Reproducibility of Brain MRI Segmentation Algorithms: Empirical Comparison of Local MAP PSTAPLE, FreeSurfer, and FSL-FIRST

Exploring Brainstem Structural Abnormalities: Potential Biomarkers for Panic Disorder

Panic disorder (PD), characterized by recurrent and intense panic attacks, presents a complex interplay between psychological and neurobiological factors. Although the amygdala and hippocampus have been studied extensively in the context of PD, the brainstem's involvement remains relatively underexplored. This study aims to address this gap by examining structural abnormalities within specific brainstem regions, including the medulla, pons, and midbrain. The study sample population comprised twenty-one adult patients diagnosed with PD and an age-gender-education-matched control group. Utilizing rigorous inclusion and exclusion criteria, confounding factors related to comorbid psychiatric conditions and brain structure abnormalities were minimized. Our findings revealed a significant reduction in medulla volume among PD patients, a finding that persisted even after correcting for individual differences in total intracranial volume. The medulla's role in cardiovascular regulation and autonomic function, coupled with its involvement in fear responses, underscores its potential significance in the pathophysiology of PD. This study elucidates the medulla's structural abnormalities as a potential biomarker for PD. Understanding the role of the brainstem in PD could pave the way for more targeted and effective interventions for this condition.

Quantification of Thalamic Atrophy in MS: From the Multicenter Italian Neuroimaging Network Initiative Data Set to Clinical Application

BACKGROUND AND PURPOSE

Thalamic atrophy occurs from the earliest phases of MS; however, this measure is not included in clinical practice. Our purpose was to obtain a reliable segmentation of the thalamus in MS by comparing existing automatic methods cross-sectionally and longitudinally.

MATERIALS AND METHODS

MR images of 141 patients with relapsing-remitting MS (mean age, 38 years; range, 19-58 years; 95 women) and 69 healthy controls (mean age, 36 years; range, 22-69 years; 47 women) were retrieved from the Italian Neuroimaging Network Initiative repository: T1WI, T2WI, and DWI at baseline and after 1 year (136 patients, 31 healthy controls). Three segmentation software programs (FSL-FIRST, FSL-MIST, FreeSurfer) were compared. At baseline, agreement among pipelines, correlations with age, disease duration, clinical score, and T2-hyperintense lesion volume were evaluated. Effect sizes in differentiating patients and controls were assessed cross-sectionally and longitudinally. Variability of longitudinal changes in controls and sample sizes were assessed. False discovery rate-adjusted P < .05 was considered significant.

RESULTS

At baseline, FSL-FIRST and FSL-MIST showed the highest agreement in the results of thalamic volume (R = 0.87, P < .001), with the highest effect size for FSL-MIST (Cohen d = 1.11); correlations with demographic and clinical variables were comparable for all software. Longitudinally, FSL-MIST showed the lowest variability in estimating thalamic volume changes for healthy controls (SD = 1.07%), the highest effect size (Cohen d = 0.44), and the smallest sample size at 80% power level (15 subjects per group).

CONCLUSIONS

Multimodal segmentation by FSL-MIST increased the robustness of the results with better capability to detect small variations in thalamic volumes.

Reduced brainstem volume is associated with mobility impairments in youth with cerebral palsy

BACKGROUND

Persons with cerebral palsy (CP) have impaired mobility that has been attributed to changes in structure and function within the nervous system. The brainstem is a region that plays a critical role in mobility by connecting the cortex and cerebellum to the spinal cord, yet this region has been largely unstudied in persons with CP.

RESEARCH QUESTION

We used high-resolution structural MRI and biomechanical analyses to examine whether the volume of the whole brainstem and its constituent elements are altered in CP and if these alterations relate to the mobility impairments within this population.

METHODS

A cohort study was conducted to assess the volume of the whole brainstem, pons, midbrain, medulla, and superior cerebellar peduncle in a cohort of persons with CP (N = 26; Age = 16.3 ± 1.0 years; GMFCS levels I-IV, Females = 12) and a cohort of neurotypical (NT) controls (N = 38; Age = 14.3 ± 0.4 years, Females = 14) using structural MR imaging of the brainstem. Outside the scanner, a digital mat was used to quantify the spatiotemporal gait biomechanics of these individuals.

RESULTS

We found a significant decrease in volume of the total brainstem, midbrain, and pons in persons with CP in comparison to the NT controls. Furthermore, we found that the altered volumes were related to reduced gait velocity and step length.

SIGNIFICANCE

The structural changes in the brainstems of persons with CP may contribute to the mobility impairments that are ubiquitous within this population.

A systematic review of repeatability and reproducibility studies of diffusion tensor imaging of cervical spinal cord

OBJECTIVES

Diffusion tensor imaging (DTI) techniques are being studied as a possible diagnostic and predictive tool for the evaluation of cervical spinal cord disease. This systematic review aims to evaluate the previous DTI studies that specifically investigated the repeatability and reproducibility of DTI in the cervical spinal cord.

METHODS AND MATERIALS

A search in the PubMed, Scopus, Web of Science and Ovid electronic databases was conducted for articles published between January 1990 and February 2022 that related to the repeatability and reproducibility of DTI in evaluating the cervical spinal cord using one of the following measurements: the intraclass correlation coefficient (ICC) and/or the coefficient of variation (CV), and/or Bland-Altman (BA) differences analysis methods. DTI studies that presented full statistical analysis of repeatability and/or reproducibility tests of the cervical spinal cord in peer-reviewed full-text publications published in journals were included. Articles that included at least one of the keywords within the titles or abstracts were identified. Additional full-text papers were found by searching the citations and reference lists of related articles. This review has followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidance. Risk of bias was evaluated with 13 criteria weighted toward methodological quality of reported studies using the QuADS assessment criteria. This assessment only included full-text articles written in English.

RESULTS

A total of 11 studies were included and assessed for different characteristics, including sample size,(3-34) re-test time interval (<1 h to >3 months), test-retest reproducibility scores and acquisition method. Six studies used ICC which ranged from poor (ICC<0.37) to excellent reproducibility (ICC 0.91-0.99). Four studies reported an overall CV lower than 40% for all DTI metrics. Three studies reported the Bland-Altman (BA) differences and reported a minimum percentage showing no strong differences between repeated measurements. Quantitative analysis was not undertaken due to heterogeneity of methods. Repeatability and reproducibility measures were generally found to be good.

CONCLUSION

This study revealed that the application of DTI and its related measures in a clinical setting in the assessment of cervical spinal cord changes is feasible and reproducible. However, cervical spinal cord DTI suffers from some existing limitations that prevent it from being routinely used in research and clinical settings.

ADVANCES IN KNOWLEDGE

DTI with its parametric maps provide broad evaluation of the tissue structure of axonal white matter and are being studied as a possible diagnostic and predictive tool for the assessment of cervical spinal cord (CSC) disease.

