We investigated a system identification method for modeling the standing posture without external perturbation. Assuming that postural sway during quiet standing is caused by internal white noise, the method adopts an autocorrelation matrix consisting of body position and velocity parameters. We showed that the method can be applied to center-of-mass fluctuation measurement data. Since we used only a force plate to minimize the burden on the subjects, the velocity could not be measured directly. Therefore, we first demonstrated through simulation that ankle joint stiffness can be estimated by numerically deriving the velocity from the displacement. The velocity was calculated using the central difference approximation, and the ankle stiffness estimation error was obtained. The system identification method was then applied to the center-of-mass measurement data, and the ankle stiffness was estimated. Simulations showed that even when the velocity was calculated through numerical differentiation, the ankle joint stiffness could still be estimated with an accuracy of less than 1% error. The ankle stiffness estimated from the measured center of mass was slightly higher than the gravitational torque coefficient, but lower than that obtained in our previous electrical stimulation study (Uchiyama T: Adv Biomed Eng. 13, 223‒229, 2024). This suggests that the control method for ankle stiffness may depend on the presence or absence of external disturbances.

Urine-related information, such as voided volume and flow rate, is important for the diagnosis of dysuria. Uroflowmeters, which are used for obtaining urine information, measure mainly the weight gain of the collection cup or the rotation of the impeller. These devices pose hygiene-related problems because of the contact between urine and the device. In addition, the patients are forced to urinate in unfamiliar environments and positions, which can lead to unusual results. Therefore, there is a need for a non-contact uroflowmeter that can measure urine information in an environment and a posture to which the user is accustomed. A uroflowmeter that measures the rise of the water level in the toilet bowl is also available; however, it requires extensive installation. To address these problems, we have developed a scale-type uroflowmeter that measures weight loss during standing urination. The purpose of this study was to evaluate the measurement accuracy of our developed scale-type uroflowmeter. For this purpose, human urination was measured simultaneously using the developed uroflowmeter and three different medical uroflowmeters (Freeflow^®, P-Flowdiary^®, and PicoFlow2^®), to compare the voided volume, maximum flow rate, and average flow rate. The results of the developed uroflowmeter showed good agreement with the weight of urine measured by an electronic balance and the voided volume measured by the medical uroflowmeters. The error rates between the true voided volume measured by the electronic balance and the four uroflowmeters ranged from 0.9% to 22.5%. Comparing the maximum and average flow rates of the three medical uroflowmeters with the scale-type uroflowmeter, the error rates ranged from 9.4% to 25.4% and 17.4% to 25.2%, respectively.

Predicting spatiotemporal neural activity could lead to early detection of seizures and better understanding of cognitive processes in the brain. Parallel reservoir computing, where spatiotemporal data are split into regions and one reservoir predicts each region, has been proposed as a method to predict large spatiotemporal data with high efficiency and accuracy. However, no studies have extended parallel reservoir computing to predict large-scale brain activity. As the accuracy of the prediction will drop dramatically if the hyperparameters are not properly set, the relationship between the hyperparameters of parallel reservoir computing and parameters of neuronal simulations such as connection length, axonal conduction speed, and excitatory and inhibitory synaptic conductance should be explored. In this study, we systematically investigated the relationship between the appropriate hyperparameters of parallel reservoir computing and parameters from neuronal simulations. Specifically, we first simulated spiking neural networks and evaluated their spatial features while varying the parameters. Then, for each parameter setting, we investigated the hyperparameters of parallel reservoir computing that reproduced the temporal features of traveling waves. Our findings indicated that axonal conduction speed and connection length drastically affected the appropriate hyperparameters, while excitatory and inhibitory synaptic weights did not. In addition, spatial features alone did not necessarily determine the appropriate hyperparameters to predict neural activity. Our study reveals that more studies are needed to find a way to assess the dynamics of neuronal population, thereby paving the way for the prediction of large-scale brain activity.

Purpose: The three-dimensional polymer gel dosimeter functions as an anthropomorphic phantom that can be applied to dose verification in clinical radiotherapy. The advantage of polymer gel dosimeters is that doses can be evaluated using multiple imaging modalities such as optical computed tomography (CT) and R₂ magnetic resonance imaging (MRI). In this study, we focused mainly on a polymer gel-optical CT system, and aimed to (1) evaluate the noise characteristics of the optical CT system, and (2) improve the image quality and reduce noise by image averaging.

Materials and Methods: Noise power spectra (NPS) were calculated from reconstructed images from an unirradiated gel dosimeter. The NPS of the unirradiated gels ranging from no averaging to averaging 128 images were calculated and compared. The signal to noise ratios (SNR) for the gel dosimeters were also calculated and compared between optical CT, MRI, and X-ray CT (XCT).

Results: The noise components of the dose image from the polymer gel-optical CT system distributed from 0 to 1.5 cycles/mm. The high frequency component of 1.0-1.5 cycles/mm is considered to reflect predominantly the quantum noise component in the optical CT system, and the noise of this frequency component could be reduced by averaging. The results indicated that increasing the number of images averaged from 2 to 128 enhanced image smoothness in visual evaluations. Specifically, the SNR improved by 1.63, 1.55 and 1.51 folds for optical CT, XCT, and MRI, respectively, when averaging 8 images compared to no averaging, with optical CT showing the greatest improvement.

Conclusion: The NPS are useful to quantitatively express the characteristics of each frequency, and we have shown that they can also be analyzed for the polymer gel-optical CT system. The quantum noise component can be improved by image averaging, providing better image quality and reducing granularity of the gel images. Image averaging achieves significantly greater improvement in SNR for optical CT compared with MRI and XCT.

