Habib Benali

Automatic classification of patients with Alzheimer's disease from structural MRI: A comparison of ten methods using the ADNI database

Recently, several high dimensional classification methods have been proposed to automatically discriminate between patients with Alzheimers disease (AD) or mild cognitive impairment (MCI) and elderly controls (CN) based on T1-weighted MRI. However, these methods were assessed on different populations, making it difficult to compare their performance. In this paper, we evaluated the performance of ten approaches (five voxel-based methods, three methods based on cortical thickness and two methods based on the hippocampus) using 509 subjects from the ADNI database. Three classification experiments were performed: CN vs AD, CN vs MCIc (MCI who had converted to AD within 18 months, MCI converters - MCIc) and MCIc vs MCInc (MCI who had not converted to AD within 18 months, MCI non-converters - MCInc). Data from 81 CN, 67 MCInc, 39 MCIc and 69 AD were used for training and hyperparameters optimization. The remaining independent samples of 81 CN, 67 MCInc, 37 MCIc and 68 AD were used to obtain an unbiased estimate of the performance of the methods. For AD vs CN, whole-brain methods (voxel-based or cortical thickness-based) achieved high accuracies (up to 81% sensitivity and 95% specificity). For the detection of prodromal AD (CN vs MCIc), the sensitivity was substantially lower. For the prediction of conversion, no classifier obtained significantly better results than chance. We also compared the results obtained using the DARTEL registration to that using SPM5 unified segmentation. DARTEL significantly improved six out of 20 classification experiments and led to lower results in only two cases. Overall, the use of feature selection did not improve the performance but substantially increased the computation times.

Contributions of the basal ganglia and functionally related brain structures to motor learning

This review discusses the cerebral plasticity, and the role of the cortico-striatal system in particular, observed as one is learning or planning to execute a newly learned motor behavior up to when the skill is consolidated or has become highly automatized. A special emphasis is given to imaging work describing the neural substrate mediating motor sequence learning and motor adaptation paradigms. These results are then put into a plausible neurobiological model of motor skill learning, which proposes an integrated view of the brain plasticity mediating this form of memory at different stages of the acquisition process.

Preclinical Alzheimer's disease: Definition, natural history, and diagnostic criteria

During the past decade, a conceptual shift occurred in the field of Alzheimers disease (AD) considering the disease as a continuum. Thanks to evolving biomarker research and substantial discoveries, it is now possible to identify the disease even at the preclinical stage before the occurrence of the first clinical symptoms. This preclinical stage of AD has become a major research focus as the field postulates that early intervention may offer the best chance of therapeutic success. To date, very little evidence is established on this “silent” stage of the disease. A clarification is needed about the definitions and lexicon, the limits, the natural history, the markers of progression, and the ethical consequence of detecting the disease at this asymptomatic stage. This article is aimed at addressing all the different issues by providing for each of them an updated review of the literature and evidence, with practical recommendations.

Partial correlation for functional brain interactivity investigation in functional MRI.

Examination of functional interactions through effective connectivity requires the determination of three distinct levels of information: (1) the regions involved in the process and forming the spatial support of the network, (2) the presence or absence of interactions between each pair of regions, and (3) the directionality of the existing interactions. While many methods exist to select regions (Step 1), very little is available to complete Step 2. The two main methods developed so far, structural equation modeling (SEM) and dynamical causal modeling (DCM), usually require precise prior information to be used, while such information is sometimes lacking. Assuming that Step 1 was successfully completed, we here propose a data-driven method to deal with Step 2 and extract functional interactions from fMRI datasets through partial correlations. Partial correlation is more closely related to effective connectivity than marginal correlation and provides a convenient graphical representation for functional interactions. As an instance of brain interactivity investigation, we consider how simple hand movements are processed by the bihemispheric cortical motor network. In the proposed framework, Bayesian analysis makes it possible to estimate and test the partial statistical dependencies between regions without any prior model on the underlying functional interactions. We demonstrate the interest of this approach on real data.

