From data to discovery: AI-guided analysis of disease-relevant molecules in spinal muscular atrophy (SMA).

Spinal Muscular Atrophy is caused by partial loss of survival of motoneuron (SMN) protein expression. The numerous interaction partners and mechanisms influenced by SMN loss result in a complex disease. Current treatments restore SMN protein levels to a certain extent, but do not cure all symptoms. The prolonged survival of patients creates an increasing need for a better understanding of SMA. Although many SMN-protein interactions, dysregulated pathways, and organ phenotypes are known, the connections among them remain largely unexplored. Monogenic diseases are ideal examples for the exploration of cause-and-effect relationships to create a network describing the disease-context. Machine learning tools can utilize such knowledge to analyze similarities between disease-relevant molecules and molecules not described in the disease so far. We used an artificial intelligence-based algorithm to predict new genes of interest. The transcriptional regulation of 8 out of 13 molecules selected from the predicted set were successfully validated in an SMA mouse model. This bioinformatic approach, using the given experimental knowledge for relevance predictions, enhances efficient targeted research in SMA and potentially in other disease settings.

Live Iterative Ptychography.

We demonstrate live-updating ptychographic reconstruction with the extended ptychographical iterative engine, an iterative ptychography method, during ongoing data acquisition. The reconstruction starts with a small subset of the total data, and as the acquisition proceeds the data used for reconstruction are extended. This creates a live-updating view of object and illumination that allows monitoring the ongoing experiment and adjusting parameters with quick turn around. This is particularly advantageous for long-running acquisitions. We show that such a gradual reconstruction yields interpretable results already with a small subset of the data. We show simulated live processing with various scan patterns, parallelized reconstruction, and real-world live processing at the hard X-ray ptychographic nanoanalytical microscope PtyNAMi at the PETRA III beamline.

Deep white matter analysis (DeepWMA): Fast and consistent tractography segmentation.

White matter tract segmentation, i.e. identifying tractography fibers (streamline trajectories) belonging to anatomically meaningful fiber tracts, is an essential step to enable tract quantification and visualization. In this study, we present a deep learning tractography segmentation method (DeepWMA) that allows fast and consistent identification of 54 major deep white matter fiber tracts from the whole brain. We create a large-scale training tractography dataset of 1 million labeled fiber samples, and we propose a novel 2D multi-channel feature descriptor (FiberMap) that encodes spatial coordinates of points along each fiber. We learn a convolutional neural network (CNN) fiber classification model based on FiberMap and obtain a high fiber classification accuracy of 90.99% on the training tractography data with ground truth fiber labels. Then, the method is evaluated on a test dataset of 597 diffusion MRI scans from six independently acquired populations across genders, the lifespan (1 day - 82 years), and different health conditions (healthy control, neuropsychiatric disorders, and brain tumor patients). We perform comparisons with two state-of-the-art tract segmentation methods. Experimental results show that our method obtains a highly consistent tract segmentation result, where on average over 99% of the fiber tracts are successfully identified across all subjects under study, most importantly, including neonates and patients with space-occupying brain tumors. We also demonstrate good generalization of the method to tractography data from multiple different fiber tracking methods. The proposed method leverages deep learning techniques and provides a fast and efficient tool for brain white matter segmentation in large diffusion MRI tractography datasets.

Deep white matter analysis: fast, consistent tractography segmentation across populations and dMRI acquisitions.

We present a deep learning tractography segmentation method that allows fast and consistent white matter fiber tract identification across healthy and disease populations and across multiple diffusion MRI (dMRI) acquisitions. We create a large-scale training tractography dataset of 1 million labeled fiber samples (54 anatomical tracts are included). To discriminate between fibers from different tracts, we propose a novel 2D multi-channel feature descriptor (FiberMap) that encodes spatial coordinates of points along each fiber. We learn a CNN tract classification model based on FiberMap and obtain a high tract classification accuracy of 90.99%. The method is evaluated on a test dataset of 374 dMRI scans from three independently acquired populations across health conditions (healthy control, neuropsychiatric disorders, and brain tumor patients). We perform comparisons with two state-of-the-art white matter tract segmentation methods. Experimental results show that our method obtains a highly consistent segmentation result, where over 99% of the fiber tracts are successfully detected across all subjects under study, most importantly, including patients with space occupying brain tumors. The proposed method leverages deep learning techniques and provides a much faster and more efficient tool for large data analysis than methods using traditional machine learning techniques.

