Biophotonics in Exercise Science, Sports Medicine, Health Monitoring Technologies, and Wearables VI

In this study, we used fNIRS to investigate cerebral responses to resistance training (RT) under high (HL) and low loads (LL) in 21 participants who performed unilateral dumbbell arm curls to failure with both arms. Our results show greater activation of the prefrontal cortex (PFC) during HL compared to LL, reflecting higher strategic planning and execution in this condition. When the task was performed with the non-dominant side, we also observed a higher activation in the motor cortex (MC), suggesting a greater demand for motor control. Additionally, different brain responses close to muscle concentric failure suggest personalized strategies to maintain performance. Overall, these results have implications for rehabilitation and training programs, which can be designed to optimize both muscular and cognitive functions.

This study investigates the potential application of near-infrared spectroscopy (NIRS) to quantify muscle contraction intensity as a surrogate for muscle spasticity assessment in healthy individuals. Using NIRS and EMG sensors, we measured changes in muscle oxygenation and electrical activity during various levels of voluntary contractions of the flexor forearm muscles. Our results indicate a significant negative correlation between muscle oxygenation levels and contraction intensity, suggesting that higher muscle spasticity correlates with lower oxygenation. This method offers a promising non-invasive alternative for assessing muscle spasticity, potentially improving diagnostic accuracy and treatment monitoring in clinical settings.

In cricket, a bowling delivery is considered illegal if the bowler’s elbow angle exceeds 15 degrees. To measure the bend angle, we developed an inline optical fiber-based interferometric sensor that uses a 2 cm long single-mode fiber (SMF) as a sensing head spliced between two 0.1 cm multimode fibers (MMF). We attach the sensor to a bowler’s arm with kinesiology tape in such a way that the sensor’s SMF portion aligns with the bowler’s elbow. When connected to a broadband source and a spectrum analyzer, the interference of SMF core mode and cladding modes is recorded on the analyzer. The interference peak’s position and full-width-half-max (FWHM) change as a function of the bend angle, forming our sensing modality's basis. We achieve a sensitivity of 0.612 nm/degrees and 0.245 nm/degrees in peak position and FWHM readings, respectively. We anticipate that the proposed sensor will find numerous applications in sports, e.g., tracking athletes' muscle movements and well-being.

Magnetomyography (MMG) could theoretically surpass EMG but faces practical challenges like the need for magnetic shielding and sensor size. These issues can be mitigated using nitrogen-vacancy (NV) diamond magnetometers. We introduce an integrated, flexible diamond-magnetometer sensor platform for MMG signals, operating with sensitivities below 20 pT/sqrt(Hz). We present initial MMG measurements, sensor architecture, and performance data. Human cardiac and muscle measurements using this platform are discussed, along with a new modeling approach for muscular activity and environmental interaction. Experiments aim to characterize sensor output and assess the force-MMG relationship in minimally-shielded environments. For sensor calibration, two simulation methods are proposed: a physiology-inspired 3D electrochemical model for simulating action potential propagation in muscles, and magnetic field strength prediction using generic shielding geometries.

The rapid, reagent-free analysis of biofluids using Raman spectroscopy is a promising and cost-effective approach to disease screening. We have found a significant application for this technology in the field of sport, particularly among high-level athletes Non-invasive biomarker tracking in high-level athletes optimizes performance, health, and recovery strategies by providing real-time insights into energy utilization and metabolic adjustments. Also, continuous monitoring helps tailor training regimens, enhance recovery strategies, and prevent overtraining, ensuring athletes maintain peak performance and overall well-being. A study of 20 non-athlete participants validated sports-related biomarkers in plasma, urine, saliva, and sweat for aerobic and anaerobic activities. Preliminary results established baseline variations and identified exercises causing significant biomarker changes. Two key activities were highlighted, and biofluid samples were collected from participants during these exercises. The study will expand to include high-level athletes across various sports disciplines.

