There is great interest in developing radiological classifiers for diagnosis, staging, and predictive modeling in progressive diseases such as Parkinson’s disease (PD), a neurodegenerative disease that is difficult to detect in its early stages. Here we leverage severity-based meta-data on the stages of disease to define a curriculum for training a deep convolutional neural network (CNN). Typically, deep learning networks are trained by randomly selecting samples in each mini-batch. By contrast, curriculum learning is a training strategy that aims to boost classifier performance by starting with examples that are easier to classify. Here we define a curriculum to progressively increase the difficulty of the training data corresponding to the Hoehn and Yahr (H&Y) staging system for PD (total N=1,012; 653 PD patients, 359 controls; age range: 20.0-84.9 years). Even with our multi-task setting using pre-trained CNNs and transfer learning, PD classification based on T1-weighted (T1-w) MRI was challenging (ROC AUC: 0.59-0.65), but curriculum training boosted performance (by 3.9%) compared to our baseline model. Future work with multimodal imaging may further boost performance.

Parkinson's disease (PD) is a progressive neurodegenerative disease that affects over 10 million people worldwide. Brain atrophy and microstructural abnormalities tend to be more subtle in PD than in other age-related conditions such as Alzheimer's disease, so there is interest in how well machine learning methods can detect PD in radiological scans. Deep learning models based on convolutional neural networks (CNNs) can automatically distil diagnostically useful features from raw MRI scans, but most CNN-based deep learning models have only been tested on T1-weighted brain MRI. Here we examine the added value of diffusion-weighted MRI (dMRI) - a variant of MRI, sensitive to microstructural tissue properties - as an additional input in CNN-based models for PD classification. Our evaluations used data from 3 separate cohorts - from Chang Gung University, the University of Pennsylvania, and the PPMI dataset. We trained CNNs on various combinations of these cohorts to find the best predictive model. Although tests on more diverse data are warranted, deep-learned models from dMRI show promise for PD classification.Clinical Relevance:This study supports the use of diffusion-weighted images as an alternative to anatomical images for AI-based detection of Parkinson's disease.

Brain amyloid and tau, and volumetric atrophy on MRI, are defining features of AD (Jack et al., 2018). Despite growing interest in these component processes, their relationship is still poorly understood. We evaluated the unique contributions of APOE4 allele (E4+/E4−) and brain amyloid (Aβ) status (Amy+/Amy−) to brain volume in the elderly to identify associations with neurodegeneration in regions vulnerable to Alzheimer’s disease (AD). Accelerated 3T T1-weighted brain MRI scans (N=1,353) obtained at screening for three distinct phases of ADNI with available information on APOE status and/or Amyloid positive/negative status were identified for the following subjects: 561 healthy controls (CN) (age: 71.8±6.3 y, 330F/231M) , 599 MCI (mild cognitive impairment; age: 71.9 ± 7.5 y, 264 F/335 M), and 193 AD (age: 74.4±8.2 y, 79 F/114 M). Amyloid status was based on standardized cut-off values in amyloid sensitive PET scans (Royse et al., 2021). We applied unbiased tensor-based morphometry (TBM; Leow et al., 2009; Hua et al., 2011) to each scan to measure relative volume differences between each subject and a common minimal deformation template (MDT), resulting in 3D Jacobian maps to represent relative expansion or contraction at each brain voxel. At each voxel, a linear regression model was fit to relate regional brain volumes to amyloid status, and APOE4, after adjusting for covariates. P-value maps were generated to visualize the pattern of voxel-wise model contributions and statistical significance while enforcing a standard 5% false discovery rate (FDR) correction for multiple statistical comparisons across the brain (Benjamini et al., 1995). Amyloid positive status was associated with significantly lower regional volume in the temporal lobes and higher volume in the ventricles, i.e., ventricular expansion. APOE4 positive status was associated with lower volume in the temporal lobes (see Fig. 1) . The highly localized effects of APOE4 status and amyloid load on brain volumes are consistent with greatest effects of AD pathologic processes in the temporal lobes and similar associations found for Amy+ and E4+ status with cortical thinning (Li et al. (2017). Continued analysis of longitudinal ADNI data will offer further insight into effects of amyloid and APOE4 allele status on atrophy estimates.

