I am interested in using machine learning and computational approaches to analyze longitudinal and multi-modal MRI to characterize how the white-matter architecture develops during adolescence to support coordinated neural activity for developing higher-order executive functions. My research also extends to characterizing the predisposing and detrimental effects of alcohol and substance use on brain structure and function. My broad interest lies in image analysis and statistical learning for the detection, diagnosis and treatment of diseases.

Academic Appointments

Administrative Appointments

  • Guest Editor, Applied Sciences (2022 - Present)
  • Associate Editor, Frontiers in Neuroimaging (2022 - Present)

Honors & Awards

  • Innovator Grant Program, Stanford Psychiatry (2022-Present)
  • K99/R00 Pathway to Independence Award, NIH/NIAAA (2021-Present)

Professional Education

  • B.S., Shanghai Jiao Tong University, Computer Science (2012)
  • Ph.D, University of North Carolina at Chapel Hill, Computer Science (2017)

All Publications

  • Disentangling Normal Aging From Severity of Disease via Weak Supervision on Longitudinal MRI IEEE TRANSACTIONS ON MEDICAL IMAGING Ouyang, J., Zhao, Q., Adeli, E., Zaharchuk, G., Pohl, K. M. 2022; 41 (10): 2558-2569


    The continuous progression of neurological diseases are often categorized into conditions according to their severity. To relate the severity to changes in brain morphometry, there is a growing interest in replacing these categories with a continuous severity scale that longitudinal MRIs are mapped onto via deep learning algorithms. However, existing methods based on supervised learning require large numbers of samples and those that do not, such as self-supervised models, fail to clearly separate the disease effect from normal aging. Here, we propose to explicitly disentangle those two factors via weak-supervision. In other words, training is based on longitudinal MRIs being labelled either normal or diseased so that the training data can be augmented with samples from disease categories that are not of primary interest to the analysis. We do so by encouraging trajectories of controls to be fully encoded by the direction associated with brain aging. Furthermore, an orthogonal direction linked to disease severity captures the residual component from normal aging in the diseased cohort. Hence, the proposed method quantifies disease severity and its progression speed in individuals without knowing their condition. We apply the proposed method on data from the Alzheimer's Disease Neuroimaging Initiative (ADNI, N =632 ). We then show that the model properly disentangled normal aging from the severity of cognitive impairment by plotting the resulting disentangled factors of each subject and generating simulated MRIs for a given chronological age and condition. Moreover, our representation obtains higher balanced accuracy when used for two downstream classification tasks compared to other pre-training approaches. The code for our weak-supervised approach is available at

    View details for DOI 10.1109/TMI.2022.3166131

    View details for Web of Science ID 000862400100002

    View details for PubMedID 35404811

    View details for PubMedCentralID PMC9578549

  • Multiple Instance Neuroimage Transformer. PRedictive Intelligence in MEdicine. PRIME (Workshop) Singla, A., Zhao, Q., Do, D. K., Zhou, Y., Pohl, K. M., Adeli, E. 2022; 13564: 36-48


    For the first time, we propose using a multiple instance learning based convolution-free transformer model, called Multiple Instance Neuroimage Transformer (MINiT), for the classification of T1-weighted (T1w) MRIs. We first present several variants of transformer models adopted for neuroimages. These models extract non-overlapping 3D blocks from the input volume and perform multi-headed self-attention on a sequence of their linear projections. MINiT, on the other hand, treats each of the non-overlapping 3D blocks of the input MRI as its own instance, splitting it further into non-overlapping 3D patches, on which multi-headed self-attention is computed. As a proof-of-concept, we evaluate the efficacy of our model by training it to identify sex from T1w-MRIs of two public datasets: Adolescent Brain Cognitive Development (ABCD) and the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA). The learned attention maps highlight voxels contributing to identifying sex differences in brain morphometry. The code is available at

    View details for DOI 10.1007/978-3-031-16919-9_4

    View details for PubMedID 36331280

    View details for PubMedCentralID PMC9629332

  • Bridging the Gap between Deep Learning and Hypothesis-Driven Analysis via Permutation Testing. PRedictive Intelligence in MEdicine. PRIME (Workshop) Paschali, M., Zhao, Q., Adeli, E., Pohl, K. M. 2022; 13564: 13-23


    A fundamental approach in neuroscience research is to test hypotheses based on neuropsychological and behavioral measures, i.e., whether certain factors (e.g., related to life events) are associated with an outcome (e.g., depression). In recent years, deep learning has become a potential alternative approach for conducting such analyses by predicting an outcome from a collection of factors and identifying the most "informative" ones driving the prediction. However, this approach has had limited impact as its findings are not linked to statistical significance of factors supporting hypotheses. In this article, we proposed a flexible and scalable approach based on the concept of permutation testing that integrates hypothesis testing into the data-driven deep learning analysis. We apply our approach to the yearly self-reported assessments of 621 adolescent participants of the National Consortium of Alcohol and Neurodevelopment in Adolescence (NCANDA) to predict negative valence, a symptom of major depressive disorder according to the NIMH Research Domain Criteria (RDoC). Our method successfully identifies categories of risk factors that further explain the symptom.

    View details for DOI 10.1007/978-3-031-16919-9_2

    View details for PubMedID 36342897

    View details for PubMedCentralID PMC9632755

  • Alcohol's effects on the mouse brain are modulated by age and sex. Addiction biology Piekarski, D. J., Zahr, N. M., Zhao, Q., Sullivan, E. V., Pfefferbaum, A. 2022; 27 (5): e13209


    Binge alcohol consumption is common among adolescents and may impair normal brain development. Emerging, longitudinal studies in adolescents suggest that the effects of binge alcohol exposure on brain structure differ between sexes. To test the hypothesis that the effects of binge alcohol exposure on developmental brain growth trajectories are influenced by age of exposure and sex, adolescent and adult, male and female C57Bl/6 mice (n = 32), were exposed to a binge-like ethanol (EtOH) exposure paradigm (i.e., 5 cycles of 2 on/2 off days of 5 g/kg EtOH intraperitoneal) or served as saline controls. Longitudinal structural magnetic resonance imaging was acquired at baseline, following binge EtOH exposure, and after 2 weeks of recovery. Alcohol treatment showed interactions with age and sex in altering whole brain volume: adolescents of both sexes demonstrated inhibited whole brain growth relative to their control counterparts, although significance was only attained in female mice which showed a larger magnitude response to EtOH compared to male mice. In region of interest analyses, the somatosensory cortex and cerebellum showed inhibited growth in male and female adolescent mice exposed to EtOH, but the difference relative to controls did not reach multiple comparison-corrected statistical significance. These data suggest that in mice exposed to binge EtOH treatment, adolescent age of exposure and female sex may confer a higher risk to the detrimental effects of EtOH on brain structure and reinforce the need for direct testing of both sexes.

    View details for DOI 10.1111/adb.13209

    View details for PubMedID 36001428

  • Alcohol's effects on the mouse brain are modulated by age and sex ADDICTION BIOLOGY Piekarski, D. J., Zahr, N. M., Zhao, Q., Sullivan, E., Pfefferbaum, A. 2022; 27 (5)

    View details for DOI 10.1111/adb.13209

    View details for Web of Science ID 000823573700001

  • A Penalty Approach for Normalizing Feature Distributions to Build Confounder-Free Models. Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention Vento, A., Zhao, Q., Paul, R., Pohl, K., Adeli, E. 2022; 13433: 387-397


    Translating the use of modern machine learning algorithms into clinical applications requires settling challenges related to explain-ability and management of nuanced confounding factors. To suitably interpret the results, removing or explaining the effect of confounding variables (or metadata) is essential. Confounding variables affect the relationship between input training data and target outputs. Accordingly, when we train a model on such data, confounding variables will bias the distribution of the learned features. A recent promising solution, Meta-Data Normalization (MDN), estimates the linear relationship between the metadata and each feature based on a non-trainable closed-form solution. However, this estimation is confined by the sample size of a mini-batch and thereby may result in an oscillating performance. In this paper, we extend the MDN method by applying a Penalty approach (referred to as PDMN). We cast the problem into a bi-level nested optimization problem. We then approximate that objective using a penalty method so that the linear parameters within the MDN layer are trainable and learned on all samples. This enables PMDN to be plugged into any architectures, even those unfit to run batch-level operations such as transformers and recurrent models. We show improvement in model accuracy and independence from the confounders using PMDN over MDN in a synthetic experiment and a multi-label, multi-site classification of magnetic resonance images.

