Thin film solar cells, ∼1 μm thick, have been fabricated from amorphous silicon deposited from a glow discharge in silane. The cells were made in a p-i-n structure by using doping gases in the discharge. The best power conversion efficiency to date is 2.4% in AM-1 sunlight. The maximum efficiency of thin-film amorphous silicon solar cells is estimated to be ∼14–15%.

Chronically implantable, closed-loop neuromodulation devices with concurrent sensing and stimulation hold promise for better understanding the nervous system and improving therapies for neurological disease. Concurrent sensing and stimulation are needed to maximize usable neural data, minimize time delays for closed-loop actuation, and investigate the instantaneous response to stimulation. Current systems lack concurrent sensing and stimulation primarily because of stimulation interference to neural signals of interest. While careful design of high performance amplifiers has proved useful to reduce disturbances in the system, stimulation continues to contaminate neural sensing due to biological effects like tissue-electrode impedance mismatch and constraints on stimulation parameters needed to deliver therapy. In this work we describe systematic methods to mitigate the effect of stimulation through a combination of sensing hardware, stimulation parameter selection, and classification algorithms that counter residual stimulation disturbances. To validate these methods we implemented and tested a completely implantable system for over one year in a large animal model of epilepsy. The system proved capable of measuring and detecting seizure activity in the hippocampus both during and after stimulation. Furthermore, we demonstrate an embedded algorithm that actuates neural modulation in response to seizure detection during stimulation, validating the capability to detect bioelectrical markers in the presence of therapy and titrate it appropriately. The capability to detect neural states in the presence of stimulation and optimally titrate therapy is a key innovation required for generalizing closed-loop neural systems for multiple disease states.

No abstract

Abstract. Low-cost particulate matter (PM) sensors are promising tools for supplementing existing air quality monitoring networks. However, the performance of the new generation of low-cost PM sensors under field conditions is not well understood. In this study, we characterized the performance capabilities of a new low-cost PM sensor model (Plantower model PMS3003) for measuring PM 2.5 at 1 min, 1 h, 6 h, 12 h, and 24 h integration times. We tested the PMS3003 sensors in both low-concentration suburban regions (Durham and Research Triangle Park (RTP), NC, US) with 1 h PM 2.5 (mean ± SD) of 9 ± 9 and 10 ± 3 µg m −3 , respectively, and a high-concentration urban location (Kanpur, India) with 1 h PM 2.5 of 36 ± 17 and 116 ± 57 µg m −3 during monsoon and post-monsoon seasons, respectively. In Durham and Kanpur, the sensors were compared to a research-grade instrument (environmental β attenuation monitor, E-BAM) to determine how these sensors perform across a range of PM 2.5 concentrations and meteorological factors (e.g., temperature and relative humidity, RH). In RTP, the sensors were compared to three Federal Equivalent Methods (FEMs) including two Teledyne model T640s and a Thermo Scientific model 5030 SHARP to demonstrate the importance of the type of reference monitor selected for sensor calibration. The decrease in 1 h mean errors of the calibrated sensors using univariate linear models from Durham (201 %) to Kanpur monsoon (46 %) and post-monsoon (35 %) seasons showed that PMS3003 performance generally improved as ambient PM 2.5 increased. The precision of reference instruments (T640: ±0.5 µg m −3 for 1 h; SHARP: ±2 µg m −3 for 24 h, better than the E-BAM) is critical in evaluating sensor performance, and β-attenuation-based monitors may not be ideal for testing PM sensors at low concentrations, as underscored by (1) the less dramatic error reduction over averaging times in RTP against optically based T640 (from 27 % for 1 h to 9 % for 24 h) than in Durham (from 201 % to 15 %); (2) the lower errors in RTP than the Kanpur post-monsoon season (from 35 % to 11 %); and (3) the higher T640-PMS3003 correlations (R 2 ≥ 0.63) than SHARP-PMS3003 (R 2 ≥ 0.25). A major RH influence was found in RTP (1 h RH = 64 ± 22 %) due to the relatively high precision of the T640 measurements that can explain up to ∼ 30 % of the variance in 1 min to 6 h PMS3003 PM 2.5 measurements. When proper RH corrections are made by empirical nonlinear equations after using a more precise reference method to calibrate the sensors, our work suggests that the PMS3003 sensors can measure PM 2.5 concentrations within ∼ 10 % of ambient values. We observed that PMS3003 sensors appeared to exhibit a nonlinear response when ambient PM 2.5 exceeded ∼ 125 µg m −3 and found that the quadratic fit is more appropriate than the univariate linear model to capture this nonlinearity and can further reduce errors by up to 11 %. Our results have substantial implications for how variability in ambient PM 2.5 concentrations, reference monitor types, and meteorologi...

