Evaluation of a Machine Learning Algorithm to Classify Ultrasonic Transducer Misalignment and Deployment Using TinyML

The challenge for ultrasonic (US) power transfer systems, in implanted/wearable medical devices, is to determine when misalignment occurs (e.g., due to body motion) and apply directional correction accordingly. In this study, a number of machine learning algorithms were evaluated to classify US transducer misalignment, based on data signal transmissions between the transmitter and receiver. Over seven hundred US signals were acquired across a range of transducer misalignments. Signal envelopes and spectrograms were used to train and evaluate machine learning (ML) algorithms, classifying misalignment extent. The algorithms included an autoencoder, convolutional neural network (CNN) and neural network (NN). The best performing algorithm, was deployed onto a TinyML device for evaluation. Such systems exploit low power microcontrollers developed specifically around edge device applications, where algorithms were configured to run on low power, restricted memory systems. TensorFlow Lite and Edge Impulse, were used to deploy trained models onto the edge device, to classify signals according to transducer misalignment extent. TinyML deployment, demonstrated near real-time (<350 ms) signal classification achieving accuracies > 99%. This opens the possibility to apply such ML alignment algorithms to US arrays (capacitive micro-machined ultrasonic transducer (CMUT), piezoelectric micro-machined ultrasonic transducer (PMUT) devices) capable of beam-steering, significantly enhancing power delivery in implanted and body worn systems.

Machine learning based canine posture estimation using inertial data

The aim of this study was to design a new canine posture estimation system specifically for working dogs. The system was composed of Inertial Measurement Units (IMUs) that are commercially available, and a supervised learning algorithm which was developed for different behaviours. Three IMUs, each containing a 3-axis accelerometer, gyroscope, and magnetometer, were attached to the dogs' chest, back, and neck. To build and test the model, data were collected during a video-recorded behaviour test where the trainee assistance dogs performed static postures (standing, sitting, lying down) and dynamic activities (walking, body shake). Advanced feature extraction techniques were employed for the first time in this field, including statistical, temporal, and spectral methods. The most important features for posture prediction were chosen using Select K Best with ANOVA F-value. The individual contributions of each IMU, sensor, and feature type were analysed using Select K Best scores and Random Forest feature importance. Results showed that the back and chest IMUs were more important than the neck IMU, and the accelerometers were more important than the gyroscopes. The addition of IMUs to the chest and back of dog harnesses is recommended to improve performance. Additionally, statistical and temporal feature domains were more important than spectral feature domains. Three novel cascade arrangements of Random Forest and Isolation Forest were fitted to the dataset. The best classifier achieved an f1-macro of 0.83 and an f1-weighted of 0.90 for the prediction of the five postures, demonstrating a better performance than previous studies. These results were attributed to the data collection methodology (number of subjects and observations, multiple IMUs, use of common working dog breeds) and novel machine learning techniques (advanced feature extraction, feature selection and modelling arrangements) employed. The dataset and code used are publicly available on Mendeley Data and GitHub, respectively.

Electrochemical sensor for enzymatic lactate detection based on laser-scribed graphitic carbon modified with platinum, chitosan and lactate oxidase

We developed a flexible laser scribed graphitic carbon based lactate biosensor fabricated using a low cost 450 nm laser. We demonstrated a facile fabrication method involving electrodeposition of platinum followed by two casting steps for modification with chitosan and lactate oxidase. The biosensor demonstrated chronoamperometric lactate detection within a linear range from 0.2 mM to 3 mM, (R² > 0.99), with a limit of detection of 0.11 mM and a sensitivity of 35.8 μA/mM/cm². The biosensor was successful in performing up to 10 consecutive measurements (one after the other) indicating good working stability (RSD <5%). Concerning storage stability, there was no decrease in signal response after 30 days of storage at 4 °C. Additionally, we demonstrate enzymatic lactate detection whilst the flexible polyimide substrates were fixed at a curvature (K) of 0.14 mm^-1. No noticeable change in signal response was observed in comparison to calibrations obtained at a curvature of 0 mm^-1, signifying potential opportunities for sensor attachment or integration with oral-care products such as mouth swabs. Both laser scribed graphitic carbon and Ag/AgCl modified-laser scribed graphitic carbon were successful as reference electrodes for chronoamperometric lactate measurements. Furthermore, using a three-electrode configuration on polyimide, lactate detection in both artificial saliva and sterile human serum samples was achieved for two spiked concentrations (0.5 mM and 1 mM).