An intravascular shunt (arteriovenous anastomosis) created during hemodialysis is a common site of stenosis. Early detection of stenosis is crucial for treatment; hence, shunt sounds are gaining attention as an indicator for simple, noninvasive screening of stenosis. However, previous studies on screening methods using shunt sounds have not achieved sufficient detection accuracy for practical application. This study aimed to elucidate the mechanism of generation of shunt sounds by conducting computational fluid dynamics analysis of the fluctuations of pressure and magnitude of vorticity as potential sound sources. This insight may contribute to improving stenosis detection methods using shunt sounds. In this analysis, a side-to-end model was used as the typical anatomical geometry for vascular access. The power spectral densities of the pressure and vorticity magnitude fluctuations were estimated using an autoregressive model. Using shunt models with and without stenosis, we calculated the time fluctuations of the pressure and vorticity magnitude in the shunt blood vessel and investigated the differences in the power spectral densities of these fluctuations due to stenosis. The source of the shunt sounds was investigated by comparing the power spectral densities of the pressure and vorticity fluctuations to those of the shunt acoustic data, and a similar trend was observed. The power spectral densities of both the pressure and vorticity magnitude fluctuations were higher in the shunt model with stenosis than in the model without stenosis as the frequency increased. This analysis suggests that pressure and vorticity magnitude fluctuations in the shunt blood flow contribute to the generation of shunt sounds. Moreover, stenosis shifted these sounds to higher frequencies due to flow separation. Additionally, when the shunt acoustic spectrum at a specific point of the blood vessel exceeded those at adjacent points, especially within the 500-1200 Hz range, stenosis was likely to be present at that location. This finding demonstrates that stenosis can be detected with high accuracy by analyzing the spectral shifts in acoustic data at multiple points along a shunt. Improving software and hardware based on these insights may further enhance the accuracy of intradialytic shunt stenosis screening devices using shunt sounds.

Atrial fibrillation (AF) is the most common arrhythmia that increases the risk of cardiac diseases such as stroke; hence, its early detection is important. Although many studies have been conducted on automatic AF detection, the accuracy of identification needs to be improved for large amounts of data such as Holter electrocardiograms (ECGs). In rare clinical cases, AF shows regular R‒R intervals (RRIs). These cases are difficult to diagnose even by experienced cardiologists or to detect automatically. To detect AF accurately with minimum number of oversights, identification of AF including regular RRIs is necessary. This paper presents a new method for improving the accuracy of AF detection. The detection is realized in two stages with different AF identification models, after eliminating noise in the ECG waveforms using a finite impulse response bandpass filter. In the first stage, a hybrid convolutional neural network (CNN)‒long short-term memory model trained with segmented ECG waveforms and their RRIs is used to identify AF with irregular RRIs. In the second stage, non-AF events are input. AF cases with regular RRIs are identified using an independent CNN model trained with waveforms that include P- or fibrillatory waves. In this study, 24-hour Holter ECG data obtained from 60 subjects were used to train both models. For evaluation, Holter ECG data from ten subjects not used for training were used, which yielded a detection accuracy of 94.7% and a sensitivity of 98.5% for AF. In addition, evaluation using another untrained dataset from two subjects whose AF had regular RRIs showed an accuracy of 96.2% and a sensitivity of 98.8% for AF. These results demonstrate the feasibility of the proposed method for AF detection.

Regular stress checks and health confirmations by supervisors are standard practices in high-risk environments such as construction sites. Therefore, effective health management of onsite workers is crucial for preventing health hazards and casualties at the workplace. Nevertheless, detection of falsified health declarations and unrecognized health issues remains a challenge. This study used facial thermography obtained by infrared thermal imaging to monitor the health conditions of workers. Thermal images of the face reflect the blood flow distribution in the skin, which is controlled by the autonomic nervous system and can serve as noninvasive health status indicators. However, ambient temperature and time-of-day variations affect skin temperature readings, making the application of facial thermal imaging challenging in uncontrolled environments outside laboratory setting. To address this issue, we proposed a method to correct the artifacts caused by environmental temperature and time-of-day variations in facial thermography. This method provides standard skin temperatures for any given environmental temperature and time of the day, allowing comparison with actual skin temperatures. We conducted a five-month longitudinal study in 279 construction workers, and collected 2,884 pairs of facial thermal images and health status data. The collected face thermal images were corrected for environmental temperature and time-of-day variations. The effectiveness of this method in assessing the health status of onsite workers was evaluated. The results showed that the proposed method effectively corrected artifacts from environmental temperature and time-of-day variations in facial thermography. Moreover, the corrected images revealed that a decline in health conditions was associated with a decrease in temperature around the nose and an increase in temperature in other areas. This is consistent with previous findings of the relationship between stress and facial skin temperature.

Phase-change nano-droplets (PCNDs) have been gaining attention for their potential applications in ultrasound-based therapeutic and diagnostic technologies. PCNDs transform from a droplet state to microbubbles in response to ultrasound or shockwaves. Although a negative pressure component contributes to this vaporization, previous studies have primarily used ultrasound with both positive and negative pressure components. Thus, the effect of the negative pressure component has not been sufficiently explored. In this study, we developed an experimental system that generates a negative pressure component using underwater shockwaves reflected at the water-air interface. Using PCNDs composed of perfluorohexane (PFH) with a boiling temperature of 59℃, we analyzed the bubble lifetime, response to pressure, and vaporization threshold when exposed to reflected waves with a rise time of 50 ns. The results showed a bubble lifetime of 11 μs, a linear response to negative peak pressure, and a threshold of vaporization between 1.5 MPa and 1.9 MPa. Moreover, the effects of vaporization of PCNDs on biological cells were evaluated by measuring cell viability before and after exposure to negative pressure.