Declining estrogen levels before, during, and after menopause can affect memory and risk for Alzheimer's disease. Undesirable side effects of hormone variations emphasize a role for hormone therapy (HT) where possible benefits include a delay in the onset of dementia-yet findings are inconsistent. Effects of HT may be mediated by estrogen receptors found throughout the brain. Effects may also depend on lifestyle factors, timing of use, and genetic risk. We studied the impact of self-reported HT use on brain volume in 562 elderly women (71-94 years) with mixed cognitive status while adjusting for aforementioned factors. Covariate-adjusted voxelwise linear regression analyses using a model with 16 predictors showed HT use as positively associated with regional brain volumes, regardless of cognitive status. Examinations of other factors related to menopause, oophorectomy and hysterectomy status independently yielded positive effects on brain volume when added to our model. One interaction term, HTxBMI, out of several examined, revealed significant negative association with overall brain volume, suggesting a greater reduction in brain volume than BMI alone. Our main findings relating HT to regional brain volume were as hypothesized, but some exploratory analyses were not in line with existing hypotheses. Studies suggest lower levels of estrogen resulting from oophorectomy and hysterectomy affect brain volume negatively, and the addition of HT modifies the relation between BMI and brain volume positively. Effects of HT may depend on the age range assessed, motivating studies with a wider age range as well as a randomized design.

This study used advanced diffusion-weighted MRI (dMRI) to examine the association between exogenous sex-hormone exposure and the brain's white matter aging trajectories in a large population-based sample of women. To investigate the effect of pre- and post-menopausal sex hormones on brain aging, cross-sectional brain dMRI data from the UK Biobank was analyzed using 3 diffusion models: conventional diffusion tensor imaging (DTI), the tensor distribution function (TDF), and neurite orientation dispersion and density imaging (NODDI). Mean skeletonized diffusivity measures were extracted and averaged across the whole brain, including fractional anisotropy, isotropic volume fraction, intracellular volume fraction and orientation dispersion index. We used general linear models and fractional polynomial regressions to characterize age-related trajectories in white matter measures following hormone therapy (HT) and oral contraceptive (OC) use in women (HT analysis: N=8,301; OC analysis: N=8,913). Sex hormone treatment (HT and OC) was statistically associated with the aging trends in white matter measures. Estrogen therapy alone appeared to exert a neuroprotective effect on age-related white matter processes, compared to HT containing both estrogen and progestin therapy - which was associated with accelerated aging-related processes in women. These results support the hypothesis that exogenous sex hormone exposure may impact white matter aging; white matter metrics may also be sensitive to sex hormone levels in women. Furthermore, we discuss the necessity to test alternative models for lifespan trajectories beyond popular linear and quadratic models, especially when dealing with large samples. Fractional polynomial models may provide a more adaptive alternative to linear or quadratic models.

Cortisol is considered the most fundamental stress hormone and is elevated in stress and multiple neuropsychiatric conditions. Prior studies have shown associations of plasma cortisol levels with total cerebral and hippocampal volumes and less consistently with the amygdala. Here, we extend our hypothesis to test associations of plasma cortisol with 1) ventricular 2) hippocampal and 3) amygdalar volumes, in two independent elderly cohorts across a broad cognitive spectrum ranging from normal cognition to Alzheimer’s disease. We demonstrate elevated cortisol to be associated with larger lateral ventricular volumes and smaller hippocampal volumes, predominantly in the right cerebral hemisphere, regardless of age, sex or cognitive status. We noted a non-significant trend of smaller amygdalar volumes with elevated cortisol. Our findings support smaller brain parenchyma volumes seen with elevated cortisol and may encourage effective strategies reducing cortisol and stress. They may also serve as imaging biomarkers for assessing therapeutic benefits of stress and cortisol lowering interventions aiming to halt or reverse the brain volume alterations and theoretically improve cognition and quality of life. Highlights Elevated cortisol associated with larger ventricular volumes and smaller hippocampal volumes Associations are predominantly noted in the right cerebral hemisphere. Similar non-significant trends noted in amygdalar volumes Cortisol and stress reducing strategies may halt brain changes and improve quality of life Imaging biomarkers may help assess efficacy of cortisol-lowering therapeutic interventions

