Project description:Neurobiological systems continually interact with the surrounding environment to refine their behaviour toward the best possible reward. Achieving such learning by experience is one of the main challenges of artificial intelligence, but currently it is hindered by the lack of hardware capable of plastic adaptation. Here, we propose a bio-inspired recurrent neural network, mastered by a digital system on chip with resistive-switching synaptic arrays of memory devices, which exploits homeostatic Hebbian learning for improved efficiency. All the results are discussed experimentally and theoretically, proposing a conceptual framework for benchmarking the main outcomes in terms of accuracy and resilience. To test the proposed architecture for reinforcement learning tasks, we study the autonomous exploration of continually evolving environments and verify the results for the Mars rover navigation. We also show that, compared to conventional deep learning techniques, our in-memory hardware has the potential to achieve a significant boost in speed and power-saving.
Project description:Recently, electronic skin that mimics human skin in measuring tactile stimuli, temperature, and humidity and having a self-healing function was developed. Furthermore, with the advances in the field of artificial intelligence and health monitoring, various materials and methods have been studied for e-skin. The limitations to work on actual human skin include device flexibility and large-area applications through array structures, and many studies are underway to overcome these problems. Polymeric materials containing ionic liquids can be used to easily fabricate devices in the solid state. They are highly sensitive to both pressure and temperature, making them suitable for multi-sensing devices. Resistive and capacitive sensors have the advantage of having a simple structure, which makes them easy to fabricate. In a single device, both types work well. For resistive sensors, the temperature sensitivity (1.1/°C) is relatively high. Conversely, capacitive sensors have a low temperature sensitivity (0.3/°C). However, they have the advantage of being uniformly variable under each condition and having a smaller error range. In the array structure, independent flex and thermo sensors are arranged repeatedly. The resistive type shows changes in temperature and bending, but in the capacitive type, it is difficult to obtain results from the pixels due to parasitic capacitance.
Project description:The inhalation, ingestion, and body absorption of noxious gases lead to severe tissue damage, ophthalmological issues, and neurodegenerative disorders; death may even occur when recognized too late. In particular, methanol gas present in traces can cause blindness, non-reversible organ failure, and even death. Even though ample materials are available for the detection of methanol in other alcoholic analogs at ppm level, their scope is very limited because of the use of either toxic or expensive raw materials or tedious fabrication procedures. In this paper, we report on a simple synthesis of fluorescent amphiphiles achieved using a starting material derived from renewable resources, this material being methyl ricinoleate in good yields. The newly synthesized bio-based amphiphiles were prone to form a gel in a broad range of solvents. The morphology of the gel and the molecular-level interaction involved in the self-assembly process were thoroughly investigated. Rheological studies were carried out to probe the stability, thermal processability, and thixotropic behavior. In order to evaluate the potential application of the self-assembled gel in the field of sensors, we performed sensor measurements. Interestingly, the twisted fibers derived from the molecular assembly could be able to display a stable and selective response towards methanol. We believe that the bottom-up assembled system holds great promise in the environmental, healthcare, medicine, and biological fields.
Project description:Mechanical cues including stretch, compression, and shear stress play a critical role in regulating the behavior of many cell types, particularly those that experience substantial mechanical stress within tissues. Devices that impart mechanical stimulation to cells in vitro have been instrumental in helping to develop a better understanding of how cells respond to mechanical forces. However, these devices often have constraints, such as cost and limited functional capabilities, that restrict their use in research or educational environments. Here, we describe a low-cost method to fabricate a uniaxial cell stretcher that would enable widespread use and facilitate engineering design and mechanobiology education for undergraduate students. The device is capable of producing consistent and reliable strain profiles through the use of a servomotor, gear, and gear rack system. The servomotor can be programmed to output various waveforms at specific frequencies and stretch amplitudes by controlling the degree of rotation, speed, and acceleration of the servogear. In addition, the stretchable membranes are easy to fabricate and can be customized, allowing for greater flexibility in culture well size. We used the custom-built stretching device to uniaxially strain macrophages and cardiomyocytes, and found that both cell types displayed functional and cell shape changes that were consistent with the previous studies using commercially available systems. Overall, this uniaxial cell stretcher provides a more cost-effective alternative to study the effects of mechanical stretch on cells, and can therefore, be widely used in research and educational environments to broaden the study and pedagogy of cell mechanobiology.