Intercellular communication is mediated by fibrous structures known as tunneling nanotubes (TNTs). TNTs are present in diseased cells, including cancer cells, and facilitate the transport of intracellular substances crucial for cell survival. They contribute to cancer survival by enhancing metastatic and invasive abilities and are implicated in the drug resistance of cancer cells. However, previous studies have primarily focused on the biochemical signaling of TNTs, and their mechanical properties are largely unknown. We hypothesized that TNTs play a mechanical role in cell survival by regulating cell-cell distance, migration direction, and mechanical signaling between cells. In this study, we investigated the mechanical properties of TNTs in a human cervical cancer cell line, HeLa cells, through nanoindentation tests using atomic force microscopy (AFM). To minimize shape variation in the TNTs and enhance the efficiency of mechanical tests, the shape of the cells and the distance between neighboring cells were controlled using a microcontact printing method that regulates cell adhesion regions. By performing AFM indentation tests on the TNTs, we determined that their internal tension was 758.3 ± 372.5 pN, axial spring constant was 32.1 ± 14.2 pN/µm, and axial elastic modulus was 696.3 ± 578.0 kPa. Furthermore, we observed that the internal tension and axial elastic modulus correlated positively with TNT length, while the axial elastic modulus correlated negatively with TNT thickness; that is, the thinner and longer the TNTs, the stiffer and higher are their tension. This implies that longer TNTs can exert greater tension on neighboring cells, thereby facilitating the proliferation of cancer cells. These findings suggest that TNTs play a critical role in the mechanical communication of cancer cells and may influence various aspects of cell survival, significantly affecting the progression of cancer and other diseases.

Near-infrared multispectral imaging (NIR-MSI) has attracted attention in the surgical field owing to its ability to penetrate tissues and analyze molecular vibrations with high spatial resolution, thus enabling the identification of deep or similarly colored tissues. However, conventional devices have limitations such as non-portability and long imaging times, rendering the surgical use of NIR-MSI challenging. This study developed a prototype NIR-MSI rigid endoscope system that can perform high-speed imaging and tissue classification. This system was constructed using a custom-made rigid endoscope that can relay NIR images, an InGaAs camera, and a lighting device containing nine high-power LEDs within the range of 940-1,535 nm. The developed device enables continuous processing of light irradiation, data acquisition, neural network analysis, and result display using customized software. The performance of the system was verified by testing tissue transparency and component classification. Tissue transparency test showed that text written on a letterboard could be recognized through a 5-mm thick layer of chicken breast meat. The system completed the imaging and neural network analysis of four color-matched transparent resins, which could not be distinguished readily under visible light, within 2 s, with an average accuracy of 96%. The proposed system may be applied for identifying deep or similarly colored tissues in the setting of non-invasive and real-time endoscopic surgery.

The neural signals of a subject listening attentively to one of two simultaneously presented speech streams can be used to decode the auditory source that the subject is focusing on. Many auditory attention decoding (AAD) studies have deciphered the attended speech envelope from the listener's delta (1-4 Hz) and theta (4-8 Hz) bands. However, other auditory source defining features, such as spatial location of the target speech relative to the listener, have not been extensively studied in the context of AAD. In this study, we systematically investigated the cross-frequency coupling (CFC) characteristics during attentive listening in a two-speaker paradigm. An open access electroencephalography (EEG) database of sixteen subjects listening attentively to one of two concurrently presented speech streams was analyzed. We evaluated CFC in the form of phase‒amplitude coupling using modified time series signal of the normalized modulation index. In this study, CFC was observed in several of the frequency pairs examined, with many pairs found in the delta and theta phase frequencies. In terms of the amplitude frequency, the 24-28 Hz frequency, corresponding to the beta band, was coupled with the 1-4 Hz, 2-6 Hz and 4-8 Hz phase frequencies. Decoding of the attended speech envelope significantly exceeded the chance level when using isolated neural frequencies and CFC measures as inputs. As reported in the literature, the isolated delta and theta frequencies contributed the most to the success of decoding the attended speech envelope, suggesting that the prominent features of the attended speech envelope are encoded in the amplitude and phase of these lower frequencies. On the other hand, the directional auditory attention decoding performance of the isolated frequencies or their combinations did not exceed chance level, but decoding of speaker location from the CFC measures was highly successful. These results indicate that the speaker's position relative to the listener is encoded in multiple CFC frequency pairs but is not encoded in any isolated frequency band, suggesting that CFC may serve as a method to enhance the neural representation of the attended speaker position.

The development of cellular structures has garnered significant interest in regenerative medicine for the restoration of lost tissues, organs, and blood vessels. Our research group focuses on the construction of cellular architectures by exploiting the tendency of spheroids to coalesce. Understanding the dynamics of spheroid fusion is pivotal for advancing the construction of larger cellular assemblies and developing innovative methodologies for three-dimensional cellular structures. Therefore, the aim of this study was to perform a morphological analysis of spheroid fusion over time. Human mesenchymal stem cells derived from adipose tissue were used to generate spheroids. Fusion experiments were performed on the second and fifth day of culture using a spheroid formation system, with the number of spheroids ranging from two to four. Subsequently, we monitored the spheroid fusion process by time-lapse imaging at 10-min intervals over three days, utilizing a fluorescent microscope that preserved the culture environment. The images acquired facilitated the development of a novel evaluation method based on the changes in neck radius and circumference of the spheroids. A quantitative assessment of these time-dependent morphological changes revealed distinct fusion behaviors that varied according to the fusion time and the number of spheroids involved.

