GSGenaro Daza Santacoloma

Genaro Daza Santacoloma

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Genaro Daza Santacoloma has a total of 18 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • An Interpretability Analysis of a Dual-Branch PPG Blood Pressure Model

    by Eduardo Martínez-Duque, Genaro Daza-Santacoloma, David Cárdenas-Peña

  • Open Access

    Predicting MCI-to-Alzheimer’s Conversion: A Systematic Benchmark of Classical, Machine Learning, and Deep Survival Models

    by Nathalia Arias Santa, Genaro Daza-Santacoloma, Walter Serna Serna

    Background/Objectives: Predicting time-to-conversion from mild cognitive impairment (MCI) to Alzheimer’s disease (AD) is a critical clinical challenge. Most survival analyses rely on static baseline variables, ignoring the temporal dynamics of serial neuroimaging. We benchmark seven survival model families using a novel temporal MRI descriptor pipeline on ADNI. Methods: A cohort of 661 MCI patients (250 conversion events, 37.8%, median follow-up 42.0 months) was constructed from ADNI. FreeSurfer volumetrics of six brain regions were summarised into three time-based descriptors each (atrophy rate, trajectory variability σ, and time-weighted cumulative mean μ) under strict anti-leakage constraints, yielding 18 neuroimaging features which, combined with six clinical variables (age, sex, education, MMSE, ADAS-13, and APOE-ε4), form a 24-dimensional covariate vector. Cox PH, Weibull AFT, Exponential, RSF, GBST, DeepSurv, and DeepHit were evaluated under stratified 5-fold cross-validation and assessed on an independent OASIS-3 cohort (n=78) as an exploratory analysis. Results: All seven models achieved C-index 0.853–0.863 (Δ=0.010), with Cox PH reaching 0.863±0.027 using only routine structural MRI. Time-dependent AUC ranged from 0.856 (12 months) to 0.938 (48 months). Risk stratification by quartile yielded highly separated Kaplan-Meier curves (log-rank p<0.001). On OASIS-3 (n=78, exploratory), classical models showed preliminary cross-cohort consistency (Weibull AFT 0.760±0.063, RSF 0.759±0.049), while deep learning degraded substantially (DeepSurv 0.551±0.116); these results are hypothesis-generating and should not be interpreted as confirmatory given the small per-fold sample size (EPV ≈ 1). Conclusions: The temporal MRI descriptor pipeline achieves a C-index of 0.863 using only routine structural MRI, without PET or CSF biomarkers, demonstrating strong discriminative performance on ADNI MCI-to-AD conversion. The narrow performance band (Δ=0.010) suggests that, within this feature-engineering paradigm, no clear performance separation was observed among the evaluated model architectures; this convergence may reflect feature-level rather than architecture-level constraints, although the underlying cause cannot be established from discriminative performance alone. Classical models showed more stable cross-cohort behaviour under data-scarce conditions; prospective validation on larger cohorts with full covariate parity is required before clinical translation can be considered.

  • Open Access

    PPG-FusionNet: A Dual-Branch Neural Architecture for Cuffless Blood Pressure Estimation from Photoplethysmography

    by Eduardo Martínez-Duque, Genaro Daza-Santacoloma, David Cárdenas-Peña

    Hypertension is a major risk factor for cardiovascular disease, yet cuffless blood pressure monitoring remains challenging because most existing methods rely on intermittent cuff-based measurements or multimodal physiological signals. This work proposes PPG-FusionNet, a dual-branch deep learning architecture for simultaneous systolic and diastolic blood pressure estimation using only photoplethysmography (PPG) signals. The model combines a local dilated convolutional encoder and a global AutoCorrelation encoder operating on a shared patch embedding, whose representations are integrated through cross-attention fusion and optimized with a constrained dual-head regression objective. The model was trained and evaluated on the PulseDB benchmark using Bayesian hyperparameter optimization and systematic ablation studies to assess each architectural component. PPG-FusionNet achieved mean absolute errors of 7.46 mmHg for systolic blood pressure and 4.72 mmHg for diastolic blood pressure, with near-zero mean errors and compliance with the ANSI/AAMI standard and BHS Grade B for diastolic estimation. Ablation experiments revealed that trend-seasonal decomposition, despite its success in long-horizon forecasting, degraded performance on short PPG windows, whereas cross-attention fusion and AutoCorrelation improved estimation accuracy. These results demonstrate that heterogeneous dual-branch representation learning provides an effective, scalable framework for cuffless blood pressure estimation from a single PPG sensor.