This study explores the use of near-infrared spectroscopy (NIRS) to monitor muscle spasticity in children with cerebral palsy undergoing Botox treatment. Conducted at BC Children’s Hospital with three participants, the study used an NIRS sensor to measure muscle oxygenation before and after Botox injections in their spastic muscle group. Preliminary results show increased muscle oxygenation up to three months post-injection, with a decline observed at six months. While initial findings are promising, further research with a larger sample is needed to validate NIRS as a reliable tool for assessing muscle spasticity and tailoring treatment plans.

Postpartum hemorrhage (PPH) is the leading cause of maternal mortality globally. Importantly, PPH has been noted as the most preventable cause of maternal mortality, and the leading factors causing preventable PPH are delays in diagnosis and treatment. Thus, there is an urgent need for an early and accurate PPH alert system that can be successfully deployed at the point of care. To this end, we developed a low-cost (<$150), wearable optical device that continuously monitors peripheral perfusion on the wrist via the laser speckle flow index (LSFI) to detect hemorrhage-induced peripheral vasoconstriction. The laser speckle imaging system was designed using a Raspberry Pi Camera (V2) and a laser diode. The system is controlled by a Raspberry Pi Zero computer. The device contains no consumables, and is reusable and compact. It can capture, process, and display LSFI signal for near real-time data visualization. The sensor has been successfully tested in a swine hemorrhage model, human blood donor model, and cesarean deliveries with high correlation (Spearman rho >0.9) to net fluids. This progress provides a framework for a novel low-cost and noninvasive technology that identifies ongoing blood loss rapidly, with the goal of reducing the unacceptably high rates of global maternal morbidity and mortality caused by hemorrhage in low- and high-resource settings alike.

This study examines the feasibility of extracting blood pressure-related information from near-infrared spectroscopy (NIRS) data. A wearable NIRS device with a data acquisition rate of 20 Hz was developed. During an experiment, NIRS device was attached to forehead and a continuous blood pressure monitoring device (Finapres, ADInstruments) was worn at a finger and opposite arm. A breath-holding task was used to stimulate blood pressure variation. In response to the breath-holding task, a similar trend was observed in both blood pressure and NIRS-extracted pulse strength. This similarity indicates the possibility of using the pulse strength to monitor blood pressure.

In recent years, the need to use contact methods has arisen to measure and monitor several physiological parameters of patients. Remote photoplethysmography (PPGr) is a noncontact method that uses a camera to record images of the skin and measure its pixel intensity changes. However, one of the main challenges faced by PPGr is that the signal is immersed in a large amount of noise coming from different sources: sensor noise itself, conditions during measurement, etc. We propose a new method to obtain high-quality signals using PPGr. The main advantages are: 1) it is possible to record in small skin areas to obtain high-quality photoplethysmography signals, 2) it is a noncontact method that is performed under uncontrolled conditions (non-laboratory conditions), and 3) it was realized with a conventional camera. This gives rise to the use of rPPG in situations involving contagious diseases to reduce the risk of propagation.

Photoplethysmography (PPG) is a non-invasive optical technique used to measure cardiovascular parameters such as heart rate and blood oxygen saturation. Recently, its integration with Deep Learning (DL) models has gained attention for detecting cardiovascular diseases. This study explores using large language models (LLMs) to enhance diagnostic accuracy from PPG signals. Our approach involves preprocessing PPG signals to reduce noise and extract features, followed by fine-tuning LLMs with these features to learn from clinical textual information. Using PPG signals and data from the MIMIC-III database, initial results show that this method significantly improves the detection of conditions like myocardial infarction, conduction disturbances, and hypertrophy.

Total-body photography (TBP) has the potential to revolutionize early detection of skin cancers by monitoring minute changes in lesions over time. However, there is no standardized Digital Imaging and Communications in Medicine (DICOM) format for TBP. In order to accommodate various TBP data types and sophisticated data preprocessing pipelines, we propose novel TBP Extended Information Object Definitions (IODs) within the DICOM framework: TBP Extended 2D Regional Image IOD, TBP Extended Dermoscopy IOD, and TBP Extended 3D Encapsulated IOD. We introduce a comprehensive pipeline integrating advanced image processing techniques, including 3D DICOM representation, super-resolution enhancement, and style transfer for dermoscopic-like visualization. Our system features deep learning models for lesion detection and tracking, stores data in a HIPAA-compliant cloud, and offers a customized DICOM viewer with various analyzing tools. This approach and proposed IODs enhance TBP interoperability and clinical utility in dermatological practice, potentially improving early skin cancer detection.