Immunosensor for Assessing the Welfare of Trainee Guide Dogs

Cortisol is a well established biomarker hormone that regulates many processes in the body and is widely referred to as the stress hormone. Cortisol can be used as a stress marker to allow for detection of stress levels in dogs during the training process. This test will indicate if they will handle the stress under the training or if they might be more suitable as an assistant or companion dog. An immunosensor for detection of cortisol was developed using electrochemical impedance spectroscopy (EIS). The sensor was characterized using chemical and topographical techniques. The sensor was calibrated and its sensitivity determined using a cortisol concentration range of 0.0005 to 50 μg/mL. The theoretical limit of detection was found to be 3.57 fg/mL. When the immunosensor was tested on canine saliva samples, cortisol was detected and measured within the relevant physiological ranges in dogs.

On-Chip Glucose Detection Based on Glucose Oxidase Immobilized on a Platinum-Modified, Gold Microband Electrode

We report the microfabrication and characterization of gold microband electrodes on silicon using standard microfabrication methods, i.e., lithography and etching techniques. A two-step electrodeposition process was carried out using the on-chip platinum reference and gold counter electrodes, thus incorporating glucose oxidase onto a platinum-modified, gold microband electrode with an o-phenylenediamine and ß-cyclodextrin mixture. The as-fabricated electrodes were studied using optical microscopy, scanning electron microscopy, and atomic force microscopy. The two-step electrodeposition process was conducted in low sample volumes (50 µL) of both solutions required for biosensor construction. Cyclic voltammetry and electrochemical impedance spectroscopy were utilised for electrochemical characterization at each stage of the deposition process. The enzymatic-based microband biosensor demonstrated a linear response to glucose from 2.5-15 mM, using both linear sweep voltammetry and chronoamperometric measurements in buffer-based solutions. The biosensor performance was examined in 30 µL volumes of fetal bovine serum. Whilst a reduction in the sensor sensitivity was evident within 100% serum samples (compared to buffer media), the sensor demonstrated linear glucose detection with increasing glucose concentrations (5-17 mM).

'Smart' wound dressings for advanced wound care: a review

Hard-to-heal wounds are a common side-effect of diabetes, obesity, pressure ulcers and age-related vascular diseases, the incidences of which are growing worldwide. The increasing financial burden of hard-to-heal wounds on global health services has provoked technological research into improving wound diagnostics and therapeutics via 'smart' dressings, within which elements such as microelectronic sensors, microprocessors and wireless communication radios are embedded. This review highlights the progress being made by research groups worldwide in producing 'smart' wound device prototypes. Significant advances have been made, for example, flexible substrates have replaced rigid circuit boards, sensors have been printed on commercial wound dressing materials and wireless communication has been demonstrated. Challenges remain, however, in the areas of power supply, disposability, low-profile components, multiparametric sensing and seamless device integration in commercial wound dressings.

A Review of Wearable Solutions for Physiological and Emotional Monitoring for Use by People with Autism Spectrum Disorder and Their Caregivers

The goal of real-time feedback on physiological changes, stress monitoring and even emotion detection is becoming a technological reality. People in their daily life experience varying emotional states, some of which are negative and which can lead to decreased attention, decreased productivity and ultimately, reduced quality of life. Therefore, having a solution that continuously monitors the physiological signals of the person and assesses his or her emotional well-being could be a very valuable tool. This paper aims to review existing physiological and motional monitoring devices, highlight their features and compare their sensing capabilities. Such technology would be particularly useful for certain populations who experience rapidly changing emotional states such as people with autism spectrum disorder and people with intellectual disabilities. Wearable sensing devices present a potential solution that can support and complement existing behavioral interventions. This paper presents a review of existing and emerging products in the market. It reviews the literature on state-of-the-art prototypes and analyzes their usefulness, clinical validity, and discusses clinical perspectives. A small number of products offer reliable physiological internal state monitoring and may be suitable for people with Autism Spectrum Disorder (ASD). It is likely that more promising solutions will be available in the near future. Therefore, caregivers should be careful in their selection of devices that meet the care-receiver's personal needs and have strong research support for reliability and validity.

Emerging technologies for point-of-care genetic testing

In the coming years, genetic test results will be increasingly used as indicators that influence medical decision making. Novel instrumentation that is able to detect relevant mutations in a point-of-care setting is being developed to facilitate this increase, frequently as a spin-off from recent research in the area of biothreat monitoring. This market review will describe the current generation of instrumentation that is most suitable for use in a point-of-care setting; it will also try to identify some of the technologies that will make-up the next generation of instrumentation currently being prepared for the market.

