FreeSurfer 5.3 versus 6.0: are volumes comparable? A Chronic Effects of Neurotrauma Consortium study

Machine learning on longitudinal multi-modal data enables the understanding and prognosis of Alzheimer's disease progression

Alzheimer's disease (AD) is a complex pathophysiological disease. Allowing for heterogeneity, not only in disease manifestations but also in different progression patterns, is critical for developing effective disease models that can be used in clinical and research settings. We introduce a machine learning model for identifying underlying patterns in Alzheimer's disease (AD) trajectory using longitudinal multi-modal data from the ADNI cohort and the AIBL cohort. Ten biologically and clinically meaningful disease-related states were identified from data, which constitute three non-overlapping stages (i.e., neocortical atrophy [NCA], medial temporal atrophy [MTA], and whole brain atrophy [WBA]) and two distinct disease progression patterns (i.e., NCA → WBA and MTA → WBA). The index of disease-related states provided a remarkable performance in predicting the time to conversion to AD dementia (C-Index: 0.923 ± 0.007). Our model shows potential for promoting the understanding of heterogeneous disease progression and early predicting the conversion time to AD dementia.

Efficacy of MRI data harmonization in the age of machine learning: a multicenter study across 36 datasets

Pooling publicly-available MRI data from multiple sites allows to assemble extensive groups of subjects, increase statistical power, and promote data reuse with machine learning techniques. The harmonization of multicenter data is necessary to reduce the confounding effect associated with non-biological sources of variability in the data. However, when applied to the entire dataset before machine learning, the harmonization leads to data leakage, because information outside the training set may affect model building, and potentially falsely overestimate performance. We propose a 1) measurement of the efficacy of data harmonization; 2) harmonizer transformer, i.e., an implementation of the ComBat harmonization allowing its encapsulation among the preprocessing steps of a machine learning pipeline, avoiding data leakage by design. We tested these tools using brain T1-weighted MRI data from 1740 healthy subjects acquired at 36 sites. After harmonization, the site effect was removed or reduced, and we showed the data leakage effect in predicting individual age from MRI data, highlighting that introducing the harmonizer transformer into a machine learning pipeline allows for avoiding data leakage by design.

Freesurfer Software Update Significantly Impacts Striatal Volumes in the Huntington's Disease Young Adult Study and Will Influence HD-ISS Staging

Background

The Huntington's Disease Integrated Staging System (HD-ISS) defined disease onset using volumetric cut-offs for caudate and putamen derived from FreeSurfer 6 (FS6). The impact of the latest software update (FS7) on volumes remains unknown. The Huntington's Disease Young Adult Study (HD-YAS) is appropriately positioned to explore differences in FS bias when detecting early atrophy.

Objective

Explore the relationships and differences between raw caudate and putamen volumes, calculated total intracranial volumes (cTICV), and adjusted caudate and putamen volumes, derived from FS6 and FS7, in HD-YAS.

Methods

Images from 123 participants were segmented and quality controlled. Relationships and differences between volumes were explored using intraclass correlation (ICC) and Bland-Altman analysis.

Results

Across the whole cohort, ICC for raw caudate and putamen was 0.99, cTICV 0.93, adjusted caudate 0.87, and adjusted putamen 0.86 (all p < 0.0005). Compared to FS6, FS7 calculated: i) larger raw caudate (+0.8%, p < 0.00005) and putamen (+1.9%, p < 0.00005), with greater difference for larger volumes; and ii) smaller cTICV (-5.1%, p < 0.00005), with greater difference for smaller volumes. The systematic and proportional difference in cTICV was greater than raw volumes. When raw volumes were adjusted for cTICV, these effects compounded (adjusted caudate +7.0%, p < 0.00005; adjusted putamen +8.2%, p < 0.00005), with greater difference for larger volumes.

Conclusions

As new software is released, it is critical that biases are explored since differences have the potential to significantly alter the findings of HD trials. Until conversion factors are defined, the HD-ISS must be applied using FS6. This should be incorporated into the HD-ISS online calculator.

