We describe an inexpensive, open-source rotating polariser, rotating analyser variable angle reflective ellipsometer (VARE), suitable for measuring the generalised ellipsometric angles of anisotropic materials. With a build cost of $800 and substantial scope for customisation, the instrument offers functionality typically associated with far costlier commercial instruments. Phase alignment of the rotating polarisers is achieved by a phase-locked loop, while the ellipsometric angles are determined by Fourier analysis of the reflected signal, using an integrated microcontroller-based lock-in amplifier. The performance of the instrument is compared against a high-end commercial instrument.

Functional neuroimaging systems are becoming more lightweight and portable in nature, which is valuable to conduct studies in naturalistic - such as homes, schools, and offices – and low-resource settings. However, most portable neuroimaging systems require skilled experts to ensure accurate functionality and reliable data collection. Our group has developed a new strategy for neuroimager alignment and coupling evaluation (NACE) that leverages augmented reality to allow lay users to independently ensure that an fNIRS device is correctly aligned and properly coupled to their scalp. The strategy would enable widespread adoption of prefrontal cortex neuroimaging data collected in naturalistic and low-resource settings.

Aptamers have shown considerable promise as chemical recognition elements in biosensor systems because they can be chemically synthesized and can deliver high stability, sensitivity, and specificity. However, since native aptamers do not inherently exhibit signaling activity upon target recognition, engineering approaches are needed to introduce signal transduction functionality so that they can generate a measurable output signal upon target binding. A number of signaling modalities have been explored to date including electrochemical aptamer-based (E-AB) sensors, which measure changes in current that result from target-binding-induced conformational changes in redox-tagged aptamer molecules. E-AB sensors have repeatedly demonstrated the potential for continuous detection over extended periods of time in vivo with high specificity and stability . However, the adaptation of existing aptamers for electrochemical platforms remains a bottleneck7, since aptamers identified through SELEX-based strategies will not necessarily undergo a meaningful conformational change upon target binding, and there is no guarantee that a promising aptamer can be readily engineered to generate an E-AB sensor that yields a sufficient signal response within the desired target concentration range. This limitation is one of the reasons why E-AB sensor implementation has been largely limited to a few target molecules. Fluorescence-based optical readouts are of particular interest as they offer the capability to achieve single-photon sensitivity with appropriate detection hardware, can readily be multiplexed through fluorophores that emit at different wavelengths, and can accelerate the transition from aptamer selection to sensor development. Aptamer beacons are a popular design strategy for fluorescence-based detection, in which an aptamer is labeled with a fluorophore and then combined with a separate, quencher-functionalized strand that directly competes with target binding, such that a fluorescence signal is only generated in response to target binding. However, this method suffers from a major drawback in the context of biosensor applications in that it cannot be used for continuous measurements, as the quenching sequence is unlikely to rehybridize once released. As an alternative, Tang et al. used a polyethylene glycol (PEG) moiety to couple aptamer sequences to a short complementary sequence, which were respectively labeled with a fluorophore and quencher. More recently, our lab devised another effective fluorescent sensor design based on intramolecular strand displacement (ISD) molecular switches, in which the complementary strand is physically coupled to the aptamer strand via a flexible linker. This strategy enables reversible switching and by modulating the lengths of the linker and the complementary region, one can achieve independent tuning of both the kinetics and thermodynamics of the resulting sensor. However, both approaches face important limitations in that they require considerable trial and error optimization to achieve an ideal balance between minimizing the background signal and ensuring efficient target binding-induced displacement of the complementary displacement strand. Optimization typically requires detailed characterization of the aptamer’s structure because post-selection engineering of the aptamer often compromises its affinity or specificity. For these reasons, most aptamer switch designs to date have been prototyped with a handful of well-characterized aptamers. Multiple rational or semi-rational approaches have been developed to date to engineer aptamers to undergo a binding-induced conformational change, including enzymatic5 and truncation approaches. For newly discovered aptamers, however, these engineering processes remain challenging as it does not always yield a sensor with a sufficient signal response across the desired target concentration range. There is a need for a universal and systematic approach for reliably endowing switching functionality into aptamers to enable the efficient generation of biosensors for diverse molecular targets. In this work, we describe a novel duplex-bubble switch (DBS) aptamer switch architecture, wherein any aptamer switch generated via Capture-SELEXcan be engineered in order to tune its kinetic and thermodynamic response without a priori knowledge of its structure or binding domain. Importantly, the DBS enables reversible switching, and we demonstrate that the architecture is suitable for use in continuous detection applications, achieving a rapid and reversible binding-induced fluorescent response at a time-scale of seconds even in complex biological matrices. We first show mathematically and experimentally that we can fine-tune the DBS design to precisely adjust the thermodynamics and kinetics of the resulting aptamer switches, both in solution and in a surface-coupled assay format. We then show that we can incorporate these DBS constructs onto a fiber-optic probe integrated with single-photon-counting hardware, exploiting a sensor design that allows us to achieve direct measurement of analytes in complex biological matrices while rejecting background autofluorescence from interferents and minimizing the impact of fouling. Specifically, we demonstrate continuous detection of dopamine in both buffer and artificial cerebrospinal fluid (aCSF) for more than a day with fast time resolution (< 3 s) and a dynamic range spanning 500 nM–500 μM. We subsequently developed a second, cortisol-specific DBS probe, achieving continuous cortisol detection in undiluted human plasma with nanomolar sensitivity and a dynamic range suitable for detecting physiologically relevant concentrations (200 nM–100 μM) over the course of multiple hours in undiluted human plasma. The DBS design thus offers a generalizable approach to accelerate the transition from aptamer selection to sensor development, and to rapidly design aptamer-based sensors for the sensitive and specific continuous optical detection of diverse small-molecule analytes in a range of biomedical applications.

