Doctorando en Interacción Persona-Computadora y Salud Digital
El propósito de esta tesis doctoral es avanzar en el estado del arte de los agentes conversacionales con interacción por voz, especialmente su integración en dispositivos inteligentes comerciales para el cribado de salud mental. La investigación se centra en el desarrollo y validación de un agente conversacional basado en voz, diseñado para administrar pruebas psicométricas orientadas a la detección temprana de la depresión, particularmente en personas mayores. Más información en el proyecto SAPIENS. El estudio se llevó a cabo con una muestra heterogénea de 97 adultos mayores y 21 cuidadores, provenientes de diversas organizaciones del área de Vigo, garantizando una amplia representación sociodemográfica.
Objetivos de investigación y metodología
El objetivo principal fue diseñar e implementar una herramienta digital de cribado utilizando altavoces inteligentes comerciales, basada en instrumentos clínicos validados como la GDS-15 (Escala de Depresión Geriátrica). La investigación se desarrolló en torno a ocho objetivos específicos, que incluyeron una revisión sistemática de la literatura (metodología PRISMA), el diseño y desarrollo de la skill DYes para Amazon Alexa, una fase piloto con aprobación ética, y la creación de un modelo de aprendizaje automático para mejorar la precisión diagnóstica.
Tecnología e implementación
DYes es una skill de voz desarrollada específicamente con base en la GDS-15. Recoge tanto las respuestas directas del usuario como datos adicionales durante la interacción (tiempos de respuesta, repeticiones), enriqueciendo así el proceso de cribado. La solución se implementó utilizando el kit de desarrollo de Amazon Alexa y está disponible públicamente en GitHub: Digi-Yesavage
Usabilidad y validación
La aceptación y usabilidad de DYes fueron evaluadas mediante el Modelo de Aceptación de Tecnología (TAM) y el Cuestionario de Usabilidad del Sistema (PSSUQ). Los resultados muestran un aumento significativo en la facilidad de uso percibida (+0,63) y una alta puntuación media de satisfacción (6 sobre 7), lo que confirma la buena acogida tanto por parte de usuarios como de profesionales. Los análisis de validación revelaron una fuerte correlación entre DYes y la GDS-15, con un coeficiente de determinación (R²) de 0,622 y un coeficiente de correlación intraclase (CCI) de 0,881.
Integración de aprendizaje automático
Se propuso un modelo predictivo basado en aprendizaje automático para mejorar la precisión del cribado. Este modelo integra variables conductuales capturadas por DYes y emplea técnicas avanzadas como la permutación de importancia de características y la optimización de hiperparámetros para afinar su capacidad predictiva.
- Existen múltiples razones para creer que este trabajo ofrece ventajas claras, especialmente en el ámbito de la salud mental digital, tales como:
- Mejora en la detección temprana de la depresión
- Mayor accesibilidad para personas mayores
- Interacción natural mediante voz que aumenta la comodidad del usuario
- Mayor precisión diagnóstica a través de datos conductuales
- Alta aceptación por parte de usuarios y cuidadores
- Escalabilidad en entornos domésticos y clínicos
- Apoyo a estrategias de salud mental basadas en datos
PhD student in Human-Computer Interaction and Digital Health
The purpose of this PhD is to advance the state of the art in voice-enabled conversational agents, particularly their integration into commercial smart devices for mental health screening. The research focuses on the development and validation of a voice-based conversational agent designed to administer psychometric tests for early detection of depression, especially in older adults. More details at the SAPIENS project. The study was carried out with a heterogeneous sample of 97 older adults and 21 caregivers from various organizations in Vigo, Spain, ensuring a diverse socio-demographic profile.
Research Goals and Methodology
The main goal was to design and implement a digital screening tool using commercial smart speakers, based on validated clinical instruments like the GDS-15. The research was developed across eight specific objectives, including a systematic literature review using PRISMA methodology, the design and implementation of the DYes skill for Amazon Alexa, pilot testing with ethical approval, and the creation of a machine learning model for improved diagnostic accuracy.
Technology and Implementation
DYes is a custom-developed voice skill based on the Geriatric Depression Scale (GDS-15). It captures user responses along with behavioral interaction data (e.g., response time, number of repetitions) to enrich screening accuracy. The solution was implemented using the Amazon Alexa development toolkit and made publicly available via GitHub: Digi-Yesavage
Usability and Validation
User acceptance and usability were evaluated using the Technology Acceptance Model (TAM) and the Post-Study System Usability Questionnaire (PSSUQ). Results show a significant increase in perceived ease of use (+0.63) and high satisfaction scores (6 out of 7), confirming the usability of the system among both users and professionals. Validation results showed a strong correlation between DYes and the GDS-15, with an R² of 0.622 and an Intraclass Correlation Coefficient (ICC) of 0.881.
Machine Learning Integration
A predictive model using machine learning was proposed to enhance screening precision. The model incorporates behavioral variables captured by DYes and uses advanced techniques like feature permutation and hyperparameter optimization to refine its accuracy.
- There are many reasons to believe that this work offers clear advantages, particularly in digital mental health innovation, such as:
- Improved early detection of depression
- Greater accessibility for older populations
- Natural, voice-based interaction for enhanced user comfort
- Increased screening accuracy through behavioral data
- Higher user and caregiver acceptance
- Scalable integration in home and clinical settings
- Support for data-driven mental health strategies
Skills
My work experience and my concern to be updated in new technologies and programming languages are the way to know my skills
E-learning Education
Experience in management, design and development.
- LMS's: Moodle, OpenEdx
- H5P, Edpuzzle
- Owncloud
Web Developer
Experience in management, design and development.