Toward More Accessible Fully Automated 3D Volumetric MRI Decision Trees for the Differential Diagnosis of Multiple System Atrophy, Related Disorders, and Age-Matched Healthy Subjects

Differentiating multiple system atrophy (MSA) from related neurodegenerative movement disorders (NMD) is challenging. MRI is widely available and automated decision-tree analysis is simple, transparent, and resistant to overfitting. Using a retrospective cohort of heterogeneous clinical MRIs broadly sourced from a tertiary hospital system, we aimed to develop readily translatable and fully automated volumetric diagnostic decision-trees to facilitate early and accurate differential diagnosis of NMDs. 3DT1 MRI from 171 NMD patients (72 MSA, 49 PSP, 50 PD) and 171 matched healthy subjects were automatically segmented using Freesurfer6.0 with brainstem module. Decision trees employing substructure volumes and a novel volumetric pons-to-midbrain ratio (3D-PMR) were produced and tenfold cross-validation performed. The optimal tree separating NMD from healthy subjects selected cerebellar white matter, thalamus, putamen, striatum, and midbrain volumes as nodes. Its sensitivity was 84%, specificity 94%, accuracy 84%, and kappa 0.69 in cross-validation. The optimal tree restricted to NMD patients selected 3D-PMR, thalamus, superior cerebellar peduncle (SCP), midbrain, pons, and putamen as nodes. It yielded sensitivities/specificities of 94/84% for MSA, 72/96% for PSP, and 73/92% PD, with 79% accuracy and 0.62 kappa. There was correct classification of 16/17 MSA, 5/8 PSP, 6/8 PD autopsy-confirmed patients, and 6/8 MRIs that preceded motor symptom onset. Fully automated decision trees utilizing volumetric MRI data distinguished NMD patients from healthy subjects and MSA from other NMDs with promising accuracy, including autopsy-confirmed and pre-symptomatic subsets. Our open-source methodology is well-suited for widespread clinical translation. Assessment in even more heterogeneous retrospective and prospective cohorts is indicated.

The effect of processing pipelines, input images and age on automatic cortical morphology estimates

BACKGROUND AND OBJECTIVE

Magnetic resonance imaging of the brain allows to enrich the study of the relationship between cortical morphology, healthy ageing, diseases and cognition. Since manual segmentation of the cerebral cortex is time consuming and subjective, many software packages have been developed. FreeSurfer (FS) and Advanced Normalization Tools (ANTs) are the most used and allow as inputs a T1-weighted (T1w) image or its combination with a T2-weighted (T2w) image. In this study we evaluated the impact of different software and input images on cortical estimates. Additionally, we investigated whether the variation of the results depending on software and inputs is also influenced by age.

METHODS

For 240 healthy subjects, cortical thickness was computed with ANTs and FreeSurfer. Estimates were derived using both the T1w image and adding the T2w image. Significant effects due to software, input images and age range were investigated with ANOVA statistical analysis. Moreover, the accuracy of the cortical thickness estimates was assessed based on their age-prediction precision.

RESULTS

Using FreeSurfer and ANTs with T1w or T1w-T2w images resulted in significant differences in the cortical thickness estimates. These differences change with the age range of the subjects. Regardless of the images used, the more recent FS version tested exhibited the best performances in terms of age prediction.

CONCLUSIONS

Our study points out the importance of i) consistently processing data using the same tool; ii) considering the software, input images and the age range of the subjects when comparing multiple studies.