Intraoperative mapping of the sensory cortex by time-resolved thermal imaging.

The resection of brain tumor requires a precise distinction between eloquent areas of the brain and pathological tumor tissue in order to improve the extent of resection as well as the patient's progression free survival time. In this study, we discuss mathematical tools necessary to recognize neural activity using thermal imaging cameras. The main contribution to thermal radiation of the exposed human cortex is regional cerebral blood flow (CBF). In fact, neurovascular coupling links neural activity to changes in regional CBF which in turn affects the cortical temperature. We propose a statistically sound framework to visualize neural activity of the primary somatosensory cortex. The framework incorporates a priori known experimental conditions such as the thermal response to neural activity as well as unrelated effects induced by random neural activity and autoregulation. These experimental conditions can be adopted to certain electrical stimulation protocols so that the framework allows to unveil arbitrary evoked neural activity. The method was applied to semisynthetic as well as two intraoperative cases with promising results as we were able to map the eloquent sensory cortex with high sensitivity. Furthermore, the results were validated by anatomical localization and electrophysiological measurements.

Framework for 2D-3D image fusion of infrared thermography with preoperative MRI.

Multimodal medical image fusion combines information of one or more images in order to improve the diagnostic value. While previous applications mainly focus on merging images from computed tomography, magnetic resonance imaging (MRI), ultrasonic and single-photon emission computed tomography, we propose a novel approach for the registration and fusion of preoperative 3D MRI with intraoperative 2D infrared thermography. Image-guided neurosurgeries are based on neuronavigation systems, which further allow us track the position and orientation of arbitrary cameras. Hereby, we are able to relate the 2D coordinate system of the infrared camera with the 3D MRI coordinate system. The registered image data are now combined by calibration-based image fusion in order to map our intraoperative 2D thermographic images onto the respective brain surface recovered from preoperative MRI. In extensive accuracy measurements, we found that the proposed framework achieves a mean accuracy of 2.46 mm.

Portrayed emotions in the movie "Forrest Gump".

Here we present a dataset with a description of portrayed emotions in the movie "Forrest Gump". A total of 12 observers independently annotated emotional episodes regarding their temporal location and duration. The nature of an emotion was characterized with basic attributes, such as arousal and valence, as well as explicit emotion category labels. In addition, annotations include a record of the perceptual evidence for the presence of an emotion. Two variants of the movie were annotated separately: 1) an audio-movie version of Forrest Gump that has been used as a stimulus for the acquisition of a large public functional brain imaging dataset, and 2) the original audio-visual movie. We present reliability and consistency estimates that suggest that both stimuli can be used to study visual and auditory emotion cue processing in real-life like situations. Raw annotations from all observers are publicly released in full in order to maximize their utility for a wide range of applications and possible future extensions. In addition, aggregate time series of inter-observer agreement with respect to particular attributes of portrayed emotions are provided to facilitate adoption of these data.

Local transient myocardial liposomal gene transfer of inducible nitric oxide synthase does not aggravate myocardial function and fibrosis and leads to moderate neovascularization in chronic myocardial ischemia in pigs.