Impact of surface nano-textured stainless steel prepared by focused ion beam on endothelial cell growth

The modification of stent surfaces with nano-structures has the potential for limiting late stent restenosis. We report here the patterning of 316L austentitic stainless steel with arrays of nano-pits of two nominal diameters: 120 and 180 nm. These nano-textured surfaces were prepared by focused ion beam milling. The influence of the ion beam current on the nano-features was investigated by scanning electron and atomic force microscopies. The optimum ion beam currents were 280 pA for 120 nm nano-pits and 920 pA for 180 nm nano-pits. The depths of the nano-pits formed were (65 +/- 24) nm (120 nm) and (84 +/- 36) nm (180 nm). This wide distribution of the depths is due to the polycrystalline nature of 316 L stainless steel, which has a strong influence on the milling rates. Endothelial cells were grown in vitro on these substrates for 1, 3 and 5 days. The cells were viable for the duration of the cell culture on the nano-textured substrates. There was no significant difference in the adhesion and the proliferation based on the nano-pit diameter.

Magnetic core-shell nanoparticles for drug delivery by nebulization

BACKGROUND

Aerosolized therapeutics hold great potential for effective treatment of various diseases including lung cancer. In this context, there is an urgent need to develop novel nanocarriers suitable for drug delivery by nebulization. To address this need, we synthesized and characterized a biocompatible drug delivery vehicle following surface coating of Fe3O4 magnetic nanoparticles (MNPs) with a polymer poly(lactic-co-glycolic acid) (PLGA). The polymeric shell of these engineered nanoparticles was loaded with a potential anti-cancer drug quercetin and their suitability for targeting lung cancer cells via nebulization was evaluated.

RESULTS

Average particle size of the developed MNPs and PLGA-MNPs as measured by electron microscopy was 9.6 and 53.2 nm, whereas their hydrodynamic swelling as determined using dynamic light scattering was 54.3 nm and 293.4 nm respectively. Utilizing a series of standardized biological tests incorporating a cell-based automated image acquisition and analysis procedure in combination with real-time impedance sensing, we confirmed that the developed MNP-based nanocarrier system was biocompatible, as no cytotoxicity was observed when up to 100 μg/ml PLGA-MNP was applied to the cultured human lung epithelial cells. Moreover, the PLGA-MNP preparation was well-tolerated in vivo in mice when applied intranasally as measured by glutathione and IL-6 secretion assays after 1, 4, or 7 days post-treatment. To imitate aerosol formation for drug delivery to the lungs, we applied quercitin loaded PLGA-MNPs to the human lung carcinoma cell line A549 following a single round of nebulization. The drug-loaded PLGA-MNPs significantly reduced the number of viable A549 cells, which was comparable when applied either by nebulization or by direct pipetting.

CONCLUSION

We have developed a magnetic core-shell nanoparticle-based nanocarrier system and evaluated the feasibility of its drug delivery capability via aerosol administration. This study has implications for targeted delivery of therapeutics and poorly soluble medicinal compounds via inhalation route.

Three-dimensional hydrogel structures as optical sensor arrays, for the detection of specific DNA sequences

The fabrication and characterization of surface-attached PEG-diacrylate hydrogel structures and their application as sensing platforms for the detection of specific target sequences are reported. Hydrogel structures were formed by a photopolymerization process, using substrate-bound Eosin Y molecules for the production of free radicals. We have demonstrated that this fabrication process allows for control over hydrogel growth down to the micrometer scale. Confocal imaging revealed relatively large pore structures for 25% (v/v) PEG-diacrylate hydrogels, which appear to lie in tightly packed layers. Our data suggest that these pore structures decrease in size for hydrogels with increasing levels of PEG-diacrylate. Surface coverage values calculated for hydrogels immobilized with 21-mer DNA probe sequences were significantly higher compared to those previously reported for 2- and 3-dimensional sensing platforms, on the order of 10(16)molecules cm(-2). Used as sensing platforms in DNA hybridization assays, a detection limit of 3.9 nM was achieved for hybridization reactions between 21-mer probe and target sequences. The ability of these hydrogel sensing platforms to discriminate between wild-type and mutant allele sequences was also demonstrated, down to target concentrations of 1-2 nM. A reduction in the hybridization time down to a period of 15 min was also achieved, while still maintaining confident results, demonstrating the potential for future integration of these sensing platforms within Lab-on-Chip or diagnostic devices.

