Computational approaches to predicting treatment response to obesity using neuroimaging.

Obesity is a worldwide disease associated with multiple severe adverse consequences and comorbid conditions. While an increased body weight is the defining feature in obesity, etiologies, clinical phenotypes and treatment responses vary between patients. These variations can be observed within individual treatment options which comprise lifestyle interventions, pharmacological treatment, and bariatric surgery. Bariatric surgery can be regarded as the most effective treatment method. However, long-term weight regain is comparably frequent even for this treatment and its application is not without risk. A prognostic tool that would help predict the effectivity of the individual treatment methods in the long term would be essential in a personalized medicine approach. In line with this objective, an increasing number of studies have combined neuroimaging and computational modeling to predict treatment outcome in obesity. In our review, we begin by outlining the central nervous mechanisms measured with neuroimaging in these studies. The mechanisms are primarily related to reward-processing and include "incentive salience" and psychobehavioral control. We then present the diverse neuroimaging methods and computational prediction techniques applied. The studies included in this review provide consistent support for the importance of incentive salience and psychobehavioral control for treatment outcome in obesity. Nevertheless, further studies comprising larger sample sizes and rigorous validation processes are necessary to answer the question of whether or not the approach is sufficiently accurate for clinical real-world application.

Immune signature of multiple sclerosis-associated depression.

Multiple neurobiological pathways have been implicated in the pathobiology of major depressive disorder (MDD). The identification of reliable biological substrates across the entire MDD spectrum, however, is hampered by a vast heterogeneity in the clinical presentation, presumably as a consequence of heterogeneous pathobiology. One way to overcome this limitation could be to explore disease subtypes based on biological similarity such as "inflammatory depression". As such a subtype may be particularly enriched in depressed patients with an underlying inflammatory condition, multiple sclerosis (MS) could provide an informative disease context for this approach. Few studies have explored immune markers of MS-associated depression and replications are missing. To address this, we analyzed data from two independent case-control studies on immune signatures of MS-associated depression, conducted at two different academic MS centers (overall sample size of n = 132). Using a stepwise data-driven approach, we identified CD4+CCR7lowTCM cell frequencies as a robust correlate of depression in MS. This signature was associated with core symptoms of depression and depression severity (but not MS severity per se) and linked to neuroinflammation as determined by magnetic resonance imaging (MRI). Furthermore, exploratory analyses of T cell polarization revealed this was largely driven by cells with a TH1-like phenotype. Our findings suggest (neuro)immune pathways linked to affective symptoms of autoimmune disorders such as MS, with potential relevance for the understanding of "inflammatory" subtypes of depression.

Delayed access to conscious processing in multiple sclerosis: Reduced cortical activation and impaired structural connectivity.

Although multiple sclerosis (MS) is frequently accompanied by visuo-cognitive impairment, especially functional brain mechanisms underlying this impairment are still not well understood. Consequently, we used a functional MRI (fMRI) backward masking task to study visual information processing stratifying unconscious and conscious in MS. Specifically, 30 persons with MS (pwMS) and 34 healthy controls (HC) were shown target stimuli followed by a mask presented 8-150 ms later and had to compare the target to a reference stimulus. Retinal integrity (via optical coherence tomography), optic tract integrity (visual evoked potential; VEP) and whole brain structural connectivity (probabilistic tractography) were assessed as complementary structural brain integrity markers. On a psychophysical level, pwMS reached conscious access later than HC (50 vs. 16 ms, p < .001). The delay increased with disease duration (p < .001, ß = .37) and disability (p < .001, ß = .24), but did not correlate with conscious information processing speed (Symbol digit modality test, ß = .07, p = .817). No association was found for VEP and retinal integrity markers. Moreover, pwMS were characterized by decreased brain activation during unconscious processing compared with HC. No group differences were found during conscious processing. Finally, a complementary structural brain integrity analysis showed that a reduced fractional anisotropy in corpus callosum and an impaired connection between right insula and primary visual areas was related to delayed conscious access in pwMS. Our study revealed slowed conscious access to visual stimulus material in MS and a complex pattern of functional and structural alterations coupled to unconscious processing of/delayed conscious access to visual stimulus material in MS.

Interactions between neural decision-making circuits predict long-term dietary treatment success in obesity.

