scholarly journals Grey matter biomarker identification in Schizophrenia: detecting regional alterations and their underlying substrates

2018 ◽  
Author(s):  
V. Chatzi ◽  
R.P. Teixeira ◽  
J. Shawe-Taylor ◽  
A. Altmann ◽  
O. O’Daly ◽  
...  
Keyword(s):  
Machine Learning ◽  
Grey Matter ◽  
Structural Mri ◽  
Brain Regions ◽  
List Type ◽  
Healthy Controls ◽  
Learning Approaches ◽  
Learning Framework ◽  
Software Packages ◽  
Biological Substrates

AbstractState-of-the-art approaches in Schizophrenia research investigate neuroanatomical biomarkers using structural Magnetic Resonance Imaging. However, current models are 1) voxel-wise, 2) difficult to interpret in biologically meaningful ways, and 3) difficult to replicate across studies. Here, we propose a machine learning framework that enables the identification of sparse, region-wise grey matter neuroanatomical biomarkers and their underlying biological substrates by integrating well-established statistical and machine learning approaches. We address the computational issues associated with application of machine learning on structural MRI data in Schizophrenia, as discussed in recent reviews, while promoting transparent science using widely available data and software. In this work, a cohort of patients with Schizophrenia and healthy controls was used. It was found that the cortical thickness in left pars orbitalis seems to be the most reliable measure for distinguishing patients with Schizophrenia from healthy controls.HighlightsWe present a sparse machine learning framework to identify biologically meaningful neuroanatomical biomarkers for SchizophreniaOur framework addresses methodological pitfalls associated with application of machine learning on structural MRI data in Schizophrenia raised by several recent reviewsOur pipeline is easy to replicate using widely available software packagesThe presented framework is geared towards identification of specific changes in brain regions that relate directly to the pathology rather than classification per se

Deep Learning Models for the Detection of Plant Leaf Diseases: A Systematic Review

2019 ◽  
Author(s):  
Malusi Sibiya ◽  
Mbuyu Sumbwanyambe
Keyword(s):  
Machine Learning ◽  
Deep Learning ◽  
Learning Systems ◽  
Learning Models ◽  
Plant Leaves ◽  
Plant Leaf ◽  
Learning Framework ◽  
Software Frameworks ◽  
Software Packages ◽  
Learning Frameworks

Machine learning systems use different algorithms to detect the diseases affecting the plant leaves. Nevertheless, selecting a suitable machine learning framework differs from study to study, depending on the features and complexity of the software packages. This paper introduces a taxonomic inspection of the literature in deep learning frameworks for the detection of plant leaf diseases. The objective of this study is to identify the dominating software frameworks in the literature for modelling machine learning plant leaf disease detecting systems.

scholarly journals Building memories on prior knowledge: behavioral and fMRI evidence of impairment in early Alzheimer’s Disease

2020 ◽  
Author(s):  
Pierre-Yves Jonin ◽  
Quentin Duché ◽  
Elise Bannier ◽  
Isabelle Corouge ◽  
Jean-Christophe Ferré ◽  
...  
Keyword(s):  
Alzheimer’S Disease ◽  
Alzheimer's Disease ◽  
Prior Knowledge ◽  
Associative Memory ◽  
Brain Regions ◽  
List Type ◽  
Healthy Controls ◽  
Experimental Knowledge ◽  
Healthy Elderly ◽  
The Right