Discrimination between Alzheimer Disease, Mild Cognitive Impairment, and Normal Aging by Using Automated Segmentation of the Hippocampus

PURPOSE To prospectively evaluate the accuracy of automated hippocampal volumetry to help distinguish between patients with Alzheimer disease (AD), patients with mild cognitive impairment (MCI), and elderly controls, by using established criteria for patients with AD and MCI as the reference standard. MATERIALS AND METHODS The regional ethics committee approved the study and written informed consent was obtained from all participants. The study included 25 patients with AD (11 men, 14 women; mean age +/- standard deviation [SD], 73 years +/- 6; Mini-Mental State Examination (MMSE) score, 24.4 +/- 2.7), 24 patients with amnestic MCI (10 men, 14 women; mean age +/- SD, 74 years +/- 8; MMSE score, 27.2 +/- 1.4) and 25 elderly healthy controls (13 men, 12 women; mean age +/- SD, 64 years +/- 8). For each participant, the hippocampi were automatically segmented on three-dimensional T1-weighted magnetic resonance (MR) images with high spatial resolution. Segmentation was performed by using recently developed software that allows fast segmentation with minimal user input. Group differences in hippocampal volume were assessed by using Student t tests. To obtain robust estimates of P values, the correct classification rate, sensitivity, and specificity, bootstrap methods were used. RESULTS Significant hippocampal volume reductions were detected in all groups of patients (-32% in AD patients vs controls, P < .001; -19% in MCI patients vs controls, P < .001; and -15% in AD patients vs MCI patients, P < .01). Individual classification on the basis of hippocampal volume resulted in 84% correct classification (sensitivity, 84%; specificity, 84%) between AD patients and controls and 73% correct classification (sensitivity, 75%; specificity, 70%) between MCI patients and controls. CONCLUSION This automated method can serve as an alternative to manual tracing and may thus prove useful in assisting with the diagnosis of AD.

Fully Automatic Hippocampus Segmentation and Classification in Alzheimer's Disease and Mild Cognitive Impairment Applied on Data from ADNI

The hippocampus is among the first structures affected in Alzheimers disease (AD). Hippocampal magnetic resonance imaging volumetry is a potential biomarker for AD but is hindered by the limitations of manual segmentation. We proposed a fully automatic method using probabilistic and anatomical priors for hippocampus segmentation. Probabilistic information is derived from 16 young controls and anatomical knowledge is modeled with automatically detected landmarks. The results were previously evaluated by comparison with manual segmentation on data from the 16 young healthy controls, with a leave‐one‐out strategy, and eight patients with AD. High accuracy was found for both groups (volume error 6 and 7%, overlap 87 and 86%, respectively). In this article, the method was used to segment 145 patients with AD, 294 patients with mild cognitive impairment (MCI), and 166 elderly normal subjects from the Alzheimers Disease Neuroimaging Initiative database. On the basis of a qualitative rating protocol, the segmentation proved acceptable in 94% of the cases. We used the obtained hippocampal volumes to automatically discriminate between AD patients, MCI patients, and elderly controls. The classification proved accurate: 76% of the patients with AD and 71% of the MCI converting to AD before 18 months were correctly classified with respect to the elderly controls, using only hippocampal volume.

Simulation of anisotropic growth of low-grade gliomas using diffusion tensor imaging

A recent computational model of brain tumor growth, developed to better describe how gliomas invade through the adjacent brain parenchyma, is based on two major elements: cell proliferation and isotropic cell diffusion. On the basis of this model, glioma growth has been simulated in a virtual brain, provided by a 3D segmented MRI atlas. However, it is commonly accepted that glial cells preferentially migrate along the direction of fiber tracts. Therefore, in this paper, the model has been improved by including anisotropic extension of gliomas. The method is based on a cell diffusion tensor derived from water diffusion tensor (as given by MRI diffusion tensor imaging). Results of simulations have been compared with two clinical examples demonstrating typical growth patterns of low‐grade gliomas centered around the insula. The shape and the kinetic evolution are better simulated with anisotropic rather than isotropic diffusion. The best fit is obtained when the anisotropy of the cell diffusion tensor is increased to greater anisotropy than the observed water diffusion tensor. The shape of the tumor is also influenced by the initial location of the tumor. Anisotropic brain tumor growth simulations provide a means to determine the initial location of a low‐grade glioma as well as its cell diffusion tensor, both of which might reflect the biological characteristics of invasion. Magn Reson Med, 2005.