FreeSurfer 7 quality control: Key problem areas and importance of manual corrections

We have studied the effects of manual quality control of brain Magnetic Resonance Imaging (MRI) images processed with Freesurfer. T1 images of first episode psychosis patients (N = 60) and healthy controls (N = 41) were inspected for gray matter boundary errors. The errors were fixed, and the effects of error correction on brain volume, thickness, and surface area were measured. It is commonplace to apply quality control to Freesurfer MRI recordings to ensure that the edges of gray and white matter are detected properly, as incorrect edge detection leads to changes in variables such as volume, cortical thickness, and cortical surface area. We find that while Freesurfer v7.1.1. does regularly make mistakes in identifying the edges of cortical gray matter, correcting these errors yields limited changes in the commonly measured variables listed above. We further find that the software makes fewer gray matter boundary errors when processing female brains. The results suggest that manually correcting gray matter boundary errors may not be worthwhile due to its small effect on the measurements, with potential exceptions for studies that focus on the areas that are more commonly affected by errors: the areas around the cerebellar tentorium, paracentral lobule, and the optic nerves, specifically the horizontal segment of the middle cerebral artery.

Detection of Alzheimer's disease onset using MRI and PET neuroimaging: longitudinal data analysis and machine learning

The scientists are dedicated to studying the detection of Alzheimer's disease onset to find a cure, or at the very least, medication that can slow the progression of the disease. This article explores the effectiveness of longitudinal data analysis, artificial intelligence, and machine learning approaches based on magnetic resonance imaging and positron emission tomography neuroimaging modalities for progression estimation and the detection of Alzheimer's disease onset. The significance of feature extraction in highly complex neuroimaging data, identification of vulnerable brain regions, and the determination of the threshold values for plaques, tangles, and neurodegeneration of these regions will extensively be evaluated. Developing automated methods to improve the aforementioned research areas would enable specialists to determine the progression of the disease and find the link between the biomarkers and more accurate detection of Alzheimer's disease onset.

Multisite test-retest reliability and compatibility of brain metrics derived from FreeSurfer versions 7.1, 6.0, and 5.3

Automatic neuroimaging processing tools provide convenient and systematic methods for extracting features from brain magnetic resonance imaging scans. One tool, FreeSurfer, provides an easy-to-use pipeline to extract cortical and subcortical morphometric measures. There have been over 25 stable releases of FreeSurfer, with different versions used across published works. The reliability and compatibility of regional morphometric metrics derived from the most recent version releases have yet to be empirically assessed. Here, we used test-retest data from three public data sets to determine within-version reliability and between-version compatibility across 42 regional outputs from FreeSurfer versions 7.1, 6.0, and 5.3. Cortical thickness from v7.1 was less compatible with that of older versions, particularly along the cingulate gyrus, where the lowest version compatibility was observed (intraclass correlation coefficient 0.37-0.61). Surface area of the temporal pole, frontal pole, and medial orbitofrontal cortex, also showed low to moderate version compatibility. We confirm low compatibility between v6.0 and v5.3 of pallidum and putamen volumes, while those from v7.1 were compatible with v6.0. Replication in an independent sample showed largely similar results for measures of surface area and subcortical volumes, but had lower overall regional thickness reliability and compatibility. Batch effect correction may adjust for some inter-version effects when most sites are run with one version, but results vary when more sites are run with different versions. Age associations in a quality controlled independent sample (N = 106) revealed version differences in results of downstream statistical analysis. We provide a reference to highlight the regional metrics that may yield recent version-related inconsistencies in published findings. An interactive viewer is provided at http://data.brainescience.org/Freesurfer_Reliability/.

Differences in gray matter volumes of subcortical nuclei between major depressive disorder with and without persistent depressive disorder

OBJECTIVE

This study aimed to compare the differences in gray matter volumes (GMVs) of subcortical nuclei between major depressive disorder (MDD) patients with and without persistent depressive disorder (PDD) at long-term follow-up.

METHODS

114 and 94 subjects with MDD, including 48 and 41 with comorbid PDD, were enrolled to undergo high-resolution T1-weighted imaging at first (FIP) and second (three years later, SIP) investigation points, respectively. FreeSurfer was used to extract the GMVs of seven subcortical nuclei, and Generalized Estimating Equation models were employed to estimate the differences in GMVs of subcortical nuclei between the two subgroups.