Purpose: In this study, we attempted to construct an improved anomaly detection system. In our previous study, we developed an anomaly detection system for the elderly. However, the detection performance may not be sufficient for relatively mild abnormalities such as fever or low peripheral oxygen saturation (SpO₂), which frequently occur in elderly people. Methods: The purposed method used 3 types of models: natural language processing model, data generation model, and anomaly detection model. Among the residents of a long-term care facility, 79 (16 males and 63 females, 88.54 ± 7.21 years of age) elderly people were selected as participants. Results: The success rate of predicting physical anomaly (mean ± standard deviation) using our previous method was 24.82 ± 32.55% and that using the method proposed in this study was 42.12 ± 39.20%, with a significant difference between the two methods. Conclusions: In this study, we succeeded to improve the performance of an anomaly detection system for the elderly. Difference in detection performance between subject was observed. Further research is needed to investigate the relationship between subject condition and detection performance.

Kinesthetic illusion induced by visual stimulation (KINVIS) is a phenomenon in which a virtual kinesthetic perception is induced by watching virtual limb movements mimicking actual limb movements. Recently, the KINVIS has been used for the rehabilitation of patients with stroke. Visual immersion is thought to be essential to achieve the KINVIS. Existing KINVIS rehabilitation equipment uses large devices, such as enclosed boxes or head-mounted displays, to improve immersion. To conveniently improve the effectiveness of the KINVIS, we combined it with the rubber hand illusion (RHI) and pulling illusion (PI), which can be induced by simple equipment. The RHI occurs when a fake hand is perceived as a real hand once tactile and visual stimuli are synchronized. The PI occurs when asymmetric vibrations induce the user to sense as though his/her body is being pulled. We developed a system that induces these three illusions by using an upper limb avatar on a laptop personal computer (PC) display and a vibrator. Participant placed the upper limb behind the laptop PC display overlapping with the image of the upper limb avatar. In this setup, participant experienced combinations of the illusions by watching the motion of the upper limb avatar and the vibrator object, and sensing the vibrations of the vibrator. Two evaluation indices; namely, sense of ownership (SoO) and sense of body motion (SoBM), were employed. SoO quantified the degree of recognition of a virtual body as one’s own, while SoBM quantified the degree of feeling that the body had moved. We conducted a questionnaire on the direction and strength of the participant’s SoBM and SoO. Results showed that the RHI increased SoO for the avatar, considered to enhance immersion, which in turn increased the effect of the KINVIS. However, there was no positive interaction between the PI and KINVIS; when the PI and KINVIS were in opposite directions, tactile and visual information did not integrate, and the KINVIS decreased the effect of the PI. The main contributions of this study are identifying effective and ineffective illusion combinations for enhancing the KINVIS and demonstrating that a simple device can achieve these illusions. These findings mark the first step toward a more accessible KINVIS.

Controlling cell orientation is of paramount importance in bioengineering processes. While several surface modification techniques have emerged to guide cell orientation, they often involve complex, repetitive procedures for each experiment. Thus, a streamlined approach for cell orientation is necessary. In this study, we present a potentially reusable metallic culture surface that induces an anisotropic cell orientation due to its unique geometric morphology. Using a femtosecond laser, periodic nanostructures were produced on the metallic culture surface, resulting in a distinctive macro stripe design composed of both laser-treated and mirrored areas. Myoblast cells cultured on this surface showed a pronounced orientation. The macro stripe design provided stronger control of cell orientation compared to the simple laser-treated surface with periodic nanostructures. Initial random cell adherence was followed by migration toward the mirror surface, culminating in the desired orientation. This shift can be attributed to the mirrored areas providing superior cell adhesion and reduced wettability compared to the laser-treated areas. This innovative culture surface has significant potential for advancing bioengineering endeavors, especially in the field of tissue engineering.

Background: Ankle sprains are very common in badminton, and chronic ankle instability (CAI) often develops in players after these injuries. CAI badminton players (CAIBP) are more susceptible to injuries during high-intensity tasks, such as jumping landings, due to decreased ankle stability. This study aims to explore the variations in lower limb biomechanics between CAIBP and normal badminton players (NBP) during single-leg medial landing tasks. Methods: Sixteen CAIBP and sixteen NBP university badminton players volunteered to participate in this experiment. The study used OpenSim open-source software to simulate and calculate lower limb joint angles, moments, and joint stiffness for the CAI group and healthy controls during a single-leg medial landing task, and utilized Delsys EMG to assess muscle pre-activation and activation levels. Independent samples t-tests and one-dimensional statistical parametric mapping were used to analyze the experimental results. Results: In terms of kinematics, before and after initial contact (IC) during landing, CAIBP showed significantly greater hip adduction and flexion angles than NBP (p < 0.001). Pre-IC, CAIBP exhibited less knee flexion (p = 0.004). Both pre- and post-IC, CAIBP exhibited significantly greater dorsiflexion angles (p = 0.045) and inversion angles (p < 0.001). In terms of muscle activation and dynamics, pre-IC, CAIBP had significantly less pre-activation of the peroneus longus than NBP (p = 0.007), and significantly more gastrocnemius lateral (p = 0.021) and gastrocnemius medial (p < 0.001) pre-activation. Post-IC, CAIBP had significantly greater muscle activation of the tibialis anterior (p < 0.001). Post-IC, peak knee extension moments (p = 0.012) and peak ankle plantarflexion moments (p = 0.001) were significantly greater in CAIBP than in NBP. In addition, CAIBP reported significantly higher knee stiffness (p = 0.001) and ankle stiffness (p < 0.001). Conclusions: During the medial landing task, CAIBP exhibited increased hip adduction and flexion, altered sagittal plane motion of the ankle, and increased activation of certain lower extremity muscles compared to NBP. Although these altered landing mechanisms contribute to enhanced stability during landing to some extent, they may also increase the potential risk of knee injury.

