AI-assisted analysis of heart sounds and interpretation of acoustic representation of brainwaves in neonates

Simple item page
dc.availability.bitstreamcontrolled
dc.check.date2023-10-31
dc.contributor.advisorPopovici, Emanuelen
dc.contributor.advisorTemko, Andriyen
dc.contributor.authorGomez Quintana, Sergi
dc.contributor.funderWellcome Trusten
dc.contributor.funderGrand Challenges Canadaen
dc.contributor.funderScience Foundation Irelanden
dc.date.accessioned2022-09-23T13:08:56Z
dc.date.available2022-09-23T13:08:56Z
dc.date.issued2022-08-18
dc.date.submitted2022-08-18
dc.description.abstractNumerous reports from World Health Organisation (WHO) consistently list the diseases of the heart and the brain among the top three causes of death across the globe. In low and low-to-middle-income countries, the neonatal stage is the most dangerous of the whole life and is a time of particular concern for medical professionals and parents. Timely detection of abnormalities during the first days of life allows medical staff to make informed decisions which have life-saving consequences. For this, continuous monitoring is required and it has several challenges in a clinical setting. First, acquiring physiological data from neonates is not trivial, often involving time-consuming processes that require specialised training. Second, specific monitoring equipment is often expensive and not affordable in low-income communities. More importantly, the complexity of the data may be difficult to interpret even for trained professionals and the required expertise might not be available 24/7. Alternative methods and tools that are low cost and require minimum training while providing the accuracy level of a specialist medical professional are required. This work deals with the development of such methods for the analysis of neonatal heart and brain signals by means of artificial intelligence (AI) and AI-guided sonification. Sound analysis can play an important role as a non-invasive, intuitive, and cost-effective tool to facilitate the interpretation of physiological signals. Heart auscultation is already part of the clinical examination routine. It uses a stethoscope, which is a low cost and reliable tool to screen for neonatal heart defects. However, heart sound interpretation is subjective, dependent on the assessor’s hearing acuity and the acquired level of expertise. Assistance from AI can provide an objective interpretation of heart sounds to complement the traditional auscultation method. A novel, accurate method for detecting congenital heart disease in phonocardiogram (PCG) signals using AI is presented. When dealing with the brain abnormalities in newborns, neonatal seizures are one of the most common neurological conditions, and they need to be treated as a medical emergency with prompt detection and intervention. Electroencephalography (EEG), the gold standard for monitoring electrical brain activity, is often difficult to interpret visually and requires a highly specialised medical professional. These professionals might not be readily available in low or medium-income settings, and even in high-income countries, they might be available only in tertiary care centres and not present 24/7. AI-driven sonification of EEG for detection of neonatal seizures, which is developed in this work, helps to improve the detection of these threatening seizure events by decreasing the level of expertise required from healthcare professionals while maintaining the same accuracy. It is shown that AI-assisted sonification can augment the medical professional to make decisions which are better than AI alone while improving the interpretability of the made decisions, which is a key requirement in the medical domain. The proposed algorithms and methodologies are validated on numerous datasets. The developed prototypes are implemented using cloud and Internet of Things technologies. It is shown that these technologies allow for an affordable, real-time analysis of heart and brain physiological signals with minimum training.en
dc.description.statusNot peer revieweden
dc.description.versionAccepted Versionen
dc.format.mimetypeapplication/pdfen
dc.identifier.citationGomez Quintana, S. 2022. AI-assisted analysis of heart sounds and interpretation of acoustic representation of brainwaves in neonates. PhD Thesis, University College Cork.en
dc.identifier.endpage159en
dc.identifier.urihttps://hdl.handle.net/10468/13659
dc.language.isoenen
dc.publisherUniversity College Corken
dc.rights© 2022, Sergi Gomez Quintana.en
dc.rights.urihttps://creativecommons.org/licenses/by-nc-nd/4.0/en
dc.subjectBiomedicalen
dc.subjectEEGen
dc.subjectPCGen
dc.subjectNeonatalen
dc.subjectBrainwave sonificationen
dc.subjectAI-assisted brainwave sonificaitonen
dc.subjectEdge implementationen
dc.subjectSignal processingen
dc.subjectMachine learningen
dc.subjectArtificial intelligenceen
dc.titleAI-assisted analysis of heart sounds and interpretation of acoustic representation of brainwaves in neonatesen
dc.typeDoctoral thesisen
dc.type.qualificationlevelDoctoralen
dc.type.qualificationnamePhD - Doctor of Philosophyen
Files
Original bundle
Now showing 1 - 2 of 2
Thumbnail Image
Name:
Sergi Gomez Quintana - PhD Thesis.pdf
Size:
3.92 MB
Format:
Adobe Portable Document Format
Description:
Full Text E-thesis
Thumbnail Image
Name:
3. 118227474 - Sergi Gomez - Submission for examination form.pdf
Size:
203.16 KB
Format:
Adobe Portable Document Format
Description:
Submission for Examination Form
License bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
license.txt
Size:
5.2 KB
Format:
Item-specific license agreed upon to submission
Description:
Download
Collections
Research Theses
CCAE: Cork School of Architecture - Doctoral Theses
College of Science, Engineering and Food Science - Doctoral Theses
Electrical and Electronic Engineering - Doctoral Theses