    View details for DOI 10.1007/978-3-031-16437-8_37

    View details for PubMedID 36331278

    View details for PubMedCentralID PMC9629333

  • Self-supervised learning of neighborhood embedding for longitudinal MRI. Medical image analysis Ouyang, J., Zhao, Q., Adeli, E., Zaharchuk, G., Pohl, K. M. 2022; 82: 102571


    In recent years, several deep learning models recommend first to represent Magnetic Resonance Imaging (MRI) as latent features before performing a downstream task of interest (such as classification or regression). The performance of the downstream task generally improves when these latent representations are explicitly associated with factors of interest. For example, we derived such a representation for capturing brain aging by applying self-supervised learning to longitudinal MRIs and then used the resulting encoding to automatically identify diseases accelerating the aging of the brain. We now propose a refinement of this representation by replacing the linear modeling of brain aging with one that is consistent in local neighborhoods in the latent space. Called Longitudinal Neighborhood Embedding (LNE), we derive an encoding so that neighborhoods are age-consistent (i.e., brain MRIs of different subjects with similar brain ages are in close proximity of each other) and progression-consistent, i.e., the latent space is defined by a smooth trajectory field where each trajectory captures changes in brain ages between a pair of MRIs extracted from a longitudinal sequence. To make the problem computationally tractable, we further propose a strategy for mini-batch sampling so that the resulting local neighborhoods accurately approximate the ones that would be defined based on the whole cohort. We evaluate LNE on three different downstream tasks: (1) to predict chronological age from T1-w MRI of 274 healthy subjects participating in a study at SRI International; (2) to distinguish Normal Control (NC) from Alzheimer's Disease (AD) and stable Mild Cognitive Impairment (sMCI) from progressive Mild Cognitive Impairment (pMCI) based on T1-w MRI of 632 participants of the Alzheimer's Disease Neuroimaging Initiative (ADNI); and (3) to distinguish no-to-low from moderate-to-heavy alcohol drinkers based on fractional anisotropy derived from diffusion tensor MRIs of 764 adolescents recruited by the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA). Across the three data sets, the visualization of the smooth trajectory vector fields and superior accuracy on downstream tasks demonstrate the strength of the proposed method over existing self-supervised methods in extracting information related to brain aging, which could help study the impact of substance use and neurodegenerative disorders. The code is available at

    View details for DOI 10.1016/

    View details for PubMedID 36115098

  • Earlier Bedtime and Effective Coping Skills Predict a Return to Low-Risk of Depression in Young Adults during the COVID-19 Pandemic. International journal of environmental research and public health Zhao, Q., Wang, K., Kiss, O., Yuksel, D., de Zambotti, M., Clark, D. B., Goldston, D. B., Nooner, K. B., Brown, S. A., Tapert, S. F., Thompson, W. K., Nagel, B. J., Pfefferbaum, A., Sullivan, E. V., Pohl, K. M., Baker, F. C. 2022; 19 (16)


    To determine the persistent effects of the pandemic on mental health in young adults, we categorized depressive symptom trajectories and sought factors that promoted a reduction in depressive symptoms in high-risk individuals. Specifically, longitudinal analysis investigated changes in the risk for depression before and during the pandemic until December 2021 in 399 young adults (57% female; age range: 22.8 ± 2.6 years) in the United States (U.S.) participating in the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA) study. The Center for Epidemiologic Studies Depression Scale (CES-D-10) was administered multiple times before and during the pandemic. A score ≥10 identified individuals at high-risk for depression. Self-reported sleep behavior, substance use, and coping skills at the start of the pandemic were assessed as predictors for returning to low-risk levels while controlling for demographic factors. The analysis identified four trajectory groups regarding depression risk, with 38% being at low-risk pre-pandemic through 2021, 14% showing persistent high-risk pre-pandemic through 2021, and the remainder converting to high-risk either in June 2020 (30%) or later (18%). Of those who became high-risk in June 2020, 51% were no longer at high-risk in 2021. Logistic regression revealed that earlier bedtime and, for the older participants (mid to late twenties), better coping skills were associated with this declining risk. Results indicate divergence in trajectories of depressive symptoms, with a considerable number of young adults developing persistent depressive symptoms. Healthy sleep behavior and specific coping skills have the potential to promote remittance from depressive symptoms in the context of the pandemic.

    View details for DOI 10.3390/ijerph191610300

    View details for PubMedID 36011934

  • Detecting negative valence symptoms in adolescents based on longitudinal self-reports and behavioral assessments. Journal of affective disorders Paschali, M., Kiss, O., Zhao, Q., Adeli, E., Podhajsky, S., Muller-Oehring, E. M., Gotlib, I. H., Pohl, K. M., Baker, F. C. 2022


    BACKGROUND: Given the high prevalence of depressive symptoms reported by adolescents and associated risk of experiencing psychiatric disorders as adults, differentiating the trajectories of the symptoms related to negative valence at an individual level could be crucial in gaining a better understanding of their effects later in life.METHODS: A longitudinal deep learning framework is presented, identifying self-reported and behavioral measurements that detect the depressive symptoms associated with the Negative Valence System domain of the NIMH Research Domain Criteria (RDoC).RESULTS: Applied to the annual records of 621 participants (age range: 12 to 17 years) of the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA), the deep learning framework identifies predictors of negative valence symptoms, which include lower extraversion, poorer sleep quality, impaired executive control function and factors related to substance use.LIMITATIONS: The results rely mainly on self-reported measures and do not provide information about the underlying neural correlates. Also, a larger sample is required to understand the role of sex and other demographics related to the risk of experiencing symptoms of negative valence.CONCLUSIONS: These results provide new information about predictors of negative valence symptoms in individuals during adolescence that could be critical in understanding the development of depression and identifying targets for intervention. Importantly, findings can inform preventive and treatment approaches for depression in adolescents, focusing on a unique predictor set of modifiable modulators to include factors such as sleep hygiene training, cognitive-emotional therapy enhancing coping and controllability experience and/or substance use interventions.

    View details for DOI 10.1016/j.jad.2022.06.002

    View details for PubMedID 35688394

  • Did the acute impact of the COVID-19 pandemic on drinking or nicotine use persist? Evidence from a cohort of emerging adults followed for up to nine years. Addictive behaviors Pelham, W. E., Yuksel, D., Tapert, S. F., Baker, F. C., Pohl, K. M., Thompson, W. K., Podhajsky, S., Reuter, C., Zhao, Q., Eberson-Shumate, S. C., Clark, D. B., Goldston, D. B., Nooner, K. B., Brown, S. A. 2022; 131: 107313


    This study examined the impact of the COVID-19 pandemic on drinking and nicotine use through June of 2021 in a community-based sample of young adults.Data were from 348 individuals (49% female) enrolled in a long-term longitudinal study with an accelerated longitudinal design: the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA) Study. Individuals completed pre-pandemic assessments biannually from 2016 to early 2020, then completed up to three web-based, during-pandemic surveys in June 2020, December 2020, and June 2021. Assessments when individuals were 18.8-22.4 years old (N = 1,458) were used to compare drinking and nicotine use pre-pandemic vs. at each of the three during-pandemic timepoints, adjusting for the age-related increases expected over time.Compared to pre-pandemic, participants were less likely to report past-month drinking in June or December 2020, but there was an increase in drinking days among drinkers in June 2020. By June 2021, both the prevalence of past-month drinking and number of drinking days among drinks were similar to pre-pandemic levels. On average, there were no statistically significant differences between pre-pandemic and during-pandemic time points for binge drinking, typical drinking quantity, or nicotine use. Young adults who reported an adverse financial impact of the pandemic showed increased nicotine use while their peers showed stable or decreased nicotine use.Initial effects of the pandemic on alcohol use faded by June 2021, and on average there was little effect of the pandemic on nicotine use.

    View details for DOI 10.1016/j.addbeh.2022.107313

    View details for PubMedID 35413486

  • Systemic Administration of the TLR7/8 Agonist Resiquimod (R848) to Mice Is Associated with Transient, In Vivo-Detectable Brain Swelling. Biology Zahr, N. M., Zhao, Q., Goodcase, R., Pfefferbaum, A. 2022; 11 (2)


    Peripheral administration of the E. coli endotoxin lipopolysaccharide (LPS) to rats promotes secretion of pro-inflammatory cytokines and in previous studies was associated with transient enlargement of cortical volumes. Here, resiquimod (R848) was administered to mice to stimulate peripheral immune activation, and the effects on brain volumes and neurometabolites determined. After baseline scans, 24 male, wild-type C57BL mice were triaged into three groups including R848 at low (50 mug) and high (100 mug) doses and saline controls. Animals were scanned again at 3 h and 24 h following treatment. Sickness indices of elevated temperature and body weight loss were observed in all R848 animals. Animals that received 50 mug R848 exhibited decreases in hippocampal N-acetylaspartate and phosphocreatine at the 3 h time point that returned to baseline levels at 24 h. Animals that received the 100 mug R848 dose demonstrated transient, localized, volume expansion (~5%) detectable at 3 h in motor, somatosensory, and olfactory cortices; and pons. A metabolic response evident at the lower dose and a volumetric change at the higher dose suggests a temporal evolution of the effect wherein the neurochemical change is demonstrable earlier than neurostructural change. Transient volume expansion in response to peripheral immune stimulation corresponds with previous results and is consistent with brain swelling that may reflect CNS edema.

    View details for DOI 10.3390/biology11020274

    View details for PubMedID 35205140

  • A Penalty Approach for Normalizing Feature Distributions to Build Confounder-Free Models Vento, A., Zhao, Q., Paul, R., Pohl, K. M., Adeli, E., Wang, L., Dou, Q., Fletcher, P. T., Speidel, S., Li, S. SPRINGER INTERNATIONAL PUBLISHING AG. 2022: 387-397
  • Thin Shell Demons for Dental Scan Registration Sahu, P., Gerber, S., Zhao, Q., Tung Nguyen, McCormick, M., Paniagua, B., Vicory, J., Colliot, O., Isgum, Landman, B. A., Loew, M. H. SPIE-INT SOC OPTICAL ENGINEERING. 2022

    View details for DOI 10.1117/12.2611514

    View details for Web of Science ID 000836295600095

  • Joint Graph Convolution for Analyzing Brain Structural and Functional Connectome Li, Y., Wei, Q., Adeli, E., Pohl, K. M., Zhao, Q., Wang, L., Dou, Q., Fletcher, P. T., Speidel, S., Li, S. SPRINGER INTERNATIONAL PUBLISHING AG. 2022: 231-240


    The white-matter (micro-)structural architecture of the brain promotes synchrony among neuronal populations, giving rise to richly patterned functional connections. A fundamental problem for systems neuroscience is determining the best way to relate structural and functional networks quantified by diffusion tensor imaging and resting-state functional MRI. As one of the state-of-the-art approaches for network analysis, graph convolutional networks (GCN) have been separately used to analyze functional and structural networks, but have not been applied to explore inter-network relationships. In this work, we propose to couple the two networks of an individual by adding inter-network edges between corresponding brain regions, so that the joint structure-function graph can be directly analyzed by a single GCN. The weights of inter-network edges are learnable, reflecting non-uniform structure-function coupling strength across the brain. We apply our Joint-GCN to predict age and sex of 662 participants from the public dataset of the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA) based on their functional and micro-structural white-matter networks. Our results support that the proposed Joint-GCN outperforms existing multi-modal graph learning approaches for analyzing structural and functional networks.