Although dietary decision-making is regulated by multiple interacting neural controllers, their impact on dietary treatment success in obesity has only been investigated individually. Here, we used fMRI to test how well interactions between the Pavlovian system (automatically triggering urges of consumption after food cue exposure) and the goal-directed system (considering long-term consequences of food decisions) predict future dietary success achieved in 39 months. Activity of the Pavlovian system was measured with a cue-reactivity task by comparing perception of food versus control pictures, activity of the goal-directed system with a food-specific delay discounting paradigm. Both tasks were applied in 30 individuals with obesity up to five times: Before a 12-week diet, immediately thereafter, and at three annual follow-up visits. Brain activity was analyzed in two steps. In the first, we searched for areas involved in Pavlovian processes and goal-directed control across the 39-month study period with voxel-wise linear mixed-effects (LME) analyses. In the second, we computed network parameters reflecting the covariation of longitudinal voxel activity (i.e. principal components) in the regions identified in the first step and used them to predict body mass changes across the 39 months with LME models. Network analyses testing the link of dietary success with activity of the individual systems as reference found a moderate negative link to Pavlovian activity primarily in left hippocampus and a moderate positive association to goal-directed activity primarily in right inferior parietal gyrus. A cross-paradigm network analysis that integrated activity measured in both tasks revealed a strong positive link for interactions between visual Pavlovian areas and goal-directed decision-making regions mainly located in right insular cortex. We conclude that adaptation of food cue processing resources to goal-directed control activity is an important prerequisite of sustained dietary weight loss, presumably since the latter activity can modulate Pavlovian urges triggered by frequent cue exposure in everyday life.

Stress-induced brain activity, brain atrophy, and clinical disability in multiple sclerosis.

Prospective clinical studies support a link between psychological stress and multiple sclerosis (MS) disease severity, and peripheral stress systems are frequently dysregulated in MS patients. However, the exact link between neurobiological stress systems and MS symptoms is unknown. To evaluate the link between neural stress responses and disease parameters, we used an arterial-spin-labeling functional MRI stress paradigm in 36 MS patients and 21 healthy controls. Specifically, we measured brain activity during a mental arithmetic paradigm with performance-adaptive task frequency and performance feedback and related this activity to disease parameters. Across all participants, stress increased heart rate, perceived stress, and neural activity in the visual, cerebellar and insular cortex areas compared with a resting condition. None of these responses was related to cognitive load (task frequency). Consistently, although performance and cognitive load were lower in patients than in controls, stress responses did not differ between groups. Insula activity elevated during stress compared with rest was negatively linked to impairment of pyramidal and cerebral functions in patients. Cerebellar activation was related negatively to gray matter (GM) atrophy (i.e., positively to GM volume) in patients. Interestingly, this link was also observed in overlapping areas in controls. Cognitive load did not contribute to these associations. The results show that our task induced psychological stress independent of cognitive load. Moreover, stress-induced brain activity reflects clinical disability in MS. Finally, the link between stress-induced activity and GM volume in patients and controls in overlapping areas suggests that this link cannot be caused by the disease alone.

Brain activity, regional gray matter loss, and decision-making in multiple sclerosis.

BACKGROUND: Decision-making (DM) abilities deteriorate with multiple sclerosis (MS) disease progression which impairs everyday life and is thus clinically important. OBJECTIVE: To investigate the underlying neurocognitive processes and their relation to regional gray matter (GM) loss induced by MS. METHODS: We used a functional magnetic resonance imaging (fMRI) Iowa Gambling Task to measure DM-related brain activity in 36 MS patients and 21 healthy controls (HC). From this activity, we determined neural parameters of two cognitive stages, a deliberation ("choice") period preceding a choice and a post-choice ("feedback") stage reporting decision outcomes. These measures were related to DM separately in intact and damaged GM areas as determined by a voxel-based morphometry analysis. RESULTS: Severely affected patients (with high lesion burden) showed worse DM-learning than HC ( t = -1.75, p = 0.045), moderately affected (low lesion burden) did not. Activity in the choice stage in intact insular ( t = 4.60, pFamily-Wise Error [FWE] corrected = 0.034), anterior cingulate ( t = 4.50, pFWE = 0.044), and dorsolateral prefrontal areas ( t = 4.43, pFWE = 0.049) and in insular areas with GM loss ( t = 3.78, pFWE = 0.011) was positively linked to DM performance across patients with severe tissue damage and HC. Furthermore, activity in intact orbitofrontal areas was positively linked to DM-learning during the feedback stage across these participants ( t = 4.49, pFWE = 0.032). During none of the stages, moderately affected patients showed higher activity than HC, which might have indicated preserved DM due to compensatory activity. CONCLUSION: We identified dysregulated activity linked to impairment in specific cognitive stages of reward-related DM. The link of brain activity and impaired DM in areas with MS-induced GM loss suggests that this deficit might be tightly coupled to MS neuropathology.

