scholarly journals Deep neural networks learn general and clinically relevant representations of the ageing brain

2021 ◽  
Author(s):  
Esten H Leonardsen ◽  
Han Peng ◽  
Tobias Kaufmann ◽  
Ingrid Agartz ◽  
Ole A Andreassen ◽  
...  
Keyword(s):  
Neural Networks ◽  
Magnetic Resonance Images ◽  
Brain Health ◽  
Unseen Data ◽  
Heterogeneous Datasets ◽  
Neuroimaging Data ◽  
Reliable Marker ◽  
Brain Age ◽  
Ageing Brain ◽  
The Brain

The discrepancy between chronological age and the apparent age of the brain based on neuroimaging data - the brain age delta - has emerged as a reliable marker of brain health. With an increasing wealth of data, approaches to tackle heterogeneity in data acquisition are vital. To this end, we compiled raw structural magnetic resonance images into one of the largest and most diverse datasets assembled (n=53542), and trained convolutional neural networks (CNNs) to predict age. We achieved state-of-the-art performance on unseen data from unknown scanners (n=2553), and showed that higher brain age delta is associated with diabetes, alcohol intake and smoking. Using transfer learning, the intermediate representations learned by our model complemented and partly outperformed brain age delta in predicting common brain disorders. Our work shows we can achieve generalizable and biologically plausible brain age predictions using CNNs trained on heterogeneous datasets, and transfer them to clinical use cases.

scholarly journals Artificial Intelligence to Analyze the Cortical Thickness Through Age

2021 ◽  
Vol 4 ◽  
Author(s):  
Sergio Ledesma ◽  
Mario-Alberto Ibarra-Manzano ◽  
Dora-Luz Almanza-Ojeda ◽  
Pascal Fallavollita ◽  
Jason Steffener
Keyword(s):  
Artificial Intelligence ◽  
Neural Networks ◽  
Artificial Neural Networks ◽  
Left Hemisphere ◽  
Cortical Thickness ◽  
Right Hemisphere ◽  
Unseen Data ◽  
Artificial Neural ◽  
The Right ◽  
The Brain

In this study, Artificial Intelligence was used to analyze a dataset containing the cortical thickness from 1,100 healthy individuals. This dataset had the cortical thickness from 31 regions in the left hemisphere of the brain as well as from 31 regions in the right hemisphere. Then, 62 artificial neural networks were trained and validated to estimate the number of neurons in the hidden layer. These neural networks were used to create a model for the cortical thickness through age for each region in the brain. Using the artificial neural networks and kernels with seven points, numerical differentiation was used to compute the derivative of the cortical thickness with respect to age. The derivative was computed to estimate the cortical thickness speed. Finally, color bands were created for each region in the brain to identify a positive derivative, that is, a part of life with an increase in cortical thickness. Likewise, the color bands were used to identify a negative derivative, that is, a lifetime period with a cortical thickness reduction. Regions of the brain with similar derivatives were organized and displayed in clusters. Computer simulations showed that some regions exhibit abrupt changes in cortical thickness at specific periods of life. The simulations also illustrated that some regions in the left hemisphere do not follow the pattern of the same region in the right hemisphere. Finally, it was concluded that each region in the brain must be dynamically modeled. One advantage of using artificial neural networks is that they can learn and model non-linear and complex relationships. Also, artificial neural networks are immune to noise in the samples and can handle unseen data. That is, the models based on artificial neural networks can predict the behavior of samples that were not used for training. Furthermore, several studies have shown that artificial neural networks are capable of deriving information from imprecise data. Because of these advantages, the results obtained in this study by the artificial neural networks provide valuable information to analyze and model the cortical thickness.

scholarly journals The Association Between Poor Sleep and Accelerated Brain Ageing

2021 ◽  
Author(s):  
Jivesh Ramduny ◽  
Matteo Bastiani ◽  
Robin Huedepohl ◽  
Stamatios Sotiropoulos N Sotiropoulos ◽  
Magdalena Chechlacz
Keyword(s):  
Older Adults ◽  
Sleep Quality ◽  
Sleep Problems ◽  
Sleep Fragmentation ◽  
Poor Sleep ◽  
Age Related ◽  
Brain Ageing ◽  
Neuroimaging Data ◽  
Brain Age ◽  
Ageing Brain