Reliability of brain atrophy measurements in multiple sclerosis using MRI: an assessment of six freely available software packages for cross-sectional analyses

PURPOSE

Volume measurement using MRI is important to assess brain atrophy in multiple sclerosis (MS). However, differences between scanners, acquisition protocols, and analysis software introduce unwanted variability of volumes. To quantify theses effects, we compared within-scanner repeatability and between-scanner reproducibility of three different MR scanners for six brain segmentation methods.

METHODS

Twenty-one people with MS underwent scanning and rescanning on three 3 T MR scanners (GE MR750, Philips Ingenuity, Toshiba Vantage Titan) to obtain 3D T1-weighted images. FreeSurfer, FSL, SAMSEG, FastSurfer, CAT-12, and SynthSeg were used to quantify brain, white matter and (deep) gray matter volumes both from lesion-filled and non-lesion-filled 3D T1-weighted images. We used intra-class correlation coefficient (ICC) to quantify agreement; repeated-measures ANOVA to analyze systematic differences; and variance component analysis to quantify the standard error of measurement (SEM) and smallest detectable change (SDC).

RESULTS

For all six software, both between-scanner agreement (ICCs ranging 0.4-1) and within-scanner agreement (ICC range: 0.6-1) were typically good, and good to excellent (ICC > 0.7) for large structures. No clear differences were found between filled and non-filled images. However, gray and white matter volumes did differ systematically between scanners for all software (p < 0.05). Variance component analysis yielded within-scanner SDC ranging from 1.02% (SAMSEG, whole-brain) to 14.55% (FreeSurfer, CSF); and between-scanner SDC ranging from 4.83% (SynthSeg, thalamus) to 29.25% (CAT12, thalamus).

CONCLUSION

Volume measurements of brain, GM and WM showed high repeatability, and high reproducibility despite substantial differences between scanners. Smallest detectable change was high, especially between different scanners, which hampers the clinical implementation of atrophy measurements.

Reproducibility and Reliability of Computing Models in Segmentation and Volumetric Measurement of Brain

Background

Segmentation and morphometric measurement of brain tissue and regions from non-invasive magnetic resonance images have clinical and research applications. Several software tools and models have been developed by different research groups which are increasingly used for segmentation and morphometric measurements. Variability in results has been observed in the imaging data processed with different neuroimaging pipelines which have increased the focus on standardization.

Purpose

The availability of several tools and models for brain morphometry poses challenges as an analysis done on the same set of data using different sets of tools and pipelines may result in different results and interpretations and there is a need for understanding the reliability and accuracy of such models.

Methods

T1-weighted (T1-w) brain volumes from the publicly available OASIS3 dataset have been analysed using recent versions of FreeSurfer, FSL-FAST, CAT12, and ANTs pipelines. grey matter (GM), white matter (WM), and estimated total intracranial volume (eTIV) have been extracted and compared for inter-method variability and accuracy.

Results

All four methods are consistent and strongly reproducible in their measurement across subjects however there is a significant degree of variability between these methods.

Conclusion

CAT12 and FreeSurfer methods have the highest degree of agreement in tissue class segmentation and are most reproducible compared to others.

Development and validation of deep learning-based automatic brain segmentation for East Asians: A comparison with Freesurfer

Purpose

To develop and validate deep learning-based automatic brain segmentation for East Asians with comparison to data for healthy controls from Freesurfer based on a ground truth.

Methods

A total of 30 healthy participants were enrolled and underwent T1-weighted magnetic resonance imaging (MRI) using a 3-tesla MRI system. Our Neuro I software was developed based on a three-dimensional convolutional neural networks (CNNs)-based, deep-learning algorithm, which was trained using data for 776 healthy Koreans with normal cognition. Dice coefficient (D) was calculated for each brain segment and compared with control data by paired t-test. The inter-method reliability was assessed by intraclass correlation coefficient (ICC) and effect size. Pearson correlation analysis was applied to assess the relationship between D values for each method and participant ages.

Results

The D values obtained from Freesurfer (ver6.0) were significantly lower than those from Neuro I. The histogram of the Freesurfer results showed remarkable differences in the distribution of D values from Neuro I. Overall, D values obtained by Freesurfer and Neuro I showed positive correlations, but the slopes and intercepts were significantly different. It was showed the largest effect sizes ranged 1.07-3.22, and ICC also showed significantly poor to moderate correlations between the two methods (0.498 ≤ ICC ≤ 0.688). For Neuro I, D values resulted in reduced residuals when fitting data to a line of best fit, and indicated consistent values corresponding to each age, even in young and older adults.

Conclusion

Freesurfer and Neuro I were not equivalent when compared to a ground truth, where Neuro I exhibited higher performance. We suggest that Neuro I is a useful alternative for the assessment of the brain volume.

Comparative validation of AI and non-AI methods in MRI volumetry to diagnose Parkinsonian syndromes

Automated segmentation and volumetry of brain magnetic resonance imaging (MRI) scans are essential for the diagnosis of Parkinson's disease (PD) and Parkinson's plus syndromes (P-plus). To enhance the diagnostic performance, we adopt deep learning (DL) models in brain MRI segmentation and compared their performance with the gold-standard non-DL method. We collected brain MRI scans of healthy controls ([Formula: see text]) and patients with PD ([Formula: see text]), multiple systemic atrophy ([Formula: see text]), and progressive supranuclear palsy ([Formula: see text]) at Samsung Medical Center from January 2017 to December 2020. Using the gold-standard non-DL model, FreeSurfer (FS), we segmented six brain structures: midbrain, pons, caudate, putamen, pallidum, and third ventricle, and considered them as annotated data for DL models, the representative convolutional neural network (CNN) and vision transformer (ViT)-based models. Dice scores and the area under the curve (AUC) for differentiating normal, PD, and P-plus cases were calculated to determine the measure to which FS performance can be reproduced as-is while increasing speed by the DL approaches. The segmentation times of CNN and ViT for the six brain structures per patient were 51.26 ± 2.50 and 1101.82 ± 22.31 s, respectively, being 14 to 300 times faster than FS (15,735 ± 1.07 s). Dice scores of both DL models were sufficiently high (> 0.85) so their AUCs for disease classification were not inferior to that of FS. For classification of normal vs. P-plus and PD vs. P-plus (except multiple systemic atrophy - Parkinsonian type) based on all brain parts, the DL models and FS showed AUCs above 0.8, demonstrating the clinical value of DL models in addition to FS. DL significantly reduces the analysis time without compromising the performance of brain segmentation and differential diagnosis. Our findings may contribute to the adoption of DL brain MRI segmentation in clinical settings and advance brain research.