Selective Release of DNA from the Surface of Indium−Tin Oxide Thin Electrode Films Using Thiol−Disulfide Exchange Chemistry

A new challenge in biointerfacial science is the development of dynamic surfaces with the ability to adjust and tune the chemical functionality at the interface between the biological and nonbiological entities. In this paper we describe fabrication of indium-tin oxide (ITO) electrodes and the design of a ligand that can be switched to enable selectively controlled interactions with DNA. Tailoring the surface composition of the ITO electrode to optimize its optical and electrical properties was also studied. The surface attachment chemistry investigated utilizes thiol-disulfide exchange chemistry. This chemistry involved the covalent attachment of a thiol-functionalized silane anchor to a hydroxyl-activated ITO electrode surface. Subsequent reaction with 2-(2-pyridinyldithio)ethanamine hydrochloride formed the disulfide bridge and provided the terminal amine group, which facilitates addition of a cross-linker. DNA was then covalently bound to the cross-linker, and hybridization with the complementary Cy3-labeled target DNA was achieved. Selective release of the attached DNA was demonstrated by both chemical and electrical reduction of the disulfide bond. The surface chemistry was then recycled, and rehybridization of the target DNA was achieved. The ability to control specific release of biomolecules has applications for the development of novel biosensor platforms and a range of medical devices.

Multiplex Analysis with Peptide-Encoded Beads

Paramagnetic beads have considerable potential as identification tags in biological analysis. For example, magnetic sensor-based arrays using the magnetic field generated by paramagnetic beads to test hybridization between interacting molecules have attracted widespread interest in recent years. However, application of paramagnetic beads as identification tags is still limited, since they do not permit differentiation between samples for multiplex analysis. Here, we report the application of a novel encoding of paramagnetic beads with peptide sequences. This strategy allows DNA samples labeled with peptide-encoded paramagnetic beads to be identified by the selective enzymatic cleavage of each peptide cross-linker.

Evaluation of silicon and polymer substrates for fabrication of integrated microfluidic microsystems for DNA extraction and amplification

This paper is presenting two different alternatives for the DNA extraction and amplification that will be carried out by two competitive research projects developing bioanalytical microsystems with microfluidics. The first project will develop the microfluidics part on polymer material and the other one on silicon. The polymer approach is currently under development based on a modular microfluidic architecture aimed to simplify the process of designing and building such a microsystem device. A silicon alternative is about to start and is expected to decrease packaging costs of the microsystem allowing future manufacturability of the device.

Microarray analysis in cystic fibrosis

DNA microarrays provide a versatile platform for applications including gene expression analysis and genotyping. In the case of cystic fibrosis (CF), DNA microarrays enable the measurement of gene expression levels of thousands of genes in parallel, and potentially therefore, to identify non-CFTR genes down- or up-regulated in CF, which could lead to insights into disease pathophysiology, as well as novel molecular markers and therapeutic strategies. Moreover, using optimised microarray protocols based on either primer extension analysis (i.e. minisequencing) or electronic hybridisation stringency control, the potential now exists to detect all relevant CFTR mutations on a single DNA microarray as a novel platform for CF screening.

A nanobiotechnology roadmap for high-throughput single nucleotide polymorphism analysis

Genetic analysis based on single nucleotide polymorphisms (SNPs) has the potential to enable identification of genes associated with disease susceptibility, to facilitate improved understanding and diagnosis of those diseases, and should ultimately contribute to the provision of new therapies. To achieve this end, new technology platforms are required that can increase genotyping throughput, while simultaneously reducing costs by as much as two orders of magnitude. Development of a variety of genotyping platforms with the potential to resolve this dilemma is already well advanced through research in the field of nanobiotechnology. Novel approaches to DNA extraction and amplification have reduced the times required for these processes to seconds. Microfluidic devices enable polymorphism detection through very rapid fragment separation using capillary electrophoresis and high-performance liquid chromatography, together with mixing and transport of reagents and biomolecules in integrated systems. The potential for application of established microelectronic fabrication processes to genetic analyses systems has been demonstrated (e.g. photolithography-based in situ synthesis of oligonucleotides on microarrays). Innovative application of state-of-the-art photonics and integrated circuitry are leading to improved detection capabilities. The diversity of genotyping applications envisaged in the future, ranging from the very high-throughput requirements for drug discovery through to rapid and cheap near-patient genotype analysis, suggests that several SNP genotyping platforms will be necessary to optimally address the different niches.