The ageing brain undergoes widespread gray (GM) and white matter (WM) degeneration. But numerous studies indicate large heterogeneity in the age-related brain changes, which can be attributed to modifiable lifestyle factors, including sleep. Inadequate sleep has been previously linked to GM atrophy and WM changes. However, the reported findings are highly inconsistent. By contrast to previous research independently characterizing patterns of either the GM or the WM changes, we used here linked independent component analysis (FLICA) to examine covariation in GM and WM in a group of older adults. Next, we employed a novel technique to estimate the brain age delta (i.e. difference between chronological and apparent brain age assessed using neuroimaging data) and study its associations with sleep quality and sleep fragmentation, hypothesizing that poor sleep accelerates brain ageing. FLICA revealed a number of multimodal (including both GM and WM) neuroimaging components, associated with age, but also with sleep quality and sleep fragmentation. Brain age delta estimates were highly sensitive in detecting the effects of sleep problems on the ageing brain. Specifically, we show significant associations between brain age delta and poor sleep quality, suggesting two years deviation above the chronological age. Our findings indicate that sleep problems in healthy older adults should be considered a risk factor for accelerated brain ageing.

scholarly journals Brain Age Prediction With Morphological Features Using Deep Neural Networks: Results From Predictive Analytic Competition 2019

2021 ◽  
Vol 11 ◽  
Author(s):  
Angela Lombardi ◽  
Alfonso Monaco ◽  
Giacinto Donvito ◽  
Nicola Amoroso ◽  
Roberto Bellotti ◽  
...  
Keyword(s):  
Neural Networks ◽  
Deep Neural Networks ◽  
Morphological Changes ◽  
Structure Characterization ◽  
Feature Importance ◽  
Mri Scans ◽  
Magnetic Resonance Imaging Mri ◽  
Brain Age ◽  
The Brain ◽  
Age Prediction

Morphological changes in the brain over the lifespan have been successfully described by using structural magnetic resonance imaging (MRI) in conjunction with machine learning (ML) algorithms. International challenges and scientific initiatives to share open access imaging datasets also contributed significantly to the advance in brain structure characterization and brain age prediction methods. In this work, we present the results of the predictive model based on deep neural networks (DNN) proposed during the Predictive Analytic Competition 2019 for brain age prediction of 2638 healthy individuals. We used FreeSurfer software to extract some morphological descriptors from the raw MRI scans of the subjects collected from 17 sites. We compared the proposed DNN architecture with other ML algorithms commonly used in the literature (RF, SVR, Lasso). Our results highlight that the DNN models achieved the best performance with MAE = 4.6 on the hold-out test, outperforming the other ML strategies. We also propose a complete ML framework to perform a robust statistical evaluation of feature importance for the clinical interpretability of the results.

scholarly journals eQTL of KCNK2 regionally influences the brain sulcal widening: evidence from 15,597 UK Biobank participants with neuroimaging data

2018 ◽  
Author(s):  
Yann Le Guen ◽  
Cathy Philippe ◽  
Denis Riviere ◽  
Hervé Lemaitre ◽  
Antoine Grigis ◽  
...  
Keyword(s):  
Immune Cell ◽  
Magnetic Resonance Images ◽  
Cortical Atrophy ◽  
Uk Biobank ◽  
Cns Inflammation ◽  
Neuroimaging Data ◽  
The Central Nervous System ◽  
Trait Locus ◽  
The Uk ◽  
The Brain

AbstractThe grey and white matter volumes are known to reduce with age. This cortical shrinkage is visible on magnetic resonance images and is conveniently identified by the increased volume of cerebrospinal fluid in the sulci between two gyri. Here, we replicated this finding using the UK Biobank dataset and studied the genetic influence on these cortical features of aging. We divided all individuals genetically confirmed of British ancestry into two sub-cohorts (12,162 and 3,435 subjects for discovery and replication samples, respectively). We found that the heritability of the sulcal opening ranges from 15 to 45% (s.e.= 4.8%). We identified 4 new loci that contribute to this opening, including one that also affects the sulci grey matter thickness. We identified the most significant variant (rs864736) on this locus as being an expression quantitative trait locus (eQTL) for the KCNK2 gene. This gene regulates the immune-cell into the central nervous system (CNS) and controls the CNS inflammation, which is implicated in cortical atrophy and cognitive decline. These results expand our knowledge of the genetic contribution to cortical shrinking and promote further investigation into these variants and genes in pathological context such as Alzheimer’s disease in which brain shrinkage is a key biomarker.

scholarly journals Smoking, but Not Other Environmental Exposures, Increases Risk for Advanced Brain Aging