RESULTS

The PDD subgroup had a significantly greater depressive severity and a higher percentage of patients undergoing pharmacotherapy at the FIP as compared with the non-PDD subgroup. These differences became insignificant at the SIP. The PDD subgroup had a significantly (p < 0.003) smaller GMV in the right putamen at the SIP and in the right nucleus accumbens (NAc) at the FIP and SIP as compared with the non-PDD subgroup. After controlling for clinical variables, PDD was independently associated with smaller GMVs in the right putamen and NAc.

LIMITATIONS

Imaging was not performed at baseline and pharmacotherapy was not controlled at the FIP and SIP.

CONCLUSIONS

MDD with PDD was associated with smaller GMVs in the right putamen and NAc as compared with MDD without PDD. Whether the two regions are biomarkers related to a poor prognosis and the chronicity of depression requires further study.

Polygenic score for Alzheimer’s disease identifies differential atrophy in hippocampal subfield volumes

Hippocampal subfield atrophy is a prime structural change in the brain, associated with cognitive aging and neurodegenerative diseases such as Alzheimer’s disease. Recent developments in genome-wide association studies (GWAS) have identified genetic loci that characterize the risk of hippocampal volume loss based on the processes of normal and abnormal aging. Polygenic risk scores are the genetic proxies mimicking the genetic role of the pre-existing vulnerabilities of the underlying mechanisms influencing these changes. Discriminating the genetic predispositions of hippocampal subfield atrophy between cognitive aging and neurodegenerative diseases will be helpful in understanding the disease etiology. In this study, we evaluated the polygenic risk of Alzheimer’s disease (AD PGRS) for hippocampal subfield atrophy in 1,086 individuals (319 cognitively normal (CN), 591 mild cognitively impaired (MCI), and 176 Alzheimer’s disease dementia (ADD)). Our results showed a stronger association of AD PGRS effect on the left hemisphere than on the right hemisphere for all the hippocampal subfield volumes in a mixed clinical population (CN+MCI+ADD). The subfields CA1, CA4, hippocampal tail, subiculum, presubiculum, molecular layer, GC-ML-DG, and HATA showed stronger AD PGRS associations with the MCI+ADD group than with the CN group. The subfields CA3, parasubiculum, and fimbria showed moderately higher AD PGRS associations with the MCI+ADD group than with the CN group. Our findings suggest that the eight subfield regions, which were strongly associated with AD PGRS are likely involved in the early stage ADD and a specific focus on the left hemisphere could enhance the early prediction of ADD.

Mothers’ Attachment Representations and Children’s Brain Structure

Ample research demonstrates that parents’ experience-based mental representations of attachment—cognitive models of close relationships—relate to their children’s social-emotional development. However, no research to date has examined how parents’ attachment representations relate to another crucial domain of children’s development: brain development. The present study is the first to integrate the separate literatures on attachment and developmental social cognitive neuroscience to examine the link between mothers’ attachment representations and 3- to 8-year-old children’s brain structure. We hypothesized that mothers’ attachment representations would relate to individual differences in children’s brain structures involved in stress regulation—specifically, amygdala and hippocampal volumes—in part via mothers’ responses to children’s distress. We assessed 52 mothers’ attachment representations (secure base script knowledge on the Attachment Script Assessment and self-reported attachment avoidance and anxiety on the Experiences in Close Relationships scale) and children’s brain structure. Mothers’ secure base script knowledge was significantly related to children’s smaller left amygdala volume but was unrelated to hippocampal volume; we found no indirect links via maternal responses to children’s distress. Exploratory analyses showed associations between mothers’ attachment representations and white matter and thalamus volumes. Together, these preliminary results suggest that mothers’ attachment representations may be linked to the development of children’s neural circuitry related to stress regulation.

Qualitative and Quantitative Comparison of Hippocampal Volumetric Software Applications: Do All Roads Lead to Rome?

