Word evolution refers to the changing meanings and associations of words throughout time, as a byproduct of human language evolution. By studying word evolution, we can infer social trends and language constructs over different periods of human history. However, traditional techniques such as word representation learning do not adequately capture the evolving language structure and vocabulary. In this paper, we develop a dynamic statistical model to learn time-aware word vector representation. We propose a model that simultaneously learns time-aware embeddings and solves the resulting "alignment problem". This model is trained on a crawled NYTimes dataset. Additionally, we develop multiple intuitive evaluation strategies of temporal word embeddings. Our qualitative and quantitative tests indicate that our method not only reliably captures this evolution over time, but also consistently outperforms state-of-the-art temporal embedding approaches on both semantic accuracy and alignment quality.

Biomaterials, which can contain appropriate biomechanical and/or biochemical cues, are increasingly being investigated as potential scaffolds for tissue regeneration and/or repair for treating myocardial infarction, heart failure, and peripheral artery disease. Specifically, injectable hydrogels are touted for their minimally invasive delivery, ability to self-assemble in situ, and capacity to encourage host tissue regeneration. Here we present detailed methods for fabricating and characterizing decellularized injectable cardiac and skeletal muscle extracellular matrix (ECM) hydrogels. The ECM derived hydrogels have low cellular and DNA content, retain sulfated glycosaminoglycans and other extracellular matrix proteins such as collagen, gel at physiologic temperature and pH, and assume a nanofibrous architecture. These injectable hydrogels are amenable to minimally invasive, tissue specific biomaterial therapies for treating myocardial infarction and peripheral artery disease.

Similar to other protein-based hydrogels, extracellular matrix (ECM) based hydrogels, derived from decellularized tissues, have a narrow range of mechanical properties and are rapidly degraded. These hydrogels contain natural cellular adhesion sites, form nanofibrous networks similar to native ECM, and are biodegradable. In this study, we expand the properties of these types of materials by incorporating poly(ethylene glycol) (PEG) into the ECM network. We use decellularized myocardial matrix as an example of a tissue specific ECM derived hydrogel. Myocardial matrix-PEG hybrids were synthesized by two different methods, cross-linking the proteins with an amine-reactive PEG-star and photo-induced radical polymerization of two different multi-armed PEG-acrylates. We show that both methods allow for conjugation of PEG to the myocardial matrix by gel electrophoresis and infrared spectroscopy. Scanning electron microscopy demonstrated that the hybrid materials still contain a nanofibrous network similar to unmodified myocardial matrix and that the fiber diameter is changed by the method of PEG incorporation and PEG molecular weight. PEG conjugation also decreased the rate of enzymatic degradation in vitro, and increased material stiffness. Hybrids synthesized with amine-reactive PEG had gelation rates of thirty minutes, similar to the unmodified myocardial matrix, and incorporation of PEG did not prevent cell adhesion and migration through the hydrogels, thus offering the possibility to have an injectable ECM hydrogel that degrades more slowly in vivo. The photo-polymerized radical systems gelled in four minutes upon irradiation allowing for 3D encapsulation and culture of cells, unlike the soft unmodified myocardial matrix. This work demonstrates PEG incorporation into ECM-based hydrogels can expand material properties, thereby opening up new possibilities for in vitro and in vivo applications.

A novel three-layer microfluidic polydimethylsiloxane (PDMS) device was constructed with two fluid chambers that holds a brain slice in place with microposts while maintaining laminar perfusate flow above and below the slice. Our fabrication technique permits rapid production of PDMS layers that can be applied to brain slices of different shapes and sizes. In this study, the device was designed to fit the shape and thickness (530-700 μm) of a medullary brain slice taken from P0-P4 neonatal rats. Medullary slices in this chamber spontaneously produced rhythmic, respiratory-related motor output for up to 3 h, thereby demonstrating that brain slice viability was maintained for prolonged periods. This design is unique in that it achieves independent control of fluids through multiple channels in two separate fluid chambers. The laminar flow exhibited by the microfluidic chamber allows controlled solutions to target specific areas of the brain slice based on the input flow rates. To demonstrate this capability, a stream of Na + -free solution was focused on one half of a medullary slice to abolish spontaneous neural activity in only that half of the brain slice, while the other half remained active. We also demonstrated that flow of different solutions can be focused over the midline of the brain slice. The multilayer brain slice chamber design can integrate several traditional types of electrophysiology tools that are commonly used to measure neurophysiological properties of brain slices. Thus, this new microfluidic chamber is advantageous for experiments that involve controlled drug or solution delivery at high spatiotemporal resolution.