Project description:An adaptive Bayesian regularized cardiac strain imaging (ABR-CSI) algorithm for in vivo murine myocardial function assessment is presented. We report on 31 BALB/CJ mice (n = 17 females, n = 14 males), randomly stratified into three surgical groups: myocardial infarction (MI, n = 10), ischemia-reperfusion (IR, n = 13) and control (sham, n = 8) imaged pre-surgery (baseline- BL), and 1, 2, 7 and 14 days post-surgery using a high frequency ultrasound imaging system (Vevo 2100). End-systole (ES) radial and longitudinal strain images were used to generate cardiac fibrosis maps using binary thresholding. Percentage fibrotic myocardium (PFM) computed from regional fibrosis maps demonstrated statistically significant differences post-surgery in scar regions. For example, the MI group had significantly higher PFMRadial (%) values in the anterior mid region (p = 0.006) at Day 14 (n = 8, 42.30 ± 14.57) compared to BL (n = 12, 1.32 ± 0.85). A random forest classifier automatically detected fibrotic regions from ground truth Masson's trichrome stained histopathology whole slide images. Both PFMRadial (r = 0.70) and PFMLongitudinal (r = 0.60) results demonstrated strong, positive correlation with PFMHistopathology (p < 0.001).
Project description:Optical control of the heart muscle is a promising strategy for cardiology because it is more specific than traditional electrical stimulation, and allows a higher temporal resolution than pharmacological interventions. Anion channelrhodopsins (ACRs) from cryptophyte algae expressed in cultured neonatal rat ventricular cardiomyocytes produced inhibitory currents at less than one-thousandth of the light intensity required by previously available optogenetic tools, such as the proton pump archaerhodopsin-3 (Arch). Because of their greater photocurrents, ACRs permitted complete inhibition of cardiomyocyte electrical activity under conditions in which Arch was inefficient. Most importantly, ACR expression allowed precisely controlled shortening of the action potential duration by switching on the light during its repolarization phase, which was not possible with previously used optogenetic tools. Optical shortening of cardiac action potentials may benefit pathophysiology research and the development of optogenetic treatments for cardiac disorders such as the long QT syndrome.
Project description:Here, we propose a Pt/HfO2/TaOx/TiN artificial synaptic device that is an excellent candidate for artificial synapses. First, XPS analysis is conducted to provide the dielectric (HfO2/TaOx/TiN) information deposited by DC sputtering and atomic layer deposition (ALD). The self-rectifying resistive switching characteristics are achieved by the asymmetric device stack, which is an advantage of the current suppression in the crossbar array structure. The results show that the programmed data are lost over time and that the decay rate, which is verified from the retention test, can be adjusted by controlling the compliance current (CC). Based on these properties, we emulate bio-synaptic characteristics, such as short-term plasticity (STP), long-term plasticity (LTP), and paired-pulse facilitation (PPF), in the self-rectifying I-V characteristics of the Pt/HfO2/TaOx/TiN bilayer memristor device. The PPF characteristics are mimicked by replacing the bio-stimulation with the interval time of paired pulse inputs. The typical potentiation and depression are also implemented by optimizing the set and reset pulse. Finally, we demonstrate the natural depression by varying the interval time between pulse inputs.
