Bonhwang Koo

An open resource for transdiagnostic research in pediatric mental health and learning disorders

Technological and methodological innovations are equipping researchers with unprecedented capabilities for detecting and characterizing pathologic processes in the developing human brain. As a result, ambitions to achieve clinically useful tools to assist in the diagnosis and management of mental health and learning disorders are gaining momentum. To this end, it is critical to accrue large-scale multimodal datasets that capture a broad range of commonly encountered clinical psychopathology. The Child Mind Institute has launched the Healthy Brain Network (HBN), an ongoing initiative focused on creating and sharing a biobank of data from 10,000 New York area participants (ages 5–21). The HBN Biobank houses data about psychiatric, behavioral, cognitive, and lifestyle phenotypes, as well as multimodal brain imaging (resting and naturalistic viewing fMRI, diffusion MRI, morphometric MRI), electroencephalography, eye-tracking, voice and video recordings, genetics and actigraphy. Here, we present the rationale, design and implementation of HBN protocols. We describe the first data release (n=664) and the potential of the biobank to advance related areas (e.g., biophysical modeling, voice analysis).

A large, open source dataset of stroke anatomical brain images and manual lesion segmentations

Stroke is the leading cause of adult disability worldwide, with up to two-thirds of individuals experiencing long-term disabilities. Large-scale neuroimaging studies have shown promise in identifying robust biomarkers (e.g., measures of brain structure) of long-term stroke recovery following rehabilitation. However, analyzing large rehabilitation-related datasets is problematic due to barriers in accurate stroke lesion segmentation. Manually-traced lesions are currently the gold standard for lesion segmentation on T1-weighted MRIs, but are labor intensive and require anatomical expertise. While algorithms have been developed to automate this process, the results often lack accuracy. Newer algorithms that employ machine-learning techniques are promising, yet these require large training datasets to optimize performance. Here we present ATLAS (Anatomical Tracings of Lesions After Stroke), an open-source dataset of 304 T1-weighted MRIs with manually segmented lesions and metadata. This large, diverse dataset can be used to train and test lesion segmentation algorithms and provides a standardized dataset for comparing the performance of different segmentation methods. We hope ATLAS release 1.1 will be a useful resource to assess and improve the accuracy of current lesion segmentation methods.

An Open Resource for Non-human Primate Imaging

Summary Non-human primate neuroimaging is a rapidly growing area of research that promises to transform and scale translational and cross-species comparative neuroscience. Unfortunately, the technological and methodological advances of the past two decades have outpaced the accrual of data, which is particularly challenging given the relatively few centers that have the necessary facilities and capabilities. The PRIMatE Data Exchange (PRIME-DE) addresses this challenge by aggregating independently acquired non-human primate magnetic resonance imaging (MRI) datasets and openly sharing them via the International Neuroimaging Data-sharing Initiative (INDI). Here, we present the rationale, design, and procedures for the PRIME-DE consortium, as well as the initial release, consisting of 25 independent data collections aggregated across 22 sites (total = 217 non-human primates). We also outline the unique pitfalls and challenges that should be considered in the analysis of non-human primate MRI datasets, including providing automated quality assessment of the contributed datasets.

The Healthy Brain Network Biobank: An open resource for transdiagnostic research in pediatric mental health and learning disorders

Innovations in methods and technologies are equipping researchers with unprecedented capabilities for detecting and characterizing pathologic processes in the developing human brain. As a result, there is growing enthusiasm about the prospect of achieving clinically useful tools that can assist in the diagnosis and management of mental health and learning disorders. For these ambitions to be realized, it is critical to accrue large-scale multimodal datasets that capture a broad range of commonly encountered clinical psychopathology. To this end, the Child Mind Institute has launched the Healthy Brain Network (HBN), an ongoing initiative focused on creating and sharing a biobank comprised of data from 10,000 New York City area children and adolescents (ages 5-21). The HBN has adopted a community-referred recruitment model. Specifically, study advertisements seek the participation of families who have concerns about one or more psychiatric symptoms in their child. The HBN Biobank houses data about psychiatric, behavioral, cognitive, and lifestyle (e.g., fitness, diet) phenotypes, as well as multimodal brain imaging, electroencephalography, digital voice and video recordings, genetics, and actigraphy. In this paper, we present the motivation, rationale and design for the HBN along with the initial implementation and evolution of the HBN protocols. We describe the first major open data release (n = 664) containing descriptive, electroencephalography, and multimodal brain imaging data (resting state and naturalistic viewing functional MRI, diffusion MRI and morphometric MRI). Beyond accelerating transdiagnostic research, we discuss the potential of the HBN Biobank to advance related areas, such as biophysical modeling, voice and speech analysis, natural viewing fMRI and EEG, and methods optimization.