Increasing the spatial resolution and sensitivity of magnetic resonance elastography by correcting for subject motion and susceptibility-induced image distortions.

PURPOSE: To improve the resolution of elasticity maps by adapting motion and distortion correction methods for phase-based magnetic resonance imaging (MRI) contrasts such as magnetic resonance elastography (MRE), a technique for measuring mechanical tissue properties in vivo. MATERIALS AND METHODS: MRE data of the brain were acquired with echo-planar imaging (EPI) at 3T (n = 14) and 7T (n = 18). Motion and distortion correction parameters were estimated using the magnitude images. The real and imaginary part of the complex MRE data were corrected separately and recombined. The width of the point-spread function (PSF) and the position variability were calculated. The images were normalized to the Montreal Neurological Institute (MNI) anatomical template. The gray-to-white matter separability of the elasticity maps was tested. RESULTS: Motion correction sharpened the |G*| maps as demonstrated by a narrowing of the PSF by 0.78 ± 0.51 mm at 7T and 0.52 ± 0.63 mm at 3T. The amount of individual head motion during MRE acquisition correlated with the decrease in the width of the PSF at 7T (r = 0.53, P = 0.025) and at 3T (r = 0.69, P = 0.006) and with the increase of gray-to-white matter separability after motion correction at 7T (r = 0.64, P = 0.0039) and at 3T (r = 0.57, P = 0.0319). Improved spatial accuracy after distortion correction results in a significant increase in separability of gray and white matter stiffness (P = 0.0067), especially in inferior parts of the brain suffering from strong B0 inhomogeneities. CONCLUSION: We demonstrate that our method leads to sharper images and higher spatial accuracy, raising the prospect of the investigation of smaller brain areas with increased sensitivity in studies using MRE. LEVEL OF EVIDENCE: 1 Technical Efficacy: Stage 1 J. MAGN. RESON. IMAGING 2017;46:134-141.

Impulse control in the dorsolateral prefrontal cortex counteracts post-diet weight regain in obesity.

A variety of studies suggest that efficient treatments to induce short-term dietary success in obesity exist. However, sustained maintenance of reduced weight is rare as a large proportion of patients start to regain weight when treatment is discontinued. Thus, from a clinical perspective, it would be desirable to identify factors that counteract post-diet weight regain across longer time-scales. To address this question, we extended our previous work on neural impulse control mechanisms of short-term dietary success in obesity and now investigated the mechanisms counteracting long-term weight regain after a diet. Specifically, we measured neural impulse control during a delay discounting task with fMRI at two time points, i.e. the beginning ('T0') and the end ('T12') of a one-year follow-up interval after a 12-week diet. Then, we tested whether activity in the dorsolateral prefrontal cortex (DLPFC) at T0 and whether activity changes across the follow-up period (T0-T12) are linked to success in weight maintenance. The analyses conducted show that control-related DLPFC activity at T0 was coupled to the degree of success in weight maintenance. Consistently, also behavioral measures of control were linked to the degree of success in maintenance. A direct comparison of neural and behavioral control parameters for prognostic weight change modeling revealed that neural signals were more informative. Taken together, neural impulse control in the DLPFC measured with fMRI directly after a diet predicts real-world diet success in obese patients across extended time periods.

Body mass, neuro-hormonal stress processing, and disease activity in lean to obese people with multiple sclerosis.

