Alireza Mehdizadeh

Background: Conventional Sacral Nerve Stimulation (SNS) devices are widely used to treat constipation but are often limited by their lack of mechanical flexibility, which can reduce long-term therapeutic efficacy. Developing advanced electrode materials may improve stimulation performance and patient outcomes. Objective: This experimental study aimed to introduce a proof-of-concept flexible cuff electrode composed of a graphene–polyacrylonitrile (PAN) composite and evaluate its electrochemical properties, their mechanical characteristics, and short-term functional performance in a rabbit SNS model. Material and Methods: In this experimental study, graphene–PAN composite electrodes were fabricated and characterized for electrochemical properties, including charge injection capacity and impedance, as well as mechanical properties, such as flexibility and tensile strength. In vivo testing was conducted on rabbits to evaluate their performance in SNS therapy and their effectiveness in relieving constipation. Results: The graphene–PAN composite electrodes demonstrated a high charge injection capacity (1062.58±73.36 µC/cm²), low impedance (31.27±2.76 Ω), and favorable mechanical flexibility (modulus: 1.31±0.45 MPa; strain at break: 27±2%). In vivo stimulation progressively improved fecal output over four days, approaching pre-injury levels. Conclusion: Graphene-based electrodes exhibited suitable electrochemical and mechanical performance, enabling effective neural stimulation and enhanced therapeutic outcomes in constipation management. As an initial proof-of-concept, these flexible graphene-based materials represent promising candidates for future neural stimulation devices.

Background: Alzheimer’s disease (AD) is a neurodegenerative condition that has no definitive treatment, and its early diagnosis can help to prevent or slow down its progress. Structural magnetic resonance imaging (sMRI) and the progress of artificial intelligence (AI) have significant attention in AD detection. This study aims to differentiate AD from NC and distinguish between LMCI and EMCI from the other two classes. Another goal is the diagnostic performance (accuracy and AUC) of sMRI for predicting AD in its early stages. Methods: In this study, 398 participants were used from the ADNI and OASIS global database of sMRI including 98 individuals with AD, 102 with early mild cognitive impairment (EMCI), 98 with late mild cognitive impairment (LMCI), and 100 normal controls (NC). Results: The proposed model achieved high area under the curve (AUC) values and an accuracy of 99.7%, which is very remarkable for all four classes: NC vs. AD: AUC = [0.985], EMCI vs. NC: AUC = [0.961], LMCI vs. NC: AUC = [0.951], LMCI vs. AD: AUC = [0.989], and EMCI vs. LMCI: AUC = [1.000]. Conclusions: The results reveal that this model incorporates DenseNet169, transfer learning, and class decomposition to classify AD stages, particularly in differentiating EMCI from LMCI. The proposed model performs well with high accuracy and area under the curve for AD diagnostics at early stages. In addition, the accurate diagnosis of EMCI and LMCI can lead to early prediction of AD or prevention and slowing down of AD before its progress.

Alzheimer's disease (AD) is a neurodegenerative condition that has no definitive treatment and its early diagnosis can help to prevent or slow down its progress. Neuroimaging in particular, structural magnetic resonance imaging (sMRI) and the progress of artificial intelligence (AI) have significant attention in AD detection. In this study, 398 participants were used from the ADNI and OASIS global database of sMRI including 98 individuals with AD, 102 with early mild cognitive impairment (EMCI), 98 with late mild cognitive impairment (LMCI), and 100 normal controls (NC). The proposed model achieved high area under the curve (AUC) values and an accuracy of 99.7%, which is very remarkable for all four classes: NC vs. AD: AUC = [0.985], EMCI vs. NC: AUC = [0.961], LMCI vs. NC: AUC = [0.951], LMCI vs. AD: AUC = [0.989], and EMCI vs. LMCI: AUC = [1.000]. The results reveal that this model incorporates DenseNet169, transfer learning, and class decomposition to classify AD stages, particularly in differentiating EMCI from LMCI. Overall, this model performs well with high accuracy and area under the curve for AD diagnostics at early stages.

