Brian Patenaude

Bayesian analysis of neuroimaging data in FSL.

Typically in neuroimaging we are looking to extract some pertinent information from imperfect, noisy images of the brain. This might be the inference of percent changes in blood flow in perfusion FMRI data, segmentation of subcortical structures from structural MRI, or inference of the probability of an anatomical connection between an area of cortex and a subthalamic nucleus using diffusion MRI. In this article we will describe how Bayesian techniques have made a significant impact in tackling problems such as these, particularly in regards to the analysis tools in the FMRIB Software Library (FSL). We shall see how Bayes provides a framework within which we can attempt to infer on models of neuroimaging data, while allowing us to incorporate our prior belief about the brain and the neuroimaging equipment in the form of biophysically informed or regularising priors. It allows us to extract probabilistic information from the data, and to probabilistically combine information from multiple modalities. Bayes can also be used to not only compare and select between models of different complexity, but also to infer on data using committees of models. Finally, we mention some analysis scenarios where Bayesian methods are impractical, and briefly discuss some practical approaches that we have taken in these cases.

A Bayesian model of shape and appearance for subcortical brain segmentation.

Automatic segmentation of subcortical structures in human brain MR images is an important but difficult task due to poor and variable intensity contrast. Clear, well-defined intensity features are absent in many places along typical structure boundaries and so extra information is required to achieve successful segmentation. A method is proposed here that uses manually labelled image data to provide anatomical training information. It utilises the principles of the Active Shape and Appearance Models but places them within a Bayesian framework, allowing probabilistic relationships between shape and intensity to be fully exploited. The model is trained for 15 different subcortical structures using 336 manually-labelled T1-weighted MR images. Using the Bayesian approach, conditional probabilities can be calculated easily and efficiently, avoiding technical problems of ill-conditioned covariance matrices, even with weak priors, and eliminating the need for fitting extra empirical scaling parameters, as is required in standard Active Appearance Models. Furthermore, differences in boundary vertex locations provide a direct, purely local measure of geometric change in structure between groups that, unlike voxel-based morphometry, is not dependent on tissue classification methods or arbitrary smoothing. In this paper the fully-automated segmentation method is presented and assessed both quantitatively, using Leave-One-Out testing on the 336 training images, and qualitatively, using an independent clinical dataset involving Alzheimers disease. Median Dice overlaps between 0.7 and 0.9 are obtained with this method, which is comparable or better than other automated methods. An implementation of this method, called FIRST, is currently distributed with the freely-available FSL package.

Prefrontal cortical regulation of brainwide circuit dynamics and reward-related behavior

