A4 Vertaisarvioitu artikkeli konferenssijulkaisussa
Ensemble-KAN: Leveraging Kolmogorov Arnold Networks to Discriminate Individuals with Psychiatric Disorders from Controls
Tekijät: De Franceschi, Gianluca; Sampaio, Inês W.; Borgwardt, Stefan; Kambeitz, Joseph; Kambeitz-Ilankovic, Lana; Meisenzahl, Eva; Salokangas, Raimo K. R.; Upthegrove, Rachel; Wood, Stephen J.; Koutsouleris, Nikolaos; Brambilla, Paolo; Maggioni, Eleonora
Toimittaja: Shandong Wu, Behrouz Shabestari, Lei Xing
Konferenssin vakiintunut nimi: Applications of Medical Artificial Intelligence
Kustantaja: Springer Nature Switzerland
Julkaisuvuosi: 2025
Journal: Lecture Notes in Computer Science
Kokoomateoksen nimi: Applications of Medical Artificial Intelligence
Tietokannassa oleva lehden nimi: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Numero: 15384
Aloitussivu: 186
Lopetussivu: 197
ISBN: 978-3-031-82006-9
eISBN: 978-3-031-82007-6
ISSN: 0302-9743
eISSN: 1611-3349
DOI: https://doi.org/10.1007/978-3-031-82007-6_18
Verkko-osoite: https://doi.org/10.1007/978-3-031-82007-6_18
Machine learning (ML) techniques are crucial for improving diagnostic accuracy in psychiatry using neuroimaging-based biomarkers. Deep learning models like Kolmogorov Arnold Networks (KANs) are particularly promising in this context but struggle with high-dimensional datasets. We propose the Ensemble-KAN (E-KAN) method to overcome these limitations, integrating multiple base learners. Our novel approach aims to advance classification especially when multiple sources of data are available. The E-KAN was tested against traditional ML models in discriminating recent-onset psychosis (ROP) or depression (ROD) from healthy controls using multimodal environmental and neuroimaging data and it underwent a rigorous ablation study to test its effectiveness. Results demonstrate enhanced performance over traditional ML models, highlighting the efficacy of E-KAN models in psychiatric diagnostics. Specifically, our E-KAN achieved an accuracy of 72.5%, outperforming single-KAN models and traditional ML algorithms. This study underscores the potential of E-KAN models in advancing psychiatric research and personalized medicine through improved diagnostic capabilities. The code is available at https://github.com/brainpolislab/E-KAN.
Julkaisussa olevat rahoitustiedot:
This study was supported by EU-FP7 project PRONIA (Personalized Prognostic Tools for Early Psychosis Management) under the Grant Agreement no 602152 (PI: NK). EM was supported by the European Union - NextGeneration EU (PRIN 2022 PNRR, grant n. P20229MFRC); GDF was supported by the Italian Ministry of Health (grant no GR-2018-12367290). IWS was supported by grants from EBRAINS-Italy, project funded under the National Recovery and Resilience Plan (NRRP), Mission 4, “Education and Research” - Component 2, “From research to Business” Investment 3.1 - Call for tender n
3264 of Dec 28, 2021 of Italian Ministry of University and Research (MUR) funded by the European Union - NextGenerationEU, with award number: Project code IR0000011, Concession Decree n
117 of June 21, 2022 adopted by the Italian Ministry of University and Research, CUP B51E22000150006, Project title “EBRAINS-Italy (European Brain ReseArch INfrastruc-tureS-Italy). PB was partially supported by the Italian Ministry of University and Research (Dipartimenti di Eccellenza Program 2023-2027 - Dept of Pathophysiology and Transplantation, University of Milan), the Italian Ministry of Health (Hub Life Science- Diagnostica Avanzata, HLS-DA, PNC-E3-2022-23683266- CUP: C43C22001630001/MI-0117; Ricerca Corrente 2024) and by the Fondazione Cariplo (grant no 2019-3416).