Reduced Brainstem Volume is Associated with Mobility Impairments in Youth with Cerebral Palsy

Persons with cerebral palsy (CP) have impaired mobility that has been attributed to changes in structure and function within the nervous system. The brainstem is a region that plays a critical role in locomotion by connecting the cortex and cerebellum to the spinal cord, yet this region has been largely unstudied in persons with CP. The objective of this investigation was to use high-resolution structural MRI and biomechanical analyses to examine whether the volume of the whole brainstem and its constituent elements are altered in CP, and if these alterations relate to the mobility impairments within this population. We assessed the volume of the pons, midbrain, medulla, and superior cerebellar peduncle (SCP) in a cohort of persons with CP (N = 26; Age = 16.3 ± 1.0 yrs; GMFCS levels I-IV, Females = 12) and a cohort of neurotypical (NT) controls (N = 38; Age = 14.3 ± 0.4 yrs, Females = 14) using structural MR imaging of the brainstem. Outside the scanner, a digital mat was used to quantify the spatiotemporal gait biomechanics of these individuals. Our MRI results revealed that there was a significant decrease in volume of the total brainstem, midbrain, and pons in persons with CP in comparison to the NT controls. Furthermore, we found that the altered volumes were related to reduced gait velocity and step length. These results suggest that there are structural changes in the brainstems of persons with CP that may contribute to the mobility impairments that are ubiquitous within this population.

Volumetric assessment and longitudinal changes of subcortical structures in formalinized Beagle brains

High field MRI is an advanced technique for diagnostic and research purposes on animal models, such as the Beagle dog. In this context, studies on neuroscience applications, e.g. aging and neuro-pathologies, are currently increasing. This led to a need for reference values, in terms of volumetric assessment, for the structures typically involved. Nowadays, several canine brain MRI atlases have been provided. However, no reports are available regarding the measurements’ reproducibility and little is known about the effect of formalin on MRI segmentation. Here, we assessed the segmentation variability of selected structures among operators (two operators segmented the same data) in a sample of 11 Beagle dogs. Then, we analyzed, for one Beagle dog, the longitudinal volumetric changes of these structures. We considered four conditions: in vivo, post mortem (after euthanasia), ex vivo (brain extracted and studied after 1 month in formalin, and after 12 months). The MRI data were collected with a 3 T scanner. Our findings suggest that the segmentation procedure was overall reproducible since only slight statistical differences were detected. In the post mortem/ ex vivo comparison, most structures showed a higher contrast, thereby leading to greater reproducibility between operators. We observed a net increase in the volume of the studied structures. This could be justified by the intrinsic relaxation time changes observed because of the formalin fixation. This led to an improvement in brain structure visualization and segmentation. To conclude, MRI-based segmentation seems to be a useful and accurate tool that allows longitudinal studies on formalin-fixed brains.

How segmentation methods affect hippocampal radiomic feature accuracy in Alzheimer's disease analysis?

OBJECTIVES

Hippocampal radiomic features (HRFs) can serve as biomarkers in Alzheimer's disease (AD). However, how different hippocampal segmentation methods affect HRFs in AD is still unknown. The aim of the study was to investigate how different segmentation methods affect HRF accuracy in AD analysis.

METHODS

A total of 1650 subjects were identified from the Alzheimer's Disease Neuroimaging Initiative database (ADNI). The mini-mental state examination (MMSE) and Alzheimer's disease assessment scale (ADAS-cog13) were also adopted. After calculating the HRFs of intensity, shape, and textural features from each side of the hippocampus in structural magnetic resonance imaging (sMRI), the consistency of HRFs calculated from 7 different hippocampal segmentation methods was validated, and the performance of machine learning-based classification of AD vs. normal control (NC) adopting the different HRFs was also examined. Additional 571 subjects from the European DTI Study on Dementia database (EDSD) were to validate the consistency of results.

RESULTS

Between different segmentations, HRFs showed a high measurement consistency (R > 0.7), a high significant consistency between NC, mild cognitive impairment (MCI), and AD (T-value plot, R > 0.8), and consistent significant correlations between HRFs and MMSE/ADAS-cog13 (p < 0.05). The best NC vs. AD classification was obtained when the hippocampus was sufficiently segmented by primitive majority voting (threshold = 0.2). High consistent results were reproduced from independent EDSD cohort.

CONCLUSIONS

HRFs exhibited high consistency across different hippocampal segmentation methods, and the best performance in AD classification was obtained when HRFs were extracted by the naïve majority voting method with a more sufficient segmentation and relatively low hippocampus segmentation accuracy.

KEY POINTS

• The hippocampal radiomic features exhibited high measurement/statistical/clinical consistency across different hippocampal segmentation methods. • The best performance in AD classification was obtained when hippocampal radiomics were extracted by the naïve majority voting method with a more sufficient segmentation and relatively low hippocampus segmentation accuracy.

Thalamic volumetric abnormalities in type 1 diabetes mellitus and 'peripheral' neuropathy

We hypothesized that thalamic volumes of patients with type 1 diabetes mellitus (DM) and nonpainful diabetic peripheral neuropathy (DPN) would be reduced relative to thalamic volumes of patients with type 1 DM and painful DPN. We calculated the standardized thalamic volumetric difference between these groups in a pilot sample to obtain a statistical power of 80% at a 5% significance level. Hence, we measured thalamic volumes from 15 patients with nonpainful DPN (10 women, mean age = 49 years, standard deviation [SD] = 11.5) and from 13 patients with painful DPN (8 women, mean age = 43 years, SD = 12.5) by using a manual segmentation approach. A volumetric difference of approximately 15% was found between the nonpainful (mean = 5072 mm³, SD = 528.1) and painful (mean = 5976 mm³, SD = 643.1) DPN groups (P < 0.001). Curiously, a volumetric difference between the left (mean = 5198 mm³, SD = 495.0) and the right (mean = 4946 mm³, SD = 590.6) thalamus was also found in patients with nonpainful DPN (P < 0.01), but not in patients with painful DPN (P = 0.97). Patients with nonpainful DPN have lower thalamic volumes than those with painful DPN, especially in the right thalamus.

Subcortical and hippocampal brain segmentation in 5-year-old children: validation of FSL-FIRST and FreeSurfer against manual segmentation

Developing accurate subcortical volumetric quantification tools is crucial for neurodevelopmental studies, as they could reduce the need for challenging and time‐consuming manual segmentation. In this study, the accuracy of two automated segmentation tools, FSL‐FIRST (with three different boundary correction settings) and FreeSurfer, were compared against manual segmentation of the hippocampus and subcortical nuclei, including the amygdala, thalamus, putamen, globus pallidus, caudate and nucleus accumbens, using volumetric and correlation analyses in 80 5‐year‐olds.

