Interpretation of Brain Morphology in Association to Alzheimer's Disease Dementia Classification Using Graph Convolutional Networks on Triangulated Meshes.

We propose a mesh-based technique to aid in the classification of Alzheimer's disease dementia (ADD) using mesh representations of the cortex and subcortical structures. Deep learning methods for classification tasks that utilize structural neuroimaging often require extensive learning parameters to optimize. Frequently, these approaches for automated medical diagnosis also lack visual interpretability for areas in the brain involved in making a diagnosis. This work: (a) analyzes brain shape using surface information of the cortex and subcortical structures, (b) proposes a residual learning framework for state-of-the-art graph convolutional networks which offer a significant reduction in learnable parameters, and (c) offers visual interpretability of the network via class-specific gradient information that localizes important regions of interest in our inputs. With our proposed method leveraging the use of cortical and subcortical surface information, we outperform other machine learning methods with a 96.35% testing accuracy for the ADD vs. healthy control problem. We confirm the validity of our model by observing its performance in a 25-trial Monte Carlo cross-validation. The generated visualization maps in our study show correspondences with current knowledge regarding the structural localization of pathological changes in the brain associated to dementia of the Alzheimer's type.

Neuroimaging modality fusion in Alzheimer's classification using convolutional neural networks.

Automated methods for Alzheimer's disease (AD) classification have the potential for great clinical benefits and may provide insight for combating the disease. Machine learning, and more specifically deep neural networks, have been shown to have great efficacy in this domain. These algorithms often use neurological imaging data such as MRI and FDG PET, but a comprehensive and balanced comparison of the MRI and amyloid PET modalities has not been performed. In order to accurately determine the relative strength of each imaging variant, this work performs a comparison study in the context of Alzheimer's dementia classification using the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset with identical neural network architectures. Furthermore, this work analyzes the benefits of using both modalities in a fusion setting and discusses how these data types may be leveraged in future AD studies using deep learning.

Temporal Characteristics of the Sleep EEG Power Spectrum in Critically Ill Children.

STUDY OBJECTIVES: Although empirical evidence is limited, critical illness in children is associated with disruption of the normal sleep-wake rhythm. The objective of the current study was to examine the temporal characteristics of the sleep electroencephalogram (EEG) in a sample of children with critical illness. METHODS: Limited montage EEG recordings were collected for at least 24 hours from 8 critically ill children on mechanical ventilation for respiratory failure in a pediatric intensive care unit (PICU) of a tertiary-care hospital. Each PICU patient was age- and gender-matched to a healthy subject from the community. Power spectral analysis with the fast Fourier transform (FFT) was used to characterize EEG spectral power and categorized into 4 frequency bands: δ (0.8 to 4.0 Hz), θ (4.1 to 8.0 Hz), α (8.1 to 13.0 Hz), and ß1/ß2 (13.1 to 20.0 Hz). RESULTS: PICU patients did not manifest the ultradian variability in EEG power spectra including the typical increase in δ-power during the first third of the night that was observed in healthy children. Differences noted included significantly lower mean nighttime δ and θ power in the PICU patients compared to healthy children (p < 0.001). Moreover, in the PICU patients, mean δ and θ power were higher during daytime hours than nighttime hours (p < 0.001). CONCLUSIONS: The results presented herein challenge the assumption that children experience restorative sleep during critical illness, highlighting the need for interventional studies to determine whether sleep promotion improves outcomes in critically ill children undergoing active neurocognitive development.

Digoxin inhibits development of hypoxic pulmonary hypertension in mice.

Chronic hypoxia is an inciting factor for the development of pulmonary arterial hypertension. The mechanisms involved in the development of hypoxic pulmonary hypertension (HPH) include hypoxia-inducible factor 1 (HIF-1)-dependent transactivation of genes controlling pulmonary arterial smooth muscle cell (PASMC) intracellular calcium concentration ([Ca(2+)](i)) and pH. Recently, digoxin was shown to inhibit HIF-1 transcriptional activity. In this study, we tested the hypothesis that digoxin could prevent and reverse the development of HPH. Mice were injected daily with saline or digoxin and exposed to room air or ambient hypoxia for 3 wk. Treatment with digoxin attenuated the development of right ventricle (RV) hypertrophy and prevented the pulmonary vascular remodeling and increases in PASMC [Ca(2+)](i), pH, and RV pressure that occur in mice exposed to chronic hypoxia. When started after pulmonary hypertension was established, digoxin attenuated the hypoxia-induced increases in RV pressure and PASMC pH and [Ca(2+)](i). These preclinical data support a role for HIF-1 inhibitors in the treatment of HPH.