    View details for DOI 10.1007/978-3-031-16431-6_22

    View details for Web of Science ID 000867524300022

    View details for PubMedID 36321855

    View details for PubMedCentralID PMC9620868

  • Bridging the Gap Between Deep Learning and Hypothesis-Driven Analysis via Permutation Testing Paschali, M., Zhao, Q., Adeli, E., Pohl, K. M., Rekik, Adeli, E., Park, S. H., Cintas, C. SPRINGER INTERNATIONAL PUBLISHING AG. 2022: 13-23
  • Multiple Instance Neuroimage Transformer Singla, A., Zhao, Q., Do, D. K., Zhou, Y., Pohl, K. M., Adeli, E., Rekik, Adeli, E., Park, S. H., Cintas, C. SPRINGER INTERNATIONAL PUBLISHING AG. 2022: 36-48
  • Risk for depression tripled during the COVID-19 pandemic in emerging adults followed for the last 8 years. Psychological medicine Alzueta, E., Podhajsky, S., Zhao, Q., Tapert, S. F., Thompson, W. K., de Zambotti, M., Yuksel, D., Kiss, O., Wang, R., Volpe, L., Prouty, D., Colrain, I. M., Clark, D. B., Goldston, D. B., Nooner, K. B., De Bellis, M. D., Brown, S. A., Nagel, B. J., Pfefferbaum, A., Sullivan, E. V., Baker, F. C., Pohl, K. M. 2021: 1-8


    BACKGROUND: The coronavirus disease 2019 (COVID-19) pandemic has significantly increased depression rates, particularly in emerging adults. The aim of this study was to examine longitudinal changes in depression risk before and during COVID-19 in a cohort of emerging adults in the U.S. and to determine whether prior drinking or sleep habits could predict the severity of depressive symptoms during the pandemic.METHODS: Participants were 525 emerging adults from the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA), a five-site community sample including moderate-to-heavy drinkers. Poisson mixed-effect models evaluated changes in the Center for Epidemiological Studies Depression Scale (CES-D-10) from before to during COVID-19, also testing for sex and age interactions. Additional analyses examined whether alcohol use frequency or sleep duration measured in the last pre-COVID assessment predicted pandemic-related increase in depressive symptoms.RESULTS: The prevalence of risk for clinical depression tripled due to a substantial and sustained increase in depressive symptoms during COVID-19 relative to pre-COVID years. Effects were strongest for younger women. Frequent alcohol use and short sleep duration during the closest pre-COVID visit predicted a greater increase in COVID-19 depressive symptoms.CONCLUSIONS: The sharp increase in depression risk among emerging adults heralds a public health crisis with alarming implications for their social and emotional functioning as this generation matures. In addition to the heightened risk for younger women, the role of alcohol use and sleep behavior should be tracked through preventive care aiming to mitigate this looming mental health crisis.

    View details for DOI 10.1017/S0033291721004062

    View details for PubMedID 34726149

  • Multi-label, multi-domain learning identifies compounding effects of HIV and cognitive impairment. Medical image analysis Zhang, J., Zhao, Q., Adeli, E., Pfefferbaum, A., Sullivan, E. V., Paul, R., Valcour, V., Pohl, K. M. 2021; 75: 102246


    Older individuals infected by Human Immunodeficiency Virus (HIV) are at risk for developing HIV-Associated Neurocognitive Disorder (HAND), i.e., from reduced cognitive functioning similar to HIV-negative individuals with Mild Cognitive Impairment (MCI) or to Alzheimer's Disease (AD) if more severely affected. Incompletely understood is how brain structure can serve to differentiate cognitive impairment (CI) in the HIV-positive (i.e., HAND) from the HIV-negative cohort (i.e., MCI and AD). To that end, we designed a multi-label classifier that labels the structural magnetic resonance images (MRI) of individuals by their HIV and CI status via two binary variables. Proper training of such an approach traditionally requires well-curated datasets containing large number of samples for each of the corresponding four cohorts (healthy controls, CI HIV-negative adults a.k.a. CI-only, HIV-positive patients without CI a.k.a. HIV-only, and HAND). Because of the rarity of such datasets, we proposed to improve training of the multi-label classifier via a multi-domain learning scheme that also incorporates domain-specific classifiers on auxiliary single-label datasets specific to either binary label. Specifically, we complement the training dataset of MRIs of the four cohorts (Control: 156, CI-only: 335, HIV-only: 37, HAND: 145) acquired by the Memory and Aging Center at the University of California - San Francisco with a CI-specific dataset only containing MRIs of HIV-negative subjects (Controls: 229, CI-only: 397) from the Alzheimer's Disease Neuroimaging Initiative and an HIV-specific dataset (Controls: 75, HIV-only: 75) provided by SRI International. Based on cross-validation on the UCSF dataset, the multi-domain and multi-label learning strategy leads to superior classification accuracy compared with one-domain or multi-class learning approaches, specifically for the undersampled HIV-only cohort. The 'prediction logits' of CI computed by the multi-label formulation also successfully stratify motor performance among the HIV-positive subjects (including HAND). Finally, brain patterns driving the subject-level predictions across all four cohorts characterize the independent and compounding effects of HIV and CI in the HAND cohort.

    View details for DOI 10.1016/

    View details for PubMedID 34706304

  • Adversarial Bayesian Optimization for Quantifying Motion Artifact Within MRI. PRedictive Intelligence in MEdicine. PRIME (Workshop) Butskova, A., Juhl, R., Zukic, D., Chaudhary, A., Pohl, K. M., Zhao, Q. 2021; 12928: 83-92


    Subject motion during an MRI sequence can cause ghosting effects or diffuse image noise in the phase-encoding direction and hence is likely to bias findings in neuroimaging studies. Detecting motion artifacts often relies on experts visually inspecting MRIs, which is subjective and expensive. To improve this detection, we develop a framework to automatically quantify the severity of motion artifact within a brain MRI. We formulate this task as a regression problem and train the regressor from a data set of MRIs with various amounts of motion artifacts. To resolve the issue of missing fine-grained ground-truth labels (level of artifacts), we propose Adversarial Bayesian Optimization (ABO) to infer the distribution of motion parameters (i.e., rotation and translation) underlying the acquired MRI data and then inject synthetic motion artifacts sampled from that estimated distribution into motion-free MRIs. After training the regressor on the synthetic data, we applied the model to quantify the motion level in 990 MRIs collected by the National Consortium on Alcohol and Neurodevelopment in Adolescence. Results show that the motion level derived by our approach is more reliable than the traditional metric based on Entropy Focus Criterion and manually defined binary labels.

    View details for DOI 10.1007/978-3-030-87602-9_8

    View details for PubMedID 35749100

  • Longitudinal Correlation Analysis for Decoding Multi-modal Brain Development. Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention Zhao, Q., Adeli, E., Pohl, K. M. 2021; 12907: 400-409


    Starting from childhood, the human brain restructures and rewires throughout life. Characterizing such complex brain development requires effective analysis of longitudinal and multi-modal neuroimaging data. Here, we propose such an analysis approach named Longitudinal Correlation Analysis (LCA). LCA couples the data of two modalities by first reducing the input from each modality to a latent representation based on autoencoders. A self-supervised strategy then relates the two latent spaces by jointly disentangling two directions, one in each space, such that the longitudinal changes in latent representations along those directions are maximally correlated between modalities. We applied LCA to analyze the longitudinal T1-weighted and diffusion-weighted MRIs of 679 youths from the National Consortium on Alcohol and Neurodevelopment in Adolescence. Unlike existing approaches that focus on either cross-sectional or single-modal modeling, LCA successfully unraveled coupled macrostructural and microstructural brain development from morphological and diffusivity features extracted from the data. A retesting of LCA on raw 3D image volumes of those subjects successfully replicated the findings from the feature-based analysis. Lastly, the developmental effects revealed by LCA were inline with the current understanding of maturational patterns of the adolescent brain.

