Luis Montesinos

Dr. Luis Montesinos

Tecnologico de Monterrey, Mexico City, Ciudad de México, Mexico
orcid
Assistant Professor of Biomedical Engineering, Tecnologico d...

Research Focus & Metrics

Academic Career Length

37Years (1990-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
17 January 2019 in BMC Medical Informatics and Decision Making
Verified
Journal Article
Open access
PDF file
CRCastaldo Rossana
Castaldo Rossana
Luis MontesinosLuis Montesinos
Luis Montesinos
PMP. Melillo
P. Melillo
CJC. James
C. James
LPLeandro Pecchia
Leandro Pecchia
This paper suggests a method to assess the extent to which ultra-short Heart Rate Variability (HRV) features (less than 5 min) can be considered as valid surrogates of short HRV features (nominally 5 min). Short term HRV analysis has been widely investigated for mental stress assessment, whereas the validity of ultra-short HRV features remains unclear. Therefore, this study proposes a method to explore the extent to which HRV excerpts can be shortened without losing their ability to automatically detect mental stress. ECGs were acquired from 42 healthy subjects during a university examination and resting condition. 23 features were extracted from HRV excerpts of different lengths (i.e., 30 s, 1 min, 2 min, 3 min, and 5 min). Significant differences between rest and stress phases were investigated using non-parametric statistical tests at different time-scales. Features extracted from each ultra-short length were compared with the standard short HRV features, assumed as the benchmark, via Spearman’s rank correlation analysis and Bland-Altman plots during rest and stress phases. Using data-driven machine learning approaches, a model aiming to detect mental stress was trained, validated and tested using short HRV features, and assessed on the ultra-short HRV features. Six out of 23 ultra-short HRV features (MeanNN, StdNN, MeanHR, StdHR, HF, and SD2) displayed consistency across all of the excerpt lengths (i.e., from 5 to 1 min) and 3 out of those 6 ultra-short HRV features (MeanNN, StdHR, and HF) achieved good performance (accuracy above 88%) when employed in a well-dimensioned automatic classifier. This study concluded that 6 ultra-short HRV features are valid surrogates of short HRV features for mental stress investigation.

20 December 2021 in Diabetology & Metabolic Syndrome
Verified
Journal Article
Open access
PDF file
LFLuis Fregoso-Aparicio
Luis Fregoso-Aparicio
JNJulieta Noguez
Julieta Noguez
Luis MontesinosLuis Montesinos
Luis Montesinos
JGJOSE ANTONIO GARCIA GARCIA
JOSE ANTONIO GARCIA GARCIA
Diabetes Mellitus is a severe, chronic disease that occurs when blood glucose levels rise above certain limits. Over the last years, machine and deep learning techniques have been used to predict diabetes and its complications. However, researchers and developers still face two main challenges when building type 2 diabetes predictive models. First, there is considerable heterogeneity in previous studies regarding techniques used, making it challenging to identify the optimal one. Second, there is a lack of transparency about the features used in the models, which reduces their interpretability. This systematic review aimed at providing answers to the above challenges. The review followed the PRISMA methodology primarily, enriched with the one proposed by Keele and Durham Universities. Ninety studies were included, and the type of model, complementary techniques, dataset, and performance parameters reported were extracted. Eighteen different types of models were compared, with tree-based algorithms showing top performances. Deep Neural Networks proved suboptimal, despite their ability to deal with big and dirty data. Balancing data and feature selection techniques proved helpful to increase the model’s efficiency. Models trained on tidy datasets achieved almost perfect models.

12 January 2024 in Education Sciences
Verified
Journal Article
Open access
PDF file
David Ernesto Salinas-NavarroDavid Ernesto Salinas-Navarro
David Ernesto Salinas-Navarro
EVEliseo Luis Vilalta-Perdomo
Eliseo Luis Vilalta-Perdomo
RMRosario Michel-Villarreal
Rosario Michel-Villarreal
Luis MontesinosLuis Montesinos
Luis Montesinos
The emergence of generative artificial intelligence (GenAI) requires innovative educational environments to leverage this technology effectively to address concerns like academic integrity, plagiarism, and others. Additionally, higher education needs effective pedagogies to achieve intended learning outcomes. This emphasizes the need to redesign active learning experiences in the GenAI era. Authentic assessment and experiential learning are two possible meaningful alternatives in this context. Accordingly, this article investigates how GenAI can enhance teaching and learning by constructively addressing study situations beyond conventional learning approaches and cultivating high-order skills and knowledge acquisition. This study employs thing ethnography to examine GenAI tools’ integration with authentic assessment and experiential learning and explore implementation alternatives. The results reveal insights into creating human-centered and GenAI-enhanced learning experiences within a constructive alignment. Specific examples are also provided to guide their implementation. Our contributions extend beyond the traditional use of GenAI tools as mere agents-to-write or agents-to-answer questions to become agents-to-support experiential learning for authentic assessment. These findings underscore the transformative role of GenAI tools in enhancing teaching and learning efficacy and effectiveness. The limitations in treating GenAI tools as subjects in thing ethnography are acknowledged, with potential for future implementation evaluation.