In this paper, we propose a two-stream convolutional neural network for breathing pattern classification (TCNNBP) for continuous monitoring of patients with infectious respiratory diseases. We have applied this method to classify different breath patterns using tissue hemodynamic responses collected from 14 participants. Four breathing patterns, namely normal, slow, rapid, and breath holding, were included in the experiment. The TCNNBP consists of a a convolutional neural network-based autoencoder and classifier. The encoder of the autoencoder generates a deep compressed feature maps, which contain the most important information constituting the data. These deep compressed feature maps are concatenate to the feature maps generated by the classifier to classify breathing patterns. The proposed TCNNBP overcomes the problem of decreasing learning performance as the layers deepen in the single-stream classification model. The proposed method achieved the highest classification accuracy of 94.46% compared to state-of-the-art classification models.

This study aimed to determine the recovery of oxyhemoglobin (O2Hb) during repeated hypotension induced by thigh cuff release. Fifteen healthy students participated in this study. We placed digital tourniquet cuffs on the participants’ thighs, inflating them to 250 mmHg for 2 min after 5 min of rest, followed by deflation for 5 min; this cycle was repeated four times. The prefrontal cortex (PFC) O2Hb levels were measured using a multichannel near-infrared spectroscopy system, and the beat-to-beat mean arterial pressure was recorded by finger pulse volume using photoplethysmography. The slope of O2Hb recovery from the nadir after deflation was calculated and compared for each episode using one-way analysis of variance. No significant changes were observed in the O2Hb recovery slope between the four measurements. These results suggest that the thigh-cuff release-induced repeated hypotension did not affect the O2Hb recovery slope in the PFC.

In response to the rising prevalence of mobility-related pathologies, this study introduces an innovative sensor for smart walkers. Integrating light-sensitive photodiodes and addressable RGB LEDs, the sensor utilizes overlapping signals from embedded emitters and receivers in a transparent waveguide layer to measure tactile deformation. This novel approach overcomes limitations of traditional sensors such as fragility and high costs, offering a more accessible and precise solution. Additionally, deep learning methods demonstrated accurate estimation of the applied normal force. Designed for seamless integration, the sensor promises to enhance functionality and reduce costs in robotic assistive devices, opening new possibilities in mobility care.

Maintaining a stable posture while responding appropriately to stimuli is vital for daily activities. This pilot study examines the neural mechanisms underlying posture-inhibitory control and mental flexibility dual-tasking and the impact of acute physical activity on dual-task performance. Eighteen healthy adults (average age 42.8) were tested using functional Near-Infrared Spectroscopy (fNIRS) to measure neural activity during tasks with different posture challenges (sitting, standing, tandem stance) and cognitive loads (none, congruent, incongruent Flanker tasks). For inhibitory control, easier conditions showed increased brain activation with higher challenges, while difficult conditions showed the opposite trend. Physical activity improved inhibitory control as indicated by reducing reaction times and altering brain activation patterns. Findings related to mental flexibility are still under analysis. Further research should further explore the impact of physical activity on dual tasks and how it might be leveraged to improve safety in real-world scenarios.

We examined the effect of gene polymorphism on seasonal variation in brown adipose tissue density (BAT-d), which was estimated using total hemoglobin concentration in the supraclavicular region by near-infrared time-resolved spectroscopy, in healthy subjects (22 men and 26 women, aged 24-54 years) both in summer and winter. The genotype of single nucleotide polymorphisms (SNPs) in uncoupling protein-1, β2- and 3- adrenergic receptors was also determined using TaqMan SNP Genotyping Assays. The results of the multiple regression analysis indicated that no SNPs affected the seasonal variation in BAT-d, but the level of BAT-d in winter significantly influenced the seasonal variation in BAT-d (β = 0.357, p = 0.013). In conclusion, SNPs determined in this study were not significantly relevant factors of seasonal variation in BAT-d, but it suggests that seasonal variation in BAT-d occurs in subjects with higher BAT-d in winter.