    View details for DOI 10.1007/978-3-030-87234-2_38

    View details for PubMedID 35253021

    View details for PubMedCentralID PMC8896397

  • Self-Supervised Longitudinal Neighbourhood Embedding. Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention Ouyang, J., Zhao, Q., Adeli, E., Sullivan, E. V., Pfefferbaum, A., Zaharchuk, G., Pohl, K. M. 2021; 12902: 80-89


    Longitudinal MRIs are often used to capture the gradual deterioration of brain structure and function caused by aging or neurological diseases. Analyzing this data via machine learning generally requires a large number of ground-truth labels, which are often missing or expensive to obtain. Reducing the need for labels, we propose a self-supervised strategy for representation learning named Longitudinal Neighborhood Embedding (LNE). Motivated by concepts in contrastive learning, LNE explicitly models the similarity between trajectory vectors across different subjects. We do so by building a graph in each training iteration defining neighborhoods in the latent space so that the progression direction of a subject follows the direction of its neighbors. This results in a smooth trajectory field that captures the global morphological change of the brain while maintaining the local continuity. We apply LNE to longitudinal T1w MRIs of two neuroimaging studies: a dataset composed of 274 healthy subjects, and Alzheimer's Disease Neuroimaging Initiative (ADNI, N = 632). The visualization of the smooth trajectory vector field and superior performance on downstream tasks demonstrate the strength of the proposed method over existing self-supervised methods in extracting information associated with normal aging and in revealing the impact of neurodegenerative disorders. The code is available at

    View details for DOI 10.1007/978-3-030-87196-3_8

    View details for PubMedID 35727732

    View details for PubMedCentralID PMC9204645

  • Representation Disentanglement for Multi-modal Brain MRI Analysis. Information processing in medical imaging : proceedings of the ... conference Ouyang, J., Adeli, E., Pohl, K. M., Zhao, Q., Zaharchuk, G. 2021; 12729: 321-333


    Multi-modal MRIs are widely used in neuroimaging applications since different MR sequences provide complementary information about brain structures. Recent works have suggested that multi-modal deep learning analysis can benefit from explicitly disentangling anatomical (shape) and modality (appearance) information into separate image presentations. In this work, we challenge mainstream strategies by showing that they do not naturally lead to representation disentanglement both in theory and in practice. To address this issue, we propose a margin loss that regularizes the similarity in relationships of the representations across subjects and modalities. To enable robust training, we further use a conditional convolution to design a single model for encoding images of all modalities. Lastly, we propose a fusion function to combine the disentangled anatomical representations as a set of modality-invariant features for downstream tasks. We evaluate the proposed method on three multi-modal neuroimaging datasets. Experiments show that our proposed method can achieve superior disentangled representations compared to existing disentanglement strategies. Results also indicate that the fused anatomical representation has potential in the downstream task of zero-dose PET reconstruction and brain tumor segmentation.

    View details for DOI 10.1007/978-3-030-78191-0_25

    View details for PubMedID 35173402

  • Longitudinal Pooling & Consistency Regularization to Model Disease Progression From MRIs. IEEE journal of biomedical and health informatics Ouyang, J., Zhao, Q., Sullivan, E. V., Pfefferbaum, A., Tapert, S. F., Adeli, E., Pohl, K. M. 2021; 25 (6): 2082-2092


    Many neurological diseases are characterized by gradual deterioration of brain structure andfunction. Large longitudinal MRI datasets have revealed such deterioration, in part, by applying machine and deep learning to predict diagnosis. A popular approach is to apply Convolutional Neural Networks (CNN) to extract informative features from each visit of the longitudinal MRI and then use those features to classify each visit via Recurrent Neural Networks (RNNs). Such modeling neglects the progressive nature of the disease, which may result in clinically implausible classifications across visits. To avoid this issue, we propose to combine features across visits by coupling feature extraction with a novel longitudinal pooling layer and enforce consistency of the classification across visits in line with disease progression. We evaluate the proposed method on the longitudinal structural MRIs from three neuroimaging datasets: Alzheimer's Disease Neuroimaging Initiative (ADNI, N=404), a dataset composed of 274 normal controls and 329 patients with Alcohol Use Disorder (AUD), and 255 youths from the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA). In allthree experiments our method is superior to other widely used approaches for longitudinal classification thus making a unique contribution towards more accurate tracking of the impact of conditions on the brain. The code is available at

    View details for DOI 10.1109/JBHI.2020.3042447

    View details for PubMedID 33270567

  • Going Beyond Saliency Maps: Training Deep Models to Interpret Deep Models. Information processing in medical imaging : proceedings of the ... conference Liu, Z., Adeli, E., Pohl, K. M., Zhao, Q. 2021; 12729: 71-82


    Interpretability is a critical factor in applying complex deep learning models to advance the understanding of brain disorders in neuroimaging studies. To interpret the decision process of a trained classifier, existing techniques typically rely on saliency maps to quantify the voxel-wise or feature-level importance for classification through partial derivatives. Despite providing some level of localization, these maps are not human-understandable from the neuroscience perspective as they often do not inform the specific type of morphological changes linked to the brain disorder. Inspired by the image-to-image translation scheme, we propose to train simulator networks to inject (or remove) patterns of the disease into a given MRI based on a warping operation, such that the classifier increases (or decreases) its confidence in labeling the simulated MRI as diseased. To increase the robustness of training, we propose to couple the two simulators into a unified model based on conditional convolution. We applied our approach to interpreting classifiers trained on a synthetic dataset and two neuroimaging datasets to visualize the effect of Alzheimer's disease and alcohol dependence. Compared to the saliency maps generated by baseline approaches, our simulations and visualizations based on the Jacobian determinants of the warping field reveal meaningful and understandable patterns related to the diseases.

    View details for DOI 10.1007/978-3-030-78191-0_6

    View details for PubMedID 34548772

  • Longitudinal self-supervised learning. Medical image analysis Zhao, Q., Liu, Z., Adeli, E., Pohl, K. M. 2021; 71: 102051


    Machine learning analysis of longitudinal neuroimaging data is typically based on supervised learning, which requires large number of ground-truth labels to be informative. As ground-truth labels are often missing or expensive to obtain in neuroscience, we avoid them in our analysis by combing factor disentanglement with self-supervised learning to identify changes and consistencies across the multiple MRIs acquired of each individual over time. Specifically, we propose a new definition of disentanglement by formulating a multivariate mapping between factors (e.g., brain age) associated with an MRI and a latent image representation. Then, factors that evolve across acquisitions of longitudinal sequences are disentangled from that mapping by self-supervised learning in such a way that changes in a single factor induce change along one direction in the representation space. We implement this model, named Longitudinal Self-Supervised Learning (LSSL), via a standard autoencoding structure with a cosine loss to disentangle brain age from the image representation. We apply LSSL to two longitudinal neuroimaging studies to highlight its strength in extracting the brain-age information from MRI and revealing informative characteristics associated with neurodegenerative and neuropsychological disorders. Moreover, the representations learned by LSSL facilitate supervised classification by recording faster convergence and higher (or similar) prediction accuracy compared to several other representation learning techniques.

    View details for DOI 10.1016/

    View details for PubMedID 33882336

  • Attenuated cerebral blood flow in frontolimbic and insular cortices in Alcohol Use Disorder: Relation to working memory. Journal of psychiatric research Sullivan, E. V., Zhao, Q. n., Pohl, K. M., Zahr, N. M., Pfefferbaum, A. n. 2021; 136: 140–48


    Chronic, excessive alcohol consumption is associated with cerebrovascular hypoperfusion, which has the potential to interfere with cognitive processes. Magnetic resonance pulsed continuous arterial spin labeling (PCASL) provides a noninvasive approach for measuring regional cerebral blood flow (CBF) and was used to study 24 men and women with Alcohol Use Disorder (AUD) and 20 age- and sex-matched controls. Two analysis approaches tested group differences: a data-driven, regionally-free method to test for group differences on a voxel-by-voxel basis and a region of interest (ROI) approach, which focused quantification on atlas-determined brain structures. Whole-brain, voxel-wise quantification identified low AUD-related cerebral perfusion in large volumes of medial frontal and cingulate cortices. The ROI analysis also identified lower CBF in the AUD group relative to the control group in medial frontal, anterior/middle cingulate, insular, and hippocampal/amygdala ROIs. Further, years of AUD diagnosis negatively correlated with temporal cortical CBF, and scores on an alcohol withdrawal scale negatively correlated with posterior cingulate and occipital gray matter CBF. Regional volume deficits did not account for AUD CBF deficits. Functional relevance of attenuated regional CBF in the AUD group emerged with positive correlations between episodic working memory test scores and anterior/middle cingulum, insula, and thalamus CBF. The frontolimbic and insular cortical neuroconstellation with dampened perfusion suggests a mechanism of dysfunction associated with these brain regions in AUD.

    View details for DOI 10.1016/j.jpsychires.2021.01.053

    View details for PubMedID 33592385

  • Longitudinal Correlation Analysis for Decoding Multi-modal Brain Development Zhao, Q., Adeli, E., Pohl, K. M., DeBruijne, M., Cattin, P. C., Cotin, S., Padoy, N., Speidel, S., Zheng, Y., Essert, C. SPRINGER INTERNATIONAL PUBLISHING AG. 2021: 400-409
  • Metadata Normalization. Proceedings. IEEE Computer Society Conference on Computer Vision and Pattern Recognition Lu, M., Zhao, Q., Zhang, J., Pohl, K. M., Fei-Fei, L., Niebles, J. C., Adeli, E. 2021; 2021: 10912-10922


    Batch Normalization (BN) and its variants have delivered tremendous success in combating the covariate shift induced by the training step of deep learning methods. While these techniques normalize feature distributions by standardizing with batch statistics, they do not correct the influence on features from extraneous variables or multiple distributions. Such extra variables, referred to as metadata here, may create bias or confounding effects (e.g., race when classifying gender from face images). We introduce the Metadata Normalization (MDN) layer, a new batch-level operation which can be used end-to-end within the training framework, to correct the influence of metadata on feature distributions. MDN adopts a regression analysis technique traditionally used for preprocessing to remove (regress out) the metadata effects on model features during training. We utilize a metric based on distance correlation to quantify the distribution bias from the metadata and demonstrate that our method successfully removes metadata effects on four diverse settings: one synthetic, one 2D image, one video, and one 3D medical image dataset.