Brain volumetric software is increasingly suggested for clinical routine. The present study quantifies the agreement across different software applications. Ten cases with and ten gender- and age-adjusted healthy controls without hippocampal atrophy (median age: 70; 25–75% range: 64–77 years and 74; 66–78 years) were retrospectively selected from a previously published cohort of Alzheimer’s dementia patients and normal ageing controls. Hippocampal volumes were computed based on 3 Tesla T1-MPRAGE-sequences with FreeSurfer (FS), Statistical-Parametric-Mapping (SPM; Neuromorphometrics and Hammers atlases), Geodesic-Information-Flows (GIF), Similarity-and-Truth-Estimation-for-Propagated-Segmentations (STEPS), and Quantib™. MTA (medial temporal lobe atrophy) scores were manually rated. Volumetric measures of each individual were compared against the mean of all applications with intraclass correlation coefficients (ICC) and Bland–Altman plots. Comparing against the mean of all methods, moderate to low agreement was present considering categorization of hippocampal volumes into quartiles. ICCs ranged noticeably between applications (left hippocampus (LH): from 0.42 (STEPS) to 0.88 (FS); right hippocampus (RH): from 0.36 (Quantib™) to 0.86 (FS). Mean differences between individual methods and the mean of all methods [mm³] were considerable (LH: FS −209, SPM-Neuromorphometrics −820; SPM-Hammers −1474; Quantib™ −680; GIF 891; STEPS 2218; RH: FS −232, SPM-Neuromorphometrics −745; SPM-Hammers −1547; Quantib™ −723; GIF 982; STEPS 2188). In this clinically relevant sample size with large spread in data ranging from normal aging to severe atrophy, hippocampal volumes derived by well-accepted applications were quantitatively different. Thus, interchangeable use is not recommended.

Different FreeSurfer versions might generate different statistical outcomes in case-control comparison studies

PURPOSE

Neuroimaging pipelines have long been known to generate mildly differing results depending on various factors, including software version. While considered generally acceptable and within the margin of reasonable error, little is known about their effect in common research scenarios such as inter-group comparisons between healthy controls and various pathological conditions. The aim of the presented study was to explore the differences in the inferences and statistical significances in a model situation comparing volumetric parameters between healthy controls and type 1 diabetes patients using various FreeSurfer versions.

METHODS

T1- and T2-weighted structural scans of healthy controls and type 1 diabetes patients were processed with FreeSurfer 5.3, FreeSurfer 5.3 HCP, FreeSurfer 6.0 and FreeSurfer 7.1, followed by inter-group statistical comparison using outputs of individual FreeSurfer versions.

RESULTS

Worryingly, FreeSurfer 5.3 detected both cortical and subcortical volume differences out of the preselected regions of interest, but newer versions such as FreeSurfer 5.3 HCP and FreeSurfer 6.0 reported only subcortical differences of lower magnitude and FreeSurfer 7.1 failed to find any statistically significant inter-group differences.

CONCLUSION

Since group averages of individual FreeSurfer versions closely matched, in keeping with previous literature, the main origin of this disparity seemed to lie in substantially higher within-group variability in the model pathological condition. Ergo, until validation in common research scenarios as case-control comparison studies is included into the development process of new software suites, confirmatory analyses utilising a similar software based on analogous, but not fully equivalent principles, might be considered as supplement to careful quality control.

Estimates of brain age for gray matter and white matter in younger and older adults: Insights into human intelligence