A way to modulate reward-seeking Which brain regions are causally involved in reward-related behavior? Ferenczi et al. combined focal, cell type-specific, optogenetic manipulations with brain imaging, behavioral testing, and in vivo electrophysiology (see the Perspective by Robbins). Stimulation of midbrain dopamine neurons increased activity in a brain region called the striatum and was correlated with reward-seeking across individual animals. However, elevated excitability of an area called the medial prefrontal cortex reduced both striatal responses to the stimulation of dopamine neurons and the behavioral drive to seek the stimulation of dopamine neurons. Finally, modulating the excitability of medial prefrontal cortex pyramidal neurons drove changes in neural circuit synchrony, as well as corresponding anhedonic behavior. These observations resemble imaging and clinical phenotypes observed in human depression, addiction, and schizophrenia. Science, this issue p. 10.1126/science.aac9698; see also p. 10.1126/science.aad9698 Optogenetic and brain imaging approaches reveal a causal brainwide dynamical mechanism for the hedonic-anhedonic transition. [Also see Perspective by Robbins] INTRODUCTION The drive to seek and experience reward is conserved across species and, in mammals, involves interactions between subcortical dopaminergic systems and limbic structures such as the striatum. Impairment of this process, observed across a number of psychiatric conditions, is the clinical symptom of anhedonia (loss of enjoyment). The neural mechanisms underlying anhedonia are unknown but could result from abnormal interactions between cortical and subcortical reward circuits. We sought to test the hypothesis that elevated medial prefrontal cortex (mPFC) excitability (a clinical feature associated with anhedonia) exerts suppressive control over the interactions between two distant subcortical regions: the dopaminergic midbrain and the striatum. RATIONALE Clinical imaging studies have detected elevated activity in the mPFC in human patients with depression, and treatment is associated with normalization of this overactivity and improvement of anhedonic symptoms. Additionally, human studies have identified areas of the brain that respond to reward anticipation and experience, and this response can be suppressed in psychiatric disease. However, the source of this reward signal and the mechanisms underlying its modulation have not been causally demonstrated. We have integrated a diverse set of chronic and acute optogenetic tools with functional magnetic resonance imaging (fMRI) to provide a bridge between the causal, cellular specificity of rodent optogenetics and the brainwide observations that characterize human neuroimaging, with the goal of locally manipulating and globally visualizing neural activity to understand the regulation of reward-seeking behavior. RESULTS We demonstrate that stimulation of midbrain dopamine neurons drives both striatal fMRI blood oxygen level–dependent (BOLD) activity and reward-seeking behavior, and we show that these are correlated across individuals. We additionally find that silencing of dopamine neurons suppresses activity in the striatum, as well as in other brain regions (such as the hypothalamus), and drives avoidance behavior. Having established this bidirectional control of reward-seeking behavior, we then tested for perturbation of this circuitry via elevation of mPFC excitability. We observed suppression of striatal responses to dopamine, as well as the behavioral drive to seek out dopamine neuron stimulation and other natural rewarding stimuli. Finally, we demonstrate that stably elevated mPFC excitability synchronizes corticolimbic BOLD and electrophysiological activity, which in turn can predict anhedonic behavior in individual animals. CONCLUSION Our findings from experiments involving local cell-specific control, simultaneously with global unbiased observation of neural activity, reveal that the mPFC exerts top-down control over midbrain dopaminergic interactions with the striatum and that, when elevated, activity in the mPFC can suppress natural reward-related behavior. Furthermore, we observe that cortical-subcortical neural dynamics work in concert to regulate reward processing. All of these findings have implications for our understanding of natural reward-related physiology and behavior, as well as the pathogenesis of anhedonia. Reward-related signaling between the dopaminergic midbrain and the striatum is under suppressive control by the mPFC. Optogenetic fMRI was used to locally manipulate and globally visualize brainwide neural activity related to reward. Habituated rats were scanned in the awake state (top photographs). We establish that striatal BOLD activity is increased by optogenetic stimulation of dopamine neurons and decreased by optogenetic neural silencing. We demonstrate that focally elevated mPFC excitability suppresses reward-seeking behavior by exerting top-down control over striatal dopamine-induced activity and drives synchrony between specific corticolimbic circuits. Motivation for reward drives adaptive behaviors, whereas impairment of reward perception and experience (anhedonia) can contribute to psychiatric diseases, including depression and schizophrenia. We sought to test the hypothesis that the medial prefrontal cortex (mPFC) controls interactions among specific subcortical regions that govern hedonic responses. By using optogenetic functional magnetic resonance imaging to locally manipulate but globally visualize neural activity in rats, we found that dopamine neuron stimulation drives striatal activity, whereas locally increased mPFC excitability reduces this striatal response and inhibits the behavioral drive for dopaminergic stimulation. This chronic mPFC overactivity also stably suppresses natural reward-motivated behaviors and induces specific new brainwide functional interactions, which predict the degree of anhedonia in individuals. These findings describe a mechanism by which mPFC modulates expression of reward-seeking behavior, by regulating the dynamical interactions between specific distant subcortical regions.

Combining shape and connectivity analysis: an MRI study of thalamic degeneration in Alzheimer's disease.

Alzheimers disease (AD) is associated with neuronal loss not only in the hippocampus and amygdala but also in the thalamus. Anterodorsal, centromedial, and pulvinar nuclei are the main sites of degeneration in AD. Here we combined shape analysis and diffusion tensor imaging (DTI) tractography to study degeneration in AD in the thalamus and its connections. Structural and diffusion tensor MRI scans were obtained from 16 AD patients and 22 demographically similar healthy volunteers. The thalamus, hippocampus, and amygdala were automatically segmented using our locally developed algorithm, and group comparisons were carried out for each surface vertex. We also employed probabilistic diffusion tractography to obtain connectivity measures between individual thalamic voxels and hippocampus/amygdala voxels and to segment the internal medullary lamina (IML). Shape analysis showed significant bilateral regional atrophy in the dorsal-medial part of the thalamus in AD patients compared to controls. Probabilistic tractography demonstrated that these regions are mainly connected with the hippocampus, temporal, and prefrontal cortex. Intrathalamic FA comparisons showed reductions in the anterodorsal region of thalamus. Intrathalamic tractography from this region revealed that the IML was significantly smaller in AD patients than in controls. We suggest that these changes can be attributed to the degeneration of the anterodorsal and intralaminar nuclei, respectively. In addition, based on previous neuropathological reports, ventral and dorsal-medial shape change in the thalamus in AD patients is likely to be driven by IML atrophy. This combined shape and connectivity analysis provides MRI evidence of regional thalamic degeneration in AD.