Summary Brain-wide fluctuations in local field potential oscillations reflect emergent network-level signals that mediate behavior. Cracking the code whereby these oscillations coordinate in time and space (spatiotemporal dynamics) to represent complex behaviors would provide fundamental insights into how the brain signals emotional pathology. Using machine learning, we discover a spatiotemporal dynamic network that predicts the emergence of major depressive disorder (MDD)-related behavioral dysfunction in mice subjected to chronic social defeat stress. Activity patterns in this network originate in prefrontal cortex and ventral striatum, relay through amygdala and ventral tegmental area, and converge in ventral hippocampus. This network is increased by acute threat, and it is also enhanced in three independent models of MDD vulnerability. Finally, we demonstrate that this vulnerability network is biologically distinct from the networks that encode dysfunction after stress. Thus, these findings reveal a convergent mechanism through which MDD vulnerability is mediated in the brain.

We propose a new task, called Story Visualization. Given a multi-sentence paragraph, the story is visualized by generating a sequence of images, one for each sentence. In contrast to video generation, story visualization focuses less on the continuity in generated images (frames), but more on the global consistency across dynamic scenes and characters -a challenge that has not been addressed by any singleimage or video generation methods. We therefore propose a new story-to-image-sequence generation model, StoryGAN, based on the sequential conditional GAN framework. Our model is unique in that it consists of a deep Context Encoder that dynamically tracks the story flow, and two discriminators at the story and image levels, to enhance the image quality and the consistency of the generated sequences. To evaluate the model, we modified existing datasets to create the CLEVR-SV and Pororo-SV datasets. Empirically, Story-GAN outperforms state-of-the-art models in image quality, contextual consistency metrics, and human evaluation.

Comprehensive descriptions of animal behavior require precise measurements of 3D whole-body movements. Although 2D approaches can track visible landmarks in restrictive environments, performance drops in freely moving animals, due to occlusions and appearance changes. Therefore, we designed DANNCE to robustly track anatomical landmarks in 3D across species and behaviors. DANNCE uses projective geometry to construct inputs to a convolutional neural network that leverages learned 3D geometric reasoning. We trained and benchmarked DANNCE using a 7-million frame dataset that relates color videos and rodent 3D poses. In rats and mice, DANNCE robustly tracked dozens of landmarks on the head, trunk, and limbs of freely moving ‡

While modulating neural activity through stimulation is an effective treatment for neurological diseases such as Parkinson's disease and essential tremor, an opportunity for improving neuromodulation therapy remains in automatically adjusting therapy to continuously optimize patient outcomes. Practical issues associated with achieving this include the paucity of human data related to disease states, poorly validated estimators of patient state, and unknown dynamic mappings of optimal stimulation parameters based on estimated states. To overcome these challenges, we present an investigational platform including: an implanted sensing and stimulation device to collect data and run automated closed-loop algorithms; an external tool to prototype classifier and control-policy algorithms; and real-time telemetry to update the implanted device firmware and monitor its state. The prototyping system was demonstrated in a chronic large animal model studying hippocampal dynamics. We used the platform to find biomarkers of the observed states and transfer functions of different stimulation amplitudes. Data showed that moderate levels of stimulation suppress hippocampal beta activity, while high levels of stimulation produce seizure-like after-discharge activity. The biomarker and transfer function observations were mapped into classifier and control-policy algorithms, which were downloaded to the implanted device to continuously titrate stimulation amplitude for the desired network effect. The platform is designed to be a flexible prototyping tool and could be used to develop improved mechanistic models and automated closed-loop systems for a variety of neurological disorders.