Secondhand smoke exposure is a major public health risk that is especially harmful to the developing brain, but it is unclear if early exposure affects brain structure during middle age and older adulthood. Here we analyzed brain MRI data from the UK Biobank in a population-based sample of individuals (ages 44-80) who were exposed (n = 2510) or unexposed (n = 6079) to smoking around birth. We used robust statistical models, including quantile regressions, to test the effect of perinatal smoke exposure (PSE) on cortical surface area (SA), thickness, and subcortical volumes. We hypothesized that PSE would be associated with cortical disruption in primary sensory areas compared to unexposed (PSE-) adults. After adjusting for multiple comparisons, SA was significantly lower in the pericalcarine (PCAL), inferior parietal (IPL), and regions of the temporal and frontal cortex of PSE+ adults; these abnormalities were associated with increased risk for several diseases, including circulatory and endocrine conditions. Sensitivity analyses conducted in a hold-out group of healthy participants (exposed, n = 109, unexposed, n = 315) replicated the effect of PSE on SA in the PCAL and IPL. Collectively our results show a negative, long term effect of PSE on sensory cortices that may increase risk for disease later in life.

Secondhand smoke exposure is a major public health risk that is especially harmful to the developing brain, but it is unclear if early exposure affects brain structure during middle age and older adulthood. Here we analyzed brain MRI data from the UK Biobank in a population-based sample of individuals (ages 44-80) who were exposed (n=2,510) or unexposed (n=6,079) to smoking around birth. We used robust statistical models, including quantile regressions, to test the effect of perinatal smoke exposure (PSE) on cortical surface area (SA), thickness, and subcortical volume. We hypothesized that PSE would be associated with cortical disruption in primary sensory areas compared to unexposed (PSE-) adults. After adjusting for multiple comparisons, SA was significantly lower in the pericalcarine (PCAL), inferior parietal (IPL), and regions of the temporal and frontal cortex of PSE+ adults; these abnormalities were associated with increased risk for several diseases, including circulatory and endocrine conditions. Sensitivity analyses conducted in a hold-out group of healthy participants (exposed, n=109, unexposed, n=315) replicated the effect of PSE on SA in the PCAL and IPL. Collectively our results show a negative, long term effect of PSE on sensory cortices that may increase risk for disease later in life.

The purpose of this study was to use a novel imaging biomarker to assess associations between physical activity (PA), body mass index (BMI), and brain structure in normal aging, mild cognitive impairment, and Alzheimer's dementia. We studied 963 participants (mean age: 74.1 +/- 4.4 years) from the multisite Cardiovascular Health Study including healthy controls (n = 724), Alzheimer's dementia patients (n = 104), and people with mild cognitive impairment (n = 135). Volumetric brain images were processed using tensor-based morphometry to analyze regional brain volumes. We regressed the local brain tissue volume on reported PA and computed BMI, and performed conjunction analyses using both variables. Covariates included age, sex, and study site. PA was independently associated with greater whole brain and regional brain volumes and reduced ventricular dilation. People with higher BMI had lower whole brain and regional brain volumes. A PA-BMI conjunction analysis showed brain preservation with PA and volume loss with increased BMI in overlapping brain regions. In one of the largest voxel-based cross-sectional studies to date, PA and lower BMI may be beneficial to the brain across the spectrum of aging and neurodegeneration. (C) 2015 Elsevier Inc. All rights reserved.