This study aimed to investigate the effect of glycerol as an Optical Clearing Agent on the temperature profile of the skin during HyperThermic Laser Lipolysis using computer simulation. In this study, a three-layer model of the skin was used to simulate HyperThermic Laser Lipolysis. The Monte Carlo MCML code was used to investigate the propagation of laser photons inside skin tissue. The energy absorbed from photons is used as a heat source to determine the increase in temperature and assess thermal damage in the layers of the skin. The finite element method in COMSOL software was used for calculation. The simulation of single-pulse radiation exposure with and without applying glycerol to the skin model was investigated to assess the impact of glycerol. Glycerol decreases the temperature and thermal damage to the epidermis layer while increasing the temperature of the fat layer. Moreover, the presence of glycerol increases the depth of fat cell destruction. Glycerol, as a supplement, significantly improves the efficacy of HyperThermic Laser Lipolysis.

Microelectrode Arrays (MEAs) neural interfaces are considered implantable devices that interact with the nervous system to monitor and/or modulate brain activity. Graphene-based materials are utilized to address some of the current challenges in neural interface design due to their desirable features, such as high conductance, large surface-to-volume ratio, suitable electrochemical properties, biocompatibility, flexibility, and ease of production.In the current study, we fabricated and characterized a type of flexible, ultrasmall, and implantable neurostimulator based on graphene fibers. In this procedure, wet-spinning was employed to create graphene fibers with diameters of 10 to 50 µm. A 10-channel polyimide Printed Circuit Board (PCB) was then custom-designed and manufactured. The fibers were attached to each channel by conductive glue and also insulated by soaking them in a polyurethane solution. The tips were subsequently exposed using a blowtorch. Microstructural information on the fibers was obtained using Scanning Electron Microscopy (SEM), and the measurements of Electrochemical Impedance Spectroscopy (EIS) were conducted for each electrode.Flexible MEAs were created using graphene fibers with diameters ranging from 10 to 50 microns with a spacing of 150 microns. This method leads to producing electrode arrays with any size of fibers and a variety of channel numbers. The flexible neural prostheses can replace conventional electrodes in both neuroscience and biomedical research.

Artificial intelligence (AI) as a branch of computer science focuses on creating intelligent computers to do activities, which generally require human intelligence. AI is considered in many aspects of the routine day, including personal assistants, aircraft, video games, and automated public transit. Moreover, AI is recently considered mostly in healthcare and medicine: disease diagnosis, referral management, service delivery, drug development, predictive modeling, and decision support, drastically leading to altering some processes of patient care [1-5]. The number of manuscripts related to AI in Medicine (AIM) submitted to the Journal of Biomedical Physics and Engineering (JBPE) has been significantly increasing since the past years, resulting in assigning this issue only to selected AIM manuscripts as a special opportunity for AIM researchers. This special issue aimed to show the most recent advances and innovative uses of AI in the medical field. Therefore, scholars and researchers are invited to submit their original research articles based on AI techniques to address health challenges. Furthermore, some of the articles were selected from the manuscripts presented at the first international congress on "Advanced Health Technologies-Artificial Intelligence in Medicine" [6], leading to the selection of 11 papers that cover a wide range of AIM: diagnosis, prediction, classification, and cancer detection using AI techniques, such as machine learning, neural network, and deep learning. This study also aimed to provide historical perspective and an up-to-date review of AIM publications. This review briefly discusses the number of publications, research trends, and hot medical topics related to AIM. From the WOS and PubMed databases, Papers were selected based on keywords like 1) AI type, 2) medicine/health from 1971 to November 25 th 2022. The conference proceedings, abstract meetings, book chapters, and articles published in languages other than English were excluded. The AI types were set as "Artificial intelligence", "Deep learning", "Machine learning", "Neural network", "Artificial neural network", "Natural language processing", "Fuzzy expert system" and "Robotics". Meanwhile, the search terms "medicine", "health*", "diagnosis", "treatment", "diseases" and "illness" were considered the medicine/health items in the research. Finally, these two queries were connected with "AND" together. Figure 1 shows the growth of some publications in WOS and PubMed, which is a significant increase in AIM publications since 2018. Figure 2 demonstrates the different hot-topic medical fields, using AI approaches, and surgery first ranks in AI usage. The number of published papers in different branches of medical sciences is compared from 2019 to 2022 (Figure 2) to illustrate the explosive growth of published papers in the last four years.