Electrical Engineering and Computer Science

Hearing loss is a widespread and disabling condition with no current cure, underscoring the urgent need for new therapeutic approaches for treatment and prevention. A recent mitochondrial therapy approach by introducing exogenous mitochondria to the cells has shown promising results in mitigating mitochondria-related disorders. Despite the essential role of mitochondria in hearing, this novel strategy has not yet been tested for the treatment of hearing loss. More importantly, whether cochlear cells take up exogenous mitochondria and its consequence on cell bioenergetics has never been tested before. Here, we showed that exogenous mitochondria from HEI-OC1 auditory cells internalize into a new set of HEI-OC1 cells through co-incubation in a dose-dependent manner without inducing toxicity. We observed that auditory cells that received exogenous mitochondria exhibited increased bioenergetics compared to the controls that received none. Furthermore, we found that mitochondrial transplantation protects cells from oxidative stress and H2O2-induced apoptosis, while partially restoring bioenergetics diminished by H2O2 exposure. These findings support initial evidence for the feasibility and potential advantages of mitochondrial therapy in auditory cells. If successful in animal models and ultimately in humans, this novel therapy offers prominent potential for the treatment of sensorineural hearing loss.

BACKGROUND AND OBJECTIVES: Culturally diverse informal caregivers of community-dwelling persons with dementia face challenges in accessing dementia care resources due to language barriers and cultural stigmas surrounding dementia. This study presents the perceived intervention experiences of a home-based approach which considers the cultural and linguistic needs of diverse family caregivers in dementia care. The intervention model includes home visits by trained bilingual, non-licensed community health workers (CHWs) whose cultural histories and understandings reflect that of the caregivers. The purpose of the present study was to understand family caregivers experience in caregiving and their feedback on the intervention, which includes caregiver support through education and skill development. METHODS: The present study thematically analyzed qualitative data from exit interviews with caregivers who participated in a CHW-led, 12-week home visit-based intervention program. RESULTS: Among 57 caregivers (mean age = 63.5, SD = 14.3) who participated in the 3-month home-visit intervention and completed the exit interviews, 33% were Korean Americans, 28% Vietnamese Americans, 21% non-Hispanic Whites, and 17.5% Latino/Hispanic. The majority were females (81%) and spouses (51%). Main themes include, (a) Individual Level: Improvements in Caregiving Self-efficacy and Self-care Awareness, (b) Relational Level: Enhanced Communications and Relationships with Persons with Dementia, and (c) Community Level: Connection and Access to Community Resources and Support. CONCLUSION: Interview data show that the culturally and linguistically tailored program supported diverse caregivers by increasing self-care awareness, improving knowledge about dementia and dementia care, strengthening communication skills, and facilitating access to community resources. Strong rapport between CHWs and caregivers enhanced the effectiveness of the intervention. Future approaches can focus on supporting caregivers with especially limited resources.

BACKGROUND: Acute pain management is critical in postoperative care, especially in vulnerable patient populations that may be unable to self-report pain levels effectively. Current methods of pain assessment often rely on subjective patient reports or behavioral pain observation tools, which can lead to inconsistencies in pain management. Multimodal pain assessment, integrating physiological and behavioral data, presents an opportunity to create more objective and accurate pain measurement systems. However, most previous work has focused on healthy subjects in controlled environments, with limited attention to real-world postoperative pain scenarios. This gap necessitates the development of robust, multimodal approaches capable of addressing the unique challenges associated with assessing pain in clinical settings, where factors like motion artifacts, imbalanced label distribution, and sparse data further complicate pain monitoring. OBJECTIVE: This study aimed to develop and evaluate a multimodal machine learning-based framework for the objective assessment of pain in postoperative patients in real clinical settings using biosignals such as electrocardiogram, electromyogram, electrodermal activity, and respiration rate (RR) signals. METHODS: The iHurt study was conducted on 25 postoperative patients at the University of California, Irvine Medical Center. The study captured multimodal biosignals during light physical activities, with concurrent self-reported pain levels using the Numerical Rating Scale. Data preprocessing involved noise filtering, feature extraction, and combining handcrafted and automatic features through convolutional and long-short-term memory autoencoders. Machine learning classifiers, including support vector machine, random forest, adaptive boosting, and k-nearest neighbors, were trained using weak supervision and minority oversampling to handle sparse and imbalanced pain labels. Pain levels were categorized into baseline and 3 levels of pain intensity (1-3). RESULTS: The multimodal pain recognition models achieved an average balanced accuracy of over 80% across the different pain levels. RR models consistently outperformed other single modalities, particularly for lower pain intensities, while facial muscle activity (electromyogram) was most effective for distinguishing higher pain intensities. Although single-modality models, especially RR, generally provided higher performance compared to multimodal approaches, our multimodal framework still delivered results that surpassed most previous works in terms of overall accuracy. CONCLUSIONS: This study presents a novel, multimodal machine learning framework for objective pain recognition in postoperative patients. The results highlight the potential of integrating multiple biosignal modalities for more accurate pain assessment, with particular value in real-world clinical settings.