2020 ◽  
Vol 4 (Supplement_1) ◽  
pp. 280-281
Author(s):  
Nathan Whitsel ◽  
Carol Franz ◽  
William Kremen
Keyword(s):  
Tobacco Smoke ◽  
Tobacco Smoking ◽  
Environmental Exposures ◽  
Magnetic Resonance Images ◽  
Agent Orange ◽  
Brain Health ◽  
Former Smokers ◽  
Harmful Substances ◽  
Brain Age ◽  
Metal Dust

Abstract Exposure to harmful substances and chemicals such as tobacco smoke, chemicals (e.g., herbicides, pesticides, Agent Orange) and metal dust has been associated with increased risk of developing cancer, cardiovascular disease, and other diseases that contribute to shorter life expectancy. Associations with brain health in relation to these exposures are less well studied. We examined the relationship between brain health and prolonged exposure to different harmful substances in 498 male participants average age 68 (range 61 to 73) from the Vietnam Era Twin Study of Aging (VETSA). For self-reported tobacco smoke, herbicides/pesticides, and metal dust we created three groups reflecting recency of exposure (current/former/never). For Agent Orange we examined two exposure groups (ever/never). Brain health, defined as predicted brain age (PBAD), was evaluated by applying Brain-Age Regression Analysis and Computation Utility software (BARACUS) to magnetic resonance images collected at age 68. Tobacco smoking (r=-0.15, p=0.0004 ) was significantly correlated with PBAD and remained significant (F=5.56, p=0.005) in multivariate analyses adjusted for age, socioeconomic status (SES), age 20 general cognitive ability, and non-independence of twins within pairs. Never smokers had significantly younger brains than current or former smokers. PBAD did not differ for current versus former smokers. In other analyses, more advanced PBAD was associated with non-amnestic MCI. In this sample, tobacco smoking had the strongest relationship with overall brain health in late midlife compared with other types of environmental exposures, reinforcing its role in pathological aging and its importance as a public health priority.

scholarly journals Decentralized Brain Age Estimation using MRI Data

2021 ◽  
Author(s):  
Sunitha Basodi ◽  
Rajikha Raja ◽  
Bhaskar Ray ◽  
Harshvardhan Gazula ◽  
Jingyu Liu ◽  
...  
Keyword(s):  
Age Estimation ◽  
Aging Process ◽  
Prediction Models ◽  
Structural Mri ◽  
Functional Changes ◽  
Feature Sets ◽  
Neuroimaging Data ◽  
Age Model ◽  
Brain Age ◽  
The Brain

Recent studies have demonstrated that neuroimaging data can be used to predict brain age, as it captures information about the neuroanatomical and functional changes the brain undergoes during development and the aging process. However, researchers often have limited access to neuroimaging data because of its challenging and expensive acquisition process, thereby limiting the effectiveness of the predictive model. Decentralized models provide a way to build more accurate and generalizable prediction models, bypassing the traditional data-sharing methodology. In this work, we propose a decentralized method for brain age estimation and evaluate it on three different feature sets, including both volumetric and voxelwise structural MRI data as well as resting functional MRI data. The results demonstrate that a decentralized brain age model can achieve similar performance compared to the models trained with all the data in one location.

Brain Tumor Detection via Asymmetry Quantification across Mid Sagittal Plane

2020 ◽  
Vol 13 ◽  
Author(s):  
Shoaib Amin Banday ◽  
Mohammad Khalid Pandit
Keyword(s):  
Brain Tumor ◽  
Brain Tumors ◽  
Sagittal Plane ◽  
Early Stage ◽  
Imaging Techniques ◽  
Complex Structure ◽  
Magnetic Resonance Images ◽  
Absolute Difference ◽  
Brain Hemispheres ◽  
The Brain