AbstractImpaired memory is a hallmark of prodromal Alzheimer’s Disease (AD). Prior knowledge associated with the memoranda has proved to have a powerful effect on memory in healthy subjects. Yet, barely nothing is known about its effect in early AD. We used functional MRI to ask whether prior knowledge enhanced memory encoding in early AD and whether the nature of prior knowledge mattered. Early AD patients and healthy controls underwent a task-based fMRI experiment, being scanned while learning face-scene associations. Famous faces carried Pre-Experimental Knowledge (PEK) while unknown faces repeatedly familiarized prior to learning carried Experimental Knowledge (EK). As expected, PEK increased subsequent memory in healthy elderly. However, patients did not benefit from PEK. Partly non-overlapping brain networks supported PEK vs. EK encoding in healthy controls. Patients displayed impaired activation in a right subhippocampal region where activity predicted successful associative memory formation of PEK stimuli. These findings call for a thorough consideration of how prior knowledge impacts learning and suggest a possible underestimation of the extent of associative memory impairment in early AD.HighlightsLearning is impaired in prodromal AD, but we currently ignore whether prior knowledge available at encoding promotes learning in AD as it does in healthy controls.Patients with AD failed to benefit from pre-experimental prior knowledge (famous faces) by comparison with experimental knowledge (unknown but familiarized faces).fMRI responses at study revealed distinct networks underlying associative encoding for both pre-experimental and experimental knowledge.A subsequent memory effect found in control subjects for associations carrying pre-experimental knowledge in the right subhippocampal structures, including the perirhinal cortex, was absent in patients.Pre-experimental knowledge-based associative encoding relies on brain regions specifically targeted by early tau pathology.Using unfamiliar materials to probe memory in early AD might underestimate learning impairment.

scholarly journals Brain age prediction of healthy subjects on anatomic MRI with deep learning: going beyond with an "explainable AI" mindset

2018 ◽  
Author(s):  
Paul Herent ◽  
Simon Jegou ◽  
Gilles Wainrib ◽  
Thomas Clozel
Keyword(s):  
Machine Learning ◽  
Deep Learning ◽  
Grey Matter ◽  
Gray Matter Volume ◽  
Brain Regions ◽  
Gradient Boosting ◽  
Tissue Segmentation ◽  
Caudate Nuclei ◽  
Skull Stripping ◽  
Brain Age

Objectives: Define a clinically usable preprocessing pipeline for MRI data. Predict brain age using various machine learning and deep learning algorithms. Define Caveat against common machine learning traps. Data and Methods: We used 1597 open-access T1 weighted MRI from 24 hospitals. Preprocessing consisted in applying: N4 bias field correction, registration to MNI152 space, white and grey stripe intensity normalization, skull stripping and brain tissue segmentation. Prediction of brain age was done with growing complexity of data input (histograms, grey matter from segmented MRI, raw data) and models for training (linear models, non linear model such as gradient boosting over decision trees, and 2D and 3D convolutional neural networks). Work on interpretability consisted in (i) proceeding on basic data visualization like correlations maps between age and voxels value, and generating (ii) weights maps of simpler models, (iii) heatmap from CNNs model with occlusion method. Results: Processing time seemed feasible in a radiological workflow: 5 min for one 3D T1 MRI. We found a significant correlation between age and gray matter volume with a correlation r = -0.74. Our best model obtained a mean absolute error of 3.60 years, with fine tuned convolution neural network (CNN) pretrained on ImageNet. We carefully analyzed and interpreted the center effect. Our work on interpretability on simpler models permitted to observe heterogeneity of prediction depending on brain regions known for being involved in ageing (grey matter, ventricles). Occlusion method of CNN showed the importance of Insula and deep grey matter (thalami, caudate nuclei) in predictions. Conclusions: Predicting the brain age using deep learning could be a standardized metric usable in daily neuroradiological reports. An explainable algorithm gives more confidence and acceptability for its use in practice. More clinical studies using this new quantitative biomarker in neurological diseases will show how to use it at its best.

scholarly journals Altered Topological Properties of Grey Matter Structural Covariance Networks in Complete Thoracic Spinal Cord Injury Patients: A Graph Theoretical Network Analysis

2021 ◽  
Vol 2021 ◽  
pp. 1-11
Author(s):  
Wen-Li Wang ◽  
Yu-Lin Li ◽  
Mou-Xiong Zheng ◽  
Xu-Yun Hua ◽  
Jia-Jia Wu ◽  
...  
Keyword(s):  
Grey Matter ◽  
Brain Regions ◽  
Healthy Controls ◽  
Age And Gender ◽  
Structural Covariance ◽  
Clinical Variables ◽  
Cord Injury ◽  
Network Properties ◽  
And Gender ◽  
The Right