Neurobiological Signatures of Anxiety and Depression in Resting-State Functional Magnetic Resonance Imaging

BACKGROUND There is increasing interest in using neurobiological measures to inform psychiatric nosology. It is unclear at the present time whether anxiety and depression are neurobiologically distinct or similar processes. It is also unknown if the best way to examine these disorders neurobiologically is by contrasting categorical definitions or by examining symptom dimensions. METHODS A cross-sectional neuroimaging study was conducted of patients with generalized anxiety disorder (GAD), major depressive disorder (MDD), comorbid GAD and MDD (GAD/MDD), or neither GAD nor MDD (control subjects). There were 90 participants, all medication-free (17 GAD, 12 MDD, 23 GAD/MDD, and 38 control subjects). Diagnosis/category and dimensions/symptoms were assessed to determine the best fit for neurobiological data. Symptoms included general distress, common to anxiety and depression, and anxiety-specific (anxious arousal) or depression-specific (anhedonia) symptoms. Low-frequency (.008-.1 Hz) signal amplitude and functional connectivity analyses of resting-state functional magnetic resonance imaging data focused on a priori cortical and subcortical regions of interest. RESULTS Support was found for effects of diagnosis above and beyond effects related to symptom levels as well as for effects of symptom levels above and beyond effects of diagnostic categories. The specific dimensional factors of general distress and anxious arousal as well as a diagnosis of MDD explained unique proportions of variance in signal amplitude or functional connectivity. CONCLUSIONS Using resting-state functional magnetic resonance imaging, our data show that a single conceptual model alone (i.e., categorical diagnoses or symptom dimensions) provides an incomplete mapping of psychopathology to neurobiology. Instead, the data support an additive model that best captures abnormal neural patterns in patients with anxiety and depression.

Comparison and Evaluation of Segmentation Techniques for Subcortical Structures in Brain MRI

The automation of segmentation of medical images is an active research area. However, there has been criticism of the standard of evaluation of methods. We have comprehensively evaluated four novel methods of automatically segmenting subcortical structures using volumetric, spatial overlap and distance-based measures. Two of the methods are atlas-based - classifier fusion and labelling (CFL) and expectation-maximisation segmentation using a dynamic brain atlas (EMS), and two model-based - profile active appearance models (PAM) and Bayesian appearance models (BAM). Each method was applied to the segmentation of 18 subcortical structures in 270 subjects from a diverse pool varying in age, disease, sex and image acquisition parameters. Our results showed that all four methods perform on par with recently published methods. CFL performed significantly better than the other three methods according to all three classes of metrics.

Age effect on subcortical structures in healthy adults

Cross-sectional age effects in normal control volunteers were investigated using magnetic resonance imaging in the following eight subcortical structures: lateral ventricles, thalamus, caudate, putamen, pallidum, hippocampus, amygdala and nucleus accumbens. Two hundred and twenty-six control subjects, ranging in age from 19 to 85 years, were scanned on a 1.5 T GE system (n=184) or a 3.0 T Siemens system (n=42). Volumes of subcortical structures, adjusted for cranium size, were estimated using FSLs FIRST software, which is fully automated. Significant age effects were found for all volumes when the entire age range was analyzed; however, the older subjects (60-85 years of age) showed a stronger correlation between age and structural volume for the ventricles, hippocampus, amygdala and accumbens than middle-aged (35-60 years of age) subjects. Middle-aged subjects were studied at both sites, and age effects in these groups were comparable, despite differences in magnet strength and acquisition systems. This agreement lends support to the validity of the image-analysis tools and procedures used in the present study.

