Improving the replicability of neuroimaging findings by thresholding effect sizes instead of p-values

Semiparametric Confidence Sets for Arbitrary Effect Sizes in Longitudinal Neuroimaging

The majority of neuroimaging inference focuses on hypothesis testing rather than effect estimation. With concerns about replicability, there is growing interest in reporting standardized effect sizes from neuroimaging group-level analyses. Confidence sets are a recently developed approach to perform inference for effect sizes in neuroimaging but are restricted to univariate effect sizes and cross-sectional data. Thus, existing methods exclude increasingly common multigroup or nonlinear longitudinal associations of biological brain measurements with inter- and intra-individual variations in diagnosis, development, or symptoms. We broadly generalize the confidence set approach by developing a method for arbitrary effect sizes in longitudinal studies. Our method involves robust estimation of the effect size image and spatial and temporal covariance function based on generalized estimating equations. We obtain more efficient effect size estimates by concurrently estimating the exchangeable working covariance and using a nonparametric bootstrap to determine the joint distribution of effect size across voxels used to construct confidence sets. These confidence sets identify regions of the image where the lower or upper simultaneous confidence interval is above or below a given threshold with high probability. We evaluate the coverage and simultaneous confidence interval width of the proposed procedures using realistic simulations and perform longitudinal analyses of aging and diagnostic differences of cortical thickness in Alzheimer's disease and diagnostic differences of resting-state hippocampal activity in psychosis. This comprehensive approach along with the visualization functions integrated into the pbj R package offers a robust tool for analyzing repeated neuroimaging measurements.

Effects of sleep-corrected social jetlag on measures of mental health, cognitive ability, and brain functional connectivity in early adolescence

Approximately half of adolescents encounter a mismatch between their sleep patterns on school days and free days, also referred to as "social jetlag." This condition has been linked to various adverse outcomes, such as poor sleep, cognitive deficits, and mental disorders. However, prior research was unsuccessful in accounting for other variables that are correlated with social jetlag, including sleep duration and quality. To address this limitation, we applied a propensity score matching method on a sample of 6335 11-12-year-olds from the 2-year follow-up (FL2) data of the Adolescent Brain Cognitive Development study. We identified 2424 pairs of participants with high sleep-corrected social jetlag (SJLsc, over 1 hour) and low SJLsc (<= 1 hour) at FL2 (1728 pairs have neuroimaging data), as well as 1626 pairs at 3-year follow-up (FL3), after matching based on 11 covariates including socioeconomic status, demographics, and sleep duration and quality. Our results showed that high SJLsc, as measured by the Munich Chronotype Questionnaire, was linked to reduced crystallized intelligence (CI), lower school performance-grades, and decreased functional connectivity between cortical networks and subcortical regions, specifically between cingulo-opercular network and right hippocampus. Further mediation and longitudinal mediation analyses revealed that this connection mediated the associations between SJLsc and CI at FL2, and between SJLsc and grades at both FL2 and FL3. We validated these findings by replicating these results using objective SJLsc measurements obtained via Fitbit watches. Overall, our study highlights the negative association between social jetlag and CI during early adolescence.

Neuroimaging Investigations of Obesity: a Review of the Treatment of Sex from 2010

PURPOSE OF REVIEW

To summarize the results of adult obesity neuroimaging studies (structural, resting-state, task-based, diffusion tensor imaging) published from 2010, with a focus on the treatment of sex as an important biological variable in the analysis, and identify gaps in sex difference research.

RECENT FINDINGS

Neuroimaging studies have shown obesity-related changes in brain structure, function, and connectivity. However, relevant factors such as sex are often not considered. We conducted a systematic review and keyword co-occurrence analysis. Literature searches identified 6281 articles, of which 199 met inclusion criteria. Among these, only 26 (13%) considered sex as an important variable in the analysis, directly comparing the sexes (n = 10; 5%) or providing single-sex/disaggregated data (n = 16, 8%); the remaining studies controlled for sex (n = 120, 60%) or did not consider sex in the analysis (n = 53, 27%). Synthesizing sex-based results, obesity-related parameters (e.g., body mass index, waist circumference, obese status) may be generally associated with more robust morphological alterations in men and more robust structural connectivity alterations in women. Additionally, women with obesity generally expressed increased reactivity in affect-related regions, while men with obesity generally expressed increased reactivity in motor-related regions; this was especially true under a fed state. The keyword co-occurrence analysis indicated that sex difference research was especially lacking in intervention studies. Thus, although sex differences in the brain associated with obesity are known to exist, a large proportion of the literature informing the research and treatment strategies of today has not specifically examined sex effects, which is needed to optimize treatment.

Effects of sleep duration on neurocognitive development in early adolescents in the USA: a propensity score matched, longitudinal, observational study

BACKGROUND

Although the American Academy of Sleep Medicine suggests at least 9 h of sleep per day for 6-12-year-olds, children in recent generations often report sleeping less than this amount. Because early adolescence is a crucial period for neurocognitive development, we aimed to investigate how insufficient sleep affects children's mental health, cognition, brain function, and brain structure over 2 years.