Multi-level bootstrap analysis of stable clusters in resting-state fMRI

A variety of methods have been developed to identify brain networks with spontaneous, coherent activity in resting-state functional magnetic resonance imaging (fMRI). We propose here a generic statistical framework to quantify the stability of such resting-state networks (RSNs), which was implemented with k-means clustering. The core of the method consists in bootstrapping the available datasets to replicate the clustering process a large number of times and quantify the stable features across all replications. This bootstrap analysis of stable clusters (BASC) has several benefits: (1) it can be implemented in a multi-level fashion to investigate stable RSNs at the level of individual subjects and at the level of a group; (2) it provides a principled measure of RSN stability; and (3) the maximization of the stability measure can be used as a natural criterion to select the number of RSNs. A simulation study validated the good performance of the multi-level BASC on purely synthetic data. Stable networks were also derived from a real resting-state study for 43 subjects. At the group level, seven RSNs were identified which exhibited a good agreement with the previous findings from the literature. The comparison between the individual and group-level stability maps demonstrated the capacity of BASC to establish successful correspondences between these two levels of analysis and at the same time retain some interesting subject-specific characteristics, e.g. the specific involvement of subcortical regions in the visual and fronto-parietal networks for some subjects.

Brain plasticity related to the consolidation of motor sequence learning and motor adaptation

This study aimed to investigate, through functional MRI (fMRI), the neuronal substrates associated with the consolidation process of two motor skills: motor sequence learning (MSL) and motor adaptation (MA). Four groups of young healthy individuals were assigned to either (i) a night/sleep condition, in which they were scanned while practicing a finger sequence learning task or an eight-target adaptation pointing task in the evening (test) and were scanned again 12 h later in the morning (retest) or (ii) a day/awake condition, in which they were scanned on the MSL or the MA tasks in the morning and were rescanned 12 h later in the evening. As expected and consistent with the behavioral results, the functional data revealed increased test–retest changes of activity in the striatum for the night/sleep group compared with the day/awake group in the MSL task. By contrast, the results of the MA task did not show any difference in test–retest activity between the night/sleep and day/awake groups. When the two MA task groups were combined, however, increased test–retest activity was found in lobule VI of the cerebellar cortex. Together, these findings highlight the presence of both functional and structural dissociations reflecting the off-line consolidation processes of MSL and MA. They suggest that MSL consolidation is sleep dependent and reflected by a differential increase of neural activity within the corticostriatal system, whereas MA consolidation necessitates either a period of daytime or sleep and is associated with increased neuronal activity within the corticocerebellar system.

Demyelination and degeneration in the injured human spinal cord detected with diffusion and magnetization transfer MRI

Characterizing demyelination/degeneration of spinal pathways in traumatic spinal cord injured (SCI) patients is crucial for assessing the prognosis of functional rehabilitation. Novel techniques based on diffusion-weighted (DW) magnetic resonance imaging (MRI) and magnetization transfer (MT) imaging provide sensitive and specific markers of white matter pathology. In this paper we combined for the first time high angular resolution diffusion-weighted imaging (HARDI), MT imaging and atrophy measurements to evaluate the cervical spinal cord of fourteen SCI patients and age-matched controls. We used high in-plane resolution to delineate dorsal and ventrolateral pathways. Significant differences were detected between patients and controls in the normal-appearing white matter for fractional anisotropy (FA, p<0.0001), axial diffusivity (p<0.05), radial diffusivity (p<0.05), generalized fractional anisotropy (GFA, p<0.0001), magnetization transfer ratio (MTR, p<0.0001) and cord area (p<0.05). No significant difference was detected in mean diffusivity (p=0.41), T1-weighted (p=0.76) and T2-weighted (p=0.09) signals. MRI metrics were remarkably well correlated with clinical disability (Pearsons correlations, FA: p<0.01, GFA: p<0.01, radial diffusivity: p=0.01, MTR: p=0.04 and atrophy: p<0.01). Stepwise linear regressions showed that measures of MTR in the dorsal spinal cord predicted the sensory disability whereas measures of MTR in the ventro-lateral spinal cord predicted the motor disability (ASIA score). However, diffusion metrics were not specific to the sensorimotor scores. Due to the specificity of axial and radial diffusivity and MT measurements, results suggest the detection of demyelination and degeneration in SCI patients. Combining HARDI with MT imaging is a promising approach to gain specificity in characterizing spinal cord pathways in traumatic injury.