Overweight and obesity can worsen disease activity in multiple sclerosis (MS). Although psychobiological stress processing is increasingly recognized as important obesity factor that is tightly connected to proinflammatory metabolic hormones and cytokines, its role for MS obesity remains unexplored. Consequently, we investigated the interplay between body mass index (BMI), neural stress processing (functional connectivity, FC), and immuno-hormonal stress parameters (salivary cortisol and T cell glucocorticoid [GC] sensitivity) in 57 people with MS (six obese, 19 over-, 28 normal-, and four underweight; 37 females, 46.4 ± 10.6 years) using an Arterial-Spin-Labeling MRI task comprising a rest and stress stage, along with quantitative PCR. Our findings revealed significant positive connections between BMI and MS disease activity (i.e., higher BMI was accompanied by higher relapse rate). BMI was positively linked to right supramarginal gyrus and anterior insula FC during rest and negatively to right superior parietal lobule and cerebellum FC during stress. BMI showed associations with GC functioning, with higher BMI associated with lower CD8+ FKBP4 expression and higher CD8+ FKBP5 expression on T cells. Finally, the expression of CD8+ FKBP4 positively correlated with the FC of right supramarginal gyrus and left superior parietal lobule during rest. Overall, our study provides evidence that body mass is tied to neuro-hormonal stress processing in people with MS. The observed pattern of associations between BMI, neural networks, and GC functioning suggests partial overlap between neuro-hormonal and neural-body mass networks. Ultimately, the study underscores the clinical importance of understanding multi-system crosstalk in MS obesity.

The role of neural impulse control mechanisms for dietary success in obesity.

Deficits in impulse control are discussed as key mechanisms for major worldwide health problems such as drug addiction and obesity. For example, obese subjects have difficulty controlling their impulses to overeat when faced with food items. Here, we investigated the role of neural impulse control mechanisms for dietary success in middle-aged obese subjects. Specifically, we used a food-specific delayed gratification paradigm and functional magnetic resonance imaging to measure eating-related impulse-control in middle-aged obese subjects just before they underwent a twelve-week low calorie diet. As expected, we found that subjects with higher behavioral impulse control subsequently lost more weight. Furthermore, brain activity before the diet in VMPFC and DLPFC correlates with subsequent weight loss. Additionally, a connectivity analysis revealed that stronger functional connectivity between these regions is associated with better dietary success and impulse control. Thus, the degree to which subjects can control their eating impulses might depend on the interplay between control regions (DLPFC) and regions signaling the reward of food (VMPFC). This could potentially constitute a general mechanism that also extends to other disorders such as drug addiction or alcohol abuse.

Similar neural pathways link psychological stress and brain-age in health and multiple sclerosis.

Clinical and neuroscientific studies suggest a link between psychological stress and reduced brain health in health and neurological disease but it is unclear whether mediating pathways are similar. Consequently, we applied an arterial-spin-labeling MRI stress task in 42 healthy persons and 56 with multiple sclerosis, and investigated regional neural stress responses, associations between functional connectivity of stress-responsive regions and the brain-age prediction error, a highly sensitive machine learning brain health biomarker, and regional brain-age constituents in both groups. Stress responsivity did not differ between groups. Although elevated brain-age prediction errors indicated worse brain health in patients, anterior insula-occipital cortex (healthy persons: occipital pole; patients: fusiform gyrus) functional connectivity correlated with brain-age prediction errors in both groups. Finally, also gray matter contributed similarly to regional brain-age across groups. These findings might suggest a common stress-brain health pathway whose impact is amplified in multiple sclerosis by disease-specific vulnerability factors.

Multi-scale classification of disease using structural MRI and wavelet transform.

Recently, multivariate analysis algorithms have become a popular tool to diagnose neurological diseases based on neuroimaging data. Most studies, however, are biased for one specific scale, namely the scale given by the spatial resolution (i.e. dimension) of the data. In the present study, we propose to use the dual-tree complex wavelet transform to extract information on different spatial scales from structural MRI data and show its relevance for disease classification. Based on the magnitude representation of the complex wavelet coefficients calculated from the MR images, we identified a new class of features taking scale, directionality and potentially local information into account simultaneously. By using a linear support vector machine, these features were shown to discriminate significantly between spatially normalized MR images of 41 patients suffering from multiple sclerosis and 26 healthy controls. Interestingly, the decoding accuracies varied strongly among the different scales and it turned out that scales containing low frequency information were partly superior to scales containing high frequency information. Usually, this type of information is neglected since most decoding studies use only the original scale of the data. In conclusion, our proposed method has not only a high potential to assist in the diagnostic process of multiple sclerosis, but can be applied to other diseases or general decoding problems in structural or functional MRI.

fMRI pattern recognition in obsessive-compulsive disorder.