The purpose of this study was to use a novel imaging biomarker to assess associations between physical activity (PA), body mass index (BMI), and brain structure in normal aging, mild cognitive impairment, and Alzheimer's dementia. We studied 963 participants (mean age: 74.1 ± 4.4 years) from the multisite Cardiovascular Health Study including healthy controls (n = 724), Alzheimer's dementia patients (n = 104), and people with mild cognitive impairment (n = 135). Volumetric brain images were processed using tensor-based morphometry to analyze regional brain volumes. We regressed the local brain tissue volume on reported PA and computed BMI, and performed conjunction analyses using both variables. Covariates included age, sex, and study site. PA was independently associated with greater whole brain and regional brain volumes and reduced ventricular dilation. People with higher BMI had lower whole brain and regional brain volumes. A PA-BMI conjunction analysis showed brain preservation with PA and volume loss with increased BMI in overlapping brain regions. In one of the largest voxel-based cross-sectional studies to date, PA and lower BMI may be beneficial to the brain across the spectrum of aging and neurodegeneration.

Physical activity influences inflammation, and both affect brain structure and Alzheimer's disease (AD) risk. We hypothesized that older adults with greater reported physical activity intensity and lower serum levels of the inflammatory marker tumor necrosis factor α (TNFα) would have larger regional brain volumes on subsequent magnetic resonance imaging (MRI) scans. In 43 cognitively intact older adults (79.3±4.8 years) and 39 patients with AD (81.9±5.1 years at the time of MRI) participating in the Cardiovascular Health Study, we examined year-1 reported physical activity intensity, year-5 blood serum TNFα measures, and year-9 volumetric brain MRI scans. We examined how prior physical activity intensity and TNFα related to subsequent total and regional brain volumes. Physical activity intensity was measured using the modified Minnesota Leisure Time Physical Activities questionnaire at year 1 of the study, when all subjects included here were cognitively intact. Stability of measures was established for exercise intensity over 9 years and TNFα over 3 years in a subset of subjects who had these measurements at multiple time points. When considered together, more intense physical activity intensity and lower serum TNFα were both associated with greater total brain volume on follow-up MRI scans. TNFα, but not physical activity, was associated with regional volumes of the inferior parietal lobule, a region previously associated with inflammation in AD patients. Physical activity and TNFα may independently influence brain structure in older adults.

Cognitive impairment and brain injury are common in people with HIV/AIDS, even when viral replication is effectively suppressed with combined antiretroviral therapies (cART). Metabolic and structural abnormalities may promote cognitive decline, but we know little about how these measures relate in people on stable cART. Here we used tensor-based morphometry (TBM) to reveal the 3D profile of regional brain volume variations in 210 HIV+ patients scanned with whole-brain MRI at 1.5 T (mean age: 48.6 +/- 8.4 years; all receiving cART). We identified brain regions where the degree of atrophy was related to HIV clinical measures and cerebral metabolite levels assessed with magnetic resonance spectroscopy (MRS). Regional brain volume reduction was linked to lower nadir CD4+ count, with a 1-2% white matter volume reduction for each 25-point reduction in nadir CD4+. Even so, brain volume measured by TBM showed no detectable association with current CD4+ count, AIDS Dementia Complex (ADC) stage, HIV RNA load in plasma or cerebrospinal fluid (CSF), duration of HIV infection, antiretroviral CNS penetration-effectiveness (CPE) scores, or years on cART, after controlling for demographic factors, and for multiple comparisons. Elevated glutamate and glutamine (Glx) and lower N-acetylaspartate (NAA) in the frontal white matter, basal ganglia, and mid frontal cortex - were associated with lower white matter, putamen and thalamus volumes, and ventricular and CSF space expansion. Reductions in brain volumes in the setting of chronic and stable disease are strongly linked to a history of immunosuppression, suggesting that delays in initiating cART may result in imminent and irreversible brain damage. (C) 2013 The Authors. Published by Elsevier Inc.