- Bootstrap, Wordpress, Prestashop
- Impresspages, Symfony
- Django
Computer Technician
Experience in Repair, replacement and troubleshooting.
- Troubleshooting hardware and software
- Installation and update of programs
- Management of Operating Systems, windows, Linux
Development and programming
Experience in management, design and development.
- PHP, CSS, HTML
- Python
- C , C#
My Portfolio
This is a list of research's works in which I've participated with Vigo's University.
Esta es una lista de trabajos de investigación en los que he participado.
Drag and Drop exercise
Some instances of how to apply h5p.
Galery NGDLE UP2U
First galery of some pictures related with UP2U NGDLE, Next Generation Digital Learning Environment.
PhD work
The purpose of my Phd is to evaluate the interaction necessities in the laboratory and in the different types of workbenches in the education sector.
ORCID
Ph. D. video presentación 2018
Ph. D. video presentación 2019
Ph. D. video presentación 2020
VR - Virtual Reality & AR - Augmented Reality
Publicaciones
Here you are some of the public work I have done.
Aquí están algunos de los trabajos públicos que he hecho.
2024
-
Comparative Analysis of MoCA and DigiMoCA Test Results: A Pilot Study
Applied Sciences 2024-06-11 | Journal article
DOI: 10.3390/app14125073
More details of this paper, Pending -
Examining Memory Performance in Senior Adults: A Comparative Cross-Sectional Study
Applied Sciences 2024-06-06 | Journal article
DOI: 10.3390/app14114930
More details of this paper, Pending -
Surfeando la ola de la desinformación y la salud mental con agentes convencionales
Contra la desinformación. Manual de herramientas y recursos didácticos para el aula 2024-03-01 | Book chapter
DOI: Surfeando la ola de la desinformación
More details of this paper, Pending 2023
-
Oral vs. Touch Interaction for Cognitive Assessment: Acceptance and Perceived Usability by Senior Adults
Electronics 2023-12-19 | Journal article
DOI: 10.3390/electronics13010013
More details of this paper, Pending -
Conversational agents for depression screening: A systematic review
International Journal of Medical Informatics 2023-10-30 | Journal article
DOI: 10.1016/j.ijmedinf.2023.105272
More details of this paper, Pending -
Open and commercial tools to generate a digital interactive story in journalism: systematic review and features analysis
New Review of Hypermedia and Multimedia 2023-01-02 | Journal article
DOI: 10.1080/13614568.2023.2175041
More details of this paper, Pending 2022
-
BRIDGING THE GAP BETWEEN THE NEW ARTIFICIAL INTELLIGENCE CAPABILITIES AND THE SKILLS OF JOURNALISTS IN DIGITAL MEDIA
EDULEARN2020 Proceedings 2022-11 | conference-paper
More details of this paper, Pending -
TOWARDS THE NEW EDUCATIONAL PARADIGMS IN THE 5TH WAVE OF ARTIFICIAL INTELLIGENCE: A FOCUS ON JOURNALISM
EDULEARN2020 Proceedings 2022-11 | conference-paper
More details of this paper, Pending -
Methods for Gastrointestinal Endoscopy Quantification: A Focus on Hands and Fingers Kinematics
Sensors 2022-11-28 | Journal article
DOI: 10.3390/s22239253
More details of this paper, Pending -
Los cimientos de la Inteligencia Artificial en el sistema productivo de contenidos periodísticos automatizados
Redmarka. Revista de Marketing Aplicado 2022-06-30 | Journal article
DOI: 10.17979/redma.2022.26.1.9056
More details of this paper, Pending 2021
-
A Systematic Review of Commercial Smart Gloves: Current Status and Applications
Sensors 2021-04-10 | Journal article
DOI: 10.3390/s21082667
More details of this paper, Pending 2020
-
OPEN EDUCATIONAL RESOURCES’ (OER) AWARENESS IN SECONDARY EDUCATION:EXPERIENCES AND LESSONS LEARNED FROM A COURSE DEDICATED TO OER
EDULEARN2020 Proceedings 2020-07-07 | conference-paper
DOI: dx.doi.org/10.21125/edulearn.2020.0304
ISBN: 978-84-09-17979-4
More details of this paper, HERE 2019
-
FIRST STEPS OF CREATING AN IMMERSIVE SCENARIO FOR IMPROVING NAVAL
ICERI2019 Proceedings 2019-11-11 | conference-paper
DOI: dx.doi.org/10.21125/iceri.2019.0945
ISBN: 978-84-09-14755-7
More details of this paper, HERE -
UP2U IN THE CLASSROOM
ICERI2019 Proceedings 2019-11-11 | conference-paper
DOI: dx.doi.org/10.21125/iceri.2019.2339
ISBN: 978-84-09-14755-7
More details of this paper, HERE -
GOOD PRACTICES ADOPTING COPYRIGHT FOR CREATING AND REUSING OER IN NGDLE ENVIRONMENT
INTED2019 Proceedings 2019-03-14 | conference-paper
DOI: dx.doi.org/10.21125/inted.2019.1359
ISBN: 978-84-09-08619-1
More details of this paper, HERE -
FACILITATING CURATION OF OPEN EDUCATIONAL RESOURCES THROUGH THE USE OF AN APPLICATION-LINKED REPOSITORY
INTED2018 Proceedings 2019-03-09 | conference-paper
DOI: dx.doi.org/10.21125/inted.2018.0190
ISBN: 978-84-69-79480-7
More details of this paper, HERE
H5P Course
Contact
You could send me a message or call me at the following Phone number, if you want to be in touch with me.