BACKGROUND: This study was designed to explore the effect of transient inducible nitric oxide synthase (iNOS) overexpression via cationic liposome-mediated gene transfer on cardiac function, fibrosis, and microvascular perfusion in a porcine model of chronic ischemia. METHODS AND RESULTS: Chronic myocardial ischemia was induced using a minimally invasive model in 23 landrace pigs. Upon demonstration of heart failure, 10 animals were treated with liposome-mediated iNOS-gene-transfer by local intramyocardial injection and 13 animals received a sham procedure to serve as control. The efficacy of this iNOS-gene-transfer was demonstrated for up to 7 days by reverse transcriptase-polymerase chain reaction in preliminary studies. Four weeks after iNOS transfer, magnetic resonance imaging showed no effect of iNOS overexpression on cardiac contractility at rest and during dobutamine stress (resting ejection fraction: control 27%, iNOS 26%; P = ns). Late enhancement, infarct size, and the amount of fibrosis were similar between groups. Although perfusion and perfusion reserve in response to adenosine and dobutamine were not significantly modified by iNOS-transfer, both vessel number and diameter were significantly increased in the ischemic area in the iNOS-treated group versus control (point score: control 15.3, iNOS 34.7; P < 0.05). CONCLUSIONS: Our findings demonstrate that transient iNOS overexpression does not aggravate cardiac dysfunction or postischemic fibrosis, while potentially contributing to neovascularization in the chronically ischemic heart.

Usefulness of MRI to differentiate between temporary and long-term coronary artery occlusion in a minimally invasive model of experimental myocardial infarction.

The surgical technique employed to determine an experimental ischemic damage is a major factor in the subsequent process of myocardial scar development. We set out to establish a minimally invasive porcine model of myocardial infarction using cardiac contrast-enhanced magnetic resonance imaging (ce-MRI) as the basic diagnostic tool. Twenty-seven domestic pigs were randomized to either temporary or permanent occlusion of the left anterior descending artery (LAD). Temporary occlusion was achieved by inflation of a percutaneous balloon in the left anterior descending artery directly beyond the second diagonal branch. Occlusion was maintained for 30 or 45 min, followed by reperfusion. Permanent occlusion was achieved via thrombin injection. Thirteen animals died peri- or postinterventionally due to arrhythmias. Fourteen animals survived the 30-min ischemia (four animals; group 1), the 45-min ischemia (six animals; group 2), or the permanent occlusion (4 animals; group 3). Coronary angiography and ce-MRI were performed 8 weeks after coronary occlusion to document the coronary flow grade and the size of myocardial scar tissue. The LAD was patent in all animals in groups 1 and 2, with normal TIMI flow; in group 3 animals, the LAD was totally occluded. Fibrosis of the left ventricle in group 1 (4.9 +/- 4.4%; p = 0.008) and group 2 (9.4 +/- 2.9%; p = 0.05) was significantly lower than in group 3 (14.5 +/- 3.9%). Wall thickness of the ischemic area was significantly lower in group 3 versus group 1 and group 2 (2.9 +/- 0.3, 5.9 +/- 0.7, and 6.1 +/- 0.7 mm; p = 0.005). The extent of late enhancement of the left ventricle was also significantly higher in group 3 (16.9 +/- 2.1%) compared to group 1 (5.3 +/- 5.4%; p = 0.003) and group 2 (9.7 +/- 3.4%, p = 0.013). In conclusion, the present model of minimally invasive infarction coupled with ce-MRI may represent a useful alternative to the open chest model for studies of myocardial infarction and scar development.

Quantification of resting myocardial blood flow in a pig model of acute ischemia based on first-pass MRI.

Qualitative and semiquantitative contrast-enhanced (CE) dynamic perfusion MRI techniques are established as noninvasive diagnostic means of assessing coronary artery disease. However, to date quantification of myocardial blood flow (MBF) has not reached the same acceptance as MBF quantification with nuclear techniques. To validate quantification of MBF at rest using the extracellular contrast agent (CA) Gd-DTPA, we performed an animal study in a pig model of acute myocardial ischemia. We quantified MBF from MRI data with a mathematical model (MMID4) of the underlying vasculature. These MBF results were subsequently compared with the results from fluorescent microspheres. The study showed a correlation of r = 0.66 between MBF estimates obtained with MRI and those obtained with fluorescent microspheres. The correlation for ischemic and nonischemic myocardium was r = 0.86 and r = 0.47, respectively. In conclusion, quantification of resting MBF using MMID4 is a valid method under conditions of acute myocardial ischemia.