Bioingeniería
Permanent URI for this collection
Browse
Browsing Bioingeniería by Author "Argüello Prada, Erick Javier (Director)"
Now showing 1 - 4 of 4
Results Per Page
Sort Options
Item Estimación de la frecuencia respiratoria a partir de la combinación de parámetros derivados de la señal fotopletismografica.(Universidad Santiago de Cali, 2023) Ruiz Hidalgo, Aura Maria; Sanchez, Yheferson Andrade; Argüello Prada, Erick Javier (Director)In the current health care setting, accurate and continuous measurement of respiratory rate (RR) is essential to effectively assess health status. It was recently reported that variations in amplitude of successive systolic peaks of the photoplethysmographic (PPG) signal, known as systolic variability (RISV), provide a more accurate estimate of RF. However, limiting this estimate to a single parameter could exclude relevant information. The main objective of this investigation was to evaluate the contribution of RISV in the estimation of RF when combined with variations in amplitude (RIAV), frequency (RIFV) and intensity (RIIV) of the PPG signal due to respiration. Public databases with PPG recordings and respiratory signal were used as reference. After extracting the series of respiration-induced variations and obtaining their representation in the frequency domain using the Lomb-Scargle periodogram, estimates were evaluated using metrics such as RMSE and MAE, with errors of 6.50 and 5.21 resp/min. In addition, Bland-Altman concordance analysis was applied, with and without "smart fusion". The results suggest that combinations incorporating RISV, especially RIAF + RIIV + RIFV + RISV, considerably reduce the error associated with RF estimation. Although the differences between estimates and true values could be significant, further research is required to validate and adapt these findings to specific clinical situations.Item Evaluación de dolor mediante conductancia cutánea: estudio de factibilidad.(Universidad Santiago de Cali, 2022) Castillo Gómez, Cristian David; Navarrete Jiménez, Carlos Mario; Argüello Prada, Erick Javier (Director)The most widely used procedure in clinical practice for pain assessment is the self-report. However, this brings with it serious limitations in patients who do not have the physical and/or cognitive skills necessary to report the origin and intensity of the pain they suffer. Based on this, several authors have examined the capabilities of multiple autonomic parameters when used for assessing pain. Despite being promising, such approaches demand a high number of sensors, which might be inappropriate when used in patients with severe pain. Thue, the present study aims to determine whether or not it is feasible to use skin conductance to assess pain in healthy people. The tests were carried out on 15 healthy volunteers, who were subjected to 2 types of stimulation : sonorous (not painful) and thermal (painful). The latter was achieved using the Cold Pressor Test (CPT) method, which consists of immersing the opposite wrist in which the biosignal is measured in a bucket of water at a temperature of 2-6°C. Once the signal was acquired, various parameters were extracted, such as the mean, median, standard deviation and entropy, which were calculated before and during the application of the stimulus. After carrying out the statistical analysis of the data, it was observed that the mean value of skin conductance increases significantly when applying a painful thermal stimulus, which does not happen when the stimulus is sound. Additionally, no significant differences were observed in the mean conductance values obtained from each of the thermal and sound stimuli, which suggests that the variations recorded are not specific to pain, but may arise as a consequence of the application of any stimulus that stands out from the environment. Therefore, although skin conductance may vary considerably as a result of the application of a highlighting stimulus; it is not a feasible parameter for theevaluation of pain since it cannot distinguish pain intensities or types of stimuli, in addition to not having been used in a clinical.Item Evaluación de métodos alternativos (phantomless) para calibración de equipos de ultrasonido dentro del contexto hospitalario colombiano: una revisión.(Universidad Santiago de Cali, 2023) Martínez Muñoz, Nicol Dayanna; Argüello Prada, Erick Javier (Director)Traditional ultrasound calibration methods use phantoms as a standard to ensure precise images. However, acquiring and managing phantoms is costly and complex. Challenges such as phantom deterioration, changes in their properties over time, and their difficulties in adapting to various equipment and environments have led to the investigation of alternative calibration methods, known as phantomless. This study aims to analyze and compare these ultrasound calibration methods that do not rely on phantoms, evaluating their effectiveness, applicability, and feasibility in the Colombian hospital context. To achieve this, a thorough search was conducted on the IEEE Xplore and Google Scholar databases, limiting the search to a 15 year period and restricting the documents to English. The identified records were organized in an Excel database. Documents referring to phantoms or hybrid calibration methods involving phantoms, as well as those related to other medical ultrasound equipment, were excluded. Furthermore, unavailable and incomplete documents were discarded. A total of four documents were obtained for analysis, two of which included the source code used for calibration. These source codes could be valuable for future research in the Colombian hospital context, providing the foundation for developing more efficient ultrasound calibration methods and overcoming the limitations associated with standard equipment, potentially leading to a more efficient ultrasound calibration process while eliminating the constraints posed by phantoms.Item Implementación de un control en línea de lazo cerrado de una interfaz cerebro computadora para la aplicación de un tercer brazo virtual(Universidad Santiago de Cali, 2023) Rendon Martinez, Yubely; Argüello Prada, Erick Javier (Director)Brain-machine interfaces (BCI) are systems that decode, order and use brain activity through a computer. Among the various challenges that can be found in BCI systems, brain stimulation or feedback suggests one of the most relevant since it facilitates the understanding and performance of the user. Rehabilitation training suggests one of the applications of the MI-BCI imagination paradigm, focused on guiding the brain to accept external prosthetic devices as its own, contributing to motor function along with physical rehabilitation of the upper extremities. This study implements the online closed-loop control of a brain-computer interface for the application of a virtual third arm, making the user take control of animations given the categories: left arm, right arm, imaginary third arm and resting position, as a means for rehabilitation seeking the perception of novelty for the brain in an immersive environment. The interactive environment was developed in Unity visualized in virtual reality (VR) glasses Oculus Quest 2 and controlled by EEG signals of unipolar or referential acquisition through Biosemi Active Two. A comparison of the contribution of the electrodes in the classification performance in Machine Learning, using Genetic Algorithms (GA), together with the Discrete Wavelet Transform (DWT) in the stages of feature selection and characterization respectively. A 77% accuracy of the model was obtained, using this methodology with the KNN algorithm and 79.66% user acceptance.