Periodic Reporting for period 1 - PARENT (PremAtuRe nEwborn motor and cogNitive impairmenTs: Early diagnosis)
Reporting period: 2020-11-01 to 2022-10-31
Preterm birth is the main cause of neurodevelopmental disabilities (NDD). Reliable neuroimaging and other clinical and biochemical markers for detecting high risk infants would be critical in order to take advantage of infant neuroplasticity and improve motor and/or cognitive outcomes through effective therapies. Instead, the classical diagnosis of neurological disfunction on premature infants, is yet based upon clinical monitoring of development (by neuroimaging data, neurological and motor assessments), pointing toward possible signs of impairments as proof of an altered neurodevelopmental trajectory.
Enhancing technologies for detection and rationalization of clinical data from premature infants and follow-ups of newborn at risk for neurological injury, is therefore a necessary step to improve the long-term quality of life and lowering the social cost for the community.
PARENT envisions a multidisciplinary approach to challenge early diagnosis of newborn motor/cognitive impairments. PARENT combines the efforts of a multidisciplinary and inter-sectorial network to develop an innovative training context for 15 Early Stage Researchers (ESRs) within the framework of European Training Network.
PARENT Scientific Objectives (SOs).
SO1 Neonatal Brain Specific Hybrid Neuroimaging Technology
SO2 Personalized Eye Tracking in Newborn at Neurological Risk
SO3 Congenital Heart Disease and Neuro-Developmental Diseases Relationships
SO4 Computational Modelling to predict ncRNA-NDD association
SO5 Multidimensional landscape characterizing neurodevelopmental diseases
Enhancing technologies for detection and rationalization of clinical data from premature infants and follow-ups of newborn at risk for neurological injury, is therefore a necessary step to improve the long-term quality of life and lowering the social cost for the community.
PARENT envisions a multidisciplinary approach to challenge early diagnosis of newborn motor/cognitive impairments. PARENT combines the efforts of a multidisciplinary and inter-sectorial network to develop an innovative training context for 15 Early Stage Researchers (ESRs) within the framework of European Training Network.
PARENT Scientific Objectives (SOs).
SO1 Neonatal Brain Specific Hybrid Neuroimaging Technology
SO2 Personalized Eye Tracking in Newborn at Neurological Risk
SO3 Congenital Heart Disease and Neuro-Developmental Diseases Relationships
SO4 Computational Modelling to predict ncRNA-NDD association
SO5 Multidimensional landscape characterizing neurodevelopmental diseases
PARENT activated 15 PhD researches to cover the five main SOs. Advancement of the activated researches are reported below
SO1
ESR3. Preterm children neurodevelopmental trajectory by neuroimaging and electric signals.
Clinical Study design and submitted protocols for the ethical committee. Investigation of specific neuroimaging markers related to brain maturation, brain injury, and neurodevelopmental outcome in preterms.
ESR6. Hybrid Neuroimaging and Electric Signals Integration by Artificial Intelligence
Deep Learning (DL)-based segmentation methods by exploiting radiomic features obtained from neuroimaging of infants (2D/3D ultrasound and MRI) to identify infants at high risk of neurological dysfunction.
ESR11. Mathematical models for predicting the evolution of cognitive status making use of MRI biomarkers
Working on a deep learning model able to automatically quantify anatomical brain structures in the complete age range (i.e. from pediatric to adult populations).
ESR15. Machine learning and deep learning on heterogeneous data sources to support early diagnosis of neurodevelopmental diseases.
Working on convolutional neural networks for discovering neurodevelopmental disease fingerprint from infant MRIs images.
SO2
ESR8. Machine learning to automatically detect motor/cognitive impairments in premature infants from various sources
Definition of eye-tracking test protocol for use with children (3 months to 2 years) and develpment of optimizated ML workflow for eye-tracking data analysis.
ESR10. Computerized neuropsychological test battery based on eye-tracking specific for preterm children.
Developed a battery of eye-tracking-based computerized neuropsychological tests to be used on preterm infants to detect cognitive or motor deficits. Artificial intelligence models of eye-tracking.