    View details for DOI 10.1109/cvpr46437.2021.01077

    View details for PubMedID 34776724

    View details for PubMedCentralID PMC8589298

  • Self-supervised Longitudinal Neighbourhood Embedding Ouyang, J., Zhao, Q., Adeli, E., Sullivan, E., Pfefferbaum, A., Zaharchuk, G., Pohl, K. M., DeBruijne, M., Cattin, P. C., Cotin, S., Padoy, N., Speidel, S., Zheng, Y., Essert, C. SPRINGER INTERNATIONAL PUBLISHING AG. 2021: 80-89
  • Representation Learning with Statistical Independence to Mitigate Bias. IEEE Winter Conference on Applications of Computer Vision. IEEE Winter Conference on Applications of Computer Vision Adeli, E., Zhao, Q., Pfefferbaum, A., Sullivan, E. V., Fei-Fei, L., Niebles, J. C., Pohl, K. M. 2021; 2021: 2512-2522


    Presence of bias (in datasets or tasks) is inarguably one of the most critical challenges in machine learning applications that has alluded to pivotal debates in recent years. Such challenges range from spurious associations between variables in medical studies to the bias of race in gender or face recognition systems. Controlling for all types of biases in the dataset curation stage is cumbersome and sometimes impossible. The alternative is to use the available data and build models incorporating fair representation learning. In this paper, we propose such a model based on adversarial training with two competing objectives to learn features that have (1) maximum discriminative power with respect to the task and (2) minimal statistical mean dependence with the protected (bias) variable(s). Our approach does so by incorporating a new adversarial loss function that encourages a vanished correlation between the bias and the learned features. We apply our method to synthetic data, medical images (containing task bias), and a dataset for gender classification (containing dataset bias). Our results show that the learned features by our method not only result in superior prediction performance but also are unbiased.

    View details for DOI 10.1109/wacv48630.2021.00256

    View details for PubMedID 34522832

  • Quantifying Parkinson's disease motor severity under uncertainty using MDS-UPDRS videos. Medical image analysis Lu, M., Zhao, Q., Poston, K. L., Sullivan, E. V., Pfefferbaum, A., Shahid, M., Katz, M., Kouhsari, L. M., Schulman, K., Milstein, A., Niebles, J. C., Henderson, V. W., Fei-Fei, L., Pohl, K. M., Adeli, E. 2021; 73: 102179


    Parkinson's disease (PD) is a brain disorder that primarily affects motor function, leading to slow movement, tremor, and stiffness, as well as postural instability and difficulty with walking/balance. The severity of PD motor impairments is clinically assessed by part III of the Movement Disorder Society Unified Parkinson's Disease Rating Scale (MDS-UPDRS), a universally-accepted rating scale. However, experts often disagree on the exact scoring of individuals. In the presence of label noise, training a machine learning model using only scores from a single rater may introduce bias, while training models with multiple noisy ratings is a challenging task due to the inter-rater variabilities. In this paper, we introduce an ordinal focal neural network to estimate the MDS-UPDRS scores from input videos, to leverage the ordinal nature of MDS-UPDRS scores and combat class imbalance. To handle multiple noisy labels per exam, the training of the network is regularized via rater confusion estimation (RCE), which encodes the rating habits and skills of raters via a confusion matrix. We apply our pipeline to estimate MDS-UPDRS test scores from their video recordings including gait (with multiple Raters, R=3) and finger tapping scores (single rater). On a sizable clinical dataset for the gait test (N=55), we obtained a classification accuracy of 72% with majority vote as ground-truth, and an accuracy of ∼84% of our model predicting at least one of the raters' scores. Our work demonstrates how computer-assisted technologies can be used to track patients and their motor impairments, even when there is uncertainty in the clinical ratings. The latest version of the code will be available at

    View details for DOI 10.1016/

    View details for PubMedID 34340101

  • Association of Heavy Drinking With Deviant Fiber Tract Development in Frontal Brain Systems in Adolescents. JAMA psychiatry Zhao, Q., Sullivan, E. V., Honnorat, N., Adeli, E., Podhajsky, S., De Bellis, M. D., Voyvodic, J., Nooner, K. B., Baker, F. C., Colrain, I. M., Tapert, S. F., Brown, S. A., Thompson, W. K., Nagel, B. J., Clark, D. B., Pfefferbaum, A., Pohl, K. M. 2020


    Importance: Maturation of white matter fiber systems subserves cognitive, behavioral, emotional, and motor development during adolescence. Hazardous drinking during this active neurodevelopmental period may alter the trajectory of white matter microstructural development, potentially increasing risk for developing alcohol-related dysfunction and alcohol use disorder in adulthood.Objective: To identify disrupted adolescent microstructural brain development linked to drinking onset and to assess whether the disruption is more pronounced in younger rather than older adolescents.Design, Setting, and Participants: This case-control study, conducted from January 13, 2013, to January 15, 2019, consisted of an analysis of 451 participants from the National Consortium on Alcohol and Neurodevelopment in Adolescence cohort. Participants were aged 12 to 21 years at baseline and had at least 2 usable magnetic resonance diffusion tensor imaging (DTI) scans and up to 5 examination visits spanning 4 years. Participants with a youth-adjusted Cahalan score of 0 were labeled as no-to-low drinkers; those with a score of greater than 1 for at least 2 consecutive visits were labeled as heavy drinkers. Exploratory analysis was conducted between no-to-low and heavy drinkers. A between-group analysis was conducted between age- and sex-matched youths, and a within-participant analysis was performed before and after drinking.Exposures: Self-reported alcohol consumption in the past year summarized by categorical drinking levels.Main Outcomes and Measures: Diffusion tensor imaging measurement of fractional anisotropy (FA) in the whole brain and fiber systems quantifying the developmental change of each participant as a slope.Results: Analysis of whole-brain FA of 451 adolescents included 291 (64.5%) no-to-low drinkers and 160 (35.5%) heavy drinkers who indicated the potential for a deleterious association of alcohol with microstructural development. Among the no-to-low drinkers, 142 (48.4%) were boys with mean (SD) age of 16.5 (2.2) years and 149 (51.2%) were girls with mean (SD) age of 16.5 (2.1) years and 192 (66.0%) were White participants. Among the heavy drinkers, 86 (53.8%) were boys with mean (SD) age of 20.1 (1.5) years and 74 (46.3%) were girls with mean (SD) age of 20.5 (2.0) years and 142 (88.8%) were White participants. A group analysis revealed FA reduction in heavy-drinking youth compared with age- and sex-matched controls (t154=-2.7, P=.008). The slope of this reduction correlated with log of days of drinking since the baseline visit (r156=-0.21, 2-tailed P=.008). A within-participant analysis contrasting developmental trajectories of youths before and after they initiated heavy drinking supported the prediction that drinking onset was associated with and potentially preceded disrupted white matter integrity. Age-alcohol interactions (t152=3.0, P=.004) observed for the FA slopes indicated that the alcohol-associated disruption was greater in younger than older adolescents and was most pronounced in the genu and body of the corpus callosum, regions known to continue developing throughout adolescence.Conclusions and Relevance: This case-control study of adolescents found a deleterious association of alcohol use with white matter microstructural integrity. These findings support the concept of heightened vulnerability to environmental agents, including alcohol, associated with attenuated development of major white matter tracts in early adolescence.

    View details for DOI 10.1001/jamapsychiatry.2020.4064

    View details for PubMedID 33377940

  • Inpainting Cropped Diffusion MRI using Deep Generative Models. PRedictive Intelligence in MEdicine. PRIME (Workshop) Ayub, R., Zhao, Q., Meloy, M. J., Sullivan, E. V., Pfefferbaum, A., Adeli, E., Pohl, K. M. 2020; 12329: 91-100


    Minor artifacts introduced during image acquisition are often negligible to the human eye, such as a confined field of view resulting in MRI missing the top of the head. This cropping artifact, however, can cause suboptimal processing of the MRI resulting in data omission or decreasing the power of subsequent analyses. We propose to avoid data or quality loss by restoring these missing regions of the head via variational autoencoders (VAE), a deep generative model that has been previously applied to high resolution image reconstruction. Based on diffusion weighted images (DWI) acquired by the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA), we evaluate the accuracy of inpainting the top of the head by common autoencoder models (U-Net, VQVAE, and VAE-GAN) and a custom model proposed herein called U-VQVAE. Our results show that U-VQVAE not only achieved the highest accuracy, but also resulted in MRI processing producing lower fractional anisotropy (FA) in the supplementary motor area than FA derived from the original MRIs. Lower FA implies that inpainting reduces noise in processing DWI and thus increase the quality of the generated results. The code is available at

    View details for DOI 10.1007/978-3-030-59354-4_9

    View details for PubMedID 33997866

  • Spatio-Temporal Graph Convolution for Resting-State fMRI Analysis. Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention Gadgil, S., Zhao, Q., Pfefferbaum, A., Sullivan, E. V., Adeli, E., Pohl, K. M. 2020; 12267: 528–38


    The Blood-Oxygen-Level-Dependent (BOLD) signal of resting-state fMRI (rs-fMRI) records the temporal dynamics of intrinsic functional networks in the brain. However, existing deep learning methods applied to rs-fMRI either neglect the functional dependency between different brain regions in a network or discard the information in the temporal dynamics of brain activity. To overcome those shortcomings, we propose to formulate functional connectivity networks within the context of spatio-temporal graphs. We train a spatio-temporal graph convolutional network (ST-GCN) on short sub-sequences of the BOLD time series to model the non-stationary nature of functional connectivity. Simultaneously, the model learns the importance of graph edges within ST-GCN to gain insight into the functional connectivities contributing to the prediction. In analyzing the rs-fMRI of the Human Connectome Project (HCP, N = 1,091) and the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA, N = 773), ST-GCN is significantly more accurate than common approaches in predicting gender and age based on BOLD signals. Furthermore, the brain regions and functional connections significantly contributing to the predictions of our model are important markers according to the neuroscience literature.

