INFANT Research Centre - Conference Items
Permanent URI for this collection
Browse
Recent Submissions
Item Detection of neurovascular coupling in full-term neonates using wavelet coherence and phase-locking value(Institute of Electrical and Electronics Engineers (IEEE), 2024-12-17) Yu, Kaiyu; Mathieson, Sean; Flynn, Andrew; Dempsey, Eugene; Garvey, Aisling; Boylan, Geraldine; Marnane, William P.; Lightbody, Gordon; Science Foundation IrelandThis study proposes a new method to monitor neurovascular coupling and to help determine the level of brain injury in full-term neonates with Hypoxic Ischemic Encephalopathy (HIE). The wavelet coherence method is used to assess the coupling between the regional cerebral oxygen saturation (rSO2) as measured by NIRS and the lower envelope of EEG power in that region of the brain. This paper also provides methods for the visualization of the phase-locking value between these signals based on both the cross-wavelet and the Hilbert Transform methods. The preliminary results presented in this paper show that the lower envelope of EEG power is much better than the standard EEG power signal for the differentiation between Normal-Mild and Abnormal-Moderate HIE cases using wavelet coherence. On our limited dataset, a clear reduction in both the phase-locking value and the mean coherence was observed in the 0.25-1 mHz frequency range with the increase in severity of the brain injury, from mild to moderate HIE.Item Assessing the effectiveness of heart rate variability as a diagnostic tool for brain injuries in infants(Institute of Electrical and Electronics Engineers (IEEE), 2024-12-17) Rezaei, Kimia; Yu, Kaiyu; Mathieson, Sean R.; Flynn, Andrew; Lightbody, Gordon; Boylan, Geraldine B.; Marnane, William P.; Science Foundation IrelandHypoxic-Ischemic Encephalopathy (HIE), marked by cerebral oxygen deprivation, prompts exploration beyond the electroencephalogram (EEG) modality. This study investigates heart rate variability (HRV) to assess its potential for seizure detection and HIE grading for neonates. This study utilizes two annotated datasets from real-world clinical settings. Heart Rate (HR) is calculated from the Electrocardiogram (ECG) signal, which are then denoised and segmented. Sixteen time and frequency domain features are extracted from each HR segment. Employing Random Forest (RF), Support Vector Machine (SVM), and Isolation Forest (IF) classifiers, the investigation addresses the detection of seizure and nonseizure segments in ECG, alongside categorizing HIE severity into two mild and normal or moderate and severe grades. While the patient-independent evaluation of the seizure detection system reveals promising outcomes for specific cases, there is a requirement for further refinement in this aspect and exploration into the correlation between HR and EEG, considering the modest AUC of 68.54 percent gained across the entire dataset. In contrast, the HIE grading results present a more promising scenario, attaining an AUC of 77.13 percent. This emphasizes the efficacy of the HIE grading system as a significant diagnostic tool, suggesting its potential for broader clinical applications.Item Neonatal hypoxic-ischemic encephalopathy grading from multi-channel EEG time-series data using a fully convolutional neural network(Institute of Electrical and Electronics Engineers (IEEE), 2023-10-18) Yu, Shuwen; Marnane, William P.; Boylan, Geraldine B.; Lightbody, Gordon; Science Foundation Ireland; Wellcome TrustA deep learning classifier is proposed for hypoxic-ischemic encephalopathy (HIE) grading in neonates. Rather than using any features, this architecture can be fed with raw EEG. Fully convolutional layers were adopted both in the feature extraction and classification blocks, which makes this architecture simpler, and deeper, but with fewer parameters. Here two large (335h and 338h respectively) multi-center neonatal continuous EEG datasets were used for training and test. The model was trained based on weak labels and channel independence. A majority vote method was used for the post-processing of the classifier results (across time and channels) to increase the robustness of the prediction. The proposed system achieved an accuracy of 86.09% (95% confidence interval: 82.41% ∼89.78%), an MCC of 0.7691, and an AUC of 86.23% on the large unseen test set. Two convolutional neural network architectures which utilized time-frequency distribution features were selected as the baseline as they had been developed or tested on the same datasets. A relative improvement of 23.65% in test accuracy was obtained as compared with the best baseline.Item Sparse-denoising methods for extracting desaturation transients in cerebral oxygenation signals of preterm Infants(IEEE, 2021-11) Ashoori, Minoo; Dempsey, Eugene M.; McDonald, Fiona B.; O'Toole, John M.; Science Foundation IrelandItem Investigation of lung volume measurements in neonates using gas in scattering media absorption spectroscopy(Optica Publishing Group, 2022-04) Pacheco, Andrea; Jayet, Bapiste; Grygoryev, Konstantin; Messina, Walter; Dehghani, Hamid; Krite Svanberh, Emilie; Dempsey, Eugene M.; Andersson-Engels, Stefan; Science Foundation IrelandWe perform phantom and numerical studies of the changes in molecular oxygen and water vapor spectroscopic signals, showing the potential of measuring pulmonary volume changes with GASMAS technique in neonates.