Aging entails a multifaceted complex of changes in macro- and micro-structural properties of human brain gray matter (GM) and white matter (WM) tissues, as well as in intellectual abilities. To better capture tissue-specific brain aging, we combined volume and distribution properties of diffusivity indices to derive subject-specific age scores for each tissue. We compared age-related variance between younger and older adults for GM and WM age scores, and tested whether tissue-specific age scores could explain different effects of aging on fluid (Gf) and crystalized (Gc) intelligence in younger and older adults. Chronological age was strongly associated with GM (R2 = 0.73) and WM (R2 = 0.57) age scores. The GM age score accounted for significantly more variance in chronological age in younger relative to older adults (p < 0.001), whereas the WM age score accounted for significantly more variance in chronological age in older compared to younger adults (p < 0.025). Consistent with existing literature, younger adults outperformed older adults in Gf while older adults outperformed younger adults in Gc. The GM age score was negatively associated with Gf in younger adults (p < 0.02), whereas the WM age score was negatively associated with Gc in older adults (p < 0.02). Our results provide evidence for differences in the effects of age on GM and WM in younger versus older adults that may contribute to age-related differences in Gf and Gc.

Neuroimaging PheWAS (Phenome-Wide Association Study): A Free Cloud-Computing Platform for Big-Data, Brain-Wide Imaging Association Studies

Large-scale, case-control genome-wide association studies (GWASs) have revealed genetic variations associated with diverse neurological and psychiatric disorders. Recent advances in neuroimaging and genomic databases of large healthy and diseased cohorts have empowered studies to characterize effects of the discovered genetic factors on brain structure and function, implicating neural pathways and genetic mechanisms in the underlying biology. However, the unprecedented scale and complexity of the imaging and genomic data requires new advanced biomedical data science tools to manage, process and analyze the data. In this work, we introduce Neuroimaging PheWAS (phenome-wide association study): a web-based system for searching over a wide variety of brain-wide imaging phenotypes to discover true system-level gene-brain relationships using a unified genotype-to-phenotype strategy. This design features a user-friendly graphical user interface (GUI) for anonymous data uploading, study definition and management, and interactive result visualizations as well as a cloud-based computational infrastructure and multiple state-of-art methods for statistical association analysis and multiple comparison correction. We demonstrated the potential of Neuroimaging PheWAS with a case study analyzing the influences of the apolipoprotein E (APOE) gene on various brain morphological properties across the brain in the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort. Benchmark tests were performed to evaluate the system's performance using data from UK Biobank. The Neuroimaging PheWAS system is freely available. It simplifies the execution of PheWAS on neuroimaging data and provides an opportunity for imaging genetics studies to elucidate routes at play for specific genetic variants on diseases in the context of detailed imaging phenotypic data.

Brain Morphology Associated With Obsessive-Compulsive Symptoms in 2,551 Children From the General Population

OBJECTIVE

Obsessive-compulsive (OC) symptoms are common in the general population, but it is unclear whether subclinical OC symptoms and obsessive-compulsive disorder (OCD) are part of a neuroanatomical continuum. The goal of this study was to investigate the relation between OC symptoms and subcortical and cortical morphology in a population-based sample of children.

METHOD

The study included 2,551 participants, aged 9-12 years, from the population-based Generation R Study. OC symptoms were measured using the 7-item caregiver-rated Short Obsessive-Compulsive Disorder Screener (SOCS). Structural (3T) magnetic resonance imaging scans were processed using FreeSurfer to study the thalamus and other subcortical volumes, intracranial volume, vertexwise cortical thickness, and surface area. We used linear regression models to investigate the association between OC symptoms and brain morphology. Emulating case-control studies from the literature, we compared children scoring above the clinical cutoff of the SOCS (probable OCD cases, n = 164) with matched children without symptoms.

RESULTS

Children with probable OCD had larger thalami compared with the control group (d 0.16, p = .044). Vertexwise analysis showed a positive association between OC symptoms and thickness of the right inferior parietal cortex, which disappeared after adjusting for total behavioral problems. SOCS scores correlated negatively with intracranial volume (B = -2444, p = .038).

CONCLUSION

Children with probable OCD showed thalamus alterations similar to those previously reported in unmedicated children with OCD. OC symptoms showed a stronger association with total intracranial volume than regional brain measures. Longitudinal studies are needed to further elucidate similarities and distinctions between neural correlates of subclinical and clinical OC symptoms.