  • Open Access

    Electrical Activation Analysis of Cortical Brain Areas in Response to Visual Stimuli: An Application in Advertising

    by Víctor Ortiz Bravo, Genaro Daza Santacoloma, Julian Echeverry Correa

    Every day, we are exposed to hundreds of advertising campaigns; however, only about 12% of all this advertising leaves a lasting impression in our brains, highlighting the importance of capturing consumer attention. Evaluating the effectiveness of an advertising campaign allows us to predict its potential success. This practice has been employed by traditional marketing companies for many years, the results can be influenced. In this research, the acquisition and pre-processing of electroencephalographic (EEG) signals generated while viewing visual advertising campaigns are conducted. These signals reflect individuals’ autonomic responses and are not consciously or voluntarily fabricated reactions to stimuli. Subsequently, an electrical activation analysis of the cortical brain and visualization of the EEG signals are performed through a three-dimensional representation on a standardized brain model. The brain regions with the highest electrical activation are analyzed and compared using two mathematical and computational techniques, one linear and one non-linear. The neural response due to the advertising images are compared against the brain representation during the performance of cognitive tasks involving selective attention and implicit memory. This allows us to infer the occurrence of these cognitive processes, evoked by marketing campaigns (visual ads), which are essential constructs for studying consumer behavior. The results indicate that visual advertising campaigns containing linguistic and cultural elements embedded in the graphic designs trigger greater brain activation, which is associated with the cognitive processes of selective attention and implicit memory. Thus, it can be concluded that this type of advertising images increases the probability of influencing purchasing decisions.

  • Open Access

    Advances in engineering: social appropriation of knowledge in neuroscience, energy and agriculture

    by Victor Alfonso Ortiz Bravo, Genaro Daza-Santacoloma, Walter Gil González, Oscar Danilo Montoya Giraldo

    Scientific and technological research, beyond its theoretical and methodological development, finds its greatest value when its outcomes are effectively integrated into social, economic, and productive contexts. In this regard, the social appropriation of knowledge is established as a strategic process to strengthen the connection between academia and society, enabling research findings to be translated into concrete solutions, applied innovations, and meaningful transformations across various sectors. This book presents two notable contributions that reflect the thematic and methodological diversity of the research developed within the Faculty of Engineering.

  • Open Access

    Predictive Model to Identify College Students with High Dropout Rates

    by Jhoan Keider Hoyos Osorio, Genaro Daza Santacoloma

    Decreasing student attrition rates is one of the main objectives of most higher education institutions. However, to achieve this goal, universities need to accurately identify and focus their efforts on students most likely to quit their studies before they graduate. This has given rise to a need to implement forecasting models to predict which students will eventually drop out. In this paper, we present an early warning system to automatically identify first-semester students at high risk of dropping out. The system is based on a machine learning model trained from historical data on first-semester students. The results show that the system can predict “at-risk” students with a sensitivity of 61.97%, which allows early intervention for those students, thereby reducing the student attrition rate.


  • Open Access

    Plataforma de analítica de datos para la caracterización poblacional y la evaluación del riesgo cardiovascular en pacientes del centro-occidente de Colombia

