{rfName}
Ma

Indexed in

License and Use

Icono OpenAccess

Citations

1

Altmetrics

Analysis of institutional authors

Feito-Casares, EnriqueCorresponding AuthorMelgarejo-Meseguer, Francisco MAuthorRojo-álvarez, José-LuisAuthor

Share

June 1, 2026
Publications
>
Article
Hybrid Gold

Manifold Learning Approaches for Characterizing Photoplethysmographic Signals

Publicated to: IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING. 73 (6): 2214-2226 - 2026-06-01, DOI: 10.1109/TBME.2025.3625858

Authors:

Feito-Casares, Enrique; Melgarejo-Meseguer, Francisco M; Cobo, Alejandro; Baumela, Luis; Rojo-Alvarez, Jose-Luis
[+]

Affiliations

Univ Politecn Madrid, Madrid, Spain - Author
Univ Rey Juan Carlos, Madrid 28933, Spain - Author

Abstract

Objective: Photoplethysmography (PPG) is widely used for cardiovascular monitoring, but its analysis is challenged by signal variability, inconsistent acquisition settings, and limited interpretability. This study investigates the use of low-dimensional embeddings to support downstream tasks, including anomaly detection, activity classification, and signal authenticity verification across diverse PPG modalities. Methods: We developed a pipeline leveraging dimensionality reduction techniques, Autoencoder (AE), Fully Connected Neural Network (FCNN), and Uniform Manifold Approximation and Projection (UMAP) to extract compact signal representations. These methods were evaluated across four datasets representing clinical (BIDMC, MIMIC-PERFORM), wearable (Wrist PPG), and remote PPG (UBFC) recordings. Performance was assessed through clustering indices, classification metrics, and anomaly detection rates under varying noise levels. Results: Quantitative evaluation demonstrated that AE-based embeddings enabled accurate discrimination between neonatal and adult signals in the MIMIC-PERFORM dataset (F1 = 0.92, AUC = 0.90), while UMAP outperformed AE and FCNN in clustering physical activities from Wrist PPG data (Davies-Bouldin Index = 5.40). In the BIDMC dataset, the framework detected synthetic anomalies with an AUC of 0.77 at 2 dB SNR, with detection rates declining consistently with reduced noise. On the UBFC dataset, UMAP embeddings supported the detection of manipulated rPPG signals with an F1 score of 0.75 and an AUC of 0.73. Conclusion: Low-dimensional representations provide a compact and task-relevant encoding of PPG signals that enhances classification and detection performance in multiple scenarios. While interpretability gains remain task-dependent, these findings support the utility of embedding-based approaches in biomedical signal analysis and their robustness across modalities and noise conditions.
[+]

Keywords

AdultAlgorithmsArterialAutoencoder (ae)AutoencodersBiomedical engineeringBiomedical monitoringBiomedical signal processingComponentsDeep learning (dl)Feature extractionFemaleHumansInfant, newbornMachine learningMaleManifold learning (mnl)ManifoldsMonitoringNeural networks, computerOptical variables measurementPhotoelectric plethysmographyPhotoplethysmographyPhotoplethysmography (ppg)PipelinesSignal processing, computer-assistedSkinWearable electronic devicesYoung adult

Quality index

Bibliometric impact. Analysis of the contribution and dissemination channel

The work has been published in the journal IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING due to its progression and the good impact it has achieved in recent years, according to the agency Scopus (SJR), it has become a reference in its field. In the year of publication of the work, 2026, it was in position , thus managing to position itself as a Q1 (Primer Cuartil), in the category Biomedical Engineering.

Independientemente del impacto esperado determinado por el canal de difusión, es importante destacar el impacto real observado de la propia aportación.

Según las diferentes agencias de indexación, el número de citas acumuladas por esta publicación hasta la fecha 2026-11-02:

  • Scopus: 1
[+]

Impact and social visibility

From the perspective of influence or social adoption, and based on metrics associated with mentions and interactions provided by agencies specializing in calculating the so-called "Alternative or Social Metrics," we can highlight as of 2026-11-02:

  • The use of this contribution in bookmarks, code forks, additions to favorite lists for recurrent reading, as well as general views, indicates that someone is using the publication as a basis for their current work. This may be a notable indicator of future more formal and academic citations. This claim is supported by the result of the "Capture" indicator, which yields a total of: 5 (PlumX).

It is essential to present evidence supporting full alignment with institutional principles and guidelines on Open Science and the Conservation and Dissemination of Intellectual Heritage. A clear example of this is:

[+]

Leadership analysis of institutional authors

There is a significant leadership presence as some of the institution’s authors appear as the first or last signer, detailed as follows: First Author (Feito Casares, Enrique) and Last Author (Rojo Álvarez, José Luis).

the author responsible for correspondence tasks has been Feito Casares, Enrique.

[+]

Project objectives

La aportación persigue los siguientes objetivos: analizar la aplicabilidad de técnicas de reducción dimensional para representar señales fotopletismográficas (PPG) en diversas modalidades; evaluar el desempeño de autoencoders, redes neuronales totalmente conectadas y UMAP en tareas de detección de anomalías, clasificación de actividades y verificación de autenticidad de señales; determinar la eficacia de estas representaciones compactas mediante índices de agrupamiento, métricas de clasificación y tasas de detección bajo distintos niveles de ruido; caracterizar la discriminación entre señales neonatales y adultas en el conjunto MIMIC-PERFORM; y comparar el rendimiento de los métodos en conjuntos clínicos, portátiles y remotos, evidenciando la robustez y utilidad de los embeddings en el análisis biomédico.
[+]

Most relevant results

Los resultados más relevantes del estudio evidencian la eficacia de técnicas de reducción dimensional para el análisis de señales fotopletismográficas (PPG). Primero, los embeddings basados en Autoencoder permitieron discriminar con alta precisión señales neonatales y adultas en el conjunto MIMIC-PERFORM, alcanzando un F1 de 0.92 y un AUC de 0.90. Segundo, UMAP superó a AE y FCNN en la agrupación de actividades físicas en datos de Wrist PPG, con un índice Davies Bouldin de 5.40. Tercero, en el conjunto BIDMC, el método detectó anomalías sintéticas con un AUC de 0.77 a 2 dB SNR, aunque la tasa de detección disminuyó con el aumento del ruido. Finalmente, en UBFC, UMAP facilitó la detección de señales rPPG manipuladas con un F1 de 0.75 y un AUC de 0.73.
[+]

Awards linked to the item

This work was supported in part by Research Grants HERMES, LATENTIA, PCardioTrials, and ETHICFACE under Grant PID2023-152331OA-I00, Grant PID2022-140786NB-C31, Grant PID2022-140553OA-C42, and Grant PID2022-137581OB-I00, in part by MICIU/AEI/10.13039/501100011033 and ERDF, FEDER/EU. in part by Rey Juan Carlos University under Grant HERMES 2024/00004/00, in part by the Programa Propio of UPM, in part by CyberFold project, funded by European Union through the NextGenerationEU instrument (Recovery, Transformation, and Resilience Plan), and managed by Instituto Nacional de Ciberseguridad de Espana (INCIBE), under Grant ETD202300129, and in part by Comunidad de Madrid to the Madrid ELLIS Unit.
[+]