Hepatocyte growth factor (HGF) has been shown to have anti-fibrotic, pro-angiogenic, and cardioprotective effects; however, it is highly unstable and expensive to manufacture, hindering its clinical translation. Recently, a HGF fragment (HGF-f), an alternative c-MET agonist, was engineered to possess increased stability and recombinant expression yields. In this study, we assessed the potential of HGF-f, delivered in an extracellular matrix (ECM)-derived hydrogel, as a potential treatment for myocardial infarction (MI). HGF-f protected cardiomyocytes from serum-starvation and induced down-regulation of fibrotic markers in whole cardiac cell isolate compared to the untreated control. The ECM hydrogel prolonged release of HGF-f compared to collagen gels, and in vivo delivery of HGF-f from ECM hydrogels mitigated negative remodeling, improved fractional area change (FAC), and increased arteriole density in rat myocardial infarction model. These results indicate that HGF-f may be a viable alternative to using recombinant HGF, and that an ECM hydrogel can be employed to increase growth factor retention and efficacy.

Abstract-In many signal processing applications, the aim is to reconstruct a signal that has a simple representation with respect to a certain basis or frame. Fundamental elements of the basis known as "atoms" allow us to define "atomic norms" that can be used to formulate convex regularizations for the reconstruction problem. Efficient algorithms are available to solve these formulations in certain special cases, but an approach that works well for general atomic norms, both in terms of speed and reconstruction accuracy, remains to be found. This paper describes an optimization algorithm called CoGEnT that produces solutions with succinct atomic representations for reconstruction problems, generally formulated with atomicnorm constraints. CoGEnT combines a greedy selection scheme based on the conditional gradient approach with a backward (or "truncation") step that exploits the quadratic nature of the objective to reduce the basis size. We establish convergence properties and validate the algorithm via extensive numerical experiments on a suite of signal processing applications. Our algorithm and analysis also allow for inexact forward steps and for occasional enhancements of the current representation to be performed. CoGEnT can outperform the basic conditional gradient method, and indeed many methods that are tailored to specific applications, when the enhancement and truncation steps are defined appropriately. We also introduce several novel applications that are enabled by the atomic-norm framework, including tensor completion, moment problems in signal processing, and graph deconvolution.

Myoblasts are precursor muscle cells that lie nascent to mature skeletal muscle. Once muscle is damaged, these cells migrate, fuse, and regenerate the muscle tissue. It is known that skeletal muscle can partially regenerate in vivo after muscle tissue damage. However, this regeneration does not always occur, especially in more severe injuries. Cellular therapy using tissue-engineering approaches has been shown to improve organ repair and function. To exploit potential benefits of using cell therapy as an avenue for skeletal muscle repair, it is important to understand the cellular dynamics underlying skeletal myocyte formation and growth. Cardiac fibroblasts have been shown to have a major influence on cardiomyocyte function, repair, and overall spatial distribution. However, little is known regarding fibroblasts’ role on skeletal myocyte function. In this study, we utilized a reconfigurable co-culture device to understand the contact and paracrine effects of fibroblasts on skeletal myocyte alignment and differentiation using murine myoblast and fibroblast cell lines. We demonstrate that myotube alignment is increased by direct contact with fibroblasts, while myotube differentiation is reduced both in the gap and contact configurations with fibroblasts after 6 days of co-culture. Furthermore, neutralizing antibodies to FGF-2 can block these effects of fibroblasts on myotube differentiation and alignment. Finally, bi-directional signaling is critical to the observed myoblast-fibroblast interactions, since conditioned media could not reproduce the same effects observed in the gap configuration. These findings could have direct implications on cell therapies for repairing skeletal muscle, which have only utilized skeletal myoblasts or stem cell populations alone.

Injection of skeletal muscle progenitors has the potential to be a minimally invasive treatment for a number of diseases that negatively affect vasculature and skeletal muscle, including peripheral artery disease (PAD). However, success with this approach has been limited because of poor transplant cell survival. This is primarily attributed to cell death due to extensional flow through the needle, the harsh ischemic environment of the host tissue, a deleterious immune cell response, and a lack of biophysical cues supporting exogenous cell viability. We show that engineering a muscle specific microenvironment, using a nanofibrous decellularized skeletal muscle extracellular matrix hydrogel (SkECM) and skeletal muscle fibroblasts, improves myoblast viability and maturation in vitro. In vivo, this translates to improved cell survival and engraftment and increased perfusion as a result of increased vascularization. Our results indicate that a combinatorial delivery system, which more fully recapitulates the tissue microenvironment, can improve cell delivery to skeletal muscle.