Low-cost biosensing methods are crucial for the early detection of diseases. The work presents a smartphone-based detection scheme based on two asymmetrically evaporated droplets on a nanofibrous membrane with the specimen sample and plasmonic nanoparticles, respectively. Leveraging deep learning algorithms, we achieve automatic detection and quantification of a range of proteins based on distinct droplet patterns to differentiate positive/negative cases, including N-protein, PCT, CEA, and PSA at concentrations as low as 10 pg/ml in 12 minutes. The method enhances the droplet pre-concentration and the generation of optically structural patterns via plasmonic particles for improved sensing sensitivity.

Low-cost slide scanners, in conjunction with remote reading and AI-powered diagnostic workflows, have been proposed to democratize digital slide scanning for histopathologic diagnosis. However, scan speed is limited because proposed systems scan only one slide at a time. We developed a high-speed, low-cost, multicamera slide scanning system using a 3d printer and a Raspberry Pi with 4 motorized focus cameras. Motorized focus results in an adjustable working distance, which allows for multi-camera fine-tuning of the focus plane without additional delay to adjust the z-position. By supporting parallel-acquisition of 4 cameras at once at a fixed z-position, scanning speed increases fourfold.

Rapid, point-of-care bacterial bioburden detection on diabetic foot ulcers, burn wounds, and surgical site wounds, and subsequent gram-type classification is a challenge that cannot be addressed in low-resource settings. This is because current gold-standard practices based on cell culture take 2-3 days for a pathogen detection and classification. Further, these methods are cumbersome, require sophisticated microbiological facilities, and trained personnel. To address this issue, we demonstrate a rapid (<2 min), point-of-care platform, Illuminate®, that combines autofluorescence intensity imaging and deep learning to detect bacterial bioburden. We use a UNET deep learning model with a DenseNet 201 backbone to detect bioburden spatially and, to further classify the gram-type. Our initial results show a >80% accuracy in detecting bioburden and a >70% accuracy in gram-type classification. We further use this platform for identifying clinically relevant bacteria such as pseudomonas aeruginosa that cause antimicrobial resistance.

Last year we introduced a new Full Field Optical Transmission Tomography (FFOTT) technique that we applied to CELL studies. This interferometric technique is based on the use of the Gouy phase shift that takes place close to a microscope objective focus. We will now show results obtained in biological tissues using both a static mode (morphology) and dynamic one (metabolic contrast). In particular we have been able with our 400 $ microscope to section in tissues. We will discuss the importance of the spatial coherence of the illumination and compare the results with those obtained with Full Field OCT.

Edible oil is an important source of macronutrients and a good vehicle for micronutrients such as Vitamin A. Quality of edible oils is determined by conventional laboratory methods that are often expensive, relying on chemicals, and slow turnaround. If quality tests are more readily available in low-resource settings, we can more effectively address human malnutrition, especially for vulnerable populations. However, there is a lack of incentive and efforts to bridge gap between efficient tests and available technology, though, so here we show initial outcomes from our edible oil quality analysis using optical approaches including ultraviolet and near-infrared spectroscopy.