Introduction: Brain tumor is among the major causes of morbidity and mortality rates worldwide. According to National Brain Tumor Foundation (NBTS), the death rate has nearly increased by as much as 300% over last couple of decades. Tumors can be categorized as benign (non-cancerous) and malignant (cancerous). The type of the brain tumor significantly depends on various factors like the site of its occurrence, its shape, the age of the subject etc. On the other hand, Computer Aided Detection (CAD) has been improving significantly in recent times. The concept, design and implementation of these systems ascend from fairly simple ones to computationally intense ones. For efficient and effective diagnosis and treatment plans in brain tumor studies, it is imperative that an abnormality is detected at an early stage as it provides a little more time for medical professionals to respond. The early detection of diseases has predominantly been possible because of medical imaging techniques developed from past many decades like CT, MRI, PET, SPECT, FMRI etc. The detection of brain tumors however, has always been a challenging task because of the complex structure of the brain, diverse tumor sizes and locations in the brain. Method: This paper proposes an algorithm that can detect the brain tumors in the presence of the Radio-Frequency (RF) inhomoginiety. The algorithm utilizes the Mid Sagittal Plane as a landmark point across which the asymmetry between the two brain hemispheres is estimated using various intensity and texture based parameters. Result: The results show the efficacy of the proposed method for the detection of the brain tumors with an acceptable detection rate. Conclusion: In this paper, we have calculated three textural features from the two hemispheres of the brain viz: Contrast (CON), Entropy (ENT) and Homogeneity (HOM) and three parameters viz: Root Mean Square Error (RMSE), Correlation Co-efficient (CC), and Integral of Absolute Difference (IAD) from the intensity distribution profiles of the two brain hemispheres to predict any presence of the pathology. First a Mid Sagittal Plane (MSP) is obtained on the Magnetic Resonance Images that virtually divides brain into two bilaterally symmetric hemispheres. The block wise texture asymmetry is estimated for these hemispheres using the above 6 parameters.

543 EEG-Based Deep Neural Network Model for Brain Age Prediction and Its Association with Patient Health Conditions

SLEEP ◽  
2021 ◽  
Vol 44 (Supplement_2) ◽  
pp. A214-A214
Author(s):  
Yoav Nygate ◽  
Sam Rusk ◽  
Chris Fernandez ◽  
Nick Glattard ◽  
Jessica Arguelles ◽  
...  
Keyword(s):  
Neural Network ◽  
Deep Neural Network ◽  
Population Sample ◽  
Health Conditions ◽  
Apnea Hypopnea Index ◽  
Healthy Population ◽  
Brain Health ◽  
Patient Health ◽  
Drug Problems ◽  
Brain Age

Abstract Introduction Electroencephalogram (EEG) provides clinically relevant information for personalized patient health evaluation and comprehensive assessment of sleep. EEG-based indices have been associated with neurodegenerative conditions, psychiatric disorders, and metabolic and cardiovascular disease, and hold promise as a biomarker for brain health. Methods A deep neural network (DNN) model was trained to predict the age of patients using raw EEG signals recorded during clinical polysomnography (PSG). The DNN was trained on N=126,241 PSGs, validated on N=6,638, and tested on a holdout set of N=1,172. The holdout dataset included several categories of patient demographic and diagnostic parameters, allowing us to examine the association between brain age and a variety of medical conditions. Brain age was assessed by subtracting the individual’s chronological brain age from their EEG-predicted brain age (Brain Age Index; BAI), and then taking the absolute value of this variable (Absolute Brain Age Index; ABAI). We then constructed two regression models to test the relationship between BAI/ABAI and the following list of patient parameters: sex, BMI, depression, alcohol/drug problems, memory/concentration problems, epilepsy/seizures, diabetes, stroke, severe excessive daytime sleepiness (e.g., Epworth Sleepiness Scale ≥ 16; EDS), apnea-hypopnea index (AHI), arousal index (ArI), and sleep efficiency (SE). Results The DNN brain age model produced a mean absolute error of 4.604 and a Pearson’s r value of 0.933 which surpass the performance of prior research. In our regression analyses, we found a statistically significant relationship between the ABAI and: epilepsy and seizure disorders, stroke, elevated AHI, elevated ArI, and low SE (all p<0.05). This demonstrates these health conditions are associated with deviations of one’s predicted brain age from their chronological brain age. We also found patients with diabetes, depression, severe EDS, hypertension, and/or memory and concentration problems showed, on average, an elevated BAI compared to the healthy population sample (all p<0.05). Conclusion We show DNNs can accurately predict the brain age of healthy patients based on their raw, PSG derived, EEG recordings. Furthermore, we reveal indices, such as BAI and ABAI, display unique characteristics within different diseased populations, highlighting their potential value as novel diagnostic biomarker and potential “vital sign” of brain health. Support (if any):

scholarly journals Sensitivity analysis for interpretation of machine learning based segmentation models in cardiac MRI