    View details for DOI 10.1007/978-3-030-59728-3_52

    View details for PubMedID 33257918

  • Deep Learning Identifies Morphological Determinants of Sex Differences in the Pre-Adolescent Brain. NeuroImage Adeli, E., Zhao, Q., Zahr, N. M., Goldstone, A., Pfefferbaum, A., Sullivan, E. V., Pohl, K. M. 2020: 117293


    The application of data-driven deep learning to identify sex differences in developing brain structures of pre-adolescents has heretofore not been accomplished. Here, the approach identifies sex differences by analyzing the minimally processed MRIs of the first 8,144 participants (age 9 and 10 years) recruited by the Adolescent Brain Cognitive Development (ABCD) study. The identified pattern accounted for confounding factors (i.e., head size, age, puberty development, socioeconomic status) and comprised cerebellar (corpus medullare, lobules III, IV/V, and VI) and subcortical (pallidum, amygdala, hippocampus, parahippocampus, insula, putamen) structures. While these have been individually linked to expressing sex differences, a novel discovery was that their grouping accurately predicted the sex in individual pre-adolescents. Another novelty was relating differences specific to the cerebellum to pubertal development. Finally, we found that reducing the pattern to a single score not only accurately predicted sex but also correlated with cognitive behavior linked to working memory. The predictive power of this score and the constellation of identified brain structures provide evidence for sex differences in pre-adolescent neurodevelopment and may augment understanding of sex-specific vulnerability or resilience to psychiatric disorders and presage sex-linked learning disabilities.

    View details for DOI 10.1016/j.neuroimage.2020.117293

    View details for PubMedID 32841716

  • Accelerated aging and motor control deficits are related to regional deformation of central cerebellar white matter in alcohol use disorder ADDICTION BIOLOGY Zhao, Q., Pfefferbaum, A., Podhajsky, S., Pohl, K. M., Sullivan, E. V. 2020; 25 (3)

    View details for DOI 10.1111/adb.12746

    View details for Web of Science ID 000528674100024

  • Training confounder-free deep learning models for medical applications. Nature communications Zhao, Q. n., Adeli, E. n., Pohl, K. M. 2020; 11 (1): 6010


    The presence of confounding effects (or biases) is one of the most critical challenges in using deep learning to advance discovery in medical imaging studies. Confounders affect the relationship between input data (e.g., brain MRIs) and output variables (e.g., diagnosis). Improper modeling of those relationships often results in spurious and biased associations. Traditional machine learning and statistical models minimize the impact of confounders by, for example, matching data sets, stratifying data, or residualizing imaging measurements. Alternative strategies are needed for state-of-the-art deep learning models that use end-to-end training to automatically extract informative features from large set of images. In this article, we introduce an end-to-end approach for deriving features invariant to confounding factors while accounting for intrinsic correlations between the confounder(s) and prediction outcome. The method does so by exploiting concepts from traditional statistical methods and recent fair machine learning schemes. We evaluate the method on predicting the diagnosis of HIV solely from Magnetic Resonance Images (MRIs), identifying morphological sex differences in adolescence from those of the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA), and determining the bone age from X-ray images of children. The results show that our method can accurately predict while reducing biases associated with confounders. The code is available at .

    View details for DOI 10.1038/s41467-020-19784-9

    View details for PubMedID 33243992

  • Structural and biochemical imaging reveals systemic LPS-induced changes in the rat brain. Journal of neuroimmunology Fritz, M. n., Klawonn, A. M., Zhao, Q. n., Sullivan, E. V., Zahr, N. M., Pfefferbaum, A. n. 2020; 348: 577367


    Despite mounting evidence for the role of inflammation in Major Depressive Disorder (MDD), in vivo preclinical investigations of inflammation-induced negative affect using whole brain imaging modalities are scarce, precluding a valid model within which to evaluate pharmacological interventions. Here we used an E. coli lipopolysaccharide (LPS)-based model of inflammation-induced depressive signs in rats to explore brain changes using multimodal neuroimaging methods. During the acute phase of the LPS response (2 h post injection), prior to the emergence of a task-quantifiable depressive phenotype, striatal glutamine levels and splenial, retrosplenial, and peri-callosal hippocampal cortex volumes were greater than at baseline. LPS-induced depressive behaviors observed at 24 h, however, occurred concurrently with lower than control levels of striatal glutamine and a reversibility of volume expansion (i.e., shrinkage of splenial, retrosplenial, and peri-callosal hippocampal cortex to baseline volumes). In both striatum and hippocampus at 24 h, mRNA expression in LPS relative to control animals demonstrated alterations in enzymes and transporters regulating glutamine homeostasis. Collectively, the observed behavioral, in vivo structural and metabolic, and mRNA expression alterations suggest a critical role for astrocytic regulation of inflammation-induced depressive behaviors.

    View details for DOI 10.1016/j.jneuroim.2020.577367

    View details for PubMedID 32866714

  • Adolescent alcohol use disrupts functional neurodevelopment in sensation seeking girls. Addiction biology Zhao, Q. n., Sullivan, E. V., Műller-Oehring, E. M., Honnorat, N. n., Adeli, E. n., Podhajsky, S. n., Baker, F. C., Colrain, I. M., Prouty, D. n., Tapert, S. F., Brown, S. A., Meloy, M. J., Brumback, T. n., Nagel, B. J., Morales, A. M., Clark, D. B., Luna, B. n., De Bellis, M. D., Voyvodic, J. T., Nooner, K. B., Pfefferbaum, A. n., Pohl, K. M. 2020: e12914


    Exogenous causes, such as alcohol use, and endogenous factors, such as temperament and sex, can modulate developmental trajectories of adolescent neurofunctional maturation. We examined how these factors affect sexual dimorphism in brain functional networks in youth drinking below diagnostic threshold for alcohol use disorder (AUD). Based on the 3-year, annually acquired, longitudinal resting-state functional magnetic resonance imaging (MRI) data of 526 adolescents (12-21 years at baseline) from the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA) cohort, developmental trajectories of 23 intrinsic functional networks (IFNs) were analyzed for (1) sexual dimorphism in 259 participants who were no-to-low drinkers throughout this period; (2) sex-alcohol interactions in two age- and sex-matched NCANDA subgroups (N = 76 each), half no-to-low, and half moderate-to-heavy drinkers; and (3) moderating effects of gender-specific alcohol dose effects and a multifactorial impulsivity measure on IFN connectivity in all NCANDA participants. Results showed that sex differences in no-to-low drinkers diminished with age in the inferior-occipital network, yet girls had weaker within-network connectivity than boys in six other networks. Effects of adolescent alcohol use were more pronounced in girls than boys in three IFNs. In particular, girls showed greater within-network connectivity in two motor networks with more alcohol consumption, and these effects were mediated by sensation-seeking only in girls. Our results implied that drinking might attenuate the naturally diminishing sexual differences by disrupting the maturation of network efficiency more severely in girls. The sex-alcohol-dose effect might explain why women are at higher risk of alcohol-related health and psychosocial consequences than men.

    View details for DOI 10.1111/adb.12914

    View details for PubMedID 32428984

  • Jacobian Mapping Reveals Converging Substrates of Disruption and Repair in Response to Ethanol Exposure and Abstinence in Two Strains of Rats. Alcoholism, clinical and experimental research Zhao, Q. n., Pohl, K. M., Sullivan, E. V., Pfefferbaum, A. n., Zahr, N. M. 2020


    In a previous study using Jacobian mapping to evaluate the morphological effects on the brain of binge (4-day) intragastric ethanol (EtOH) on wild-type Wistar rats, we reported reversible thalamic shrinkage and lateral ventricular enlargement, but persistent superior and inferior colliculi shrinkage in response to binge EtOH treatment.Herein, we used similar voxel-based comparisons of Magnetic Resonance Images collected in EtOH-exposed relative to control animals to test the hypothesis that regardless of the intoxication protocol or the rat strain, the hippocampi, thalami, and colliculi would be affected.Two experiments [binge (4-day) intragastric EtOH in Fisher 344 rats and chronic (1-month) vaporized EtOH in Wistar rats] showed similarly affected brain regions including retrosplenial and cingulate cortices, dorsal hippocampi, central and ventroposterior thalami, superior and inferior colliculi, periaqueductal gray, and corpus callosum. While most of these regions showed significant recovery, volumes of the colliculi and periaqueductal gray continued to show response to each proximal alcohol exposure but at diminished levels with repeated exposures.Given the high metabolic rate of these enduringly affected regions, the current findings suggest that EtOH per se may affect cellular respiration leading to brain volume deficits. Further, responsivity greatly diminished likely reflecting neuroadaptation to repeated alcohol exposure. In summary, this unbiased, in vivo based approach demonstrating convergent brain systems responsive to two EtOH exposure protocols in two rat strains highlights regions that warrant further investigation in both animal models of alcoholism and in humans with Alcohol Use Disorder.

