An Effective Deep Learning Framework for Fall Detection: Model Development and Study Design.

BACKGROUND: Fall detection is of great significance in safeguarding human health. By monitoring the motion data, a fall detection system (FDS) can detect a fall accident. Recently, wearable sensors-based FDSs have become the mainstream of research, which can be categorized into threshold-based FDSs using experience, machine learning-based FDSs using manual feature extraction, and deep learning (DL)-based FDSs using automatic feature extraction. However, most FDSs focus on the global information of sensor data, neglecting the fact that different segments of the data contribute variably to fall detection. This shortcoming makes it challenging for FDSs to accurately distinguish between similar human motion patterns of actual falls and fall-like actions, leading to a decrease in detection accuracy. OBJECTIVE: This study aims to develop and validate a DL framework to accurately detect falls using acceleration and gyroscope data from wearable sensors. We aim to explore the essential contributing features extracted from sensor data to distinguish falls from activities of daily life. The significance of this study lies in reforming the FDS by designing a weighted feature representation using DL methods to effectively differentiate between fall events and fall-like activities. METHODS: Based on the 3-axis acceleration and gyroscope data, we proposed a new DL architecture, the dual-stream convolutional neural network self-attention (DSCS) model. Unlike previous studies, the used architecture can extract global feature information from acceleration and gyroscope data. Additionally, we incorporated a self-attention module to assign different weights to the original feature vector, enabling the model to learn the contribution effect of the sensor data and enhance classification accuracy. The proposed model was trained and tested on 2 public data sets: SisFall and MobiFall. In addition, 10 participants were recruited to carry out practical validation of the DSCS model. A total of 1700 trials were performed to test the generalization ability of the model. RESULTS: The fall detection accuracy of the DSCS model was 99.32% (recall=99.15%; precision=98.58%) and 99.65% (recall=100%; precision=98.39%) on the test sets of SisFall and MobiFall, respectively. In the ablation experiment, we compared the DSCS model with state-of-the-art machine learning and DL models. On the SisFall data set, the DSCS model achieved the second-best accuracy; on the MobiFall data set, the DSCS model achieved the best accuracy, recall, and precision. In practical validation, the accuracy of the DSCS model was 96.41% (recall=95.12%; specificity=97.55%). CONCLUSIONS: This study demonstrates that the DSCS model can significantly improve the accuracy of fall detection on 2 publicly available data sets and performs robustly in practical validation.

Biodegradable Flexible Electronic Device with Controlled Drug Release for Cancer Treatment.

Nowadays, controllable drug release is a vitally important strategy for cancer treatment and usually realized using implanting biocompatible devices. However, these devices need to be removed by another surgery after the function fails, which brings the risks of inflammation or potential death. In this article, a biodegradable flexible electronic device with controllable drug (paclitaxel) release was proposed for cancer treatment. The device is powered by an external alternating magnetic field to generate internal resistance heat and promote drug release loaded on the substrate. Moreover, the device temperature can even reach to 65 °C, which was sufficient for controllable drug release. This device also has similar mechanical properties to human tissues and can autonomously degrade due to the structure design of the circuit and degradable compositions. Finally, it is confirmed that the device has a good inhibitory effect on the proliferation of breast cancer cells (MCF-7) and could be completely degraded in vitro. Thus, its great biodegradability and conformity can relieve patients of second operation, and the device proposed in this paper provides a promising solution to complete conquest of cancer in situ.

Climbing-inspired twining electrodes using shape memory for peripheral nerve stimulation and recording.

Peripheral neuromodulation has been widely used throughout clinical practices and basic neuroscience research. However, the mechanical and geometrical mismatches at current electrode-nerve interfaces and complicated surgical implantation often induce irreversible neural damage, such as axonal degradation. Here, compatible with traditional 2D planar processing, we propose a 3D twining electrode by integrating stretchable mesh serpentine wires onto a flexible shape memory substrate, which has permanent shape reconfigurability (from 2D to 3D), distinct elastic modulus controllability (from ~100 MPa to ~300 kPa), and shape memory recoverability at body temperature. Similar to the climbing process of twining plants, the temporarily flattened 2D stiff twining electrode can naturally self-climb onto nerves driven by 37°C normal saline and form 3D flexible neural interfaces with minimal constraint on the deforming nerves. In vivo animal experiments, including right vagus nerve stimulation for reducing the heart rate and action potential recording of the sciatic nerve, demonstrate the potential clinical utility.

Reversible Semicrystalline Polymer as Actuators Driven by Organic Solvent Vapor.

A semicrystalline polymer actuator, which is responsive to solvent vapor with fast and large scale locomotion, is described. The thermoset semicrystalline polymer can be easily synthesized from crystallizable polyester segment poly (ε-caprolactone) and isophorone diisocyanate trimer. Organic solvent vapor is used to induce the reversible swelling-crystallization conversion of the crystallizable polyester segment, resulting in its expansion/shrinkage. The contraction of the polymer actuator (1 mm thick) needs only ≈4 s in room temperature. When exposed to air the polymer actuator can exhibit a fast self-oscillation. Then, a soft crawler based on this polymer is demonstrated. Driven by organic solvent it walks rapidly and steadily. The microscope images show the fast swelling-crystallization conversion that gives rise to reversible shape changes of the polymer.

[High sensitivity conherent anti-stokes Raman scattering for linearly probing low concentration materials].

The authors made a theoretical analysis and experiment research on the relation of time-resolved coherent anti-Stokes Raman scattering (T-CARS) intensity and the sample concentrations in this paper. It was proved experimentally that the T-CARS intensity is quadratic at the concentration higher than 35%, but is linear with the sample concentration at the concentration lower than 20%, which fits with theoretical analysis. And the research results correct inaccurate previous perceptions, which is conducive to better interpretation and application of the CARS process. The linear relation between the intensity of the CARS with the sample concentration at low concentrations indicates that the CARS is allowed for direct and precise concentration measurements, therefore it will be of great importance in biology and biochemistry.