Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference最新文献

Newborns communicate with the outside world primarily by crying. Infant cry-based verification can reduce the risk of mix-ups in hospital obstetrics. Recent studies have explored the potential of using infant cries for identity verification. Yet, model performance remains limited by training with variable-length clips and evaluating the complete audio recording from a single view. To this end, we propose a novel unified training and evaluation framework that uses fixed-length segments during training to ensure input consistency and incorporates a multi-view joint evaluation strategy by associating the audio recording with its local segments. Extensive experiments conducted on the public CryCeleb2023 dataset show that our framework leads to consistent improvements on different verification models. Specifically, the Equal Error Rate (EER) exhibited a reduction of 10.29% for the whisper-PMFA model, 6.63% for the X-Vector model, and 5.91% for the ECAPA-TDNN model. These results demonstrate the effectiveness of our fixed-length segment training and slice-based multi-view evaluation strategy in enhancing the model stability and evaluation accuracy, providing a more robust framework for newborn voice verification. The source code is released at https://github.com/contactless-healthcare/Unified-Infant-Cry-Verification.

Ocular artifacts, particularly blinks, significantly affect the integrity of electroencephalographic (EEG) signals, posing a challenge for real-time applications. Traditional correction methods often require a calibration phase or additional electrooculogram (EOG) channels, limiting their applicability in mobile and real-world settings. This study presents a novel detection and correction method, designed for online ocular artifact correction without the need for prior calibration: the CFo-CLEAN. The proposed method integrates an Enhanced Adaptive Data-driven Algorithm (eADA) for real-time identification and correction of ocular artifacts directly from EEG signals. Unlike conventional approaches, this implementation adapts dynamically to ongoing EEG variations, enhancing flexibility and performance. The study evaluates the CFo-CLEAN method using EEG data recorded from 38 participants during real-world driving scenarios. Performance comparisons were conducted against established correction techniques, including Independent Component Analysis (ICA), regression-based methods, and subspace reconstruction approaches. The evaluation considered both artifact removal efficiency and EEG signal preservation across different experimental conditions. Results demonstrated that the method effectively reduced ocular artifact contamination while preserving neurophysiological content. Specifically, two implementations of the method, utilizing 60-second and 90-second time windows, were analyzed, revealing that longer windows provided superior EEG signal preservation, particularly in higher frequency bands. These findings validate the effectiveness of the CFo-CLEAN method for real-time applications, making it a valuable tool for brain-computer interfaces (BCIs), neuroergonomics, and cognitive state monitoring. By avoiding the need for a calibration phase and incorporating adaptive processing, this method represents a significant advancement in real-time EEG artifact correction, facilitating its deployment in dynamic, real-world environments.

Tumor-associated macrophages (TAMs) are critical to tumor progression. Quantifying their interactions with biomaterial surfaces is crucial for developing effective cancer therapies. Traditionally, manual cell counting has been used to assess macrophage adhesion, a labor-intensive and subjective process. To address these limitations and enable unbiased analysis, we developed an automated in-situ system to quantify TAM attachment to antifouling polymer brushes. Bland-Altman analysis indicated a high agreement between our automated method and traditional manual cell counting. For M1 macrophages, the mean difference was less than 4 cells, with limits of agreement (LoA) ranging from -70.18% to 80.16%. For M2 macrophages, the mean difference was 25 cells, with LoA ranging from -51.61% to 72.71%. These results were consistent across different experimental conditions, including Unspecific Binding, Specific Antibody, and IgG Control. Our analysis revealed no systematic differences in cell counts and holds significant potential for point-of-care applications, potentially enhancing personalized treatment strategies.Clinical relevance- This approach could enhance personalized treatment strategies by providing real-time assessment of the tumor microenvironment through minimally invasive liquid biopsies.

This work introduces a novel phenomenological model designed to replicate the deterministic aspects of artifacts that affect retinal imaging. To validate the model's ability to reproduce real artifacts, we utilized the CORD database, which contains retinal images affected by artifacts, along with corresponding clinically standard-quality images, serving as ground truth. The model was implemented in a Matlab script recreating various artifact distortions. The results demonstrate a robust correlation between the quality attributes of simulated artifact-affected images and real-world artifacts, with ANOVA tests yielding p-values > 0.05 across the most discriminative features (e.g., mean, IQR, and BVC). Furthermore, a quality classification analysis using Neighbourhood Components Analysis showed overlapping distributions between real and generated artifacts, supporting the model's ability to mimic realistic quality deterioration. This underscores the model's utility as a tool for generating synthetic artifacts, addressing the current lack of available datasets, with potential impact on the development of quality retrieval algorithms and modeling in digital retinal images.Clinical Relevance-The proposed model enables the generation of synthetic retinal imaging artifacts that closely resemble real-world distortions, providing a valuable tool for developing and evaluating quality enhancement algorithms in clinical ophthalmic imaging.

In magnetic resonance imaging (MRI), susceptibility-induced distortions pose a significant challenge for images acquired using echo planar imaging (EPI). Classical methods use two EPI images acquired in reverse phaseencoding (PE) directions to correct susceptibility artifacts. However, these methods suffer from long computation times, making them impractical for clinical usage. Recently, deep learning-based approaches have been proposed to enable a leap in computation efficiency for EPI susceptibility artifact correction. A vital consideration in reverse-PE-based correction is the need to take into account any potential subject motion between reversed-PE acquisitions. In this work, we propose an alignment-guided forward distortion network (agFD-Net) that accounts for subject motion during correction of susceptibility artifacts. Similar to its predecessor FD-Net, agFD-Net is trained in a physics-driven unsupervised fashion to estimate a single corrected image and a displacement field. In agFD-Net, a new pre-trained alignment network called AlignNet is plugged into the network architecture to facilitate motion correction. The results on experimental NIH dataset featuring realistic levels of motion demonstrate that agFD-Net provides rapid and high-fidelity artifact correction, while successfully accounting for subject motion.Clinical Relevance-EPI is the most commonly used sequence for diffusion MRI and functional MRI. While susceptibility artifacts in EPI require correction before any downstream evaluation, the long computation time of classical correction methods make them impractical for use in clinical settings. The proposed agFD-Net provides more than two orders of magnitude speed up in computational efficiency, making it a highly promising approach for use in clinical settings.