    by Genaro Daza, Jhon Jairo Castañeda, Jorge Iván Castaño

    Introducción: las enfermedades cardiovasculares son la primera causa de muerte en el mundo. Por tanto, muchas investigaciones han sido dirigidas hacia la predicción del riesgo cardiovascular, con el fin de poder evitarlo. Asimismo, se ha buscado la implementación de sistemas que involucren el análisis de datos automatizados que permita que la información se ponga a disposición, no solo del personal administrativo y directivo, sino también del personal clínico, para mejorar el control de las patologías. Objetivo: construir una herramienta para la caracterización poblacional y la evaluación del riesgo cardiovascular en pacientes del centro-occidente de Colombia. Materiales y métodos: se propone la construcción de una plataforma de análisis de datos sociodemográficos y clínicos. El modelo general de diseño de la plataforma es el desarrollo evolutivo, que entrelaza actividades de especificación, desarrollo y validación. La plataforma presenta un modelo vista-controlador que permite la creación de plantillas dinámicas distribuidas en módulos de acceso controlados por perfiles de usuario. Resultados: se implementó el cálculo automatizado del riesgo de enfermedad cardiovascular y la emisión de alertas tempranas, lo cual mejoró la gestión de los procesos clínicos, así como el apoyo a la toma de decisiones administrativas, a través de la conformación de dos módulos interactivos en la plataforma. Conclusiones: la unión de saberes clínicos, administrativos y de ingeniería permitió la consolidación de una herramienta que contribuye en el monitoreo y trazabilidad de los pacientes, orientando la priorización de posibles intervenciones que impacten en la salud de estos.

  • Open Access

    Comparación sistemática de metodologías basadas en información mutua para el registro multimodal de imágenes médicas

    by Gustavo Adolfo Ospina Torres, Walter Serna Serna, Genaro Daza Santacoloma

    El registro de imágenes sobre todo en el ámbito de la medicina, han cobrado una gran importancia en las últimas décadas pues esto se mejora la planeación, intervención y posterior análisis de dichos procedimientos médicos, como es el caso de las cirugías guiadas por imágenes donde se requiere una especial precisión y confiabilidad. Registrar dos o más imágenes consiste en alinearlas de tal manera que queden orientadas de la misma forma, así se puede combinar la información de ambas en un mismo espacio y en una misma representación visual. El objetivo principal de este trabajo es comparar de forma sistemática algunas metodologías para el registro de imágenes multimodales utilizando medidas de similitud basadas en información mutua (IM).Se obtuvo que el registro de imágenes basada en IM es capaz de registrar imágenes multimodales y monomodales con poca susceptibilidad a los cambios de brillo o contraste en la imagen, pero con alta sensibilidad al ruido. También se estableció que las técnicas de optimización basadas en enjambres de partículas tienen un mejor desempeño para el resultado final de registro en comparación con las otras técnicas analizadas. Además, la técnica de interpolación cúbica (transformación espacial) mejora la confiabilidad del algoritmo en imágenes multimodales.


  • Open Access

    Modelado y comportamiento de la simulación de propagación eléctrica durante la estimulación cerebral profunda

    by Pablo Alejandro Alvarado, Cristian Alejandro Torres Valencia, Álvaro Ángel Orozco Gutiérrez, Mauricio Alexander Álvarez López, Genaro Daza Santacoloma, Hans Carmona Vilada

    La Estimulación Cerebral Profunda (DBS) es un tratamiento efectivo para la enfermedad de Parkinson. Gran variedad de modelos matemáticos y computacionales para describir la propagación eléctrica debido a la DBS han sido propuestos, desafortunadamente, no existe claridad sobre las razones que justifican el uso de un modelo específico. En el presente trabajo se presenta una formulación matemática detallada de la propagación eléctrica debido a DBS que soporta un modelo basado en la ecuación de Laplace. Se realizan simulaciones para diferentes modelos geométricos del cerebro para determinar si la geometría, el tamaño y la ubicación de la tierra del modelo afectan la predicción de la estimulación eléctrica mediante el uso del Método de Elementos Finitos (FEM). Los análisis teórico y experimental muestran en primera instancia que la ecuación de Laplace es adecuada para describir la propagación eléctrica en el cerebro, y en segunda instancia que la estructura geométrica, tamaño y ubicación de la tierra afectan la magnitud del potencial eléctrico, particularmente para modos de estimulación monopolar. Los resultados muestran que para modelos básicos y más realistas pueden existir diferencias en la propagación de hasta un 2900%.