networks, typically altered in affective disorders.