A Cognitive–Emotional Biomarker for Predicting Remission with Antidepressant Medications: A Report from the iSPOT-D Trial

Depression involves impairments in a range of cognitive and emotional capacities. It is unknown whether these functions can inform medication choice when considered as a composite predictive biomarker. We tested whether behavioral tests, grounded in the neurobiology of cognitive and emotional functions, predict outcome with common antidepressants. Medication-free outpatients with nonpsychotic major depressive disorder (N=1008; 665 completers) were assessed before treatment using 13 computerized tests of psychomotor, executive, memory–attention, processing speed, inhibitory, and emotional functions. Matched healthy controls (N=336) provided a normative reference sample for test performance. Depressed participants were then randomized to escitalopram, sertraline, or venlafaxine–extended release, and were assessed using the 16-item Quick Inventory of Depressive Symptomatology (QIDS-SR16) and the 17-item Hamilton Rating Scale for Depression. Given the heterogeneity of depression, analyses were furthermore stratified by pretreatment performance. We then used pattern classification with cross-validation to determine individual patient-level composite predictive biomarkers of antidepressant outcome based on test performance. A subgroup of depressed participants (approximately one-quarter of patients) were found to be impaired across most cognitive tests relative to the healthy norm, from which they could be discriminated with 91% accuracy. These patients with generally impaired cognitive task performance had poorer treatment outcomes. For this impaired subgroup, task performance furthermore predicted remission on the QIDS-SR16 at 72% accuracy specifically following treatment with escitalopram but not the other medications. Therefore, tests of cognitive and emotional functions can form a clinically meaningful composite biomarker that may help drive general treatment outcome prediction for optimal treatment selection in depression, particularly for escitalopram.

Subcortical Volumetric Reductions in Adult Niemann-Pick Disease Type C: A Cross-Sectional Study

BACKGROUND AND PURPOSE: Voxel-based analysis has suggested that deep gray matter rather than cortical regions is initially affected in adult Niemann-Pick type C. We sought to examine a range of deep gray matter structures in adults with NPC and relate these to clinical variables. MATERIALS AND METHODS: Ten adult patients with NPC (18–49 years of age) were compared with 27 age- and sex-matched controls, and subcortical structures were automatically segmented from normalized T1-weighted MR images. Absolute volumes (in cubic millimeters) were generated for a range of deep gray matter structures and were compared between groups and correlated with illness variables. RESULTS: Most structures were smaller in patients with NPC compared with controls. The thalamus, hippocampus, and striatum showed the greatest and most significant reductions, and left hippocampal volume correlated with symptom score and cognition. Vertex analysis of the thalamus, hippocampus, and caudate implicated regions involved in memory, executive function, and motor control. CONCLUSIONS: Thalamic and hippocampal reductions may underpin the memory and executive deficits seen in adult NPC. Volume losses in other subcortical regions may also be involved in the characteristic range of motor, psychiatric, and cognitive deficits seen in the disease.

Subcortical Volumes in Long-Term Abstinent Alcoholics: Associations With Psychiatric Comorbidity

BACKGROUND Research in chronic alcoholics on memory, decision-making, learning, stress, and reward circuitry has increasingly highlighted the importance of subcortical brain structures. In addition, epidemiological studies have established the pervasiveness of co-occurring psychiatric diagnoses in alcoholism. Subcortical structures have been implicated in externalizing pathology, including alcohol dependence, and in dysregulated stress and reward circuitry in anxiety and mood disorders and alcohol dependence. Most studies have focused on active or recently detoxified alcoholics, while subcortical structures in long-term abstinent alcoholics (LTAA) have remained relatively uninvestigated. METHODS Structural MRI was used to compare volumes of 8 subcortical structures (lateral ventricles, thalamus, caudate, putamen, pallidum, hippocampus, amygdala, and nucleus accumbens) in 24 female and 28 male LTAA (mean abstinence=6.3 years, mean age= 46.6 years) and 23 female and 25 male nonalcoholic controls (NAC) (mean age=45.6 years) to explore relations between subcortical brain volumes and alcohol use measures in LTAA and relations between subcortical volumes and psychiatric diagnoses and symptom counts in LTAA and NAC. RESULTS We found minimal differences between LTAA and NAC in subcortical volumes. However, in LTAA, but not NAC, volumes of targeted subcortical structures were smaller in individuals with versus without comorbid lifetime or current psychiatric diagnoses, independent of lifetime alcohol consumption. CONCLUSIONS Our finding of minimal differences in subcortical volumes between LTAA and NAC is consistent with LTAA never having had volume deficits in these regions. However, given that imaging studies have frequently reported smaller subcortical volumes in active and recently detoxified alcoholics compared to controls, our results are also consistent with the recovery of subcortical volumes with sustained abstinence. The finding of persistent smaller subcortical volumes in LTAA, but not NAC, with comorbid psychiatric diagnoses, suggests that the smaller volumes are a result of the combined effects of chronic alcohol dependence and psychiatric morbidity and suggests that a comorbid psychiatric disorder (even if not current) interferes with the recovery of subcortical volumes.