Patients suffering from obsessive-compulsive disorder (OCD) are characterized by dysregulated neuronal processing of disorder-specific and also unspecific affective stimuli. In the present study, we investigated whether generic fear-inducing, disgust-inducing, and neutral stimuli can be decoded from brain patterns of single fMRI time samples of individual OCD patients and healthy controls. Furthermore, we tested whether differences in the underlying encoding provide information to classify subjects into groups (OCD patients or healthy controls). Two pattern classification analyses were conducted. In analysis 1, we used a classifier to decode the category of a currently viewed picture from extended fMRI patterns of single time samples (TR=3s) in individual subjects for several pairs of categories. In analysis 2, we used a searchlight approach to predict subjects' diagnostic status based on local brain patterns. In analysis 1, we obtained significant accuracies for the separation of fear-eliciting from neutral pictures in OCD patients and healthy controls. Separation of disgust-inducing from neutral pictures was significant in healthy controls. In analysis 2, we identified diagnostic information for the presence of OCD in the orbitofrontal cortex, and in the caudate nucleus. Accuracy obtained in these regions was 100% (p<10(-6)). To summarize our findings, by using multivariate pattern classification techniques we were able to identify neurobiological markers providing reliable diagnostic information about OCD. The classifier-based fMRI paradigms proposed here might be integrated in future diagnostic procedures and treatment concepts.

Diagnosing different binge-eating disorders based on reward-related brain activation patterns.

This study addresses how visual food cues are encoded in reward related brain areas and whether this encoding might provide information to differentiate between patients suffering from eating disorders [binge-eating disorder (BED) and bulimia nervosa (BN)], overweight controls (C-OW), and normal-weight controls (C-NW). Participants passively viewed pictures of food stimuli and neutral stimuli in a cue reactivity design. Two classification analyses were conducted. First, we used multivariate pattern recognition techniques to decode the category of a currently viewed picture from local brain activity patterns. In the second analysis, we applied an ensemble classifier to predict the clinical status of subjects (BED, BN, C-OW, and C-NW) based on food-related brain response patterns. The left insular cortex separated between food and neutral contents in all four groups. Patterns in the right insular cortex provided a maximum diagnostic accuracy for the separation of BED patients and C-NW (86% accuracy, P < 10(-5) , 82% sensitivity, and 90% specificity) as well as BN patients and C-NW (78% accuracy, P = 0.001, 86% sensitivity, and 70% specificity). The right ventral striatum separated maximally between BED patients and C-OW (71% accuracy, P = 0.013, 59% sensitivity, and 82% specificity). The right lateral orbitofrontal cortex separated maximally between BN patients and C-OW (86% accuracy, P < 10(-4) , 79% sensitivity, and 94% specificity). The best differential diagnostic separation between BED and BN patients was obtained in the left ventral striatum (84% accuracy, P < 10(-3) , 82% sensitivity, and 86% specificity). Our results indicate that pattern recognition techniques are able to contribute to a reliable differential diagnosis of BN and BED.

Prefrontal-amygdala emotion regulation and depression in multiple sclerosis.

Depression is among the most common comorbidities in multiple sclerosis and has severe psychosocial consequences. Alterations in neural emotion regulation in amygdala and prefrontal cortex have been recognized as key mechanism of depression but never been investigated in multiple sclerosis depression. In this cross-sectional observational study, we employed a functional MRI task investigating neural emotion regulation by contrasting regulated versus unregulated negative stimulus perception in 16 persons with multiple sclerosis and depression (47.9 ± 11.8 years; 14 female) and 26 persons with multiple sclerosis but without depression (47.3 ± 11.7 years; 14 female). We tested the impact of depression and its interaction with lesions in amygdala-prefrontal fibre tracts on brain activity reflecting emotion regulation. A potential impact of sex, age, information processing speed, disease duration, overall lesion load, grey matter fraction, and treatment was taken into account in these analyses. Patients with depression were less able (i) to downregulate negative emotions than those without (t = -2.25, P = 0.012, ß = -0.33) on a behavioural level according to self-report data and (ii) to downregulate activity in a left amygdala coordinate (t = 3.03, P Family-wise error [FWE]-corrected = 0.017, ß = 0.39). Moreover, (iii) an interdependent effect of depression and lesions in amygdala-prefrontal tracts on activity was found in two left amygdala coordinates (t = 3.53, pFWE = 0.007, ß = 0.48; t = 3.21, pFWE = 0.0158, ß = 0.49) and one right amygdala coordinate (t = 3.41, pFWE = 0.009, ß = 0.51). Compatible with key elements of the cognitive depression theory formulated for idiopathic depression, our study demonstrates that depression in multiple sclerosis is characterized by impaired neurobehavioural emotion regulation. Complementing these findings, it shows that the relation between neural emotion regulation and depression is affected by lesion load, a key pathological feature of multiple sclerosis, located in amygdala-prefrontal tracts.