Transcanal endoscopic ear surgery (TEES) is a minimally invasive approach for treating ear diseases. TEES requires simultaneous insertion of an endoscope and surgical instruments through the narrow external auditory canal (less than 1 cm in diameter). Due to this anatomical constraint, conventional TEES has limitations including one-handed instrument manipulation and endoscope instability. We developed a compact, bed rail-mounted endoscope manipulator to enable two-handed surgery. The manipulator features three degrees of freedom: yaw and pitch movements controlled by gimbal rotational linkages, and insertion facilitated by a linear guide rail. We evaluated the mechanical performance of the prototype using an optical displacement sensor, and demonstrated positional accuracy within 0.1 mm across all axes under a 6-N load. Functionality of the manipulator was verified by cadaver trials and an ear canal model simulation. Five otolaryngologists performed a simulated surgical task using both manual and robot-assisted endoscopic procedures. While there was no significant difference in task completion time between manual and robot-assisted methods, less experienced surgeons showed improved performance and reduced time variability with the robot. The manipulator demonstrated satisfactory performance, meeting the dexterity and precision required for ear surgery. This novel endoscope-holding robot shows potential to enhance surgical precision and efficiency in TEES, particularly benefiting less experienced surgeons. Future iterations will aim to optimize its form factor and functional complexity to enhance clinical utility.

A coronary artery aneurysm (CAA) occurs when the wall of the coronary artery that surrounds and nourishes the heart becomes locally dilated. Coronary aneurysms are asymptomatic; however, a serious complication is myocardial infarction caused by thrombosis, which can lead to sudden death. Computational fluid dynamics analysis is used to evaluate the risk of thrombosis in CAA. However, a rigid wall model is generally assumed; thus, this analysis does not accurately reproduce the hemodynamics of coronary arteries adjacent to the heart. This study investigated the effect of wall elasticity on the hemodynamics of CAA caused by Kawasaki disease. A fluid-structure interaction (FSI) analysis considering the heartbeat was performed to compare the time-averaged wall shear stress (TAWSS), highly oscillatory low magnitude shear (HOLMES), and oscillatory shear index (OSI) using a rigid wall model. A computational model of a branching coronary artery with an aneurysm was created from the vessel model repository provided by the open source cardiovascular modelling platform SimVascular, with a vessel wall thickness of 0.9 mm. For the boundary conditions, a physiological flow rate waveform was imposed at the inlet. The distal branches at the outlet were connected to a lumped parameter model considering the intramyocardial pressure. As a result, considering elastic walls, the OSI in the CAA increased by 42.8%. In contrast, the TAWSS decreased by 3.6%, and the HOLMES showed only a 5.4% reduction. These findings suggest that elastic wall properties significantly impact the OSI, indicating that simulations assuming rigid vessel walls may overlook the progression of atherosclerosis.

Research reporting significant cognitive improvement in patients with severe dementia following non-pharmacological interventions is scarce. Furthermore, few studies have focused on improvement of Mini-mental State Examination (MMSE) score in individuals with severe dementia following interventions implemented during the coronavirus 2019 (COVID-19) lockdown. COVID-19-related quarantine significantly worsened cognitive function and depression in residents of a Japanese geriatric health facility. The MMSE Japanese version (MMSE-J) scores of four participants with moderate to severe dementia (aged 88.3 ± 2.9 years) were 14.0 ± 2.2 before quarantine and 9.0 ± 1.9 one week after quarantine was started. Individualized active music therapy using Cymis (Cymis-MT), which is a new electronic musical instrument, was administered to four participants for 5 days (two sessions per day) during quarantine. Cognitive function during Cymis-MT improved significantly (P < 0.05; Cohen's d > 2.0). MMSE-J scores at 2 and 5 days after initiation of intervention did not differ significantly from the score before quarantine. In addition, Cymis-MT significantly improved depression, as measured by the Geriatric Depression Scale, short version (Cohen's d = 2.9). In conclusion, short-term Cymis-MT administered to non-infected patients with severe dementia during COVID-19 quarantine restored their cognitive function to pre-quarantine level. Moreover, their level of depression also improved during Cymis-MT. Therefore, Cymis-MT may be recommended as an intervention for individuals with moderate to severe dementia in seclusion situations.

Motor imagery (MI)-based brain-computer interfaces (BCIs) have received increasing attention in recent years. While most of the existing electroencephalogram (EEG) based MI-BCIs have demonstrated satisfactory performance in binary and four-class classifications, it is essential to develop intuitive systems with higher dimensionality for object-centered control to support daily living. However, increasing the dimensionality in MI-BCI has been a major challenge. One of the main reasons is the difficulty in selecting appropriate MI tasks. To solve this problem, we proposed a novel MI paradigm that selected six real life movements that have underlying intents aligned with the object-centered control of external devices along three axes: front-back, left-right, and up-down. The users' perspective of intuitiveness was evaluated using an intuitiveness questionnaire. The results from twelve participants indicated that using the proposed MI tasks achieved closer cognitive matches compared to the traditional tasks. Information-preserved data conversion was implemented to preserve temporal, spatial and band power features in video-like data. A spatiotemporal convolutional neural network (CNN) was then applied for six-class classification. The highest accuracy achieved by the proposed system in six-class within-participant classification was 46.88% for the hold-out validation dataset and 39.66% for the test dataset. The average validation accuracy was 33.99% and the test accuracy was 26.26% for seven healthy participants, which were significantly higher than the chance level of 16.67%. Furthermore, the key metrics obtained from the confusion matrices suggested that both the similarity of somatotopic representations and the symmetry involved in the MI signals significantly affected the classification performance. In this study, we introduced the use of intuitive movements in intuitive MI-BCI paradigms and demonstrated their suitability for controlling external objects in 3D space. Our results indicate that compound MI signals involving multiple joints are separable using deep neural networks, which would greatly expand the available tasks for multiclass MI-BCIs. Although large individual differences were observed, the possibility of adopting spatiotemporal CNN classifiers to decode complex MI signals was demonstrated. Based on our findings, future research with larger sample sizes should consider task-specific modifications of the architecture of the classification model.

