To date, most functional Magnetic Resonance Imaging (fMRI) studies have assumed that the functional connectivity (FC) between time series from distinct brain regions is constant across time. However, recently, there has been increased interest in quantifying possible dynamic changes in FC during fMRI experiments, as it is thought this may provide insight into the fundamental workings of brain networks. In this work we focus on the specific problem of estimating the dynamic behavior of pair-wise correlations between time courses extracted from two different regions of the brain. We critique the commonly used sliding-windows technique, and discuss some alternative methods used to model volatility in the finance literature that could also prove useful in the neuroimaging setting. In particular, we focus on the Dynamic Conditional Correlation (DCC) model, which provides a model-based approach towards estimating dynamic correlations. We investigate the properties of several techniques in a series of simulation studies and find that DCC achieves the best overall balance between sensitivity and specificity in detecting dynamic changes in correlations. We also investigate its scalability beyond the bivariate case to demonstrate its utility for studying dynamic correlations between more than two brain regions. Finally, we illustrate its performance in an application to test-retest resting state fMRI data.

Deep neural networks (DNNs) are complex computational models that have found great success in many artificial intelligence applications, such as computer vision1,2 and natural language processing.3,4 In the past four years, DNNs have also generated promising results for quantitative structure-activity relationship (QSAR) tasks.5,6 Previous work showed that DNNs can routinely make better predictions than traditional methods, such as random forests, on a diverse collection of QSAR data sets. It was also found that multitask DNN models-those trained on and predicting multiple QSAR properties simultaneously-outperform DNNs trained separately on the individual data sets in many, but not all, tasks. To date there has been no satisfactory explanation of why the QSAR of one task embedded in a multitask DNN can borrow information from other unrelated QSAR tasks. Thus, using multitask DNNs in a way that consistently provides a predictive advantage becomes a challenge. In this work, we explored why multitask DNNs make a difference in predictive performance. Our results show that during prediction a multitask DNN does borrow "signal" from molecules with similar structures in the training sets of the other tasks. However, whether this borrowing leads to better or worse predictive performance depends on whether the activities are correlated. On the basis of this, we have developed a strategy to use multitask DNNs that incorporate prior domain knowledge to select training sets with correlated activities, and we demonstrate its effectiveness on several examples.

Due to the dynamic, condition-dependent nature of brain activity, interest in estimating rapid functional connectivity (FC) changes that occur during resting-state functional magnetic resonance imaging (rs-fMRI) has recently soared. However, studying dynamic FC is methodologically challenging, due to the low signal-to-noise ratio of the blood oxygen level dependent (BOLD) signal in fMRI and the massive number of data points generated during the analysis. Thus, it is important to establish methods and summary measures that maximize reliability and the utility of dynamic FC to provide insight into brain function. In this study, we investigated the reliability of dynamic FC summary measures derived using three commonly used estimation methods - sliding window (SW), tapered sliding window (TSW), and dynamic conditional correlations (DCC) methods. We applied each of these techniques to two publicly available rs-fMRI test-retest data sets - the Multi-Modal MRI Reproducibility Resource (Kirby Data) and the Human Connectome Project (HCP Data). The reliability of two categories of dynamic FC summary measures were assessed, specifically basic summary statistics of the dynamic correlations and summary measures derived from recurring whole-brain patterns of FC (“brain states”). The results provide evidence that dynamic correlations are reliably detected in both test-retest data sets, and the DCC method outperforms SW methods in terms of the reliability of summary statistics. However, across all estimation methods, reliability of the brain state-derived measures was low. Notably, the results also show that the DCC-derived dynamic correlation variances are significantly more reliable than those derived using the non-parametric estimation methods. This is important, as the fluctuations of dynamic FC (i.e., its variance) has a strong potential to provide summary measures that can be used to find meaningful individual differences in dynamic FC. We therefore conclude that utilizing the variance of the dynamic connectivity is an important component in any dynamic FC-derived summary measure.

Protein redesign and engineering has become an important task in pharmaceutical research and development. Recent advances in technology have enabled efficient protein redesign by mimicking natural evolutionary mutation, selection, and amplification steps in the laboratory environment. For any given protein, the number of possible mutations is astronomical. It is impractical to synthesize all sequences or even to investigate all functionally interesting variants. Recently, there has been an increased interest in using machine learning to assist protein redesign, since prediction models can be used to virtually screen a large number of novel sequences. However, many state-of-the-art machine learning models, especially deep learning models, have not been extensively explored. Moreover, only a small selection of protein sequence descriptors has been considered. In this work, the performance of prediction models built using an array of machine learning methods and protein descriptor types, including two novel, single amino acid descriptors and one structure-based three-dimensional descriptor, is benchmarked. The predictions were evaluated on a diverse collection of public and proprietary data sets, using a variety of evaluation metrics. The results of this comparison suggest that Convolution Neural Network models built with amino acid property descriptors are the most widely applicable to the types of protein redesign problems faced in the pharmaceutical industry.

Sb 3+ with stereochemically active lone pair 5s 2 electrons is overwhelming in the doping engineering of the luminescent metal halides, and it usually leads to extrinsic self-trapped excitons (STEs) with tunable emissions. However, the photoluminescence enhancement mechanism of Sb 3+ doped metal halides compared to the pristine host remains unclear. Herein, we doped Sb 3+ into all-inorganic non-emissive Rb 4 CdCl 6 , realizing bright green emission peaking at 525 nm with a photoluminescence quantum yield of 70.2%. A comparison of Raman spectra, as well as the Debye temperature, was utilized to elucidate the STEs mechanism, verifying that the doping of Sb 3+ softens the structural lattice. Thus, strong electron−phonon interactions enable highly efficient photoluminescence originating from STEs emission in Rb 4 CdCl 6 :Sb 3+ . This work demonstrates solid evidence that the efficient emissions of metal halides can be triggered by Sb 3+ doping, and the design principle involved will guide the future studies for emerging luminescence material exploration.