Central stress processing, T-cell responsivity to stress hormones and disease severity in multiple sclerosis.

Epidemiological, clinical and neuroscientific studies support a link between psychobiological stress and multiple sclerosis. Neuroimaging suggests that blunted central stress processing goes along with higher multiple sclerosis severity, neuroendocrine studies suggest that blunted immune system sensitivity to stress hormones is linked to stronger neuroinflammation. Until now, however, no effort has been made to elucidate whether central stress processing and immune system sensitivity to stress hormones are related in a disease-specific fashion, and if so, whether this relation is clinically meaningful. Consequently, we conducted two functional MRI analyses based on a total of 39 persons with multiple sclerosis and 25 healthy persons. Motivated by findings of an altered interplay between neuroendocrine stress processing and T-cell glucocorticoid sensitivity in multiple sclerosis, we searched for neural networks whose stress task-evoked activity is differentially linked to peripheral T-cell glucocorticoid signalling in patients versus healthy persons as a potential indicator of disease-specific CNS-immune crosstalk. Subsequently, we tested whether this activity is simultaneously related to disease severity. We found that activity of a network comprising right anterior insula, right fusiform gyrus, left midcingulate and lingual gyrus was differentially coupled to T-cell glucocorticoid signalling across groups. This network's activity was simultaneously linked to patients' lesion volume, clinical disability and information-processing speed. Complementary analyses revealed that T-cell glucocorticoid signalling was not directly linked to disease severity. Our findings show that alterations in the coupling between central stress processing and T-cell stress hormone sensitivity are related to key severity measures of multiple sclerosis.

Emotion modulates the effects of endogenous attention on retinotopic visual processing.

A fundamental challenge for organisms is how to focus on perceptual information relevant to current goals while remaining able to respond to goal-irrelevant stimuli that signal potential threat. Here, we studied how visual threat signals influence the effects of goal-directed spatial attention on the retinotopic distribution of processing resources in early visual cortex. We used a combined blocked and event-related functional magnetic resonance imaging paradigm with target displays comprising diagonal pairs of intact and scrambled faces presented simultaneously in the four visual field quadrants. Faces were male or female and had fearful or neutral emotional expressions. Participants attended covertly to a pair of two diagonally opposite stimuli and performed a gender-discrimination task on the attended intact face. In contrast to the fusiform face area, where attention and fearful emotional expression had additive effects, neural responses to attended and unattended fearful faces were indistinguishable in early retinotopic visual areas: When attended, fearful face expression did not further enhance responses, whereas when unattended, fearful expression increased responses to the level of attended face stimuli. Remarkably, the presence of fearful stimuli augmented the enhancing effect of attention on retinotopic responses to neutral faces in remote visual field locations. We conclude that this redistribution of neural activity in retinotopic visual cortex may serve the purpose of allocating processing resources to task-irrelevant threat-signaling stimuli while at the same time increasing resources for task-relevant stimuli as required for the maintenance of goal-directed behavior.

Patch individual filter layers in CNNs to harness the spatial homogeneity of neuroimaging data.