Early detection of osteoporosis is becoming imperative in an aging society, as low bone mineral density (BMD) poses an elevated risk of bone injury. We develop an optical bone densitometer (OBD) combined with deep learning to predict BMD in specific regions of human body. The OBD utilized a near-infrared light source to obtain photos with optical information from distal radius by emitting the light through the wrist, the relatively thin region of human body.To precisely capture the position of wrist, we employed U-net for biomedical image segmentation, which generates a mask for the original wrist image, leaving behind images without background noise for the subsequent deep learning analysis. The algorithm considers the preprocessed images and various physiological parameters to predict BMD in the target regions, thus providing a reliable result for both orthopedic surgeons and patients.

Bacterial biofilms on wounds lead to severe infections because of their resistance to antibiotics and host defence mechanisms. These bacterial colonies comprising of bacterial cells embedded in an extracellular matrix can easily develop on wounds, are hard to detect, and significantly delay wound healing leading to chronic infections especially in diabetic foot ulcer patients. Current methods to detect bacterial biofilms are extremely cumbersome and time-consuming (2-3 days) and, therefore, pose a challenge to low-resource implementations. We demonstrate a machine learning aided rapid wound blot detection method (<10 min) that combines the wound blotting technique using nitrocellulose membranes, white light imaging and machine learning-based models to accurately infer the presence of biofilms. We validate our method against the standard test-tube method that utilizes dye staining of the wound-swab culture to infer biofilm presence and demonstrate a detection accuracy in excess of 85%.

Lesion count is a key outcome to assess disease progression and the efficacy of new treatments for mpox. Accurately counting the hundreds to thousands of skin lesions typical of clade I mpox in endemic areas poses considerable human resource challenges. To document lesion appearance, evolution, and resolution during the PALM007 randomized controlled clinical trial of tecovirimat for mpox, we established smartphone-based protocols for standardized total body photography at remote sites in the Democratic Republic of the Congo. The trial is supported by the Institut National de la Recherche Biomédicale, NIAID, and funded by NCI Contract No.75N910D00024, Task Order No.75N91020F00025.

The U.S. faced a record 110,000 drug overdose deaths in 2022, primarily due to opioids, underscoring the need for efficient detection. Current methods suffer from poor accuracy or high costs. We aim to address this with a portable, high-throughput nanophotonic sensor that uses Ag/Au nanoparticles decorated on ZnO nanoarray to detect multiple drugs, like opioids, cocaine, and methadone from biofluids such as urine or blood. Amplified signals from surface-enhanced Raman scattering (SERS) are interpreted by machine learning, allowing automated, multiplex drug detection without expert supervision. This breakthrough approach could significantly advance understanding and control of the drug overdose crisis.

We report a point-of-care sensor for multiplexed quantification of three cardiac biomarkers, i.e., myoglobin, creatine kinase-MB (CK-MB) and heart-type fatty acid binding protein (FABP) from human serum. The sensor uses a paper-based fluorescent vertical-flow assay (fxVFA), and the assay operation takes <15 min and requires 50 µL of serum. After the assay, an image of the sensing membrane is captured by a cellphone-based reader, and a deep learning-based algorithm infers the concentrations of the 3 cardiac biomarkers from the captured fluorescence image. Our fxVFA achieved a limit-of-detection of <0.52 ng/mL, a coefficient-of-determination of >0.9, and a coefficient-of-variation (CV) of <15%.

Lyme disease (LD) is a tick-borne illness and can lead to severe health complications if left undiagnosed/untreated. This study introduces a novel approach to point-of-care testing for LD, utilizing deep-learning-enabled peptide-based serodiagnosis. 82 unique patient serum samples were used for validation, obtained from the Bay Area Lyme Disease Biobank and the CDC Lyme Serum Repository. Blinded test results demonstrated outstanding diagnostic performance, with 84.0% sensitivity and 96.3% specificity. By integrating deep learning techniques with a peptide-based sensing panel, this serodiagnosis offers a rapid (< 15 minutes) and cost-effective (< $0.5/test) platform for LD diagnostics, while minimizing cross-reactivity.