    View details for DOI 10.1111/acer.14496

    View details for PubMedID 33119896

  • Confounder-Aware Visualization of ConvNets. Machine learning in medical imaging. MLMI (Workshop) Zhao, Q., Adeli, E., Pfefferbaum, A., Sullivan, E. V., Pohl, K. M. 2019; 11861: 328–36


    With recent advances in deep learning, neuroimaging studies increasingly rely on convolutional networks (ConvNets) to predict diagnosis based on MR images. To gain a better understanding of how a disease impacts the brain, the studies visualize the salience maps of the ConvNet highlighting voxels within the brain majorly contributing to the prediction. However, these salience maps are generally confounded, i.e., some salient regions are more predictive of confounding variables (such as age) than the diagnosis. To avoid such misinterpretation, we propose in this paper an approach that aims to visualize confounder-free saliency maps that only highlight voxels predictive of the diagnosis. The approach incorporates univariate statistical tests to identify confounding effects within the intermediate features learned by ConvNet. The influence from the subset of confounded features is then removed by a novel partial back-propagation procedure. We use this two-step approach to visualize confounder-free saliency maps extracted from synthetic and two real datasets. These experiments reveal the potential of our visualization in producing unbiased model-interpretation.

    View details for DOI 10.1007/978-3-030-32692-0_38

    View details for PubMedID 32549051

  • Covariance Shrinkage for Dynamic Functional Connectivity. Connectomics in neuroImaging : third International Workshop, CNI 2019, held in conjunction with MICCAI 2019, Shenzhen, China, October 13, 2019, Proceedings. CNI (Workshop) (3rd : 2019 : Shenzhen Shi, China) Honnorat, N., Adeli, E., Zhao, Q., Pfefferbaum, A., Sullivan, E. V., Pohl, K. 2019; 11848: 32–41


    The tracking of dynamic functional connectivity (dFC) states in resting-state fMRI scans aims to reveal how the brain sequentially processes stimuli and thoughts. Despite the recent advances in statistical methods, estimating the high dimensional dFC states from a small number of available time points remains a challenge. This paper shows that the challenge is reduced by linear covariance shrinkage, a statistical method used for the estimation of large covariance matrices from small number of samples. We present a computationally efficient formulation of our approach that scales dFC analysis up to full resolution resting-state fMRI scans. Experiments on synthetic data demonstrate that our approach produces dFC estimates that are closer to the ground-truth than state-of-the-art estimation approaches. When comparing methods on the rs-fMRI scans of 162 subjects, we found that our approach is better at extracting functional networks and capturing differences in rs-fMRI acquisition and diagnosis.

    View details for DOI 10.1007/978-3-030-32391-2_4

    View details for PubMedID 32924030

  • Data Augmentation Based on Substituting Regional MRIs Volume Scores. Large-Scale Annotation of Biomedical Data and Expert Label Synthesis and Hardware Aware Learning for Medical Imaging and Computer Assisted Intervention : International Workshops, LABELS 2019, HAL-MICCAI 2019, and CuRIOUS 2019, held in c... Leng, T., Zhao, Q., Yang, C., Lu, Z., Adeli, E., Pohl, K. M. 2019; 11851: 32–41


    Due to difficulties in collecting sufficient training data, recent advances in neural-network-based methods have not been fully explored in the analysis of brain Magnetic Resonance Imaging (MRI). A possible solution to the limited-data issue is to augment the training set with synthetically generated data. In this paper, we propose a data augmentation strategy based on regional feature substitution. We demonstrate the advantages of this strategy with respect to training a simple neural-network-based classifier in predicting when individual youth transition from no-to-low to medium-to-heavy alcohol drinkers solely based on their volumetric MRI measurements. Based on 20-fold cross-validation, we generate more than one million synthetic samples from less than 500 subjects for each training run. The classifier achieves an accuracy of 74.1% in correctly distinguishing non-drinkers from drinkers at baseline and a 43.2% weighted accuracy in predicting the transition over a three year period (5-group classification task). Both accuracy scores are significantly better than training the classifier on the original dataset.

    View details for DOI 10.1007/978-3-030-33642-4_4

    View details for PubMedID 32924031

  • On discrete Wirtinger-Northcott problems LINEAR ALGEBRA AND ITS APPLICATIONS Leng, T., Zhao, Q., Qin, X. 2019; 575: 141–58
  • Variational Autoencoder with Truncated Mixture of Gaussians for Functional Connectivity Analysis. Information processing in medical imaging : proceedings of the ... conference Zhao, Q., Honnorat, N., Adeli, E., Pohl, K. M. 2019; 11492: 867-879


    Resting-state functional connectivity states are often identified as clusters of dynamic connectivity patterns. However, existing clustering approaches do not distinguish major states from rarely occurring minor states and hence are sensitive to noise. To address this issue, we propose to model major states using a non-linear generative process guided by a Gaussian-mixture distribution in a low-dimensional latent space, while separately modeling the connectivity patterns of minor states by a non-informative uniform distribution. We embed this truncated Gaussian-Mixture model in a Variational Autoencoder framework to obtain a general joint clustering and outlier detection approach, tGM-VAE. When applied to synthetic data with known ground-truth, tGM-VAE is more accurate in clustering connectivity patterns than existing approaches. On the rs-fMRI of 593 healthy adolescents, tGM-VAE identifies meaningful major connectivity states. The dwell time of these states significantly correlates with age.

    View details for DOI 10.1007/978-3-030-20351-1_68

    View details for PubMedID 32699491

    View details for PubMedCentralID PMC7375028

  • Accelerated aging and motor control deficits are related to regional deformation of central cerebellar white matter in alcohol use disorder. Addiction biology Zhao, Q., Pfefferbaum, A., Podhajsky, S., Pohl, K. M., Sullivan, E. V. 2019


    The World Health Organization estimates a 12-month prevalence rate of 8+% for an alcohol use disorder (AUD) diagnosis in people age 15years and older in the United States and Europe, presenting significant health risks that have the potential of accelerating age-related functional decline. According to neuropathological studies, white matter systems of the cerebellum are vulnerable to chronic alcohol dependence. To pursue the effect of AUD on white matter structure and functions in vivo, this study used T1-weighted, magnetic resonance imaging (MRI) to quantify the total corpus medullare of the cerebellum and a finely grained analysis of its surface in 135 men and women with AUD (mean duration of abstinence, 248d) and 128 age- and sex-matched control participants; subsets of these participants completed motor testing. We identified an AUD-related volume deficit and accelerated aging in the total corpus medullare. Novel deformation-based surface morphometry revealed regional shrinkage of surfaces adjacent to lobules I-V, lobule IX, and vermian lobule X. In addition, accelerated aging was detected in the regional surface areas adjacent to lobules I-V, lobule VI, lobule VIIB, and lobules VIII, IX, and X. Sex differences were not identified for any measure. For both volume-based and surface-based analyses, poorer performance in gait and balance, manual dexterity, and grip strength were linked to greater regional white matter structural deficits. Our results suggest that local deformation of the corpus medullare has the potential of identifying structurally and functionally segregated networks affected in AUD.

    View details for PubMedID 30932270

  • PREDICTION OF TREATMENT OUTCOME FOR AUTISM FROM STRUCTURE OF THE BRAIN BASED ON SURE INDEPENDENCE SCREENING. Proceedings. IEEE International Symposium on Biomedical Imaging Zhuang, J., Dvornek, N. C., Zhao, Q., Li, X., Ventola, P., Duncan, J. S. 2019; 2019: 404-408


    Autism spectrum disorder (ASD) is a complex neurodevelopmental disorder, and behavioral treatment interventions have shown promise for young children with ASD. However, there is limited progress in understanding the effect of each type of treatment. In this project, we aim to detect structural changes in the brain after treatment and select structural features associated with treatment outcomes. The difficulty in building large databases of patients who have received specific treatments and the high dimensionality of medical image analysis problems are the challenges in this work. To select predictive features and build accurate models, we use the sure independence screening (SIS) method. SIS is a theoretically and empirically validated method for ultra-high dimensional general linear models, and it achieves both predictive accuracy and correct feature selection by iterative feature selection. Compared with step-wise feature selection methods, SIS removes multiple features in each iteration and is computationally efficient. Compared with other linear models such as elastic-net regression, support vector regression (SVR) and partial least squares regression (PSLR), SIS achieves higher accuracy. We validated the superior performance of SIS in various experiments: First, we extract brain structural features from FreeSurfer, including cortical thickness, surface area, mean curvature and cortical volume. Next, we predict different measures of treatment outcomes based on structural features. We show that SIS achieves the highest correlation between prediction and measurements in all tasks. Furthermore, we report regions selected by SIS as biomarkers for ASD.

    View details for DOI 10.1109/ISBI.2019.8759156

    View details for PubMedID 32256966

    View details for PubMedCentralID PMC7119202

  • Longitudinally consistent estimates of intrinsic functional networks. Human brain mapping Zhao, Q., Kwon, D., Müller-Oehring, E. M., Le Berre, A. P., Pfefferbaum, A., Sullivan, E. V., Pohl, K. M. 2019


    Increasing numbers of neuroimaging studies are acquiring data to examine changes in brain architecture by investigating intrinsic functional networks (IFN) from longitudinal resting-state functional MRI (rs-fMRI). At the subject level, these IFNs are determined by cross-sectional procedures, which neglect intra-subject dependencies and result in suboptimal estimates of the networks. Here, a novel longitudinal approach simultaneously extracts subject-specific IFNs across multiple visits by explicitly modeling functional brain development as an essential context for seeking change. On data generated by an innovative simulation based on real rs-fMRI, the method was more accurate in estimating subject-specific IFNs than cross-sectional approaches. Furthermore, only group-analysis based on longitudinally consistent estimates identified significant developmental effects within IFNs of 246 adolescents from the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA) study. The findings were confirmed by the cross-sectional estimates when the corresponding group analysis was confined to the developmental effects. Those effects also converged with current concepts of neurodevelopment.