For the application of brain-computer interfaces in fields such as medicine, a method of classifying brain activity with high accuracy using small amount of training data is needed. To apply machine learning to biological signal processing, it is important to develop a method based on a model that matches the in vivo mechanism. While a high dimensional and high degree of freedom model can reproduce complex dynamics in a living body, it tends to overfit to the training data and may estimate activities that cannot exist in a living body when encountering novel data. Therefore, to improve the accuracy of classification and prediction, it is necessary to build a model that fits the nonlinearity and time delay of the neural population. In this study, we propose a machine learning method for the data from brain activity measurements by computing the time evolution of nodes in the reservoir computing network using the macroscopic neural population model. Since reservoir computing reproduces signals by linear summation of the time evolution of nodes, the outputs of the proposed method are limited to the waveforms that are based on the dynamics of the neuron population model. To verify the effect of the proposed method, the brain response to visual stimuli following a pseudorandom pattern was measured by magnetoencephalography, and the response of brain activities evoked by different stimulus patterns were estimated. The correlation coefficient between the measured waveform and the estimated waveform generated by the proposed method was higher than that of the conventional method. Comparison of individual node responses suggests that the non-linearity and time delay of the firing rate model, and the interaction between the excitatory and inhibitory neuronal populations contributed to the improvement in the adaptation to in vivo dynamics. Analytical approaches that incorporate knowledge of life phenomena for specialized machine learning models suitable for simulating biological phenomena will become more important in the future.

Microneedle arrays (MNAs) consisting of numerous tiny needles have attracted considerable attention for selective intradermal and subcutaneous delivery of pharmaceutical agents. Since reliable skin puncture and insertion are essential for dose control, applicators that push MNAs into the skin using springs or other mechanisms have been extensively researched. The puncture and insertion capabilities of an MNA depend on the compressive load generated when it is pressed against the skin. The compressive load is considered to be influenced not only by the MNA per se and the application method, but also by the structures and mechanical properties of the biological tissues at the application site. Nevertheless, most previous studies on evaluating the puncture and insertion abilities of MNAs did not consider the impact of subcutaneous soft tissues. In this study, the compressive loads generated by an applicator were measured at various sites of the body in 15 healthy adults. We used a proprietary test applicator that employed a spring to impact the skin at high speed, and a system that can measure the short-term load generated during impact. Furthermore, the structures and mechanical properties of the skin and subcutaneous soft tissues at the same body locations among the 15 individuals were evaluated and correlated with the measured compressive loads. The compressive load induced by the impact varied significantly with location, and correlated highly with the distance to the bone. There was no correlation with age, sex, or body mass index, or with the structure or mechanical properties of the skin. These results indicate that the structure of the subcutaneous tissue has a strong influence on the compression load and the ability to puncture and insert the MNA. The findings of this study can be useful for developing MNAs and applicators, determining their essential physical characteristics, selecting the appropriate administration locations, and developing non-clinical methods for evaluation.

Demand is increasing for foods that are both palatable and beneficial for physical and mental health. This study investigates the influences of flavored jelly beverage consumption on emotional sensitivity by examining the heart rate variability (HRV) of 11 participants (six men and five women; age range: 20-50 years). To investigate the effects of different food quality factors, four samples were tested: a standard flavored jelly beverage, flavored jelly beverage without fruit juice and flavoring, flavored jelly beverage without the gelling agent, and water. The study compared the effects of flavor and gelling agents by analyzing HRV before, during, and after consuming the test samples. Analysis of variance of the HRV data indicated that the presence of gelling agents suppressed the activity of the parasympathetic nervous system during consumption, whereas flavoring enhanced the activity of the parasympathetic nervous system after consumption, suggesting different effects of the samples on autonomic nervous system functions during and after ingestion. Furthermore, when examining correlation with subjective mood assessment, strong associations were observed among several parameters. Notably, the activity of the parasympathetic nervous system was suppressed, and the arousal level was elevated during consumption, while the activity of the parasympathetic nervous system and feeling of pleasure were enhanced after consumption. These correlations imply a potential causal relationship between the sensory properties of the jelly beverage and physiological responses.