Passive Audio Vocal Capture and Measurement in the Evaluation of Selective Mutism

Selective Mutism (SM) is an anxiety disorder often diagnosed in early childhood and characterized by persistent failure to speak in certain social situations but not others. Diagnosing SM and monitoring treatment response can be quite complex, due in part to changing definitions of and scarcity of research about the disorder. Subjective self-reports and parent/teacher interviews can complicate SM diagnosis and therapy, given that similar speech problems of etiologically heterogeneous origin can be attributed to SM. The present perspective discusses the potential for passive audio capture to help overcome psychiatry’s current lack of objective and quantifiable assessments in the context of SM. We present evidence from two pilot studies indicating the feasibility of using a digital wearable device to quantify child vocalization features affected by SM. We also highlight limitations in the design and implementation of this preliminary work that can help guide future efforts.

Assessment of the impact of shared brain imaging data on the scientific literature

Data sharing is increasingly recommended as a means of accelerating science by facilitating collaboration, transparency, and reproducibility. While few oppose data sharing philosophically, a range of barriers deter most researchers from implementing it in practice. To justify the significant effort required for sharing data, funding agencies, institutions, and investigators need clear evidence of benefit. Here, using the International Neuroimaging Data-sharing Initiative, we present a case study that provides direct evidence of the impact of open sharing on brain imaging data use and resulting peer-reviewed publications. We demonstrate that openly shared data can increase the scale of scientific studies conducted by data contributors, and can recruit scientists from a broader range of disciplines. These findings dispel the myth that scientific findings using shared data cannot be published in high-impact journals, suggest the transformative power of data sharing for accelerating science, and underscore the need for implementing data sharing universally.Data sharing is recognized as a way to promote scientific collaboration and reproducibility, but some are concerned over whether research based on shared data can achieve high impact. Here, the authors show that neuroimaging papers using shared data are no less likely to appear in top-ranked journals.

Clinical Perspective on Passive Audio Vocal Measurement in the Evaluation of Selective Mutism

Selective Mutism (SM) is an anxiety disorder often diagnosed in early childhood and characterized by persistent failure to speak in certain social situations but not others. Diagnosing SM and monitoring treatment response can be quite complex, due in part to changing definitions of and scarcity of research about the disorder. Subjective self-reports and parent/teacher interviews can complicate SM diagnosis and therapy, given that similar speech problems of etiologically heterogeneous origin can be attributed to SM. The present perspective discusses the potential for passive audio capture to help overcome psychiatrys current lack of objective and quantifiable assessments in the context of SM. We present supportive evidence from two pilot studies indicating the feasibility of using a digital wearable device to quantify child vocalization features affected by SM. We also highlight comparative analyses of passive audio capture and its potential to enhance diagnostic characterizations for SM, as well as possible limitations of such technologies.