12 December 2018 in Journal of NeuroEngineering and Rehabilitation
Verified
Journal Article
Open access
PDF file
Luis MontesinosLuis Montesinos
Luis Montesinos
CRCastaldo Rossana
Castaldo Rossana
LPLeandro Pecchia
Leandro Pecchia
Approximate entropy (ApEn) and sample entropy (SampEn) have been previously used to quantify the regularity in centre of pressure (COP) time-series in different experimental groups and/or conditions. ApEn and SampEn are very sensitive to their input parameters: m (subseries length), r (tolerance) and N (data length). Yet, the effects of changing those parameters have been scarcely investigated in the analysis of COP time-series. This study aimed to investigate the effects of changing parameters m, r and N on ApEn and SampEn values in COP time-series, as well as the ability of these entropy measures to discriminate between groups. A public dataset of COP time-series was used. ApEn and SampEn were calculated for m = {2, 3, 4, 5}, r = {0.1, 0.15, 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5} and N = {600, 1200} (30 and 60 s, respectively). Subjects were stratified in young adults (age < 60, n = 85), and older adults (age ≥ 60) with (n = 18) and without (n = 56) falls in the last year. The effects of changing parameters m, r and N on ApEn and SampEn were investigated with a three-way ANOVA. The ability of ApEn and SampEn to discriminate between groups was investigated with a mixed ANOVA (within-subject factors: m, r and N; between-subject factor: group). Specific combinations of m, r and N producing significant differences between groups were identified using the Tukey’s honest significant difference procedure. A significant three-way interaction between m, r and N confirmed the sensitivity of ApEn and SampEn to the input parameters. SampEn showed a higher consistency and ability to discriminate between groups than ApEn. Significant differences between groups were mostly observed in longer (N = 1200) COP time-series in the anterior-posterior direction. Those differences were observed for specific combinations of m and r, highlighting the importance of an adequate selection of input parameters. Future studies should favour SampEn over ApEn and longer time-series (≥ 60 s) over shorter ones (e.g. 30 s). The use of parameter combinations such as SampEn (m = {4, 5}, r = {0.25, 0.3, 0.35}) is recommended.

1 July 2021 in Biomedical Signal Processing and Control
Verified
Journal Article
Open access
ERErick Axel Martinez Rios
Erick Axel Martinez Rios
Luis MontesinosLuis Montesinos
Luis Montesinos
MAMariel Alfaro-Ponce
Mariel Alfaro-Ponce
LPLeandro Pecchia
Leandro Pecchia

18 January 2018 in IEEE Transactions on Neural Systems and Rehabilitation Engineering
Verified
Journal Article
Open access
Luis MontesinosLuis Montesinos
Luis Montesinos
CRCastaldo Rossana
Castaldo Rossana
LPLeandro Pecchia
Leandro Pecchia
Wearable inertial sensors have been widely investigated for fall risk assessment and prediction in older adults. However, heterogeneity in published studies in terms of sensor location, task assessed and features extracted is high, making challenging evidence-based design of new studies and/or real-life applications. We conducted a systematic review and meta-analysis to appraise the best available evidence in the field. Namely, we applied established statistical methods for the analysis of categorical data to identify optimal combinations of sensor locations, tasks, and feature categories. We also conducted a meta-analysis on sensor-based features to identify a set of significant features and their pivot values. The results demonstrated that with a walking test, the most effective feature to assess the risk of falling was the velocity with the sensor placed on the shins. Conversely, during quite standing, linear acceleration measured at the lower back was the most effective combination of feature-placement. Similarly, during the sit-to-stand and/or the stand-to-sit tests, linear acceleration measured at the lower back seems to be the most effective feature-placement combination. The meta-analysis demonstrated that four features resulted significantly higher in fallers: the root-mean-square acceleration in the mediolateral direction during quiet standing with eyes closed [Mean Difference (MD): 0.01 g; 95% Confidence Interval (CI95%): 0.006 to 0.014]; the number of steps (MD: 1.638 steps; CI95%: 0.384 to 2.892) and total time (MD: 2.274 seconds; CI95%: 0.531 to 4.017) to complete the timed up and go test; and the step time (MD: 0.053; CI95%: 0.012 to 0.095; p = 0.01) during walking.