Modality-Level Obstacles and Initiatives to Improve Representation in Fetal, Infant, and Toddler Neuroimaging Research Samples

Developmental Cognitive Neuroscience, Journal Year: 2025, Volume and Issue: 72, P. 101505 - 101505

Published: Jan. 5, 2025

Fetal, infant, and toddler (FIT) neuroimaging researchers study early brain development to gain insights into neurodevelopmental processes identify markers of neurobiological vulnerabilities target for intervention. However, the field has historically excluded people from global majority countries marginalized communities in FIT research. Inclusive representative samples are essential generalizing findings across modalities, such as magnetic resonance imaging, magnetoencephalography, electroencephalography, functional near-infrared spectroscopy, cranial ultrasonography. These techniques pose unique overlapping challenges equitable representation research through sampling bias, technical constraints, limited accessibility, insufficient resources. The present article adds conversation around need improve inclusivity by highlighting modality-specific historical current obstacles ongoing initiatives. We conclude discussing tangible solutions that transcend individual ultimately providing recommendations promote neuroscience.

Language: Английский

---

Structural MRI of brain similarity networks

Nature reviews. Neuroscience, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 28, 2024

Language: Английский

---

Feasibility and Usability of Low-Field Magnetic Resonance Imaging for Pediatric Neuroimaging in Low- and Middle-Income Countries: A Qualitative Study

Medical Devices Evidence and Research, Journal Year: 2025, Volume and Issue: Volume 18, P. 107 - 121

Published: Feb. 1, 2025

The burden of neurological disorders in low- and middle-income countries (LMICs) may be underestimated due to the limited number diagnostic imaging devices trained specialists operate interpret scans. Recent advancements low-field (<100 milliteslas) magnetic resonance (LFMRI) hold significant promise for improving access pediatric neuroimaging technology's lower costs, portability, reduced infrastructure training requirements. Explore user needs experiences on use a portable LFMRI LMICs. We conducted qualitative interviews with end users systems across 11 sites Bangladesh, Ethiopia, Ghana, Malawi, Pakistan, South Africa, Uganda, Zambia. A semi-structured questionnaire open-ended questions usability feasibility was used encourage participants share their opinions ease use, satisfaction, integration into local health systems. Among 46 participants, key challenges were reported infant positioning, power stability, internet connectivity. Suggestions included developing reference materials content format tailored contexts, conducting refresher trainings, providing education that includes technical maintenance support crucial appropriate utilization implementation sustainability. This study underscores importance incorporating human-centered design principles feedback identifying resolving issues, sharing insights successful within existing care infrastructures LMICs, optimizing populations.

Language: Английский

---

Magnetization transfer imaging using non‐balanced SSFP at ultra‐low field

Magnetic Resonance in Medicine, Journal Year: 2025, Volume and Issue: unknown

Published: March 17, 2025

Abstract Purpose Ultra‐low field MRI scanners have the potential to improve health care delivery, both through improved access in areas where there are few and allowing more frequent monitoring of disease progression treatment response. This may be particularly true white matter disorders, including leukodystrophies multiple sclerosis, which myelin‐sensitive imaging, such as magnetization transfer (MT) might clinical patient outcomes. Methods We implemented an on‐resonance approach MT imaging on a commercial point‐of‐care 64 mT scanner using non‐balanced steady‐state free precession sequence. Phantom vivo experiments were used evaluate optimize sequence sensitivity reproducibility, demonstrate performance inter‐site reproducibility. Results From phantom experiments, T 1 2 effects determined negligible effect differential weighting. ratio (MTR) values 23.1 ± 1.0% from 10 healthy volunteers, with average reproducibility coefficient variation 1.04%. Normal‐appearing MTR sclerosis participant (21.5 6.2%) lower, but similar spread values, compared age‐matched volunteer (23.3 6.2%). Conclusion An was developed at that can performed little 4 min. A semi‐quantitative biomarker this strength is available for assessing myelination demyelination.

Language: Английский

---

Integration of multimodal imaging data with machine learning for improved diagnosis and prognosis in neuroimaging

Frontiers in Human Neuroscience, Journal Year: 2025, Volume and Issue: 19

Published: March 21, 2025

Introduction Combining many types of imaging data—especially structural MRI (sMRI) and functional (fMRI)—may greatly assist in the diagnosis treatment brain disorders like Alzheimer’s. Current approaches are less helpful for forecasting, however, as they do not always blend spatial temporal patterns from different sources properly. This work presents a novel mixed deep learning (DL) method combining data using CNN, GRU, attention techniques. introduces hybrid Dynamic Cross-Modality Attention Module to help more efficiently data. Through working around issues with current multimodal fusion techniques, our approach increases accuracy readability diagnoses. Methods Utilizing CNNs models dynamics fMRI connection measures utilizing GRUs, proposed extracts characteristics sMRI. Strong integration is made possible by including an mechanism give diagnostically important features top priority. Training evaluation model took place Human Connectome Project (HCP) dataset behavioral data, fMRI, Measures include accuracy, recall, precision F1-score used evaluate performance. Results It was correct 96.79% time combined structure. Regarding identification disorders, successful than existing ones. Discussion These findings indicate that strategy makes sense complimentary information several kinds photos. detail helped one choose which aspects concentrate on, thereby enhancing diagnostic accuracy. Conclusion The offers fresh benchmark neuroimaging analysis has great potential use real-world assessment prediction. Researchers will investigate future applications this technique new picture clinical