The Anatomical Tracings of Lesions After Stroke (ATLAS) Dataset - Release 1.1

Stroke is the leading cause of adult disability worldwide, with up to two-thirds of individuals experiencing long-term disabilities. Large-scale neuroimaging studies have shown promise in identifying robust biomarkers (e.g., measures of brain structure) of long-term stroke recovery following rehabilitation. However, analyzing large rehabilitation-related datasets is problematic due to barriers in accurate stroke lesion segmentation. Manually-traced lesions are currently the gold standard for lesion segmentation on T1-weighted MRIs, but are labor intensive and require anatomical expertise. While algorithms have been developed to automate this process, the results often lack accuracy. Newer algorithms that employ machine-learning techniques are promising, yet these require large training datasets to optimize performance. Here we present ATLAS (Anatomical Tracings of Lesions After Stroke), an open-source dataset of 304 T1-weighted MRIs with manually segmented lesions and metadata. This large, diverse dataset can be used to train and test lesion segmentation algorithms and provides a standardized dataset for comparing the performance of different segmentation methods. We hope ATLAS release 1.1 will be a useful resource to assess and improve the accuracy of current lesion segmentation methods.Stroke is the leading cause of adult disability worldwide, with up to two-thirds of individuals experiencing long-term disabilities. Large-scale neuroimaging studies have shown promise in identifying robust biomarkers (e.g., measures of brain structure) of stroke recovery. However, analyzing large datasets is problematic due to barriers in accurate stroke lesion segmentation. Manually-traced lesions are currently the gold standard for lesion segmentation, but are labor intensive and require anatomical expertise. While algorithms have been developed to automate this process, the results often lack accuracy. Newer algorithms that employ machine-learning techniques are promising, yet these require large training datasets to optimize performance. Here we present ATLAS (Anatomical Tracings of Lesions After Stroke), an open-source dataset of 304 T1-weighted MRIs with manually segmented lesions and metadata. This large, diverse dataset can be used to train and test lesion segmentation algorithms and provides a standardized dataset for comparing the performance of different segmentation methods. We hope ATLAS R1.1 will be a useful resource to assess and improve the accuracy of current lesion segmentation methods.

Assessment of the impact of shared data on the scientific literature

Data sharing is increasingly recommended as a means of accelerating science by facilitating collaboration, transparency, and reproducibility. While few oppose data sharing philosophically, a range of barriers deter most researchers from implementing it in practice (e.g., workforce and infrastructural demands, sociocultural and privacy concerns, lack of standardization). To justify the significant effort required for sharing data (e.g., organization, curation, distribution), funding agencies, institutions, and investigators need clear evidence of benefit. Here, using the International Neuroimaging Data-sharing Initiative, we present a brain imaging case study that provides direct evidence of the impact of open sharing on data use and resulting publications over a seven-year period (2010-2017). We dispel the myth that scientific findings using shared data cannot be published in high-impact journals and demonstrate rapid growth in the publication of such journal articles, scholarly theses, and conference proceedings. In contrast to commonly used ‘pay to play’ models, we demonstrate that openly shared data can increase the scale (i.e., sample size) of scientific studies conducted by data contributors, and can recruit scientists from a broader range of disciplines. These findings suggest the transformative power of data sharing for accelerating science and underscore the need for the scientific ecosystem to embrace the challenge of implementing data sharing universally.

Assessing actimeters for inclusion in the Healthy Brain Network

Background The Healthy Brain Network is an openly shared pediatric psychiatric biobank with a target of 10,000 participants between the ages of 5 and 21, inclusively In adding ecological actimetry to the Healthy Brain Network, we intend to use appropriate, accurate, reliable tools. Currently a wide range of personal activity trackers are commercially available, providing a wide variety of sensor configurations. For many of these devices, accelerometry provides the basis of measuring both physical activity and sleep with comparable derivative measures. Results In order to include an ecological biotracker in the Healthy Brain Network protocol, we first evaluated the specifications of a variety of actimeters available for purchase. We then acquired physical instances of 5 of these devices (ActiGraph wGT3X-BT, Empatica Embrace, Empatica E4, GENEActiv Original, and Wavelet Wristband) and wore each of them in our daily lives, annotating our activities and evaluating the reasonableness of the data from each device and the logistical affordances of each device. Conclusions We decided that the ActiGraph wGT3X-BT is the most appropriate device for inclusion in the Healthy Brain Network. However, none of the devices we evaluated was clearly superior or inferior to the rest; rather, each device seems to have use cases in which that device excels beyond the others.