Both FSL‐FIRST and FreeSurfer overestimated the volume on all structures except the caudate, and the accuracy varied depending on the structure. Small structures such as the amygdala and nucleus accumbens, which are visually difficult to distinguish, produced significant overestimations and weaker correlations with all automated methods. Larger and more readily distinguishable structures such as the caudate and putamen produced notably lower overestimations and stronger correlations. Overall, the segmentations performed by FSL‐FIRST's default pipeline were the most accurate, whereas FreeSurfer's results were weaker across the structures.

In line with prior studies, the accuracy of automated segmentation tools was imperfect with respect to manually defined structures. However, apart from amygdala and nucleus accumbens, FSL‐FIRST's agreement could be considered satisfactory (Pearson correlation > 0.74, intraclass correlation coefficient (ICC) > 0.68 and Dice score coefficient (DSC) > 0.87) with highest values for the striatal structures (putamen, globus pallidus, caudate) (Pearson correlation > 0.77, ICC > 0.87 and DSC > 0.88, respectively). Overall, automated segmentation tools do not always provide satisfactory results, and careful visual inspection of the automated segmentations is strongly advised.

Amygdala and subregion volumes are associated with photoperiod and seasonal depressive symptoms: A cross-sectional study in the UK Biobank cohort

Although seasonal changes in amygdala volume have been demonstrated in animals, seasonal differences in human amygdala subregion volumes have yet to be investigated. Amygdala volume has also been linked to depressed mood. Therefore, we hypothesised that differences in photoperiod would predict differences in amygdala or subregion volumes and that this association would be linked to depressed mood. 10,033 participants ranging in age from 45 to 79 years were scanned by MRI in a single location. Amygdala subregion volumes were obtained using automated processing and segmentation algorithms. A mediation analysis tested whether amygdala volume mediated the relationship between photoperiod and mood. Photoperiod was positively associated with total amygdala volume (p < .001). Multivariate (GLM) analyses revealed significant effects of photoperiod across all amygdala subregion volumes for both hemispheres (p < .001). Post hoc univariate regression analyses revealed significant associations of photoperiod with each amygdala subregion volume (p < .001). PLS showed the highest loadings of amygdala subregions in lateral nucleus, ABN, basal nucleus, CAT, PLN, AAA, central nucleus, cortical nucleus and medial nucleus for left hemisphere and ABN, lateral nucleus, CAT, PLN, cortical nucleus, AAA, central nucleus and medial nucleus for right hemisphere. There were no significant associations between photoperiod and mood nor between mood scores and amygdala volumes, and due to the lack of these associations, the mediation hypothesis was not supported. This study is the first to demonstrate an association between photoperiod and amygdala volume. These findings add to the evidence supporting the role of photoperiod on brain structural plasticity.

Hippocampal structural alterations in early-stage psychosis: Specificity and relationship to clinical outcomes

•

Individuals with early-stage psychosis show reduced hippocampal volumes.

•

FEP show bilateral and widespread changes, while left hemisphere is affected in CHR-P.

•

However, hippocampal changes do not show a relationship with clinical outcomes.

Hippocampal dysfunctions are a core feature of schizophrenia, but conflicting evidence exists whether volumetric and morphological changes are present in early-stage psychosis and to what extent these deficits are related to clinical trajectories. In this study, we recruited individuals at clinical high risk for psychosis (CHR-P) (n = 108), patients with a first episode of psychosis (FEP) (n = 37), healthy controls (HC) (n = 70) as well as a psychiatric control group with substance abuse and affective disorders (CHR-N: n = 38). MRI-data at baseline were obtained and volumetric as well as vertex analyses of the hippocampus were carried out. Moreover, volumetric changes were examined in the amygdala, caudate, nucleus accumbens, pallidum, putamen and thalamus. In addition, we obtained follow-up functional and symptomatic assessments in CHR-P individuals to examine the question whether anatomical deficits at baseline predicted clinical trajectories. Our results show that the hippocampus is the only structure showing significant volumetric decrease in early-stage psychosis, with FEPs showing significantly smaller hippocampal volumes bilaterally alongside widespread shape changes in the vertex analysis. For the CHR-P group, volumetric decreases were confined to the left hippocampus. However, hippocampal alterations in the CHR-P group were not robustly associated with clinical outcomes, including the persistence of attenuated psychotic symptoms and functional trajectories. Accordingly, our findings highlight that dysfunctions in hippocampal anatomy are an important feature of early-stage psychosis which may, however, not be related to clinical outcomes in CHR-P participants.

Rapid processing and quantitative evaluation of structural brain scans for adaptive multimodal imaging

Current neuroimaging acquisition and processing approaches tend to be optimised for quality rather than speed. However, rapid acquisition and processing of neuroimaging data can lead to novel neuroimaging paradigms, such as adaptive acquisition, where rapidly processed data is used to inform subsequent image acquisition steps. Here we first evaluate the impact of several processing steps on the processing time and quality of registration of manually labelled T₁ -weighted MRI scans. Subsequently, we apply the selected rapid processing pipeline both to rapidly acquired multicontrast EPImix scans of 95 participants (which include T₁ -FLAIR, T₂ , T₂ *, T₂ -FLAIR, DWI and ADC contrasts, acquired in ~1 min), as well as to slower, more standard single-contrast T₁ -weighted scans of a subset of 66 participants. We quantify the correspondence between EPImix T₁ -FLAIR and single-contrast T₁ -weighted scans, using correlations between voxels and regions of interest across participants, measures of within- and between-participant identifiability as well as regional structural covariance networks. Furthermore, we explore the use of EPImix for the rapid construction of morphometric similarity networks. Finally, we quantify the reliability of EPImix-derived data using test-retest scans of 10 participants. Our results demonstrate that quantitative information can be derived from a neuroimaging scan acquired and processed within minutes, which could further be used to implement adaptive multimodal imaging and tailor neuroimaging examinations to individual patients.