Exercise, insulin-like growth factor 1 (IGF1), and inflammation moderate each other and relate to structural brain integrity. We hypothesized that in older adults, more exercise and higher serum IGF1 levels would correlate with greater brain volume while higher levels of the inflammatory marker tumor necrosis factor α (TNFα) would relate to smaller brain volumes. We further hypothesized that IGF1 and TNFa would partly mediate the exercise effect on brain volume. We examined 43 controls (79.3 ± 4.8 years) and 39 Alzheimer's disease (AD) patients (81.9 ± 5.1 years) from the Cardiovascular Health Study. Subjects had reported exercise intensity at year 1 of the study, blood serum TNF α and IGF1 measures at year 5, and volumetric MRI scans at year 9. We automatically segmented skull-stripped images into brain and cerebrospinal fluid. We evaluated how brain as a percentage of intracranial volume (brain %) related to exercise, TNF α, and IGF1 individually, while controlling for age, sex, education, diagnosis, apolipoprotein E genotype, scan location, body mass index, and depression rating. We refined significant results by including the same variables in a voxelwise tensor based morphometry (TBM) analysis. The false discovery rate (FDR; q=0.05) controlled for multiple voxelwise comparisons. Baseline exercise intensity (overall p = 3.95 x 10 -9; exercise partial p = 0.03) and TNF α (overall p = 4.99 x 10 -9; TNF α partial p = 0.006) (but not IGF1) were significantly associated with brain % individually and together (overall p = 4.68 x 10 -9; TNF α partial p = 0.007; exercise partial p = 0.018). Exercise intensity and TNF α were not significantly associated with each other (overall model p = 0.19). Using TBM, TNF α, but not exercise, was significantly associated with brain structure in the left supramarginal gyrus (FDR critical p value = 3.2 x10 -5).

Obesity and major depressive disorder (MDD) are highly prevalent and often comorbid health conditions. Both are associated with differences in brain structure and are genetically influenced. Yet, little is known about how obesity, MDD, and known risk genotypes might interact in the brain. Subjects were 81 patients with MDD (mean age 48.6 years) and 69 matched healthy controls (mean age 51.2 years). Subjects underwent 1.5T magnetic resonance imaging, genotyping for the fat mass and obesity associated (FTO) gene rs3751812 polymorphism, and measurements for body mass index (BMI). We conducted a whole brain voxelwise analysis using tensor-based morphometry (TBM) to examine the main and interaction effects of diagnosis, BMI and FTO genotype. Significant effects of BMI were observed across widespread brain regions, indicating reductions in predominantly subcortical and white matter areas associated with increased BMI, but there was no influence of MDD or FTO rs3751812 genotype. There were no significant interaction effects. Within MDD patients, there was no effect of current depressive symptoms; however the use of antidepressant medication was associated with reductions in brain volume in the frontal lobe and cerebellum. Obesity affects brain structure in both healthy participants and MDD patients; this influence may account for some of the brain changes previously associated with MDD. BMI and the use of medication should ideally be measured and controlled for when conducting structural brain imaging research in MDD.

Various neuroimaging measures are being evaluated for tracking Alzheimer's disease (AD) progression in therapeutic trials, including measures of structural brain change based on repeated scanning of patients with magnetic resonance imaging (MRI). Methods to compute brain change must be robust to scan quality. Biases may arise if any scans are thrown out, as this can lead to the true changes being overestimated or underestimated. Here we analyzed the full MRI dataset from the first phase of Alzheimer's Disease Neuroimaging Initiative (ADNI-1) from the first phase of Alzheimer's Disease Neuroimaging Initiative (ADNI-1) and assessed several sources of bias that can arise when tracking brain changes with structural brain imaging methods, as part of a pipeline for tensor-based morphometry (TBM). In all healthy subjects who completed MRI scanning at screening, 6, 12, and 24months, brain atrophy was essentially linear with no detectable bias in longitudinal measures. In power analyses for clinical trials based on these change measures, only 39AD patients and 95 mild cognitive impairment (MCI) subjects were needed for a 24-month trial to detect a 25% reduction in the average rate of change using a two-sided test (α=0.05, power=80%). Further sample size reductions were achieved by stratifying the data into Apolipoprotein E (ApoE) ε4 carriers versus non-carriers. We show how selective data exclusion affects sample size estimates, motivating an objective comparison of different analysis techniques based on statistical power and robustness. TBM is an unbiased, robust, high-throughput imaging surrogate marker for large, multi-site neuroimaging studies and clinical trials of AD and MCI.