Language: Английский

---

Bahir Dar Child Development Cross-Sectional Study, Ethiopia: study protocol

BMJ Paediatrics Open, Journal Year: 2025, Volume and Issue: 9(1), P. e003173 - e003173

Published: April 1, 2025

Introduction Foundational preacademic skills are crucial for academic success and serve as predictors of socioeconomic status, income access to healthcare. However, there is a gap in our understanding neurodevelopmental patterns underlying children across low-income middle-income countries (LMICs). It essential identify primary global regional factors that drive children’s neurodevelopment LMICs. This study aims characterise the typical development healthy influence child Bahir Dar, Ethiopia. Methods analysis The Dar Child Development Study cross-sectional implemented two health centres, Shimbit Abaymado Felege Hiwot Comprehensive Specialized Hospital (FHCSH) Amhara, Healthy between 6 60 months age will be recruited from centres during vaccination visits or via community outreach. Young aged 6–36 complete Global Scale Early Developmen t . A battery paper tablet-based assessments neurocognitive outcomes including visual verbal reasoning, executive functions school readiness completed 48–60 months. Caregivers respond surveys covering sociodemographic information, child’s medical history nutrition, psychosocial experiences parental stress mental health. During second visit, participants undergo low-field MRI scan using ultra-low-field point-of-care Hyperfine machine at FHCSH. Analyses examine relationships risk protective factors, brain volumes neurocognitive/developmental outcomes. Ethics dissemination approved by Institutional Review Boards Addis Continental Institute Public Health (ACIPH/lRERC/004/2023/Al/05-2024), Mass General Brigham (2022P002539) Brown University (STUDY00000474). Findings disseminated local events, international conferences publications. Trial Registeration number NCT06648863

Language: Английский

---

Ultra-low-field brain MRI morphometry: test-retest reliability and correspondence to high-field MRI

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Aug. 19, 2024

Magnetic resonance imaging (MRI) enables non-invasive monitoring of healthy brain development and disease. Widely used higher field (>1.5 T) MRI systems are associated with high energy infrastructure requirements, costs. Recent ultra-low-field (<0.1T) provide a more accessible cost-effective alternative. However, it is not known whether anatomical neuroimaging can be to extract quantitative measures morphometry, what extent such correspond high-field MRI. Here we scanned 23 adults aged 20-69 years on two identical 64 mT 3 T system, using 1 w 2 scans across range (64 mT) resolutions. We segmented images into 4 global tissue types 98 local structures, systematically evaluated between-scanner reliability morphometry correspondence measurements, correlations volume Dice spatial overlap segmentations. report scan contrasts resolutions, highest performance shown by combining three low through-plane resolution single higher-resolution multi-resolution registration. Larger structures show T. Finally, showcase the potential for mapping neuroanatomical changes lifespan, relevant neurological disorders. Raw code publicly available ( upon publication ), enabling systematic validation pre-processing analysis approaches neuroimaging.

Language: Английский

---

Deep learning super-resolution of paediatric ultra-low-field MRI without paired high-field scans

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 30, 2024

Brain magnetic resonance imaging (MRI) is essential for diagnosis and neurodevelopmental research, but the high cost infrastructure demands of high-field MRI scanners restrict their use to high-income settings. To address this, more affordable energy-efficient ultra-low-field have been developed. However, reduced resolution signal-to-noise ratio resulting scans limit clinical utility, motivating development super-resolution techniques. The current state-of-the-art methods require either three anisotropic acquired at different orientations (axial, coronal, sagittal) reconstruct a higher-resolution image using multi-resolution registration (MRR), or training deep learning models paired ultra-low- scans. Since acquiring high-quality not always feasible, data may be available target population, this study explores efficacy model, 3D UNet, generate brain from just one scan. model was trained receive single scan 6-month-old infants produce MRR quality. Results showed significant improvement in quality output compared input scans, including increased metrics, stronger correlations tissue volume estimates across participants, greater Dice overlap underlying segmentations those demonstrates that UNet effectively enhances infant Generating without needing data, reduces scanning time supports wider low- middle-income countries. Additionally, approach allows easier on site- population-specific basis, enhancing adaptability diverse

Language: Английский

---

Ultra‐Low‐Field Paediatric MRI in Low‐ and Middle‐Income Countries: Super‐Resolution Using a Multi‐Orientation U‐Net

Human Brain Mapping, Journal Year: 2024, Volume and Issue: 46(1)

Published: Dec. 30, 2024

ABSTRACT Owing to the high cost of modern magnetic resonance imaging (MRI) systems, their use in clinical care and neurodevelopmental research is limited hospitals universities income countries. Ultra‐low‐field systems with significantly lower scanning costs present a promising avenue towards global MRI accessibility; however, reduced SNR compared 1.5 or 3 T limits applicability for use. In this paper, we describe deep learning‐based super‐resolution approach generate high‐resolution isotropic 2 ‐weighted scans from low‐resolution paediatric input scans. We train ‘multi‐orientation U‐Net’, which uses multiple anisotropic images acquired orthogonal orientations construct super‐resolved output. Our exhibits improved quality outputs current state‐of‐the‐art methods ultra‐low‐field populations. Crucially development, our improves reconstruction brain structures greatest improvement volume estimates caudate, where model upon in: linear correlation ( r = 0.94 vs. 0.84 using existing methods), exact agreement (Lin's concordance 0.80) mean error (0.05 cm 0.36 ). serves as proof‐of‐principle viability training deep‐learning based models presents first trained exclusively on paired high‐field data infants.

Language: Английский

---