Artificial Neural Network (ANN)-Based Determination of Fractional Contributions from Mixed Fluorophores using Fluorescence Lifetime Measurements.

Here we present an artificial neural network (ANN)-approach to determine the fractional contributions Pi from fluorophores to a multi-exponential fluorescence decay in time-resolved lifetime measurements. Conventionally, Pi are determined by extracting two parameters (amplitude and lifetime) for each underlying mono-exponential decay using non-linear fitting. However, in this case parameter estimation is highly sensitive to initial guesses and weighting. In contrast, the ANN-based approach robustly gives the Pi without knowledge of amplitudes and lifetimes. By experimental measurements and Monte-Carlo simulations, we comprehensively show that accuracy and precision of Pi determination with ANNs and hence the number of distinguishable fluorophores depend on the fluorescence lifetimes' differences. For mixtures of up to five fluorophores, we determined the minimum uniform spacing Δτmin between lifetimes to obtain fractional contributions with a standard deviation of 5%. In example, five lifetimes can be distinguished with a respective minimum uniform spacing of approx. 10 ns even when the fluorophores' emission spectra are overlapping. This study underlines the enormous potential of ANN-based analysis for multi-fluorophore applications in fluorescence lifetime measurements.

ComplexEye: a multi-lens array microscope for high-throughput embedded immune cell migration analysis.

Autonomous migration is essential for the function of immune cells such as neutrophils and plays an important role in numerous diseases. The ability to routinely measure or target it would offer a wealth of clinical applications. Video microscopy of live cells is ideal for migration analysis, but cannot be performed at sufficiently high-throughput (HT). Here we introduce ComplexEye, an array microscope with 16 independent aberration-corrected glass lenses spaced at the pitch of a 96-well plate to produce high-resolution movies of migrating cells. With the system, we enable HT migration analysis of immune cells in 96- and 384-well plates with very energy-efficient performance. We demonstrate that the system can measure multiple clinical samples simultaneously. Furthermore, we screen 1000 compounds and identify 17 modifiers of migration in human neutrophils in just 4 days, a task that requires 60-times longer with a conventional video microscope. ComplexEye thus opens the field of phenotypic HT migration screens and enables routine migration analysis for the clinical setting.

A Cell-Type Selective Stimulation and Recording System for Retinal Ganglion Cells.

Future retinal implants will require a stimulation selectivity between different sub-types of Retinal Ganglion Cells (RGCs) to evoke natural perceptions rather than phosphenes in patients. To achieve this, a cell-type specific stimulation pipeline is required that identifies target RGC sub-types from recorded input images and extracts the specific stimulation parameters to activate this cell-type selectively. Promising biological experiments showed that ON-/OFF- sustained/transient RGCs could be selectively activated by modulating repetition rate and amplitude of an electrical stimulation current in the kilohertz range. This research presents a 42 channel current controlled stimulation and recording system on chip (SoC) with parameter input from a real time target RGC selection algorithm. The SoC is able to stimulate retinal tissue with sinusoidal frequencies higher than 1 kHz at amplitudes of up to 200 µA at a supply voltage of 1.8 V. It also includes tunable recording units with an integrated action potential detection pipeline that are able to amplify signals between 1 Hz and 50 kHz. The required area of one stimulator is 0.0071 mm2, while one recording unit consumes an area of 0.0092 mm2. The application of sinusoidal stimulation currents in the kilohertz range towards retinal tissue leads to a suppressive response of only certain RGC sub-types that has not been oberved before, using electrical stimulation. Because this response is very similar to the natural light response of RGCs, this stimulation approach can lead to a more genuine visual perception for patients using retinal implants.

Label-Free, Impedance-Based Biosensor for Kidney Disease Biomarker Uromodulin.

We demonstrate the development of a label-free, impedance-based biosensor by using a passivation layer of 50-nm tantalum pentoxide (Ta2O5) on interdigitated electrodes (IDE). This layer was fabricated by atomic layer deposition (ALD) and has a high dielectric constant (high-κ), which improves the capacitive property of the IDE. We validate the biosensor's performance by measuring uromodulin, a urine biomarker for kidney tubular damage, from artificial urine samples. The passivation layer is functionalized with uromodulin antibodies for selective binding. The passivated IDE enables the non-faradaic impedance measurement of uromodulin concentrations with a measurement range from 0.5 ng/mL to 8 ng/mL and with a relative change in impedance of 15 % per ng/mL at a frequency of 150 Hz (log scale). This work presents a concept for point-of-care biosensing applications for disease biomarkers.

Current methods for contactless optical patient diagnosis: a systematic review.