Purpose. This study is aimed at investigating brain structural changes and structural network properties in complete spinal cord injury (SCI) patients, as well as their relationship with clinical variables. Materials and Methods. Structural MRI of brain was acquired in 24 complete thoracic SCI patients ( 38.50 ± 11.19 years, 22 males) within the first postinjury year, while 26 age- and gender-matched healthy participants ( 38.38 ± 10.63 years, 24 males) were enrolled as control. The voxel-based morphometry (VBM) approach and graph theoretical network analysis based on cross-subject grey matter volume- (GMV-) based structural covariance networks (SCNs) were conducted to investigate the impact of SCI on brain structure. Partial correlation analysis was performed to explore the relationship between the GMV of structurally changed brain regions and SCI patients’ clinical variables, including injury duration, injury level, Visual Analog Scale (VAS), American Spinal Injury Association Impairment Scale (AIS), International Classification of Functioning, Disability and Health (ICF) scale, Self-rating Depression Scale (SDS), and Self-rating Anxiety Scale (SAS), after removing the effects of age and gender. Results. Compared with healthy controls, SCI patients showed higher SDS score ( t = 4.392 and p < 0.001 ). In the VBM analysis, significant GMV reduction was found in the left middle frontal cortex, right superior orbital frontal cortex (OFC), and left inferior OFC. No significant difference was found in global network properties between SCI patients and healthy controls. In the regional network properties, significantly higher betweenness centrality (BC) was noted in the right anterior cingulum cortex (ACC) and left inferior OFC and higher nodal degree and efficiency in bilateral middle OFCs, while decreased BC was noted in the right putamen in SCI patients. Only negative correlation was found between GMV of right middle OFC and SDS score in SCI patients ( r = − 0.503 and p = 0.017 ), while no significant correlation between other abnormal brain regions and any of the clinical variables (all p > 0.05 ). Conclusions. SCI patients would experience depressive and/or anxious feelings at the early stage. Their GMV reduction mainly involved psychology-cognition related rather than sensorimotor brain regions. The efficiency of regional information transmission in psychology-cognition regions increased. Greater GMV reduction in psychology region was related with more severe depressive feelings. Therefore, early neuropsychological intervention is suggested to prevent psychological and cognitive dysfunction as well as irreversible brain structure damage.

scholarly journals A simple classification framework for predicting Alzheimer’s disease from region-based grey matter volume and APOE genotype status

2020 ◽  
Vol 8 (2) ◽  
pp. 15
Author(s):  
Reyhaneh Ghoreishiamiri ◽  
Graham Little ◽  
Matthew R. G. Brown ◽  
Russell Greiner
Keyword(s):  
Alzheimer’S Disease ◽  
Alzheimer's Disease ◽  
Grey Matter ◽  
Learning Algorithm ◽  
Apoe Genotype ◽  
Structural Mri ◽  
Grey Matter Volume ◽  
Healthy Controls ◽  
Diagnostic Systems ◽  
Initial Dataset

Alzheimer’s Disease (AD) is a prevalent neurodegenerative disease currently affecting more than 47 million people in the world. There are now many complex classifiers that can accurately distinguish AD patients from healthy controls, based on the subject’s structural magnetic resonance imaging (MRI) brain scan. Most such automated diagnostic systems are blackboxes: While their predictions are accurate, it is difficult for clinicians to interpret those predictions, due to the large number of features used by the classifier, and/or by the complexity of that classifier. This work demonstrates that an automated learning algorithm can produce a simple classifier that can correctly distinguish AD patients from healthy controls (HC) similar to its more-complex counterparts. Here we buildthis classifier from the data in the Alzheimer’s Disease Neuroimaging Initiative database, using a fairly small set of features, including grey matter volumes of 33 regions of interest derived from structural MRI, as well as the APOE genotype. We first considered three simple base-learners that each produce a classifier that is simple and interpretable. Running our overall learner, involving standard feature selection processes and these simple base-learners, on these features, produced a 7-feature elastic net model, EN7, that achieved accuracy of 89.28% on the test set. Next, we ran the same overall learner using two more-complex base-learners over the same initial dataset. The accuracy of the best model here was 90.47%, which was not statistically different from the performance of our much simpler EN7 model.