ESR12. Neurological factors determining visual deficits and visuomotor control in children with unilateral Cerebral Palsy.
Design and submission of clinical study protocol for the ethics committee for retrospective, and prospective patient study.
SO3
ESR 13. Neurological biomarkers in predicting neurodevelopment disability driven by Congenital Heart Diseases.
The research is also focusing on patients with Fontan circulation. Design of clinical study protocol for the ethics committee for the retrospective, prospective patient study related to CHD in infants.
ESR 2. Artificial Intelligence driven supporting tools to predict pathology trajectories.
Automated system capable of reading an ECG to predict neurological outcomes. Automatic alignment of 3D ultrasound (US) images with the help of a well-oriented standard ultrasound image.
SO4
ESR 1. Modeling of biochemical pathways and prediction of target interactions in neuro-diseases with the help of bioinformatics and machine learning methods.
Developmed a hybrid multi-objective evolutionary tool using an XGBoost classifier for biomarker discovery on biomedical datasets.
ESR 4. Preterm children neurodevelopmental trajectory by molecular biomarker investigation.
Experimental protocols for studying miRNA changes in the developing brain of the preterm infant and exploring the impact of these on preterm neurodevelopment.
SO5
ESR 5. Monitoring and Integrating Neonates Behavioral and Physiological Parameters.
Design and development of NRP (Neonates Recording Platform), a multi-source heterogeneous and automatic data collection system for infants in neonatal intensive care units.
ESR 9. Predictor of neurodevelopmental trajectory and risk factors from behavioural and physiological parameters acquired by IoT sensors.
Design and development of ML tools for human emotion recognition from voice analysis to be applied to infants.
ESR 7. Ortega Arantxa (UCA): Machine learning to identify significant biomarkers from heterogeneous data for early diagnose of neurodevelopmental diseases in premature infants.
Review on multimodal data integration to enhance prediction of preterm infants at 2 years. Machine Learning classification algorithms trained on clinical data (antenatal and neonatal) to predict abnormal results on MRI in preterm infants.
ESR 14. Romero Pablo (7HC): Integration of AI driven supporting decision tools for diagnosis prognosis and therapy: application to pathology dynamics and neurological pathologies also related to newborn.
Designing software architectures, semantics and decision workflows system capable of integrating AI based predictors and rationalizing heterogeneous data.
SO1
ESR3. Preterm children neurodevelopmental trajectory by neuroimaging and electric signals.
Clinical Study design and submitted protocols for the ethical committee. Investigation of specific neuroimaging markers related to brain maturation, brain injury, and neurodevelopmental outcome in preterms.
ESR6. Hybrid Neuroimaging and Electric Signals Integration by Artificial Intelligence
Deep Learning (DL)-based segmentation methods by exploiting radiomic features obtained from neuroimaging of infants (2D/3D ultrasound and MRI) to identify infants at high risk of neurological dysfunction.
ESR11. Mathematical models for predicting the evolution of cognitive status making use of MRI biomarkers
Working on a deep learning model able to automatically quantify anatomical brain structures in the complete age range (i.e. from pediatric to adult populations).
ESR15. Machine learning and deep learning on heterogeneous data sources to support early diagnosis of neurodevelopmental diseases.
Working on convolutional neural networks for discovering neurodevelopmental disease fingerprint from infant MRIs images.
SO2
ESR8. Machine learning to automatically detect motor/cognitive impairments in premature infants from various sources
Definition of eye-tracking test protocol for use with children (3 months to 2 years) and develpment of optimizated ML workflow for eye-tracking data analysis.
ESR10. Computerized neuropsychological test battery based on eye-tracking specific for preterm children.
Developed a battery of eye-tracking-based computerized neuropsychological tests to be used on preterm infants to detect cognitive or motor deficits. Artificial intelligence models of eye-tracking.
ESR12. Neurological factors determining visual deficits and visuomotor control in children with unilateral Cerebral Palsy.
Design and submission of clinical study protocol for the ethics committee for retrospective, and prospective patient study.