Project description:After the recent cloning of light-sensitive ion channels and their expression in mammalian cells, a new field, optogenetics, emerged in neuroscience, allowing for precise perturbations of neural circuits by light. However, functionality of optogenetic tools has not been fully explored outside neuroscience, and a nonviral, nonembryogenesis-based strategy for optogenetics has not been shown before.We demonstrate the utility of optogenetics to cardiac muscle by a tandem cell unit (TCU) strategy, in which nonexcitable cells carry exogenous light-sensitive ion channels, and, when electrically coupled to cardiomyocytes, produce optically excitable heart tissue. A stable channelrhodopsin2 (ChR2)-expressing cell line was developed, characterized, and used as a cell delivery system. The TCU strategy was validated in vitro in cell pairs with adult canine myocytes (for a wide range of coupling strengths) and in cardiac syncytium with neonatal rat cardiomyocytes. For the first time, we combined optical excitation and optical imaging to capture light-triggered muscle contractions and high-resolution propagation maps of light-triggered electric waves, found to be quantitatively indistinguishable from electrically triggered waves.Our results demonstrate feasibility to control excitation and contraction in cardiac muscle by light, using the TCU approach. Optical pacing in this case uses less energy, offers superior spatiotemporal control and remote access and can serve not only as an elegant tool in arrhythmia research but may form the basis for a new generation of light-driven cardiac pacemakers and muscle actuators. The TCU strategy is extendable to (nonviral) stem cell therapy and is directly relevant to in vivo applications.
Project description:Harvesting water vapor from desert, arid environments by metal-organic framework (MOF) based devices to deliver clean liquid water is critically dependent on environment and climate conditions. However, reported devices have yet been developed to adapt in real-time to such conditions during their operation, which severely limits water production efficiency and unnecessarily increases power consumption. Herein, we report and detail a mode of water harvesting operation, termed ‘adaptive water harvesting’, from which a MOF-based device is proven capable of adapting the adsorption and desorption phases of its water harvesting cycle to weather fluctuations throughout a given day, week, and month such that its water production efficiency is continuously optimized. In performance evaluation experiments in a desert, arid climate (17–32% relative humidity), the adaptive water harvesting device achieves a 169% increase in water production (3.5 LH2O kgMOF−1 d−1) when compared to the best-performing, reported active device (0.7–1.3 LH2O kgMOF−1 d−1 at 10–32% relative humidity), a lower power consumption (1.67–5.25 kWh LH2O−1), and saves time by requiring nearly 1.5 cycles less than a counterpart active device. Furthermore, the produced water meets the national drinking standards of a potential technology-adopting country. Atmospheric water harvesting by MOF-based devices is critically dependent on weather conditions. Here, the authors report a MOF-based adaptive water harvesting system capable of adapting its cycles to temperature and relative humidity fluctuations throughout the day, continuously optimizing water production.
Project description:All-trans-Retinal (ATR) is a photosensitizer, serving as the chromophore for depolarizing and hyperpolarizing light-sensitive ion channels and pumps (opsins), recently employed as fast optical actuators. In mammalian optogenetic applications (in brain and heart), endogenous ATR availability is not considered a limiting factor, yet it is unclear how ATR modulation may affect the response to optical stimulation. We hypothesized that exogenous ATR may improve light responsiveness of cardiac cells modified by Channelrhodopsin2 (ChR2), hence lowering the optical pacing energy. In virally-transduced (Ad-ChR2(H134R)-eYFP) light-sensitive cardiac syncytium in vitro, ATR supplements ≤2 μM improved cardiomyocyte viability and augmented ChR2 membrane expression several-fold, while >4 μM was toxic. Employing integrated optical actuation (470 nm) and optical mapping, we found that 1-2 μM ATR dramatically reduced optical pacing energy (over 30 times) to several μW/mm(2), lowest values reported to date, but also caused action potential prolongation, minor changes in calcium transients and no change in conduction. Theoretical analysis helped explain ATR-caused reduction of optical excitation threshold in cardiomyocytes. We conclude that cardiomyocytes operate at non-saturating retinal levels, and carefully-dosed exogenous ATR can enhance the performance of ChR2 in cardiac cells and yield energy benefits over orders of magnitude for optogenetic stimulation.