Many countries around the world face a shortage of medical personnel, leading to work overload or even burnout. This calls for political and scientific solutions to relieve the medical personnel. The measurement of vital signs in hospitals is still predominately carried out manually with traditional contact-based methods, taking over a substantial share of the medical personnel's workload. The introduction of contactless methods for vital sign monitoring (e.g., with a camera) has great potential to relieve the medical personnel. This systematic review's objective is to analyze the state of the art in the field of contactless optical patient diagnosis. This review distinguishes itself from already existing reviews by considering studies that do not only propose the contactless measurement of vital signs but also include an automatic diagnosis of the patient's condition. This means that the included studies incorporate the physician's reasoning and evaluation of vital signs into their algorithms, allowing an automated patient diagnosis. The literature screening of two independent reviewers resulted in a total of five eligible studies. The highest number of studies (three) introduce methods for the risk assessment of infectious diseases, one study introduces a method for the risk assessment of cardiovascular diseases, and one study introduces a method for the diagnosis of obstructive sleep apnea. Overall, high heterogeneity in relevant study parameters is reported among the included studies. The low number of included studies indicates a large research gap and emphasizes the demand for further research on this emerging topic.

Design of Hardware Accelerators for Optimized and Quantized Neural Networks to Detect Atrial Fibrillation in Patch ECG Device with RISC-V.

Atrial Fibrillation (AF) is one of the most common heart arrhythmias. It is known to cause up to 15% of all strokes. In current times, modern detection systems for arrhythmias, such as single-use patch electrocardiogram (ECG) devices, have to be energy efficient, small, and affordable. In this work, specialized hardware accelerators were developed. First, an artificial neural network (NN) for the detection of AF was optimized. Special attention was paid to the minimum requirements for the inference on a RISC-V-based microcontroller. Hence, a 32-bit floating-point-based NN was analyzed. To reduce the silicon area needed, the NN was quantized to an 8-bit fixed-point datatype (Q7). Based on this datatype, specialized accelerators were developed. Those accelerators included single-instruction multiple-data (SIMD) hardware as well as accelerators for activation functions such as sigmoid and hyperbolic tangents. To accelerate activation functions that require the e-function as part of their computation (e.g., softmax), an e-function accelerator was implemented in the hardware. To compensate for the losses of quantization, the network was expanded and optimized for run-time and memory requirements. The resulting NN has a 7.5% lower run-time in clock cycles (cc) without the accelerators and 2.2 percentage points (pp) lower accuracy compared to a floating-point-based net, while requiring 65% less memory. With the specialized accelerators, the inference run-time was lowered by 87.2% while the F1-Score decreased by 6.1 pp. Implementing the Q7 accelerators instead of the floating-point unit (FPU), the silicon area needed for the microcontroller in 180 nm-technology is below 1 mm2.

High Sensitivity Near-Infrared Imaging of Fluorescent Nanosensors.

Biochemical processes are fast and occur on small-length scales, which makes them difficult to measure. Optical nanosensors based on single-wall carbon nanotubes (SWCNTs) are able to capture such dynamics. They fluoresce in the near-infrared (NIR, 850-1700 nm) tissue transparency window and the emission wavelength depends on their chirality. However, NIR imaging requires specialized indium gallium arsenide (InGaAs) cameras with a typically low resolution because the quantum yield of normal Si-based cameras rapidly decreases in the NIR. Here, an efficient one-step phase separation approach to isolate monochiral (6,4)-SWCNTs (880 nm emission) from mixed SWCNT samples is developed. It enables imaging them in the NIR with high-resolution standard Si-based cameras (>50× more pixels). (6,4)-SWCNTs modified with (GT)10 -ssDNA become highly sensitive to the important neurotransmitter dopamine. These sensors are 1.7× brighter and 7.5× more sensitive and allow fast imaging (<50 ms). They enable high-resolution imaging of dopamine release from cells. Thus, the assembly of biosensors from (6,4)-SWCNTs combines the advantages of nanosensors working in the NIR with the sensitivity of (Si-based) cameras and enables broad usage of these nanomaterials.

Depth-Based Measurement of Respiratory Volumes: A Review.

Depth-based plethysmography (DPG) for the measurement of respiratory parameters is a mobile and cost-effective alternative to spirometry and body plethysmography. In addition, natural breathing can be measured without a mouthpiece, and breathing mechanics can be visualized. This paper aims at showing further improvements for DPG by analyzing recent developments regarding the individual components of a DPG measurement. Starting from the advantages and application scenarios, measurement scenarios and recording devices, selection algorithms and location of a region of interest (ROI) on the upper body, signal processing steps, models for error minimization with a reference measurement device, and final evaluation procedures are presented and discussed. It is shown that ROI selection has an impact on signal quality. Adaptive methods and dynamic referencing of body points to select the ROI can allow more accurate placement and thus lead to better signal quality. Multiple different ROIs can be used to assess breathing mechanics and distinguish patient groups. Signal acquisition can be performed quickly using arithmetic calculations and is not inferior to complex 3D reconstruction algorithms. It is shown that linear models provide a good approximation of the signal. However, further dependencies, such as personal characteristics, may lead to non-linear models in the future. Finally, it is pointed out to focus developments with respect to single-camera systems and to focus on independence from an individual calibration in the evaluation.