    View details for DOI 10.1002/hbm.24541

    View details for PubMedID 30806009

  • Longitudinally consistent estimates of intrinsic functional networks Human Brain Mapping Zhao, Q., Kwon, D., Müller-Oehring, E. M., Le Berre, A., Pfefferbaum, A., Sullivan, E. V., Pohl, K. M. 2019

    View details for DOI 10.1002/hbm.24541

  • Variational Autoencoder with Truncated Mixture of Gaussians for Functional Connectivity Analysis Zhao, Q., Honnorat, N., Adeli, E., Pfefferbaum, A., Sullivan, E. V., Pohl, K. M., Chung, A. C., Gee, J. C., Yushkevich, P. A., Bao, S. SPRINGER INTERNATIONAL PUBLISHING AG. 2019: 867–79
  • Variational AutoEncoder for Regression: Application to Brain Aging Analysis Zhao, Q., Adeli, E., Honnorat, N., Leng, T., Pohl, K. M., Shen, D., Liu, T., Peters, T. M., Staib, L. H., Essert, C., Zhou, S., Yap, P. T., Khan, A. SPRINGER INTERNATIONAL PUBLISHING AG. 2019: 823–31


    While unsupervised variational autoencoders (VAE) have become a powerful tool in neuroimage analysis, their application to supervised learning is under-explored. We aim to close this gap by proposing a unified probabilistic model for learning the latent space of imaging data and performing supervised regression. Based on recent advances in learning disentangled representations, the novel generative process explicitly models the conditional distribution of latent representations with respect to the regression target variable. Performing a variational inference procedure on this model leads to joint regularization between the VAE and a neural-network regressor. In predicting the age of 245 subjects from their structural Magnetic Resonance (MR) images, our model is more accurate than state-of-the-art methods when applied to either region-of-interest (ROI) measurements or raw 3D volume images. More importantly, unlike simple feed-forward neural-networks, disentanglement of age in latent representations allows for intuitive interpretation of the structural developmental patterns of the human brain.

    View details for DOI 10.1007/978-3-030-32245-8_91

    View details for Web of Science ID 000548438900091

    View details for PubMedID 32705091

    View details for PubMedCentralID PMC7377006

  • Jacobian Maps Reveal Under-reported Brain Regions Sensitive to Extreme Binge Ethanol Intoxication in the Rat FRONTIERS IN NEUROANATOMY Zhao, Q., Fritz, M., Pfefferbaum, A., Sullivan, E. V., Pohl, K. M., Zahr, N. M. 2018; 12
  • Jacobian Maps Reveal Under-reported Brain Regions Sensitive to Extreme Binge Ethanol Intoxication in the Rat. Frontiers in neuroanatomy Zhao, Q., Fritz, M., Pfefferbaum, A., Sullivan, E. V., Pohl, K. M., Zahr, N. M. 2018; 12: 108


    Individuals aged 12-20 years drink 11% of all alcohol consumed in the United States with more than 90% consumed in the form of binge drinking. Early onset alcohol use is a strong predictor of future alcohol dependence. The study of the effects of excessive alcohol use on the human brain is hampered by limited information regarding the quantity and frequency of exposure to alcohol. Animal models can control for age at alcohol exposure onset and enable isolation of neural substrates of exposure to different patterns and quantities of ethanol (EtOH). As with humans, a frequently used binge exposure model is thought to produce dependence and affect predominantly corticolimbic brain regions. in vivo neuroimaging enables animals models to be examined longitudinally, allowing for each animal to serve as its own control. Accordingly, we conducted 3 magnetic resonance imaging (MRI) sessions (baseline, binge, recovery) to track structure throughout the brains of wild type Wistar rats to test the hypothesis that binge EtOH exposure affects specific brain regions in addition to corticolimbic circuitry. Voxel-based comparisons of 13 EtOH- vs. 12 water- exposed animals identified significant thalamic shrinkage and lateral ventricular enlargement as occurring with EtOH exposure, but recovering with a week of abstinence. By contrast, pretectal nuclei and superior and inferior colliculi shrank in response to binge EtOH treatment but did not recover with abstinence. These results identify brainstem structures that have been relatively underreported but are relevant for localizing neurocircuitry relevant to the dynamic course of alcoholism.

    View details for DOI 10.3389/fnana.2018.00108

    View details for PubMedID 30618652

    View details for PubMedCentralID PMC6297262

  • Chained regularization for identifying brain patterns specific to HIV infection NEUROIMAGE Adeli, E., Kwon, D., Zhao, Q., Pfefferbaum, A., Zahr, N. M., Sullivan, E. V., Pohl, K. M. 2018; 183: 425–37
  • A Riemannian Framework for Longitudinal Analysis of Resting-State Functional Connectivity. Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention Zhao, Q., Kwon, D., Pohl, K. M. 2018; 11072: 145-153


    Even though the number of longitudinal resting-state-fMRI studies is increasing, accurately characterizing the changes in functional connectivity across visits is a largely unexplored topic. To improve characterization, we design a Riemannian framework that represents the functional connectivity pattern of a subject at a visit as a point on a Riemannian manifold. Geodesic regression across the 'sample' points of a subject on that manifold then defines the longitudinal trajectory of their connectivity pattern. To identify group differences specific to regions of interest (ROI), we map the resulting trajectories of all subjects to a common tangent space via the Lie group action. We account for the uncertainty in choosing the common tangent space by proposing a test procedure based on the theory of latent p-values. Unlike existing methods, our proposed approach identifies sex differences across 246 subjects, each of them being characterized by three rs-fMRI scans.

    View details for DOI 10.1007/978-3-030-00931-1_17

    View details for PubMedID 33005907

    View details for PubMedCentralID PMC7526985

  • Chained regularization for identifying brain patterns specific to HIV infection. NeuroImage Adeli, E., Kwon, D., Zhao, Q., Pfefferbaum, A., Zahr, N. M., Sullivan, E. V., Pohl, K. M. 2018


    Human Immunodeficiency Virus (HIV) infection continues to have major adverse public health and clinical consequences despite the effectiveness of combination Antiretroviral Therapy (cART) in reducing HIV viral load and improving immune function. As successfully treated individuals with HIV infection age, their cognition declines faster than reported for normal aging. This phenomenon underlines the importance of improving long-term care, which requires better understanding of the impact of HIV on the brain. In this paper, automated identification of patients and brain regions affected by HIV infection are modeled as a classification problem, whose solution is determined in two steps within our proposed Chained-Regularization framework. The first step focuses on selecting the HIV pattern (i.e., the most informative constellation of brain region measurements for distinguishing HIV infected subjects from healthy controls) by constraining the search for the optimal parameter setting of the classifier via group sparsity (ℓ2,1-norm). The second step improves classification accuracy by constraining the parameterization with respect to the selected measurements and the Euclidean regularization (ℓ2-norm). When applied to the cortical and subcortical structural Magnetic Resonance Images (MRI) measurements of 65 controls and 65 HIV infected individuals, this approach is more accurate in distinguishing the two cohorts than more common models. Finally, the brain regions of the identified HIV pattern concur with the HIV literature that uses traditional group analysis models.

    View details for PubMedID 30138676

  • Alcohol use effects on adolescent brain development revealed by simultaneously removing confounding factors, identifying morphometric patterns, and classifying individuals SCIENTIFIC REPORTS Park, S., Zhang, Y., Kwon, D., Zhao, Q., Zahr, N. M., Pfefferbaum, A., Sullivan, E. V., Pohl, K. M. 2018; 8: 8297


    Group analysis of brain magnetic resonance imaging (MRI) metrics frequently employs generalized additive models (GAM) to remove contributions of confounding factors before identifying cohort specific characteristics. For example, the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA) used such an approach to identify effects of alcohol misuse on the developing brain. Here, we hypothesized that considering confounding factors before group analysis removes information relevant for distinguishing adolescents with drinking history from those without. To test this hypothesis, we introduce a machine-learning model that identifies cohort-specific, neuromorphometric patterns by simultaneously training a GAM and generic classifier on macrostructural MRI and microstructural diffusion tensor imaging (DTI) metrics and compare it to more traditional group analysis and machine-learning approaches. Using a baseline NCANDA MR dataset (N = 705), the proposed machine learning approach identified a pattern of eight brain regions unique to adolescents who misuse alcohol. Classifying high-drinking adolescents was more accurate with that pattern than using regions identified with alternative approaches. The findings of the joint model approach thus were (1) impartial to confounding factors; (2) relevant to drinking behaviors; and (3) in concurrence with the alcohol literature.

    View details for PubMedID 29844507

  • A Riemannian Framework for Longitudinal Analysis of Resting-State Functional Connectivity Zhao, Q., Kwon, D., Pohl, K. M., Frangi, A. F., Schnabel, J. A., Davatzikos, C., AlberolaLopez, C., Fichtinger, G. SPRINGER INTERNATIONAL PUBLISHING AG. 2018: 145–53