Early prediction of the risk of postoperative recurrence can help formulate treatment strategies for patients with lung cancer; however, few medical techniques accurately predict cancer recurrence, which remains a challenge in clinical practice. In addition, no previous reports on prediction of postoperative recurrence of lung cancer by computed tomography (CT) imaging have discussed the performance when using non-contrast-enhanced CT images. The aim of this study was to propose a computer-aided diagnostic (CAD) system using an ensemble of multiple deep convolutional neural networks (DCNNs) for predicting postoperative recurrence in patients who had undergone lung adenocarcinoma surgery, from contrast-enhanced and non-contrast-enhanced CT images obtained preoperatively. This retrospective study included 190 patients who underwent surgery for primary lung adenocarcinoma. The patients included in the study underwent preoperative CT scanning with a mixture of methods: both contrast-enhanced and non-contrast-enhanced, only contrast-enhanced, and only non-contrast-enhanced. In this study, data augmentation increased the number of images to 20 per patient for contrast-enhanced and non-contrast-enhanced images. This CAD system was constructed through transfer learning with five pretrained DCNNs using an ensemble method. The average of the probabilities output by these DCNN models was used as the probability of the ensemble method. We defined recurrence as reappearance of lesion within 5 years after surgical resection. Individual DCNN had the best prediction performance with 72% sensitivity, 88% specificity, 76% accuracy, and area under the receiver operating characteristic curve (AUC) of 0.80. The ensemble method performed best, with 70% sensitivity, 90% specificity, 75% accuracy, and AUC of 0.82. The ensemble method using multiple DCNN models is effective in improving the AUC for predicting postoperative recurrence of lung cancer using preoperative CT images, regardless of whether the images were contrast-enhanced, non-contrast-enhanced, or a combination of contrast-enhanced and non-contrast-enhanced. The ensemble method achieves optimal results when combining a large number of non-contrast-enhanced CT images with a small number of contrast-enhanced CT images. This approach achieves highly accurate prediction of postoperative recurrence using preoperative non-contrast-enhanced CT images, allowing noninvasive preoperative assessment.

The purpose of this study was to clarify the effect of reducing somatosensory feedback by foot cooling on the medial gastrocnemius muscle. Using a system identification technique, we estimated the natural frequencies of the muscle and the ankle joint. The natural frequencies are proportional to the square root of the muscle and joint stiffnesses and indicate the frequency characteristics of quiet standing. We measured the center of pressure (COP) of eight young volunteers aged 22-24 years before and after foot-cooling by immersing the feet in water with ice. Electrical stimulation was applied to the medial gastrocnemius muscle, and electrically induced COP fluctuation was extracted using a Kalman filter and synchronously averaged based on the stimulation. The synchronously averaged COP fluctuation was considered an impulse response, and the transfer function was identified using a singular value decomposition method. The natural frequencies were estimated from the poles of the transfer functions. The natural frequencies of the muscle before foot cooling ranged from 1.96 to 3.09 Hz, and those after foot cooling decreased to a range of 1.40 to 2.19 Hz. There was a significant difference (p < 0.01) between the natural frequencies before and after foot cooling. The natural frequencies of the ankle joint tended to decrease after foot cooling, although the difference was not significant. The natural frequency of the medial gastrocnemius muscle decreased significantly when foot cooling reduced somatosensory feedback in the feet.

Quality of life (QOL) measurement is essential in evaluating effective interventions for persons with dementia. Essentially, QOL is assessed based on self-ratings. As the disease progresses, persons with dementia cannot perform these self-ratings; however, their emotional expressions can be observed. Thus, proxy-rating is used, which involves another person such as a caregiver, who answers from the perspective of the person studied or evaluates the emotional state from the person’s facial expressions. However, the results of proxy-rating differ from those of self-rating. This is considered to be partly due to the subjectivity of the caregiver who performs the rating. Thus, an objective method is needed to evaluate the QOL of persons with dementia. A pilot study was conducted to detect objectively the smiles of persons with dementia, because conventional QOL assessments include facial expression-related items, and smile is strongly related to pleasantness or friendliness. To detect smiles, we developed an image processing method including tilt correction of the facial image and face detection of a specified person among many faces regardless of the facial expressions presented. To verify the effectiveness of the proposed method, we measured smiles before and after a day care intervention program that involved 1 hour of physical exercises and 2 hours of cognitive stimulation, which has been shown to improve QOL. We developed an experimental protocol to quantify the changes in the smiles of 11 persons with dementia at a day care facility. A facility staff member induced each person to smile using simple words while recording video. Each recorded video was then processed to calculate the increase in detection of smiles evoked by the staff member. By comparing the differences before and after participating in the day care program, we confirmed that the frequency of smile detection increased significantly after the program.

The objective of this study was to design and develop a tongue model capable of accurately replicating the tongue movements observed during infant suckling, including jaw elevation, front-tip tongue pressing against the nipple, and milk expulsion through peristaltic movements. In a previous study, our group utilized imaging techniques to analyze tongue movements and combined the data with tongue force measurements to create a distinctive model known as the Bio-Inspired Breastfeeding Simulator (BIBS). While the existing tongue design in the BIBS system is suitable for studying milk flow during breastfeeding, a new tongue design is necessary to investigate the forces involved in the interactions between the tongue and nipple and breast models. To address this need, we designed and developed a novel tongue model based on force and frequency measurements obtained from tongue movement data collected from over 100 infants. These measurements were acquired using a bottle-type device equipped with integrated force sensors. The newly developed tongue model was inspired by clinical data collected and cross-validated with sample infant suckling data, resulting in a model that accurately replicates an infant's feeding dynamics.

Background: The impact of instability resistance training (IRT) on the biomechanical properties and muscle activation patterns of persons with different arch heights is not well researched, despite its popularity in the domains of sports and rehabilitation. The purpose of this study was to investigate the effects of IRT on the lower limb biomechanics, muscle activation patterns, and motor control abilities of individuals with different arch heights and to provide scientific evidence for personalized exercise training and rehabilitation strategies.

Methods: Sixty-six healthy adult male subjects were selected and categorized into three groups based on their arch height. They performed Bulgarian squats on an unstable surface. The kinematics data obtained from the Vicon system were imported into an OpenSim musculoskeletal model and analyzed using SPM1D one-way ANOVA. The electrical data collected from the electromyographic (EMG) device was utilized for EMGworks Analysis, which involves processing and calculating integrated EMG (IEMG) and root mean square (RMS) contributions, as well as analyzing muscle activation. SPSS was used to perform one-way ANOVA.

