The inventors have developed a new flexible motion sensing method that exploits nonlinear intermodulation of MRI receiver coil preamplifiers to sense the motion of a subject in an MRI scanner without on-subject hardware. The method transmits two tones at two different frequencies, f1 and f2, designed to be received at frequency f_BPT by the receiver via intermodulation, where f1 and f2 are much greater than the MRI center frequency. These signals are picked up by the receiver coils, mixed at the pre-amplification stage by intermodulation, then digitized by the receiver chain. The method is 20 times more sensitive to motion than the state-of-the-art Pilot Tone (PT) method of motion sensing. The inventors have demonstrated the method with second order intermodulation. Additionally, more transmitters can be used, each with a different set of frequencies. Higher frequency tones enable greater sensitivity to subject motion. This method enables the detection of motion at multiple temporal and spatial scales, for example, breathing and rigid motion of the head. The method is used simultaneously with conventional MR imaging and does not adversely impact the signal-to-noise ratio (SNR) of the acquired MR image. The method has been demonstrated using inexpensive consumer grade hardware for the 2.4GHz ISM band as a proof-of-concept. Since the MR signal is small (< -30dBm), little transmit power is necessary to induce an intermodulation signal similar in amplitude to the MR signal.

Scanning tunneling microscopy (STM) techniques and associated spectroscopic (STS) methods, such as dI/dV point spectroscopy, have been widely used to measure electronic structures and local density of states of molecules and materials with unprecedented spatial and energy resolutions. However, the quality of dI/dV spectra highly depends on the shape of the probe tips, and atomically sharp tips with well-defined apex structures are required for obtaining reliable spectra. In most cases, STS measurements are performed in ultra-high vacuum  and low temperature (4 K) to minimize disturbances. Advance tip preparation and constant in situ tip conditioning are required before and during the characterization of target molecules and materials. A common way to prepare STM tips is to repetitively poke them on known and bare substrates (i.e. coinage metals or silicon) to remove contaminations and to potentially coat the tip with substrate atoms. The standard dI/dV spectra of the substrate is then used as a reference to determine whether the tip is available for further experiments. However, tip geometry changes during the poking process are unpredictable, and consequently tip conditioning is typically slow and needs to be constantly monitored. Therefore, it restricts the speed of high-quality STM spectroscopic studies. In order to make efficient use of instrument idle time and minimize the research time wasted on tip conditioning, UC Berkeley researchers developed software based on Python and machine learning that can automate the time-consuming tip conditioning processes. The program is designed to do tip conditioning on Au(111) surfaces that are clean or with low molecular coverage with little human intervention. By just one click, the program is capable of continued poking until the tip can generate near-publication quality spectroscopic data on gold surfaces. It can control the operation of a Scienta Omicron STM and automatically analyze the collected topographic images to find bare Au areas that are large enough for tip conditioning. It will then collect dI/dV spectra at selected positions and use machine learning models to determine their quality compared to standard dI/dV spectra for Au20 and determine if the tip is good enough for further STS measurements. If the tip condition is not ideal, the program will control the STM to poke at the identified positions until the machine learning model predicts the tip to be in good condition.

Researchers at the University of California, Davis have identified earth abundant and other relatively inexpensive materials that form the basis of novel molecules (anolytes), with long lifecycles and high energy densities, to be used in redox flow batteries.