Electroacoustic imaging is an imaging modality used to detect electric field energy distribution during electroporation, offering valuable guidance for clinical procedures, particularly in deep tissues. Traditionally, single-element piezoelectric transducers or arrays have been employed for this purpose. However, these piezoelectric sensors are sensitive to electromagnetic interference and require physical contact with the sample through a coupling medium, raising concerns for both clinical and preclinical applications. To overcome these limitations, a multi-channel random quadrature ultrasonics system has been developed, enabling non-contact detection of electroacoustic signals. In this study, we demonstrated that this non-contact technique effectively detects electroacoustic signals, identifies electroporation regions, and reconstructs electric energy distribution, offering a promising approach for monitoring electroporation therapy.

Conventionally, the size, shape, and biomechanics of cartilages are determined by their voluminous extracellular matrix. By contrast, we found that multiple murine cartilages consist of lipid-filled cells called lipochondrocytes. Despite resembling adipocytes, lipochondrocytes were molecularly distinct and produced lipids exclusively through de novo lipogenesis. Consequently, lipochondrocytes grew uniform lipid droplets that resisted systemic lipid surges and did not enlarge upon obesity. Lipochondrocytes also lacked lipid mobilization factors, which enabled exceptional vacuole stability and protected cartilage from shrinking upon starvation. Lipid droplets modulated lipocartilage biomechanics by decreasing the tissue's stiffness, strength, and resilience. Lipochondrocytes were found in multiple mammals, including humans, but not in nonmammalian tetrapods. Thus, analogous to bubble wrap, superstable lipid vacuoles confer skeletal tissue with cartilage-like properties without "packing foam-like" extracellular matrix.

Large language models (LLMs) are fundamentally transforming human-facing applications in the health and well-being domains: boosting patient engagement, accelerating clinical decision-making, and facilitating medical education. Although state-of-the-art LLMs have shown superior performance in several conversational applications, evaluations within nutrition and diet applications are still insufficient. In this paper, we propose to employ the Registered Dietitian (RD) exam to conduct a standard and comprehensive evaluation of state-of-the-art LLMs, GPT-4o, Claude 3.5 Sonnet, and Gemini 1.5 Pro, assessing both accuracy and consistency in nutrition queries. Our evaluation includes 1050 RD exam questions encompassing several nutrition topics and proficiency levels. In addition, for the first time, we examine the impact of Zero-Shot (ZS), Chain of Thought (CoT), Chain of Thought with Self Consistency (CoT-SC), and Retrieval Augmented Prompting (RAP) on both accuracy and consistency of the responses. Our findings revealed that while these LLMs obtained acceptable overall performance, their results varied considerably with different prompts and question domains. GPT-4o with CoT-SC prompting outperformed the other approaches, whereas Gemini 1.5 Pro with ZS recorded the highest consistency. For GPT-4o and Claude 3.5, CoT improved the accuracy, and CoT-SC improved both accuracy and consistency. RAP was particularly effective for GPT-4o to answer Expert level questions. Consequently, choosing the appropriate LLM and prompting technique, tailored to the proficiency level and specific domain, can mitigate errors and potential risks in diet and nutrition chatbots.

Background

Objective

Methods

Results

Conclusion

Many contemporary imaging systems seek tunable focusing components with minimal form factors and versatile functionalities; however, existing solutions are typically limited in size, efficiency, and tuning speed. Here, low-loss all-dielectric metasurfaces integrated with liquid crystals (LCs) are used to demonstrate highly compact multifunctional zoom components. The phase profiles imparted by the metalens are modulated in real time by means of field-dependent LCs, enabling electrically driven continuous focal length variation and active bifocal imaging with low applied voltages (<10 V). These applications are achieved through the systematic design and validation of resonant metasurface elements that ensure the desired metalens response in each LC state. We engineer and fabricate a high-contrast voltage-actuated continuous-zoom LC-metalens with up to 18% total shift in focal length. Additionally, we fabricate simplified large-diameter LC-metalenses, composed of only a few resonator types, that facilitate electrically tunable multidepth imaging. These results demonstrate the promise of electrically controlled LC-embedded zoom-metasurfaces to serve as lightweight and ultrathin multifunctional focusing components, with prospective uses in next-generation imaging devices.

Conventional refractive microscope objective lenses have limited applicability to a range of imaging modalities due to the dispersive nature of their optical elements. Designing a conventional refractive microscope objective that provides well-corrected imaging over a broad spectral range can be challenging, if not impossible. In contrast, reflective optics are inherently achromatic, so a system composed entirely of reflective elements is free from chromatic aberrations and, as a result, can image over an ultra-wide spectral range with perfect color correction. This study explores the design space of unobscured high numerical aperture, allreflective microscope objectives. In particular, using freeform optical elements we obviate the need for a center obscuration, rendering the objective's modulation transfer function comparable to that of refractive lens systems of similar numerical aperture. We detail the design process of the reflective objective, from determining the design specifications to the system optimization and sensitivity analysis. The outcome is an all-reflective freeform microscope objective lens with a 0.65 numerical aperture that provides diffraction-limited imaging and is compatible with the geometric constraints of commercial microscope systems.