The degree to which posture confounds hemodynamic data collected via fNIRS is currently unknown, especially for Fowler’s and semi-Fowler’s postures. This lack of information presents a problem in accurately interpreting and processing fNIRS data in environments where posture is constantly changing. We explored the effects of five postures commonly assumed in healthcare settings – upright standing, upright sitting, Fowler’s, semi-Fowler’s and supine – on fNIRS global data associated with auditory stimuli. The results of the study could be leveraged to enhance fNIRS data processing and more accurately define cortical and functional relationships in naturalistic or healthcare settings.

Oral squamous cell carcinoma (OSCC) is the most prevalent cancer in the oral and maxillofacial region, with a 75% 5-year survival rate dependent on early detection. This study aimed to non-invasively detect OSCC by measuring the thermal difference between carcinogenic tissue and healthy mucosa using an infrared sensor and assess the accuracy of this diagnostic approach. A novel intraoral infrared device was designed to measure intraoral tissue temperature in 20 participants (10 OSCC patients and 10 healthy individuals). Results indicated a significant temperature difference between tumoral and healthy tissues (P<0.001). The device demonstrated high accuracy, with a temperature differential greater than 0.97°C indicating potential OSCC presence (sensitivity=1, specificity=1). Temperatures exceeding 38.42°C suggested malignant lesions (sensitivity=1, specificity=0.9). This feasibility study highlights thermography's potential as a non-invasive diagnostic tool for early OSCC detection. Further research with larger samples is needed to validate these findings and the device's performance.

We have developed a portable, highly accurate optical sensor chip for minimally invasive training load monitoring in sports, to track athlete’s stress and recovery cycle. The goal is to simultaneously detect critical stress markers in blood and saliva. Currently, blood samples are collected on-site, frozen, and transported to laboratories for conventional ELISA testing. This process is time-consuming and prone to sample loss. Portable monitoring could provide faster and more cost-effective results. The miniaturized optical sensor and biochip will enable rapid point-of-care diagnostics. This can be further developed and integrated with mobile phones.

Speckle plethysmography (SPG), based on laser speckle imaging methods, is an optical method that holds great promise for measuring vital signs remotely and through contact. This study evaluates the ability of breathing rate (BR) monitoring through SPG using a contact sensor on the finger, including a near infrared laser in reflection mode. An algorithm is developed for BR extraction from the SPG signal using the concept of respiration induced modulation on the cardiac signal. The results show good accuracy of SPG-derived respiration compared to the reference.

Near-infrared spectroscopy (NIRS) is a non-invasive method that leverages NIR light penetration in tissues to measure tissue oxygenation level1. NIRS sensors monitor tissue oxygenation in patients by estimating oxygen saturation within the tissue; this is a parameter known as regional tissue oxygenation (RSO2). Ten bilateral regions, in addition to the Sternum and tongue, were selected for the data collection. These sites include the bilateral Tibia, Quadriceps, Thenar, Ramus, Orbital rim, Temporal, TMJ, and Forehead, as well as Sternum and Tongue in the Supine position. NIRS sensors (Root with O3 Regional Oximetry) were affixed to the participant measurement zones using double-sided tape. Data were collected continuously at a sampling rate of 0.5 Hz for 30 seconds per site. Skyndex skinfold calliper was used to measure subcutaneous tissue thickness. Nix Mini 3 Color Sensor was used to capture tissue pigmentation. Twenty participants (15 males and 5 Females), with a mean age of 29.5 participated in this study. The average body RSO2 is 74.2 ±1.6 among 20 participants. The average thickness of the subcutaneous fat layer at the rim orbit was the lowest at 3.87 ±2.81.