Advantages and Limitations of Fluorescence Lifetime Measurements Using Single-Photon Avalanche Diode (SPAD) Array Detector: A Comprehensive Theoretical and Experimental Study.

Fast fluorescence lifetime (FL) determination is a major factor for studying dynamic processes. To achieve a required precision and accuracy a certain number of photon counts must be detected. FL methods based on single-photon counting have strongly limited count rates because of the detector's pile-up issue and are suffering from long measurement times in the order of tens of seconds. Here, we present an experimental and Monte Carlo simulation-based study of how this limitation can be overcome using array detectors based on single-photon avalanche diodes (SPADs). We investigated the maximum count rate per pixel to determine FL with a certain precision and accuracy before pile-up occurs. Based on that, we derived an analytical expression to calculate the total measurement time which is proportional to the FL and inversely proportional to the number of pixels. However, a higher number of pixels drastically increases data rate. This can be counteracted by lowering the time resolution. We found that even with a time resolution of four times the FL, an accuracy of 10% can be achieved. Taken all together, FLs between 10 ns and 3 ns can be determined with a 300-pixel SPAD array detector with a measurement time and data rate less than 1 µs and 700 Mbit/s, respectively. This shows the enormous potential of SPAD array detector for high-speed applications requiring continuous data read out.

In vivo validation of the electronic depth control probes.

In this article, we evaluated the electrophysiological performance of a novel, high-complexity silicon probe array. This brain-implantable probe implements a dynamically reconfigurable voltage-recording device, coordinating large numbers of electronically switchable recording sites, referred to as electronic depth control (EDC). Our results show the potential of the EDC devices to record good-quality local field potentials, and single- and multiple-unit activities in cortical regions during pharmacologically induced cortical slow wave activity in an animal model.

Approaches for drug delivery with intracortical probes.

Intracortical microprobes allow the precise monitoring of electrical and chemical signaling and are widely used in neuroscience. Microelectromechanical system (MEMS) technologies have greatly enhanced the integration of multifunctional probes by facilitating the combination of multiple recording electrodes and drug delivery channels in a single probe. Depending on the neuroscientific application, various assembly strategies are required in addition to the microprobe fabrication itself. This paper summarizes recent advances in the fabrication and assembly of micromachined silicon probes for drug delivery achieved within the EU-funded research project NeuroProbes. The described fabrication process combines a two-wafer silicon bonding process with deep reactive ion etching, wafer grinding, and thin film patterning and offers a maximum in design flexibility. By applying this process, three general comb-like microprobe designs featuring up to four 8-mm-long shafts, cross sections from 150×200 to 250×250 µm², and different electrode and fluidic channel configurations are realized. Furthermore, we discuss the development and application of different probe assemblies for acute, semichronic, and chronic applications, including comb and array assemblies, floating microprobe arrays, as well as the complete drug delivery system NeuroMedicator for small animal research.

Enhancing the yield of high-density electrode arrays through automated electrode selection.

Recently developed CMOS-based microprobes contain hundreds of electrodes on a single shaft with inter-electrode distances as small as 30 µm. So far, neuroscientists needed to select electrodes manually from hundreds of electrodes. Here we present an electronic depth control algorithm that allows to select electrodes automatically, hereby allowing to reduce the amount of data and locating those electrodes that are close to neurons. The electrodes are selected according to a new penalized signal-to-noise ratio (PSNR) criterion that demotes electrodes from becoming selected if their signals are redundant with previously selected electrodes. It is shown that, using the PSNR, interneurons generating smaller spikes are also selected. We developed a model that aims to evaluate algorithms for electronic depth control, but also generates benchmark data for testing spike sorting and spike detection algorithms. The model comprises a realistic tufted pyramidal cell, non-tufted pyramidal cells and inhibitory interneurons. All neurons are synaptically activated by hundreds of fibers. This arrangement allows the algorithms to be tested in more realistic conditions, including backgrounds of synaptic potentials, varying spike rates with bursting and spike amplitude attenuation.

Control and data acquisition software for high-density CMOS-based microprobe arrays implementing electronic depth control.

This paper presents the NeuroSelect software for managing the electronic depth control of cerebral CMOS-based microprobes for extracellular in vivo recordings. These microprobes contain up to 500 electronically switchable electrodes which can be appropriately selected with regard to specific neuron locations in the course of a recording experiment. NeuroSelect makes it possible to scan the electrodes electronically and to (re)select those electrodes of best signal quality resulting in a closed-loop design of a neural acquisition system. The signal quality is calculated by the relative power of the spikes compared with the background noise. The spikes are detected by an adaptive threshold using a robust estimator of the standard deviation. Electrodes can be selected in a manual or semi-automatic mode based on the signal quality. This electronic depth control constitutes a significant improvement for multielectrode probes, given that so far the only alternative has been the fine positioning by mechanical probe translation. In addition to managing communication with the hardware controller of the probe array, the software also controls acquisition, processing, display and storage of the neural signals for further analysis.