Convolutional neural networks (CNNs)-as a type of deep learning-have been specifically designed for highly heterogeneous data, such as natural images. Neuroimaging data, however, is comparably homogeneous due to (1) the uniform structure of the brain and (2) additional efforts to spatially normalize the data to a standard template using linear and non-linear transformations. To harness spatial homogeneity of neuroimaging data, we suggest here a new CNN architecture that combines the idea of hierarchical abstraction in CNNs with a prior on the spatial homogeneity of neuroimaging data. Whereas early layers are trained globally using standard convolutional layers, we introduce patch individual filters (PIF) for higher, more abstract layers. By learning filters in individual latent space patches without sharing weights, PIF layers can learn abstract features faster and specific to regions. We thoroughly evaluated PIF layers for three different tasks and data sets, namely sex classification on UK Biobank data, Alzheimer's disease detection on ADNI data and multiple sclerosis detection on private hospital data, and compared it with two baseline models, a standard CNN and a patch-based CNN. We obtained two main results: First, CNNs using PIF layers converge consistently faster, measured in run time in seconds and number of iterations than both baseline models. Second, both the standard CNN and the PIF model outperformed the patch-based CNN in terms of balanced accuracy and receiver operating characteristic area under the curve (ROC AUC) with a maximal balanced accuracy (ROC AUC) of 94.21% (99.10%) for the sex classification task (PIF model), and 81.24% and 80.48% (88.89% and 87.35%) respectively for the Alzheimer's disease and multiple sclerosis detection tasks (standard CNN model). In conclusion, we demonstrated that CNNs using PIF layers result in faster convergence while obtaining the same predictive performance as a standard CNN. To the best of our knowledge, this is the first study that introduces a prior in form of an inductive bias to harness spatial homogeneity of neuroimaging data.

Neural Processes of Psychological Stress and Relaxation Predict the Future Evolution of Quality of Life in Multiple Sclerosis.

Health-related quality of life (HRQoL) is an essential complementary parameter in the assessment of disease burden and treatment outcome in multiple sclerosis (MS) and can be affected by neuropsychiatric symptoms, which in turn are sensitive to psychological stress. However, until now, the impact of neurobiological stress and relaxation on HRQoL in MS has not been investigated. We thus evaluated whether the activity of neural networks triggered by mild psychological stress (elicited in an fMRI task comprising mental arithmetic with feedback) or by stress termination (i.e., relaxation) at baseline (T0) predicts HRQoL variations occurring between T0 and a follow-up visit (T1) in 28 patients using a robust regression and permutation testing. The median delay between T0 and T1 was 902 (range: 363-1,169) days. We assessed HRQoL based on the Hamburg Quality of Life Questionnaire in MS (HAQUAMS) and accounted for the impact of established HRQoL predictors and the cognitive performance of the participants. Relaxation-triggered activity of a widespread neural network predicted future variations in overall HRQoL (t = 3.68, p family-wise error [FWE]-corrected = 0.008). Complementary analyses showed that relaxation-triggered activity of the same network at baseline was associated with variations in the HAQUAMS mood subscale on an αFWE = 0.1 level (t = 3.37, p FWE = 0.087). Finally, stress-induced activity of a prefronto-limbic network predicted future variations in the HAQUAMS lower limb mobility subscale (t = -3.62, p FWE = 0.020). Functional neural network measures of psychological stress and relaxation contain prognostic information for future HRQoL evolution in MS independent of clinical predictors.

Viscoelasticity of striatal brain areas reflects variations in body mass index of lean to overweight male adults.

Although a variety of MRI studies investigated the link between body mass index (BMI) and parameters of neural gray matter (GM), the technique applied in most of these studies, voxel-based morphometry (VBM), focusses on the regional GM volume, a macroscopic tissue property. Thus, the studies were not able to exploit the BMI-related information contained in the GM microstructure although PET studies suggest that these factors are important. Here, we used cerebral MR Elastography (MRE) to characterize features of tissue microstructure by evaluating the propagation of shear waves applied to the skull and to assess local tissue viscoelasticity to test the link between this parameter and BMI in 22 lean to overweight males. Unlike the majority of existing MRE studies investigating neural viscoelasticity signals averaged across large brain regions, we used the viscoelasticity of individual voxels for our experiment. Our technique revealed a negative link between BMI and viscoelasticity of two areas of the striatal reward system, i.e., right putamen (t = -8.2; pFWE-corrected = 0.005) and left globus pallidus (t = -7.1; pFWE = 0.037) which was independent of GM volume at these coordinates. Finally, comparison of BMI models based on individual voxels vs. on signals averaged across brain atlas regions demonstrates that voxel-based models explain a significantly higher proportion of variance. Consequently, our findings show that cerebral MRE is suitable to identify medically relevant microstructural tissue properties. Using a voxel-wise analysis approach, we were able to utilize the high spatial resolution of MRE for mapping BMI-related information in the brain.