SO3
ESR 13. Neurological biomarkers in predicting neurodevelopment disability driven by Congenital Heart Diseases.
The research is also focusing on patients with Fontan circulation. Design of clinical study protocol for the ethics committee for the retrospective, prospective patient study related to CHD in infants.
ESR 2. Artificial Intelligence driven supporting tools to predict pathology trajectories.
Automated system capable of reading an ECG to predict neurological outcomes. Automatic alignment of 3D ultrasound (US) images with the help of a well-oriented standard ultrasound image.
SO4
ESR 1. Modeling of biochemical pathways and prediction of target interactions in neuro-diseases with the help of bioinformatics and machine learning methods.
Developmed a hybrid multi-objective evolutionary tool using an XGBoost classifier for biomarker discovery on biomedical datasets.
ESR 4. Preterm children neurodevelopmental trajectory by molecular biomarker investigation.
Experimental protocols for studying miRNA changes in the developing brain of the preterm infant and exploring the impact of these on preterm neurodevelopment.
SO5
ESR 5. Monitoring and Integrating Neonates Behavioral and Physiological Parameters.
Design and development of NRP (Neonates Recording Platform), a multi-source heterogeneous and automatic data collection system for infants in neonatal intensive care units.
ESR 9. Predictor of neurodevelopmental trajectory and risk factors from behavioural and physiological parameters acquired by IoT sensors.
Design and development of ML tools for human emotion recognition from voice analysis to be applied to infants.
ESR 7. Ortega Arantxa (UCA): Machine learning to identify significant biomarkers from heterogeneous data for early diagnose of neurodevelopmental diseases in premature infants.
Review on multimodal data integration to enhance prediction of preterm infants at 2 years. Machine Learning classification algorithms trained on clinical data (antenatal and neonatal) to predict abnormal results on MRI in preterm infants.
ESR 14. Romero Pablo (7HC): Integration of AI driven supporting decision tools for diagnosis prognosis and therapy: application to pathology dynamics and neurological pathologies also related to newborn.
Designing software architectures, semantics and decision workflows system capable of integrating AI based predictors and rationalizing heterogeneous data.
PARENT advances on the state of the art across the spectrum of neuroimaging and eye tracking techniques, e.g. automatic hybrid US/MRI imaging coupled to fine-grained computational analysis and laboratory data.
The use of predictive modelling fed by artificial intelligence will improve early diagnosis and personalisation of patient evaluation. The follow-up and the and long term therapies will be supported by a huge number of quantitative data compressed in more simple and easy to use parameters for clinicians. PARENT maintain clinicians at the centre of biomedical research: they will not be considered only final users, but main players driving problem investigation and using supporting decision for boosting new ideas.
Early diagnosis means also better preparation to therapies. Families aware of the exact path that their children will have to follow, will also be more responsive and prepared to face the important burden of the therapy. The reduction of parents’ anxiety will be itself a positive factor for children life quality.
PARENT foster the design of flexible technologies and strategies. The abandonment to a ‘dedicated’ approach in favour of a more ‘universal’ one, will encourage open-concurrence in the market. The innovation in technological assessment will affect positively the profit margins, with new beneficial investments.
The use of predictive modelling fed by artificial intelligence will improve early diagnosis and personalisation of patient evaluation. The follow-up and the and long term therapies will be supported by a huge number of quantitative data compressed in more simple and easy to use parameters for clinicians. PARENT maintain clinicians at the centre of biomedical research: they will not be considered only final users, but main players driving problem investigation and using supporting decision for boosting new ideas.
Early diagnosis means also better preparation to therapies. Families aware of the exact path that their children will have to follow, will also be more responsive and prepared to face the important burden of the therapy. The reduction of parents’ anxiety will be itself a positive factor for children life quality.
PARENT foster the design of flexible technologies and strategies. The abandonment to a ‘dedicated’ approach in favour of a more ‘universal’ one, will encourage open-concurrence in the market. The innovation in technological assessment will affect positively the profit margins, with new beneficial investments.