Prof. Anandasankar Ray’s lab at the University of California, Riverside (UCR) has developed machine learning models to predict the inhibitory activity of molecules that bind to potential and actual SARS-CoV-2 interacting human proteins as targets.  An important target is the human ACE2 receptor which is used for viral entry. A recent and published systems-level analysis of protein-protein interactions with peptides encoded in the SARS-CoV-2 genome identified ~66 additional human proteins were considered suitable candidates for the identification of COVID-19 therapeutics. This information was used to train machine learning models and then used to screen most FDA registered chemicals and approved drugs (~100,000) and an additional ~14 million chemicals. Predictions were filtered according to estimated mammalian toxicity and vapor pressure. Prospective volatile candidates identified may be used as novel inhaled therapeutics since the nasal cavity and respiratory tracts are early bottlenecks for infection. Fig. 1 Overview of the pipeline to predict chemicals for 65 SARS-CoV-2 human targets and using bioassay data from publicly available databases

Prof. Alexander Balandin, Dr. Fariborz Kargar and colleagues from the University of California, Riverside have developed novel composites with fillers comprised of graphene and/or quasi-1D van der Waals materials that provide efficient EMI shielding. These unique composites can block EM radiation and are also electrically insulating. These composites may be added to adhesives used in the packaging of electronic components so that the resulting electronic devices will have EMI shielding properties. This technology ensures that EMI shielding is achieved so that a variety of electronics operate reliably and without detrimental effects on human health. Fig 1: Coefficients of absorption for composites made with the UCR graphene-based composition.  

Researchers at the University of California, Irvine have developed artificial red blood cells consisting of hemoglobin that is tethered to polyethylene glycol (PEG) molecules and formed into microspheres.

With the emergence of the Internet of Things (IoT), many applications run machine learning algorithms to perform cognitive tasks. The learning algorithms have been shown effectiveness for many tasks, e.g., object tracking, speech recognition, image classification, etc. However, since sensory and embedded devices are generating massive data streams, it poses huge technical challenges due to limited device resources. For example, although Deep Neural Networks (DNNs) such as AlexNet and GoogleNet have provided high classification accuracy for complex image classification tasks, their high computational complexity and memory requirement hinder usability to a broad variety of real-life (embedded) applications where the device resources and power budget is limited. Furthermore, in IoT systems, sending all the data to the powerful computing environment, e.g., cloud, cannot guarantee scalability and real-time response. It is also often undesirable due to privacy and security concerns. Thus, we need alternative computing methods that can run the large amount of data at least partly on the less-powerful IoT devices. Brain-inspired Hyperdimensional (HD) computing has been proposed as the alternative computing method that processes the cognitive tasks in a more light-weight way.  The HD computing is developed based on the fact that brains compute with patterns of neural activity which are not readily associated with numerical numbers. Recent research instead have utilized high dimension vectors (e.g., more than a thousand dimension), called hypervectors, to represent the neural activities, and showed successful progress for many cognitive tasks such as activity recognition, object recognition, language recognition, and bio-signal classification. 

resent strategies aimed to target and manipulate RNA in living cells mainly rely on the use of antisense oligonucleotides (ASO) or engineered RNA binding proteins (RBP). Although ASO therapies have been shown great promise in eliminating pathogenic transcripts or modulating RBP binding, they are synthetic in construction and thus cannot be encoded within DNA. This complicates potential gene therapy strategies, which would rely on regular administration of ASOs throughout the lifetime of the patient. Furthermore, they are incapable of modulating the genetic sequence of RNA. Although engineered RBPs such as PUF proteins can be designed to recognize target transcripts and fused to RNA modifying effectors to allow for specific recognition and manipulation, these constructs require extensive protein engineering for each target and may prove to be laborious and costly. Normal 0 false false false EN-US X-NONE X-NONE /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-parent:""; mso-padding-alt:0in 5.4pt 0in 5.4pt; mso-para-margin-top:0in; mso-para-margin-right:0in; mso-para-margin-bottom:8.0pt; mso-para-margin-left:0in; line-height:107%; mso-pagination:widow-orphan; font-size:11.0pt; font-family:"Calibri",sans-serif; mso-ascii-font-family:Calibri; mso-ascii-theme-font:minor-latin; mso-hansi-font-family:Calibri; mso-hansi-theme-font:minor-latin; mso-bidi-font-family:"Times New Roman"; mso-bidi-theme-font:minor-bidi;}