Cardiovascular complications account for nearly half of US and global maternal deaths. Monitoring cardiac output (CO), systemic vascular resistance (SVR), and mean arterial pressure (MAP) would enable early detection of cardiac complications. A low-cost wearable laser speckle flow index (LSFI) sensor that captures pulsatile waveforms across the cardiac cycle was developed to accurately estimate CO, SVR, and MAP. Waveforms were collected from swine undergoing blood removal and human subjects undergoing exercise. From these waveforms, features were extracted and used to train multivariate regression models to predict CO, MAP, and SVR. In swine, average absolute errors of 2.13 L/min, 9.0 mmHg, and 464 dynes/sec/cm-5 were produced, respectively. The correlation coefficients for a regression between true and predicted values were r=0.79, r=0.88, and r=0.68, respectively. In human subjects CO and SVR were predicted and produced absolute errors and correlation coefficients of 1.52 L/min and r=0.80 and 278 dynes/sec/cm-5 and r=0.66, respectively.

To address the challenge of blurriness caused by overlapping spectra in conventional thermal imaging and to broaden biomedical applications of thermal biophotonics, we developed an innovative phasor-based hyperspectral thermal imaging system, phasor thermography (PTG). PTG delivers vivid texture details, accurate material classification, and precise temperature extraction, enhancing the clarity and quality of thermography with our advanced phasor thermographic vision. Through multiple imaging scenarios and a functional study, we demonstrate that PTG is a reliable and robust image-based technique for contactless vital sign detection, including body temperature, respiration rate, and pulse rate, thereby significantly advancing health monitoring.

Wearable devices are becoming more significant in remote healthcare, fitness tracking, and athletics through advancing sensor technologies. These devices employ light to non-invasively monitor physiological changes allowing for the estimation of heart rate and other vital cardiac metrics. A significant challenge in the utility of these optical signals is the presence of motion artifacts, which diminish the usefulness of data from wearable devices during high-movement activities. Here, we propose integrating a pressure channel and multiple light wavelengths within the wearable to support the denoising process. This multimodal data can be combined with deep learning models to reconstruct motion-free physiological waveforms. We developed a multisensory device that combines force and multiwavelength optical measurements to capture relative motion at the sensor interface. Deep learning models are then trained to reconstruct the photoplethysmography waveform. Initial testing and training consist of controlled lab experiments with plans to translate into real-world examples of motion.

The visual detection of atypical or changing pigmented lesions is the cornerstone of melanoma diagnosis. Dermatoscopes are available as independent devices, cellular device attachments, connections to a digital camera via coupling adaptors, or direct lens attachments. Sequential digital dermoscopy imaging (SDDI) involves the capture and assessment of successive dermoscopic images, separated by a time interval, to detect suspicious changes. However, the current crop of dermatoscopy attachments is inefficient for lesion documentation because they require the dermatologist to manually and pianistically record the location of the lesions of interest. Flacara allows clinicians to efficiently compare temporally separated dermoscopic images of a lesion. This integration not only assists the clinical decision but also facilitates a proactive approach to skin health management, which is crucial for effective treatment outcomes and improved patient care in telemedicine settings.

A photoplethysmography (PPG) signal can detect volume changes within blood vessels during a cardiac pulse. Such change directly relates to the change in local hemoglobin concentration. Utilizing this principle, we will create a simple, gel-based Hemoglobin phantom model for quick calibration of PPG sensors. The wavelength-specific absorption coefficients of the phantoms are scaled to that of a PPG pulse shape and are calculated based on Beer-Lambert’s law. We will use an in-house fiber-based diffuse reflectance spectroscopy to measure the PPG pulse of these phantoms. In addition, the effect of the phantom’s top layer thickness on PPG will also be evaluated.

Purpose: This study aims to develop six units of panoramic VR interactive teaching materials and incorporate action research methods to measure the learning outcomes of fourth-year junior college students before and one week after completing the course. Methods: A quasi-experimental pretest-posttest design with two groups was adopted, using purposive sampling at a nursing college in southern Taiwan. Students from two classes taking the emergency nursing course were divided into a control group and an experimental group. The control group received team-oriented scenario-based teaching plans and traditional PPT presentations. In addition to the control group's content, the experimental group was integrated with six units of panoramic VR interactive teaching materials. Through the action research phases of problem identification, planning, action, observation, and reflection, iterative adjustments were made before, during, and after the course. Results: After the intervention, the experimental group's communication and critical thinking scores improved significantly compared to the pretest. The experimental group's team emergency skills were significantly higher than control group.