scholarly journals MLCA—A Machine Learning Framework for INS Coarse Alignment

Sensors ◽  
2020 ◽  
Vol 20 (23) ◽  
pp. 6959
Author(s):  
Idan Zak ◽  
Reuven Katz ◽  
Itzik Klein
Keyword(s):  
Machine Learning ◽  
Field Experiments ◽  
Inertial Sensors ◽  
Initial Conditions ◽  
Low Cost ◽  
Navigation Systems ◽  
Learning Approaches ◽  
Alignment Procedure ◽  
Navigation Solution ◽  
Learning Framework

Inertial navigation systems provides the platform’s position, velocity, and attitude during its operation. As a dead-reckoning system, it requires initial conditions to calculate the navigation solution. While initial position and velocity vectors are provided by external means, the initial attitude can be determined using the system’s inertial sensors in a process known as coarse alignment. When considering low-cost inertial sensors, only the initial roll and pitch angles can be determined using the accelerometers measurements. The accuracy, as well as time required for the for the coarse alignment process are critical for the navigation solution accuracy, particularly for pure-inertial scenarios, because of the navigation solution drift. In this paper, a machine learning framework for the stationary coarse alignment stage is proposed. To that end, classical machine learning approaches are used in a two-stage approach to regress the roll and pitch angles. Alignment results obtained both in simulations and field experiments, using a smartphone, shows the benefits of using the proposed approach instead of the commonly used analytical coarse alignment procedure.

Abnormal cerebellar functional MRI connectivity in patients with paediatric multiple sclerosis

2015 ◽  
Vol 22 (3) ◽  
pp. 292-301 ◽  
Author(s):  
Sara Cirillo ◽  
Maria A Rocca ◽  
Angelo Ghezzi ◽  
Paola Valsasina ◽  
Lucia Moiola ◽  
...  
Keyword(s):  
Disease Duration ◽  
Structural Mri ◽  
Brain Regions ◽  
Lesion Volume ◽  
Seed Region ◽  
Healthy Controls ◽  
Resting State Functional Connectivity ◽  
Focal Lesions ◽  
Caudate Nuclei ◽  
The Right

Objectives: We investigated resting state functional connectivity (RSFC) of the cerebellar dentate nuclei in paediatric MS patients and its correlations with clinical, neuropsychological and structural MRI measures. Methods: RSFC analysis was performed using a seed-region correlation approach and SPM8 from 48 paediatric MS patients and 27 matched healthy controls. Results: In both groups, dentate nuclei RSFC was significantly correlated with RSFC of several cerebellar and extra-cerebellar brain regions. Compared with healthy controls, paediatric MS patients had reduced RSFC between the right dentate nuclei and the bilateral caudate nuclei and left thalamus as well as increased RSFC between the right dentate nuclei and the left precentral and postcentral gyri. Cognitively impaired patients showed a reduced RSFC between the dentate nuclei and bilateral regions located in the parietal, frontal and temporal lobes. Decreased RSFC was correlated with longer disease duration and higher T2 lesion volumes, whereas increased RSFC correlated with shorter disease duration, lower T2 lesion volume and a better motor performance. Conclusions: Modifications of cerebellar RSFC occur in paediatric MS and are influenced by the duration of the disease and brain focal lesions. Decreased RSFC may reflect early maladaptive plasticity contributing to cognitive impairment.

scholarly journals Machine Learning Framework for Fully Automatic Quality Checking of Rigid and Affine Registrations in Big Data Brain MRI

