Revista Mexicana de Ingenieria Biomedica

Evaluation of PLA and PETG filaments for the addition of hydroxyapatite in 3D-printed dental models

Jorge Carlos Ríos-Hurtado — 2026-06-16

Additive manufacturing has established itself as a key technology in dentistry, enabling the manufacture of customized devices with precision and in reduced times. Among the most widely used filaments are polylactic acid (PLA) and glycol-modified polyethylene terephthalate (PETG), both with different properties that influence clinical performance. This study presents a comparative evaluation of PLA and PETG filaments in the generation of hydroxyapatite on 3D-printed pieces through hydrothermal treatment in simulated body fluid (SBF) solution. Dental models were printed with PLA and PETG filaments under controlled conditions and immersed in SBF for 7, 14, and 21 days. Modified pieces were characterized by infrared spectroscopy (IR-TF), X-ray diffraction (XRD), and scanning electron microscopy (SEM) to evaluate surface modifications and mineral formation. The results showed that PLA, due to its greater porosity and roughness, favored early hydroxyapatite nucleation, presenting a stable layer at 21 days. PETG showed slow nucleation, but at 21 days it showed characteristic hydroxyapatite agglomerates. Cytotoxicity tests with 3T3 fibroblasts confirmed that both materials maintained cell viability above 70%.

Evaluation of Classification Methods for EEG Signals in Older Adults for Right-Hand Motor Imagery Movements

Ivan Carrillo — 2026-06-30

Cognitive decline, characterized by the progressive loss of functions such as memory, attention, and speech, significantly affects the well-being of older adults, a population that is growing rapidly worldwide. In response to this issue, brain-computer interfaces (BCIs) represent a promising technological alternative to facilitate interaction with digital devices, particularly for individuals experiencing reduced motor abilities. This study proposes a methodology for classifying five imagined right-hand movements in older adults using EEG signals. Exhaustive experimentation was conducted, evaluating eight distinct representational features extracted from the EEG data and applying various machine learning algorithms, including ensemble methods, to develop a computational model that achieved an accuracy of 93.7%. Additionally, subsets of features capable of maintaining classification accuracy above 90% were identified. These findings support the feasibility of integrating BCI solutions tailored to the needs of older adults in assistive and rehabilitation applications.

Predictive models of anthropometric parameters for primary screening of sarcopenia based on Machine Learning

Santiago Arceo Díaz — 2026-06-29

This work reports a free-access primary screening system for detecting sarcopenia risk in older Mexican adults, using machine learning and anthropometric variables obtained through accessible instruments such as measuring tapes. An observational, retrospective, and analytical study was conducted based on records from beneficiaries of the Mexican Social Security Institute from the year 2019, with a sample of 1,678 participants. The models, developed using data from individuals without comorbidities, followed a structured machine learning workflow that included data preprocessing, variable transformation and clustering, and supervised classification using decision-tree-based models. The optimal variable combinations for men and women achieved F1-scores above 0.94, accurately classifying the risk levels of sarcopenia and severe sarcopenia. The current models need to be expanded to include individuals with comorbidities such as type 2 diabetes, hypertension, and arthritis, which have been associated with greater muscle mass loss. This proposal does not replace clinical diagnostic testing but serves as a complementary tool to rule out low-risk individuals and prioritize specialized evaluation for those who may be affected by sarcopenia.

Techani: Use of technology as an ally in the management, monitoring, and control of Type 1 diabetes mellitus.

Rocío Contreras-Jiménez — 2026-06-23

Type 1 Diabetes Mellitus (T1DM) is an autoimmune, unpreventable, and incurable disease in which insulin production ceases due to pancreatic beta-cell destruction. This work presents Techani, a web and mobile system designed to enhance treatment adherence and patient monitoring through a culturally adapted interface tailored to the Mexican context. The system integrates Artificial Intelligence tools and incorporates a broader range of variables than current applications, including mood, hydration, sleep, and atypical days, in addition to traditional clinical data. A mixed-methodology approach combining Scrum, UI/UX design, and empirical data collection through physician surveys was used to identify relevant determinants of glycemic control. The proposed system aims to strengthen patient autonomy, facilitate physician follow-up, and improve adherence by automating carbohydrate counting using the Mexican Food Equivalents System (SMAE). Compared to existing mHealth tools, Techani increases the number of monitored variables by more than 40% and enables potential data integration for predictive glucose modeling. Limitations include manual data entry and a small convenience sample for usability testing. Future validation studies will evaluate its clinical impact on glycemic control and treatment adherence, and integration with CGM data to assess its clinical impact on glycemic control.