PCR amplification of a polymorphic minisatellite VNTR locus in whiting (Merlangius merlangus L.)

An approach has been developed for the screening of allelic variation of minisatellite DNA loci that substantially reduces the time and hazards involved. Primers were designed for a minisatellite region isolated from a gadoid fish species (Merlangius merlangus L.), enabling amplification by polymerase chain reaction, so that differences in the number of minisatellite repeat units (allelic variability) were detectable by ethidium bromide fluorescence (over UV light) following separation by agarose gel electrophoresis. This amplifiable minisatellite variable number tandem repeat region, the first non-primate marker of its kind can be used successfully with DNA extracted by a rapid Chelex protocol. From a sample of 97 individuals, 24 alleles were resolved (750-2200 kb) and heterozygosity was estimated at 0.94.

Impaired glucose tolerance after endurance exercise is associated with reduced insulin secretion rather than altered insulin sensitivity

Paired frequently sampled intravenous glucose tolerance tests (FSIGT) were performed on five highly trained athletes within 2 hours of completing a 6-day ultramarathon run (E) and after 2 weeks of complete rest (R). Severe exercise increased free fatty acid (FFA) levels (E 1.2 +/- 0.16 v 0.42 +/- 0.07 mmol/L, P < .01) and norepinephrine levels (E 573 +/- 141 v 224 +/- 33 pg/mL, P < .01), with only moderate reductions in glucose tolerance (glucose disappearance [Kg] E 1.06 +/- 0.2 v R 1.7 +/- 0.3 min-1 x 10(2), P < .05). The minimal model analysis of FSIGT data using the method of Bergman et al (Endocr Rev 6:45-86, 1985) showed a reduced second-phase insulin secretion ([phi 2] E 5.2 +/- 1.3 v 13 +/- 2.2 microU/mL.min-2 per mg/dL, P < .05) and glucose disposition index ([SI x phi 2] E 33.8 +/- 10 v 73.9 +/- 11 mg-1.dL.min-3 x 10(4), P < .02). Insulin sensitivity (SI) and glucose-mediated glucose disposal (SG) were unchanged (SI E 6.9 +/- 1.0 v 6.0 +/- 0.6 min-1 per microU/mL x 10(4); SG E 1.8 +/- 0.6 v 1.4 +/- 0.3 min-1 x 10(2)). Reduced glucose tolerance after prolonged extreme physical exercise was accompanied by reduced phi 2 and not by alterations of SI or SG, despite the marked increase of FFA levels. Elevated norepinephrine levels, reflecting activation of the sympathetic noradrenergic system, was also associated with the reduction in Kg. The reduction in phi 2 would promote mobilization of FFA, the predominant metabolic substrate in these endurance events.

A simple method for quantitation of insulin sensitivity and insulin release from an intravenous glucose tolerance test

Both insulin secretion and insulin sensitivity are important in the development of diabetes but current methods used for their measurements are complex and cannot be used for epidemiological surveys. This study describes a simplified approach for the estimation of first phase insulin release and insulin sensitivity from a standard 40-min intravenous glucose tolerance test (IVGTT), and compares these parameter estimations with the sophisticated minimal model analysis of a frequently sampled 3-h IVGTT and the euglycaemic clamp technique. For the simplified IVGTT, first phase insulin release was measured as the insulin area above basal post glucose load unit-1 incremental change (i.e. peak rise) in plasma glucose over 0-10 min, and insulin sensitivity as a rate of glucose disappearance (Kg) unit-1 insulin increase above basal from 0-40 min post-glucose load in 18 subjects who were studied twice, either basally or in a perturbed pathophysiological state (i.e. pre- and post-ultramarathon race, n = 5; pre- and post-20 h pulsatile hyperinsulinaemia, n = 8; pre- and post-thyrotoxic state, n = 5). A further 12 subjects were compared by IVGTT, and glucose clamp. In addition, seven dogs were studied three times by IVGTT during normal saline infusion and after short-term (1/2 hour) or long-term (72 hour) adrenaline infusions. First phase insulin release and insulin sensitivity estimated from the simplified IVGTT as calculated by the two methods correlated closely (rs = 0.89 and rs = 0.87, respectively), although less precisely in markedly insulin-resistant subjects and the slopes and y intercepts of the linear regression lines were similar in the basal and perturbed states.(ABSTRACT TRUNCATED AT 250 WORDS)