Structural Brain Network Reproducibility: Influence of Different Diffusion Acquisition and Tractography Reconstruction Schemes on Graph Metrics

Background: Graph metrics of structural brain networks demonstrate to be a powerful tool for investigating brain topology at a large scale. However, the variability of the results related to applying different magnetic resonance acquisition schemes and tractography reconstruction techniques is not fully characterized. Materials and Methods: The present work aims to evaluate the influence of different combinations of diffusion acquisition schemes (single and multishell), diffusion models (tensor and spherical deconvolution), and tractography reconstruction approaches (deterministic and probabilistic) on the reproducibility of graph metrics derived from structural connectome on test/retest (TRT) data released by the Human Connectome Project. From each implemented experimental setup, both global and local graph metrics were evaluated and their reproducibility was estimated by the intraclass correlation coefficient (ICC). Moreover, the percentage relative standard deviation (pRSD) from the ICC values of local graph metrics was calculated to quantify how much the reproducibility varied across nodes within each experimental setup. Results: The presented results show that different combinations of diffusion acquisition schemes, diffusion models, and tractography algorithms can strongly affect the reproducibility of global and local graph metrics. The combination of constrained spherical deconvolution (CSD) and deterministic tractography gave generally high reproducibility (ICCs >0.75) and lowest pRSD for the considered graph metrics, meanwhile probabilistic CSD with a high b-value returned the highest reproducibility. Notably, the introduction of streamline selection filters on CSD can substantially affect the reproducibility. Discussion: This work demonstrates that the TRT reproducibility of graph metrics is generally high but can vary substantially with different combinations of acquisition and reconstruction schemes. Impact statement This work demonstrates the influence of different diffusion acquisition schemes, diffusion models, and tractography reconstruction approaches on the reproducibility of graph metrics derived from structural connectome. The presented findings impact on the choice of both acquisition protocol and processing pipeline for topological analyses to produce reproducible measurements for brain network studies.

MRI-based thalamic volumetry in multiple sclerosis using FSL-FIRST: Systematic assessment of common error modes

BACKGROUND AND PURPOSE

FSL's FMRIB's Integrated Registration and Segmentation Tool (FSL-FIRST) is a widely used and well-validated tool. Automated thalamic segmentation is a common application and an important longitudinal measure for multiple sclerosis (MS). However, FSL-FIRST's algorithm is based on shape models derived from non-MS groups. As such, the present study sought to systematically assess common thalamic segmentation errors made by FSL-FIRST on MRIs from people with multiple sclerosis (PwMS).

METHODS

FSL-FIRST was applied to generate thalamic segmentation masks for 890 MR images in PwMS. Images and masks were reviewed systematically to classify and quantify errors, as well as associated anatomical variations and MRI abnormalities. For cases with overt errors (n = 362), thalamic masks were corrected and quantitative volumetric differences were calculated.

RESULTS

In the entire quantitative volumetric group, the mean volumetric error of FSL-FIRST was 2.74% (0.360 ml): among only corrected cases, the mean volumetric error was 6.79% (0.894 ml). The average percent volumetric error associated with seven error types, two anatomical variants, and motions artifacts are reported. Additional analyses showed that the presence of motion artifacts or anatomical variations significantly increased the probability of error (χ²  = 18.14, p < .01 and χ²  = 64.89, p < .001, respectively). Finally, thalamus volume error was negatively associated with degree of atrophy, such that smaller thalami were systematically overestimated (r = -.28, p < .001).

CONCLUSIONS

In PwMS, FSL-FIRST thalamic segmentation miscalculates thalamic volumetry in a predictable fashion, and may be biased to overestimate highly atrophic thalami. As such, it is recommended that segmentations be reviewed and corrected manually when appropriate for specific studies.

Brainstem Correlates of Pathological Laughter and Crying Frequency in ALS

Pseudobulbar affect is a disorder of emotional expression commonly observed in amyotrophic lateral sclerosis (ALS), presenting as episodes of involuntary laughter, or crying. The objective of the current study was to determine the association between frequency of pathological laughter and crying (PLC) episodes with clinical features, cognitive impairment, and brainstem pathology. Thirty-five sporadic ALS patients underwent neuropsychological assessment, with a subset also undergoing brain imaging. The Center for Neurological Study Lability Scale (CNS-LS) was used to screen for presence and severity of pseudobulbar affect (CNS-LS ≥ 13) and frequency of PLC episodes. Presence of pseudobulbar affect was significantly higher in bulbar onset ALS (p = 0.02). Frequency of PLC episodes was differentially associated with cognitive performance and brainstem integrity. Notably pathological laughter frequency, but not crying, showed a significant positive association with executive dysfunction on the Trail Making Test B-A (R² = 0.14, p = 0.04). Similarly, only pathological laughter frequency demonstrated a significant negative correlation with gray matter volume of the brainstem (R² = 0.46, p < 0.01), and mean fractional anisotropy of the superior cerebellar peduncles (left: R² = 0.44, p < 0.01; right: R² = 0.44, p < 0.01). Hierarchical regression indicated brainstem imaging in combination with site of symptom onset explained 73% of the variance in pathological laughter frequency in ALS. The current findings suggest emotional lability is underpinned by degeneration across distinct neural circuits, with brainstem integrity critical in the emergence of pathological laughter.

A structural brain network of genetic vulnerability to psychiatric illness

Psychiatry is undergoing a paradigm shift from the acceptance of distinct diagnoses to a representation of psychiatric illness that crosses diagnostic boundaries. How this transition is supported by a shared neurobiology remains largely unknown. In this study, we first identify single nucleotide polymorphisms (SNPs) associated with psychiatric disorders based on 136 genome-wide association studies. We then conduct a joint analysis of these SNPs and brain structural connectomes in 678 healthy children in the PING study. We discovered a strong, robust, and transdiagnostic mode of genome-connectome covariation which is positively and specifically correlated with genetic risk for psychiatric illness at the level of individual SNPs. Similarly, this mode is also significantly positively correlated with polygenic risk scores for schizophrenia, alcohol use disorder, major depressive disorder, a combined bipolar disorder-schizophrenia phenotype, and a broader cross-disorder phenotype, and significantly negatively correlated with a polygenic risk score for educational attainment. The resulting "vulnerability network" is shown to mediate the influence of genetic risks onto behaviors related to psychiatric vulnerability (e.g., marijuana, alcohol, and caffeine misuse, perceived stress, and impulsive behavior). Its anatomy overlaps with the default-mode network, with a network of cognitive control, and with the occipital cortex. These findings suggest that the brain vulnerability network represents an endophenotype funneling genetic risks for various psychiatric illnesses through a common neurobiological root. It may form part of the neural underpinning of the well-recognized but poorly explained overlap and comorbidity between psychiatric disorders.