2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Markus J. Ankenbrand ◽  
Liliia Shainberg ◽  
Michael Hock ◽  
David Lohr ◽  
Laura M. Schreiber
Keyword(s):  
Neural Networks ◽  
Sensitivity Analysis ◽  
Cardiac Magnetic Resonance ◽  
Magnetic Resonance ◽  
Cardiac Mri ◽  
Magnetic Resonance Images ◽  
Model Interpretation ◽  
First Case ◽  
Segmentation Models

Abstract Background Image segmentation is a common task in medical imaging e.g., for volumetry analysis in cardiac MRI. Artificial neural networks are used to automate this task with performance similar to manual operators. However, this performance is only achieved in the narrow tasks networks are trained on. Performance drops dramatically when data characteristics differ from the training set properties. Moreover, neural networks are commonly considered black boxes, because it is hard to understand how they make decisions and why they fail. Therefore, it is also hard to predict whether they will generalize and work well with new data. Here we present a generic method for segmentation model interpretation. Sensitivity analysis is an approach where model input is modified in a controlled manner and the effect of these modifications on the model output is evaluated. This method yields insights into the sensitivity of the model to these alterations and therefore to the importance of certain features on segmentation performance. Results We present an open-source Python library (misas), that facilitates the use of sensitivity analysis with arbitrary data and models. We show that this method is a suitable approach to answer practical questions regarding use and functionality of segmentation models. We demonstrate this in two case studies on cardiac magnetic resonance imaging. The first case study explores the suitability of a published network for use on a public dataset the network has not been trained on. The second case study demonstrates how sensitivity analysis can be used to evaluate the robustness of a newly trained model. Conclusions Sensitivity analysis is a useful tool for deep learning developers as well as users such as clinicians. It extends their toolbox, enabling and improving interpretability of segmentation models. Enhancing our understanding of neural networks through sensitivity analysis also assists in decision making. Although demonstrated only on cardiac magnetic resonance images this approach and software are much more broadly applicable.

scholarly journals The Brain Resting-State Functional Connectivity Underlying Violence Proneness: Is It a Reliable Marker for Neurocriminology? A Systematic Review

2019 ◽  
Vol 9 (1) ◽  
pp. 11 ◽  
Author(s):  
Ángel Romero-Martínez ◽  
Macarena González ◽  
Marisol Lila ◽  
Enrique Gracia ◽  
Luis Martí-Bonmatí ◽  
...  
Keyword(s):  
Functional Connectivity ◽  
Effective Treatment ◽  
Resting State ◽  
Brain Structures ◽  
Prevention Programs ◽  
Angular Gyrus ◽  
Resting State Functional Connectivity ◽  
Treatment And Prevention ◽  
Reliable Marker ◽  
The Brain

Introduction: There is growing scientific interest in understanding the biological mechanisms affecting and/or underlying violent behaviors in order to develop effective treatment and prevention programs. In recent years, neuroscientific research has tried to demonstrate whether the intrinsic activity within the brain at rest in the absence of any external stimulation (resting-state functional connectivity; RSFC) could be employed as a reliable marker for several cognitive abilities and personality traits that are important in behavior regulation, particularly, proneness to violence. Aims: This review aims to highlight the association between the RSFC among specific brain structures and the predisposition to experiencing anger and/or responding to stressful and distressing situations with anger in several populations. Methods: The scientific literature was reviewed following the PRISMA quality criteria for reviews, using the following digital databases: PubMed, PsycINFO, Psicodoc, and Dialnet. Results: The identification of 181 abstracts and retrieval of 34 full texts led to the inclusion of 17 papers. The results described in our study offer a better understanding of the brain networks that might explain the tendency to experience anger. The majority of the studies highlighted that diminished RSFC between the prefrontal cortex and the amygdala might make people prone to reactive violence, but that it is also necessary to contemplate additional cortical (i.e. insula, gyrus [angular, supramarginal, temporal, fusiform, superior, and middle frontal], anterior and posterior cingulated cortex) and subcortical brain structures (i.e. hippocampus, cerebellum, ventral striatum, and nucleus centralis superior) in order to explain a phenomenon as complex as violence. Moreover, we also described the neural pathways that might underlie proactive violence and feelings of revenge, highlighting the RSFC between the OFC, ventral striatal, angular gyrus, mid-occipital cortex, and cerebellum. Conclusions. The results from this synthesis and critical analysis of RSFC findings in several populations offer guidelines for future research and for developing a more accurate model of proneness to violence, in order to create effective treatment and prevention programs.

Sign in / Sign up

Export Citation Format

Share Document