Results: Ankle inversion and eversion showed significant differences between individuals with low arches and those with high arches (0%-16% phase, p < 0.001; 66%-95% phase, p < 0.001) as well as those with normal arches (81%-89% phase, p = 0.011). The gastrocnemius medial (GL) RMS was found to be considerably lower in individuals with high arches compared to those with low arches (p = 0.040) and normal arches (p < 0.001). The IEMG of the rectus femoris (RF) muscle was considerably reduced in individuals with high arches compared to those with low arches (p = 0.007) and normal arches (p = 0.014).

Conclusions: Individuals with high arches encounter higher joint strain and reduced muscular engagement while performing IRT activities. Customized exercise that targets specific muscle groups, in conjunction with IRT, enhances joint stability. Individuals with low arches may experience decreased foot stability and muscle fatigue during intense physical activity, such as IRT. These individuals can benefit from improved foot support. This method enhances muscle coordination and modulation, hence improving sports performance and providing helpful exercise suggestions to prevent injuries in the general population.

A cluster of coronavirus disease (COVID-19) cases occurred in a 40-bed private hospital in Oita City, Japan. This study examined the effect of heat and moisture exchangers (HMEs) with electrostatic filters on infection control. The study included 32 hospitalized patients, including four patients with normal breathing (non-tracheostomized patients), two tracheostomized patients with spontaneous breathing, and 26 tracheostomized patients on ventilators. HMEs without filters were used for airway management in the two tracheostomized patients with spontaneous breathing and one tracheostomized patient on ventilator, and HMEs with electrostatic filters were used in 25 tracheostomized patients on ventilators. All patients with negative test results for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) antigen were further tested by reverse transcription polymerase chain reaction to confirm the absence of infection. All non-tracheostomized patients and tracheostomized patients with spontaneous breathing (100%) and two of the 26 (7.6%) tracheostomized patients on ventilators were confirmed to have SARS-CoV-2 infection. Both patients (100%) who used HMEs without filters were infected, whereas only one of the 25 patients (4%) using HMEs with electrostatic filters was infected, showing a significantly lower rate of infection in the latter group. The present results suggest that in situations where the level of virus contamination in the environment is high, such as during a large-scale COVID-19 outbreak, use of HMEs with electrostatic filters would be effective for controlling infection.

Background: Fractional flow reserve (FFR) is the gold standard for invasive assessment of myocardial ischemia in the diagnosis of coronary artery disease (CAD). Traditionally, FFR is measured using pressure wires during coronary angiography. Recently, computational fluid dynamics (CFD) has been applied to calculate FFR noninvasively and efficiently. However, conventional CFD methods often require complex computations. This study aimed to evaluate simplified analytical conditions for CFD-based FFR calculations.

Methods and Results: We enrolled 25 patients with angina who underwent coronary computed tomography (CT), coronary angiography, and invasive FFR measurement. Using 3D data reconstructed from CT, CFD analysis was conducted with a vascular volume ratio-based boundary condition. Results demonstrated a Pearson correlation coefficient r of 0.98 and a concordance correlation coefficient CCC of 0.982, with an error margin within ± 5%. Additionally, one case with signal drift during invasive measurement achieved comparable results using this CFD approach.

Conclusion: This study confirmed the utility of a CFD method utilizing vascular volume ratios for non-invasive FFR calculations. This simplified approach facilitates straightforward and accurate FFR assessment, offering a potential novel tool for early CAD detection and treatment planning.

The demand for telemedicine systems has increased following the COVID-19 pandemic. Conventional telemedicine systems provide video and acoustic transferring functions using smartphones and other communication systems. However, it is difficult to use conventional telemedicine systems in rehabilitation fields because they require real-time motion data from patients. In this study, we introduce an innovative system to transfer three-dimensional (3D) motion data to remote places in real time, which is referred to as the “remote motion capture protocol (RMCP).” The RMCP runs on a conventional game engine (Unity) and multiple operation systems such as Windows, Mac OS, and Android. The transferring delay was estimated to be 230 ms at the maximum using remote procedure call. In addition, any type of motion capture can be used in this system if it is compatible with the conventional game engine. This technical note describes the basis of the system, a sample application, and usage of RMCP.

In long-term care facilities for the elderly, understanding physical anomalies or daily living conditions is important to improve the quality of care. Internet of Things (IoT) sensors are used to monitor the conditions of users, but it remains difficult to understand the full aspects of users' daily lives. Facilities often maintain care records as text data. Therefore, it is difficult to use sensor data systematically to understand physical anomalies or living conditions. In this study, we constructed a system that uses natural language processing artificial intelligence to analyze nursing care records and automatically determine whether an abnormality exists and the type of abnormality. This study aimed to understand the condition of the users using both care records and IoT sensors. We also studied a system for creating summaries of care records based on a condition judgment system. We collected care records from one nursing home between January 2021 and November 2022; built a natural language processing model that classifies the presence or absence of abnormalities and the types of abnormalities into multiple classes using XLNet; and evaluated the classification performance. The results of normal cases achieved recall of 0.70 and precision of 0.92. In cases of dementia-related symptoms (disquiet), the recall and precision were 0.80 and 0.39, respectively. There were many cases in which records from another data class contaminated the disquiet group. We plan to develop methods to improve judgment performance. In addition, we compared the generated summaries with the actual notes. In some cases, the system was able to generate almost the same contents as the summary records prepared by humans.