Quality control strategies for brain MRI segmentation and parcellation: Practical approaches and recommendations - insights from the Maastricht study

Quality control of brain segmentation is a fundamental step to ensure data quality. Manual quality control strategies are the current gold standard, although these may be unfeasible for large neuroimaging samples. Several options for automated quality control have been proposed, providing potential time efficient and reproducible alternatives. However, those have never been compared side to side, which prevents consensus in the appropriate quality control strategy to use. This study aimed to elucidate the changes manual editing of brain segmentations produce in morphological estimates, and to analyze and compare the effects of different quality control strategies on the reduction of the measurement error. Structural brain MRI from 259 participants of The Maastricht Study were used. Morphological estimates were automatically extracted using FreeSurfer 6.0. Segmentations with inaccuracies were manually edited, and morphological estimates were compared before and after editing. In parallel, 12 quality control strategies were applied to the full sample. Those included: two manual strategies, in which images were visually inspected and either excluded or manually edited; five automated strategies, where outliers were excluded based on the tools "MRIQC" and "Qoala-T", and the metrics "morphological global measures", "Euler numbers" and "Contrast-to-Noise ratio"; and five semi-automated strategies, where the outliers detected through the mentioned tools and metrics were not excluded, but visually inspected and manually edited. In order to quantify the effects of each quality control strategy, the proportion of unexplained variance relative to the total variance was extracted after the application of each strategy, and the resulting differences compared. Manually editing brain surfaces produced particularly large changes in subcortical brain volumes and moderate changes in cortical surface area, thickness and hippocampal volumes. The performance of the quality control strategies depended on the morphological measure of interest. Overall, manual quality control strategies yielded the largest reduction in relative unexplained variance. The best performing automated alternatives were those based on Euler numbers and MRIQC scores. The exclusion of outliers based on global morphological measures produced an increase of relative unexplained variance. Manual quality control strategies are the most reliable solution for quality control of brain segmentation and parcellation. However, measures must be taken to prevent the subjectivity associated with these strategies. The detection of inaccurate segmentations based on Euler numbers or MRIQC provides a time efficient and reproducible alternative. The exclusion of outliers based on global morphological estimates must be avoided.

Replicability, Repeatability, and Long-term Reproducibility of Cerebellar Morphometry

To identify robust and reproducible methods of cerebellar morphometry that can be used in future large-scale structural MRI studies, we investigated the replicability, repeatability, and long-term reproducibility of three fully automated software tools: FreeSurfer, CEREbellum Segmentation (CERES), and automatic cerebellum anatomical parcellation using U-Net with locally constrained optimization (ACAPULCO). Replicability was defined as computational replicability, determined by comparing two analyses of the same high-resolution MRI data set performed with identical analysis software and computer hardware. Repeatability was determined by comparing the analyses of two MRI scans of the same participant taken during two independent MRI sessions on the same day for the Kirby-21 study. Long-term reproducibility was assessed by analyzing two MRI scans of the same participant in the longitudinal OASIS-2 study. We determined percent difference, the image intraclass correlation coefficient, the coefficient of variation, and the intraclass correlation coefficient between two analyses. Our results show that CERES and ACAPULCO use stochastic algorithms that result in surprisingly high differences between identical analyses for ACAPULCO and small differences for CERES. Changes between two consecutive scans from the Kirby-21 study were less than ± 5% in most cases for FreeSurfer and CERES (i.e., demonstrating high repeatability). As expected, long-term reproducibility was lower than repeatability for all software tools. In summary, CERES is an accurate, as demonstrated before, and reproducible tool for fully automated segmentation and parcellation of the cerebellum. We conclude with recommendations for the assessment of replicability, repeatability, and long-term reproducibility in future studies on cerebellar structure.

A Review of Publicly Available Automatic Brain Segmentation Methodologies, Machine Learning Models, Recent Advancements, and Their Comparison

BACKGROUND

The noninvasive study of the structure and functions of the brain using neuroimaging techniques is increasingly being used for its clinical and research perspective. The morphological and volumetric changes in several regions and structures of brains are associated with the prognosis of neurological disorders such as Alzheimer's disease, epilepsy, schizophrenia, etc. and the early identification of such changes can have huge clinical significance. The accurate segmentation of three-dimensional brain magnetic resonance images into tissue types (i.e., grey matter, white matter, cerebrospinal fluid) and brain structures, thus, has huge importance as they can act as early biomarkers. The manual segmentation though considered the "gold standard" is time-consuming, subjective, and not suitable for bigger neuroimaging studies. Several automatic segmentation tools and algorithms have been developed over the years; the machine learning models particularly those using deep convolutional neural network (CNN) architecture are increasingly being applied to improve the accuracy of automatic methods.

PURPOSE

The purpose of the study is to understand the current and emerging state of automatic segmentation tools, their comparison, machine learning models, their reliability, and shortcomings with an intent to focus on the development of improved methods and algorithms.

METHODS

The study focuses on the review of publicly available neuroimaging tools, their comparison, and emerging machine learning models particularly those based on CNN architecture developed and published during the last five years.

CONCLUSION

Several software tools developed by various research groups and made publicly available for automatic segmentation of the brain show variability in their results in several comparison studies and have not attained the level of reliability required for clinical studies. The machine learning models particularly three dimensional fully convolutional network models can provide a robust and efficient alternative with relation to publicly available tools but perform poorly on unseen datasets. The challenges related to training, computation cost, reproducibility, and validation across distinct scanning modalities for machine learning models need to be addressed.

Evaluation of Reproducibility of Brain Volumetry between Commercial Software, Inbrain and Established Research Purpose Method, FreeSurfer

Background and Purpose

We aimed to determine the intermethod reproducibility between the commercial software Inbrain (MIDAS IT) and the established research-purpose method FreeSurfer, as well as the effect of MRI resolution and the pathological condition of subjects on their intermethod reproducibility.

Methods

This study included 45 healthy volunteers and 85 patients with mild cognitive impairment (MCI). In 43 of the 85 patients with MCI, three-dimensional, T1-weighted MRI data were obtained at an in-plane resolution of 1.2 mm. The data of the remaining 42 patients with MCI and the healthy volunteers were obtained at an in-plane resolution of 1.0 mm. The within-subject coefficient of variation (CoV), intraclass correlation coefficient (ICC), and effect size were calculated, and means were compared using paired t-tests. The parameters obtained at 1.0-mm and 1.2-mm resolutions in patients with MCI were compared to evaluate the effect of the in-plane resolution on the intermethod reproducibility. The parameters obtained at a 1.0-mm in-plane resolution in patients with MCI and healthy volunteers were used to analyze the effect of subject condition on intermethod reproducibility.

Results

Overall the two methods showed excellent reproducibility across all regions of the brain (CoV=0.5–3.9, ICC=0.93 to >0.99). In the subgroup of healthy volunteers, the intermethod reliability was only good in some regions (frontal, temporal, cingulate, and insular). The intermethod reproducibility was better in the 1.0-mm group than the 1.2-mm group in all regions other than the nucleus accumbens.

Conclusions

Inbrain and FreeSurfer showed good-to-excellent intermethod reproducibility for volumetric measurements. Nevertheless, some noticeable differences were found based on subject condition, image resolution, and brain region.

Inter-scanner reproducibility of brain volumetry: influence of automated brain segmentation software

Background

The inter-scanner reproducibility of brain volumetry is important in multi-site neuroimaging studies, where the reliability of automated brain segmentation (ABS) tools plays an important role. This study aimed to evaluate the influence of ABS tools on the consistency and reproducibility of the quantified brain volumetry from different scanners.