Extracorporeal shock wave therapy (ESWT) is a commonly used physical therapy method in clinical and rehabilitation fields. This paper assesses the effect of ESWT in the rapid relief of exercise-induced muscle fatigue, based on fatigue protocol and ESWT intervention experiments, combined with subjective and objective fatigue evaluation methods. After 22 subjects experienced exercise-induced biceps brachii fatigue, the experimental group received a single 250-second ESWT intervention, while the control group underwent placebo treatment. During the experiment, the Borg Rating of Perceived Exertion (RPE) scores, maximum voluntary contraction (MVC), and electromyographic (EMG) data were recorded, and the values of EMG time-domain and frequency-domain parameters were calculated. After intervention, the MVC values of the experimental group significantly recovered, and the EMG parameters returned to baseline levels, all of which showed significant differences compared to the control group. The RPE scores also significantly recovered, but the recovery effect was not significantly better than that of the control group. A single session of ESWT intervention can effectively improve exercise-induced muscle fatigue, characterized by the recovery of muscle strength and the improvement of electrophysiological status.

Objective and precise measurement of facial muscle activity is crucial for understanding emotional expressions across diverse stimuli. However, traditional methods often require controlled lab environments, limiting real-world applicability. In this study, we introduce an automated framework that integrates wireless facial surface electromyography (sEMG) with a robust facial muscle Atlas for comprehensive facial expression analysis. This approach enables cross-subject and inter-subject measurements, facilitating applications in psychology, affective computing, and clinical research. We validate our framework by analyzing facial muscle activity in response to olfactory, visual, and auditory emotional stimuli, demonstrating its efficacy in capturing nuanced expression dynamics.Clinical relevance- This work advances emotion research by providing a scalable and objective tool for facial expression analysis, with potential applications in psychology and medical diagnostics, particularly in conditions where facial expressions play a crucial role.

Traditional gel-based electrodes are widely used for bioelectric signal acquisition, but they come with drawbacks such as skin irritation, signal degradation over time, and the need for frequent replacements, making them less ideal for long-term health monitoring scenarios. This study explores the development of 3D-printed dry electrodes as a gel-free, reusable alternative for ECG monitoring. Gel-less electrodes of six distinct electrode geometries namely flat, dome, flat needle, curved needle, half pointed, and Pointed were fabricated using fused deposition modeling (FDM) with conductive polylactic acid (PLA) to evaluate the impact of shape on signal quality. ECG data collected in both resting and exercise conditions was evaluated by measuring the electrodes' electrical conductivity and analyzing key parameters, including kurtosis, skewness, signal-to-noise ratio (SNR), and standard deviation of Normal-to-Normal Intervals (SDNN), to compare their performance against conventional gel-based electrodes. The results demonstrate that 3D-printed dry electrodes can provide high-quality signals comparable to conventional electrodes in wild scenarios. Among the tested designs, the flat and FN electrodes exhibited the best performance making them particularly effective under both static and dynamic conditions. Future work will focus on enhancing flexibility, stretchability, and adhesion to improve comfort and long-term usability, making them more suitable for continuous, real-time monitoring in wearable healthcare applications.

Disorders of Consciousness (DoC) are clinical conditions characterized by different levels of arousal and awareness, including coma, Unresponsive Wakefulness Syndrome and Minimally Conscious State (MCS). A Brain-Computer Interface (BCI) employs brain signals to establish a non-muscular outward channel, representing a key frontier in the clinical care of individuals in MCS, with high potential to enhance communication and quality of life. The P300-based BCIs, which use the P300 ERP as a control signal, are the most investigated to emulate communication in MCS. However, a reliable control by MCS patients of these BCIs still remains matter of question. One major challenge could be the across trials variability of P300 characteristics, possibly related to attentional fluctuations in this population. The trial-by-trial instability of the P300 peak latency, known as latency jitter, negatively impacts classification performance, and an approach to mitigating this issue involves template matching algorithms (e.g. the Adaptive Wavelet Filtering, AWF) which detect and realign the P300 latency at the single-trial level. This study investigated the offline classification performance using Stepwise Linear Discriminant Analysis (SWLDA) models trained with progressively larger training sets, to discriminate target from non-target stimuli during an active auditory oddball paradigm. Performance from raw and jitter-corrected data, collected from a control group and a group of patients diagnosed as MCS, were compared. Results highlighted the key role of latency jitter correction in the enhancement of performance and classification speed.Clinical Relevance- The findings suggest that jitter correction could improve real-world applicability of P300-BCI systems for individuals with DoC.

Different brain rhythms are often signatures of different behavioral and perceptual states of the brain. As resonance at neurons has a close connection with brain rhythms, characterizing and analyzing the neuronal resonance promote the understanding of brain rhythms and dynamical organization of the brain. In this paper, we study the resonant behaviors of both the four-dimensional (4D) Hodgkin-Huxley model and a reduced-order (2D) model at the subthreshold level. We analyze the frequency responses of a neuron and characterize resonance through investigating the neuron's transfer functions, frequency response functions, and root locus plots. The 2D model, especially, allows us to visualize the state space, perform phase plane analysis, and derive a closed-form formula for the resonant frequency. These analyses will be generalized to study resonance at the spiking regime and network level in future work.