METHODS

In this propensity score matched, longitudinal, observational cohort study, we obtained data from a population-based sample of 9-10-year-olds from 21 US study sites in the ongoing Adolescent Brain Cognitive Development (ABCD) study. Participants were categorised as having sufficient sleep or insufficient sleep on the basis of a cutoff of 9 h sleep per day. Using propensity score matching, we matched these two groups of participants on 11 key covariates, including sex, socioeconomic status, and puberty status. Participants were excluded from our analysis if they did not pass a baseline resting-state functional MRI quality check or had missing data for the covariates involved in propensity score matching. Outcome measures retrieved from the ABCD study were behavioural problems, mental health, cognition, and structural and resting-state functional brain measures, assessed at baseline and at 2-year follow-up. We examined group differences on these outcomes over those 2 years among all eligible participants. We then did mediation analyses of the neural correlates of behavioural changes induced by insufficient sleep.

FINDINGS

Between Sept 1, 2016, and Oct 15, 2018, 11 878 individuals had baseline data collected for the ABCD study, of whom 8323 were eligible and included in this study (4142 participants in the sufficient sleep group and 4181 in the insufficient sleep group). Follow-up data were collected from July 30, 2018, to Jan 15, 2020. We identified 3021 matched sufficient sleep-insufficient sleep pairs at baseline and 749 matched pairs at 2-year follow-up, and observed similar differences between the groups in behaviour and neural measures at both timepoints; the effect sizes of between-group differences in behavioural measures at these two timepoints were significantly correlated with each other (r=0·85, 95% CI 0·73-0·92; p<0·0001). A similar pattern was observed in resting-state functional connectivity (r=0·54, 0·45-0·61; p<0·0001) and in structural measures (eg, in grey matter volume r=0·61, 0·51-0·69; p<0·0001). We found that cortico-basal ganglia functional connections mediate the effects of insufficient sleep on depression, thought problems, and crystallised intelligence, and that structural properties of the anterior temporal lobe mediate the effect of insufficient sleep on crystallised intelligence.

INTERPRETATION

These results provide population-level evidence for the long-lasting effect of insufficient sleep on neurocognitive development in early adolescence. These findings highlight the value of early sleep intervention to improve early adolescents' long-term developmental outcomes.

FUNDING

National Institutes of Health.

Non-parametric combination of multimodal MRI for lesion detection in focal epilepsy

•

Multivariate voxel-based analysis useful for lesion detection in focal epilepsy.

•

Non-parametric combination algorithm used to combine data from various MR sequences.

•

Successful lesion detection demonstrated in MRI-positive and MRI-negative patients.

•

Multimodal analysis detected abnormalities from diverse epileptogenic pathologies.

•

Sensitivity of multivariate analysis notably higher than univariate analyses.

One third of patients with medically refractory focal epilepsy have normal-appearing MRI scans. This poses a problem as identification of the epileptogenic region is required for surgical treatment. This study performs a multimodal voxel-based analysis (VBA) to identify brain abnormalities in MRI-negative focal epilepsy. Data was collected from 69 focal epilepsy patients (42 with discrete lesions on MRI scans, 27 with no visible findings on scans), and 62 healthy controls. MR images comprised T1-weighted, fluid-attenuated inversion recovery (FLAIR), fractional anisotropy (FA) and mean diffusivity (MD) from diffusion tensor imaging, and neurite density index (NDI) from neurite orientation dispersion and density imaging. These multimodal images were coregistered to T1-weighted scans, normalized to a standard space, and smoothed with 8 mm FWHM. Initial analysis performed voxel-wise one-tailed t-tests separately on grey matter concentration (GMC), FLAIR, FA, MD, and NDI, comparing patients with epilepsy to controls. A multimodal non-parametric combination (NPC) analysis was also performed simultaneously on FLAIR, FA, MD, and NDI. Resulting p-maps were family-wise error rate corrected, threshold-free cluster enhanced, and thresholded at p < 0.05. Sensitivity was established through visual comparison of results to manually drawn lesion masks or seizure onset zone (SOZ) from stereoelectroencephalography. A leave-one-out cross-validation with the same analysis protocols was performed on controls to determine specificity. NDI was the best performing individual modality, detecting focal abnormalities in 38% of patients with normal MRI and conclusive SOZ. GMC demonstrated the lowest sensitivity at 19%. NPC provided superior performance to univariate analyses with 50% sensitivity. Specificity in controls ranged between 96 and 100% for all analyses. This study demonstrated the utility of a multimodal VBA utilizing NPC for detecting epileptogenic lesions in MRI-negative focal epilepsy. Future work will apply this approach to datasets from other centres and will experiment with different combinations of MR sequences.

Improving the replicability of neuroimaging findings by thresholding effect sizes instead of p-values

The classical approach for testing statistical images using spatial extent inference (SEI) thresholds the statistical image based on the p‐value. This approach has an unfortunate consequence on the replicability of neuroimaging findings because the targeted brain regions are affected by the sample size—larger studies have more power to detect smaller effects. Here, we use simulations based on the preprocessed Autism Brain Imaging Data Exchange (ABIDE) to show that thresholding statistical images by effect sizes has more consistent estimates of activated regions across studies than thresholding by p‐values. Using a constant effect size threshold means that the p‐value threshold naturally scales with the sample size to ensure that the target set is similar across repetitions of the study that use different sample sizes. As a consequence of thresholding by the effect size, the type 1 and type 2 error rates go to zero as the sample size gets larger. We use a newly proposed robust effect size index that is defined for an arbitrary statistical image so that effect size thresholding can be used regardless of the test statistic or model.