CONTEXT Nonpsychotic siblings of patients with childhood-onset schizophrenia (COS) share cortical gray matter abnormalities with their probands at an early age; these normalize by the time the siblings are aged 18 years, suggesting that the gray matter abnormalities in schizophrenia could be an age-specific endophenotype. Patients with COS also show significant white matter (WM) growth deficits, which have not yet been explored in nonpsychotic siblings. OBJECTIVE To study WM growth differences in nonpsychotic siblings of patients with COS. DESIGN Longitudinal (5-year) anatomic magnetic resonance imaging study mapping WM growth using a novel tensor-based morphometry analysis. SETTING National Institutes of Health Clinical Center, Bethesda, Maryland. PARTICIPANTS Forty-nine healthy siblings of patients with COS (mean [SD] age, 16.1 [5.3] years; 19 male, 30 female) and 57 healthy persons serving as controls (age, 16.9 [5.3] years; 29 male, 28 female). INTERVENTION Magnetic resonance imaging. MAIN OUTCOME MEASURE White matter growth rates. RESULTS We compared the WM growth rates in 3 age ranges. In the youngest age group (7 to <14 years), we found a significant difference in growth rates, with siblings of patients with COS showing slower WM growth rates in the parietal lobes of the brain than age-matched healthy controls (false discovery rate, q = 0.05; critical P = .001 in the bilateral parietal WM; a post hoc analysis identified growth rate differences only on the left side, critical P = .004). A growth rate difference was not detectable at older ages. In 3-dimensional maps, growth rates in the siblings even appeared to surpass those of healthy individuals at later ages, at least locally in the brain, but this effect did not survive a multiple comparisons correction. CONCLUSIONS In this first longitudinal study of nonpsychotic siblings of patients with COS, the siblings showed early WM growth deficits, which normalized with age. As reported before for gray matter, WM growth may also be an age-specific endophenotype that shows compensatory normalization with age.

Callosal volume reduction has been observed in patients with bipolar disorder, but whether these deficits reflect genetic vulnerability to the illness remains unresolved. Here, we used computational methods to map corpus callosum abnormalities in a population-based sample of twin pairs discordant for bipolar disorder. Twenty-one probands with bipolar I disorder (mean age 44.4 ± 7.5 years; 48% female), 19 of their non-bipolar co-twins, and 34 demographically matched control twin individuals underwent magnetic resonance imaging. Three-dimensional callosal surface models were created to visualize its morphologic variability and to localize group differences. Neurocognitive correlates of callosal area differences were additionally investigated in a subsample of study participants. Bipolar (BPI) probands, but not their co-twins, showed significant callosal thinning and area reduction, most pronounced in the genu and splenium, relative to healthy twins. Altered callosal curvature was additionally observed in BPI probands. In bipolar probands and co-twins, genu and splenium midsagittal areas were significantly correlated with verbal processing speed and response inhibition. These findings suggest that aberrant connections between cortical regions—possibly reflecting decreased myelination of white matter tracts—may be involved in bipolar pathophysiology. However, findings of callosal thinning appear to be disease related, rather than reflecting genetic vulnerability to bipolar illness.

This paper responds to Thompson and Holland (2011), who challenged our tensor-based morphometry (TBM) method for estimating rates of brain changes in serial MRI from 431 subjects scanned every 6months, for 2years. Thompson and Holland noted an unexplained jump in our atrophy rate estimates: an offset between 0 and 6months that may bias clinical trial power calculations. We identified why this jump occurs and propose a solution. By enforcing inverse-consistency in our TBM method, the offset dropped from 1.4% to 0.28%, giving plausible anatomical trajectories. Transitivity error accounted for the minimal remaining offset. Drug trial sample size estimates with the revised TBM-derived metrics are highly competitive with other methods, though higher than previously reported sample size estimates by a factor of 1.6 to 2.4. Importantly, estimates are far below those given in the critique. To demonstrate a 25% slowing of atrophic rates with 80% power, 62 AD and 129 MCI subjects would be required for a 2-year trial, and 91 AD and 192 MCI subjects for a 1-year trial.