2020 ◽  
Author(s):  
Sudhakar Tummala ◽  
Niels K. Focke
Keyword(s):  
Machine Learning ◽  
Data Exchange ◽  
Brain Mri ◽  
Structural Mri ◽  
Imaging Data ◽  
Normalized Mutual Information ◽  
Learning Framework ◽  
Fully Automatic ◽  
Magnetic Resonance Imaging Mri ◽  
Brain Imaging Data

ABSTRACTRigid and affine registrations to a common template are the essential steps during pre-processing of brain structural magnetic resonance imaging (MRI) data. Manual quality check (QC) of these registrations is quite tedious if the data contains several thousands of images. Therefore, we propose a machine learning (ML) framework for fully automatic QC of these registrations via local computation of the similarity functions such as normalized cross-correlation, normalized mutual-information, and correlation ratio, and using these as features for training of different ML classifiers. To facilitate supervised learning, misaligned images are generated. A structural MRI dataset consisting of 215 subjects from autism brain imaging data exchange is used for 5-fold cross-validation and testing. Few classifiers such as kNN, AdaBoost, and random forest reached testing F1-scores of 0.98 for QC of both rigid and affine registrations. These tested ML models could be deployed for practical use.

scholarly journals Distinctive Oculomotor Behaviors in Alzheimer's Disease and Frontotemporal Dementia

2021 ◽  
Vol 12 ◽  
Author(s):  
Carmen Lage ◽  
Sara López-García ◽  
Alexandre Bejanin ◽  
Martha Kazimierczak ◽  
Ignacio Aracil-Bolaños ◽  
...  
Keyword(s):  
Machine Learning ◽  
Alzheimer’S Disease ◽  
Alzheimer's Disease ◽  
Eye Tracking ◽  
Frontotemporal Dementia ◽  
Brain Regions ◽  
Middle Temporal Gyrus ◽  
Classification Algorithms ◽  
Learning Approaches ◽  
Machine Learning Classification

Oculomotor behavior can provide insight into the integrity of widespread cortical networks, which may contribute to the differential diagnosis between Alzheimer's disease and frontotemporal dementia. Three groups of patients with Alzheimer's disease, behavioral variant of frontotemporal dementia (bvFTD) and semantic variant of primary progressive aphasia (svPPA) and a sample of cognitively unimpaired elders underwent an eye-tracking evaluation. All participants in the discovery sample, including controls, had a biomarker-supported diagnosis. Oculomotor correlates of neuropsychology and brain metabolism evaluated with 18F-FDG PET were explored. Machine-learning classification algorithms were trained for the differentiation between Alzheimer's disease, bvFTD and controls. A total of 93 subjects (33 Alzheimer's disease, 24 bvFTD, seven svPPA, and 29 controls) were included in the study. Alzheimer's disease was the most impaired group in all tests and displayed specific abnormalities in some visually-guided saccade parameters, as pursuit error and horizontal prosaccade latency, which are theoretically closely linked to posterior brain regions. BvFTD patients showed deficits especially in the most cognitively demanding tasks, the antisaccade and memory saccade tests, which require a fine control from frontal lobe regions. SvPPA patients performed similarly to controls in most parameters except for a lower number of correct memory saccades. Pursuit error was significantly correlated with cognitive measures of constructional praxis and executive function and metabolism in right posterior middle temporal gyrus. The classification algorithms yielded an area under the curve of 97.5% for the differentiation of Alzheimer's disease vs. controls, 96.7% for bvFTD vs. controls, and 92.5% for Alzheimer's disease vs. bvFTD. In conclusion, patients with Alzheimer's disease, bvFTD and svPPA exhibit differentiating oculomotor patterns which reflect the characteristic neuroanatomical distribution of pathology of each disease, and therefore its assessment can be useful in their diagnostic work-up. Machine learning approaches can facilitate the applicability of eye-tracking in clinical practice.

Sign in / Sign up

Export Citation Format

Share Document