Methods

We included fifteen healthy volunteers who were scanned with 3D isotropic brain T1-weighted sequence on three different 3.0 Tesla MRI scanners (GE, Siemens and Philips). For each individual, the time span between image acquisitions on different scanners was limited to 1 h. All the T1-weighted images were processed with FreeSurfer v6.0, FSL v5.0 and AccuBrain^® with default settings to obtain volumetry of brain tissues (e.g. gray matter) and substructures (e.g. basal ganglia structures) if available. Coefficient of variation (CV) was calculated to test inter-scanner variability in brain volumetry of various structures as quantified by these ABS tools.

Results

The mean inter-scanner CV values per brain structure among three MRI scanners ranged from 6.946 to 12.29% (mean, 9.577%) for FreeSurfer, 7.245 to 20.98% (mean, 12.60%) for FSL and 1.348 to 8.800% (mean value, 3.546%) for AccuBrain^®. In addition, AccuBrain^® and FreeSurfer achieved the lowest mean values of region-specific CV between GE and Siemens scanners (from 0.818 to 5.958% for AccuBrain^®, and from 0.903 to 7.977% for FreeSurfer), while FSL-FIRST had the lowest mean values of region-specific CV between GE and Philips scanners (from 2.603 to 16.310%). AccuBrain^® also had the lowest mean values of region-specific CV between Siemens and Philips scanners (from 1.138 to 6.615%).

Conclusion

There is a large discrepancy in the inter-scanner reproducibility of brain volumetry when using different processing software. Image acquisition protocols and selection of ABS tool for brain volumetry quantification have impact on the robustness of results in multi-site studies.

Whole Brain and Cranial Size Adjustments in Volumetric Brain Analyses of Sex- and Age-Related Trends

Our goal was to determine the influence of sex, age and the head/brain size on the compartmental brain volumes in the radiologically verified healthy population (96 subjects; 54 women and 42 men) from the Upper Silesia region in Poland. The MRI examinations were done using 3T Philips Achieva with the same T1-weighted and T2-weighted protocols. The image segmentation procedures were performed with SPM (Statistical Parameter Mapping) and FSL-FIRST software. The volumes of 14 subcortical structures for the left and right hemispheres and 4 overall volumes were calculated. The General Linear Models (GLM) analysis was used with and without the Total Brain Volume (TBV) and Intracranial Volume (ICV) parameters as the covariates to study the regional vs. global brain atrophy. After the ICV/TBV adjustments, the majority of sex differences in the specific volumes of interest (VOIs) revealed to be linked to the difference in the head/brain size parameters. The analysis also confirmed the significant effect of the aging process on the brain loss. After the TBV adjustment, the age- and sex-related volumetric trends for the gray and white matter volumes were observed: the negative age dependence of the gray matter volume is more pronounced in the males, while in case of the white matter the positive age-related trend in the female group is weaker. The local losses of the left caudate nucleus and the right thalamus are more advanced than the global brain atrophy. Different head-size correction strategies are not interchangeable and may yield various volumetric results, but when used together, facilitate studies on the regional dependencies inherent to a healthy, but aging, brain.

Brainstem volume mediates seasonal variation in depressive symptoms: A cross sectional study in the UK Biobank cohort

Seasonal differences in mood and depressive symptoms affect a large percentage of the general population, with seasonal affective disorder (SAD) representing the most common presentation. SAD affects up to 3% of the world’s population, and it tends to be more predominant in females than males. The brainstem has been shown to be affected by photoperiodic changes, and that longer photoperiods are associated with higher neuronal density and decreased depressive-like behaviours. We predict that longer photoperiod days are associated with larger brainstem volumes and lower depressive scores, and that brainstem volume mediates the seasonality of depressive symptoms. Participants (N = 9289, 51.8% females and 48.1% males) ranging in age from 44 to 79 years were scanned by MRI at a single location. Photoperiod was found to be negatively correlated with low mood and anhedonia in females while photoperiod was found to be positively correlated with brainstem volumes. In females, whole brainstem, pons and medulla volumes individually mediated the relationship between photoperiod and both anhedonia and low mood, while midbrain volume mediated the relationship between photoperiod and anhedonia. No mediation effects were seen in males. Our study extends the understanding of the neurobiological factors that contribute to the pathophysiology of seasonal mood variations.

Biomechanical modeling and computer simulation of the brain during neurosurgery

Computational biomechanics of the brain for neurosurgery is an emerging area of research recently gaining in importance and practical applications. This review paper presents the contributions of the Intelligent Systems for Medicine Laboratory and its collaborators to this field, discussing the modeling approaches adopted and the methods developed for obtaining the numerical solutions. We adopt a physics-based modeling approach and describe the brain deformation in mechanical terms (such as displacements, strains, and stresses), which can be computed using a biomechanical model, by solving a continuum mechanics problem. We present our modeling approaches related to geometry creation, boundary conditions, loading, and material properties. From the point of view of solution methods, we advocate the use of fully nonlinear modeling approaches, capable of capturing very large deformations and nonlinear material behavior. We discuss finite element and meshless domain discretization, the use of the total Lagrangian formulation of continuum mechanics, and explicit time integration for solving both time-accurate and steady-state problems. We present the methods developed for handling contacts and for warping 3D medical images using the results of our simulations. We present two examples to showcase these methods: brain shift estimation for image registration and brain deformation computation for neuronavigation in epilepsy treatment.

Reproducibility of Structural and Diffusion Tensor Imaging in the TACERN Multi-Center Study

BACKGROUND

Multi-site MRI studies are often necessary for recruiting sufficiently sized samples when studying rare conditions. However, they require pooling data from multiple scanners into a single data set, and therefore it is critical to evaluate the variability of quantitative MRI measures within and across scanners used in multi-site studies. The aim of this study was to evaluate the reproducibility of structural and diffusion weighted (DW) MRI measurements acquired on seven scanners at five medical centers as part of the Tuberous Sclerosis Complex Autism Center of Excellence Research Network (TACERN) multisite study.

METHODS

The American College of Radiology (ACR) phantom was imaged monthly to measure reproducibility of signal intensity and uniformity within and across seven 3T scanners from General Electric, Philips, and Siemens vendors. One healthy adult male volunteer was imaged repeatedly on all seven scanners under the TACERN structural and DW protocol (5 b = 0 s/mm2 and 30 b = 1000 s/mm2) over a period of 5 years (age 22-27 years). Reproducibility of inter- and intra-scanner brain segmentation volumes and diffusion tensor imaging metrics fractional anisotropy (FA) and mean diffusivity (MD) within white matter regions was quantified with coefficient of variation.

RESULTS

The American College of Radiology Phantom signal intensity and uniformity were similar across scanners and changed little over time, with a mean intra-scanner coefficient of variation of 3.6 and 1.8%, respectively. The mean inter- and intra-scanner coefficients of variation of brain structure volumes derived from T1-weighted (T1w) images of the human phantom were 3.3 and 1.1%, respectively. The mean inter- and intra-scanner coefficients of variation of FA in white matter regions were 4.5 and 2.5%, while the mean inter- and intra-scanner coefficients of variation of MD in white matter regions were 5.4 and 1.5%.

CONCLUSION

Our results suggest that volumetric and diffusion tensor imaging (DTI) measurements are highly reproducible between and within scanners and provide typical variation amplitudes that can be used as references to interpret future findings in the TACERN network.

Quantifying deep grey matter atrophy using automated segmentation approaches: A systematic review of structural MRI studies

The deep grey matter (DGM) nuclei of the brain play a crucial role in learning, behaviour, cognition, movement and memory. Although automated segmentation strategies can provide insight into the impact of multiple neurological conditions affecting these structures, such as Multiple Sclerosis (MS), Huntington's disease (HD), Alzheimer's disease (AD), Parkinson's disease (PD) and Cerebral Palsy (CP), there are a number of technical challenges limiting an accurate automated segmentation of the DGM. Namely, the insufficient contrast of T1 sequences to completely identify the boundaries of these structures, as well as the presence of iso-intense white matter lesions or extensive tissue loss caused by brain injury. Therefore in this systematic review, 269 eligible studies were analysed and compared to determine the optimal approaches for addressing these technical challenges. The automated approaches used among the reviewed studies fall into three broad categories, atlas-based approaches focusing on the accurate alignment of atlas priors, algorithmic approaches which utilise intensity information to a greater extent, and learning-based approaches that require an annotated training set. Studies that utilise freely available software packages such as FIRST, FreeSurfer and LesionTOADS were also eligible, and their performance compared. Overall, deep learning approaches achieved the best overall performance, however these strategies are currently hampered by the lack of large-scale annotated data. Improving model generalisability to new datasets could be achieved in future studies with data augmentation and transfer learning. Multi-atlas approaches provided the second-best performance overall, and may be utilised to construct a "silver standard" annotated training set for deep learning. To address the technical challenges, providing robustness to injury can be improved by using multiple channels, highly elastic diffeomorphic transformations such as LDDMM, and by following atlas-based approaches with an intensity driven refinement of the segmentation, which has been done with the Expectation Maximisation (EM) and level sets methods. Accounting for potential lesions should be achieved with a separate lesion segmentation approach, as in LesionTOADS. Finally, to address the issue of limited contrast, R2*, T2* and QSM sequences could be used to better highlight the DGM due to its higher iron content. Future studies could look to additionally acquire these sequences by retaining the phase information from standard structural scans, or alternatively acquiring these sequences for only a training set, allowing models to learn the "improved" segmentation from T1-sequences alone.

Accurate, rapid and reliable, fully automated MRI brainstem segmentation for application in multiple sclerosis and neurodegenerative diseases

Neurodegenerative disorders, such as Alzheimer's disease (AD) and progressive forms of multiple sclerosis (MS), can affect the brainstem and are associated with atrophy that can be visualized by MRI. Anatomically accurate, large-scale assessments of brainstem atrophy are challenging due to lack of automated, accurate segmentation methods. We present a novel method for brainstem volumetry using a fully-automated segmentation approach based on multi-dimensional gated recurrent units (MD-GRU), a deep learning based semantic segmentation approach employing a convolutional adaptation of gated recurrent units. The neural network was trained on 67 3D-high resolution T1-weighted MRI scans from MS patients and healthy controls (HC) and refined using segmentations of 20 independent MS patients' scans. Reproducibility was assessed in MR test-retest experiments in 33 HC. Accuracy and robustness were examined by Dice scores comparing MD-GRU to FreeSurfer and manual brainstem segmentations in independent MS and AD datasets. The mean %-change/SD between test-retest brainstem volumes were 0.45%/0.005 (MD-GRU), 0.95%/0.009 (FreeSurfer), 0.86%/0.007 (manually edited segmentations). Comparing MD-GRU to manually edited segmentations the mean Dice scores/SD were: 0.97/0.005 (brainstem), 0.95/0.013 (mesencephalon), 0.98/0.006 (pons), 0.95/0.015 (medulla oblongata). Compared to the manual gold standard, MD-GRU brainstem segmentations were more accurate than FreeSurfer segmentations (p < .001). In the multi-centric acquired AD data, the mean Dice score/SD for the MD-GRU-manual segmentation comparison was 0.97/0.006. The fully automated brainstem segmentation method MD-GRU provides accurate, highly reproducible, and robust segmentations in HC and patients with MS and AD in 200 s/scan on an Nvidia GeForce GTX 1080 GPU and shows potential for application in large and longitudinal datasets.

Brain Segmentation From Computed Tomography of Healthy Aging and Geriatric Concussion at Variable Spatial Resolutions

When properly implemented and processed, anatomic T ₁-weighted magnetic resonance imaging (MRI) can be ideal for the noninvasive quantification of white matter (WM) and gray matter (GM) in the living human brain. Although MRI is more suitable for distinguishing GM from WM than computed tomography (CT), the growing clinical use of the latter technique has renewed interest in head CT segmentation. Such interest is particularly strong in settings where MRI is unavailable, logistically unfeasible or prohibitively expensive. Nevertheless, whereas MRI segmentation is a sophisticated and technically-mature research field, the task of automatically classifying soft brain tissues from CT remains largely unexplored. Furthermore, brain segmentation methods for MRI hold considerable potential for adaptation and application to CT image processing. Here we demonstrate this by combining probabilistic, atlas-based classification with topologically-constrained tissue boundary refinement to delineate WM, GM and cerebrospinal fluid (CSF) from head CT images. The feasibility and utility of this approach are revealed by comparison of MRI-only vs. CT-only segmentations in geriatric concussion victims with both MRI and CT scans. Comparison of the two segmentations yields mean Sørensen-Dice coefficients of 85.5 ± 4.6% (WM), 86.7 ± 5.6% (GM) and 91.3 ± 2.8% (CSF), as well as average Hausdorff distances of 3.76 ± 1.85 mm (WM), 3.43 ± 1.53 mm (GM) and 2.46 ± 1.27 mm (CSF). Bootstrapping results suggest that the segmentation approach is sensitive enough to yield WM, GM and CSF volume estimates within ~5%, ~4%, and ~3% of their MRI-based estimates, respectively. To our knowledge, this is the first 3D segmentation approach for CT to undergo rigorous within-subject comparison with high-resolution MRI. Results suggest that (1) standard-quality CT allows WM/GM/CSF segmentation with reasonable accuracy, and that (2) the task of soft brain tissue classification from CT merits further attention from neuroimaging researchers.