Evaluation of Functional Decline in Alzheimer’s Dementia Using 3D Deep Learning and Group ICA for rs-fMRI Measurements

Qureshi, Muhammad Naveed Iqbal; Ryu, Seungjun; Song, Joonyoung; Lee, Kun Ho; Lee, Boreom

doi:10.3389/fnagi.2019.00008

Cited by 35 publications

(33 citation statements)

References 56 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…A different approach for classification, which is becoming very popular, is to use deep learning algorithms, which differ from ML for their ability to learn features automatically at multiple levels, therefore allowing the system to learn complex functions mapping the input to the output directly from data (Goodfellow et al, 2016;Qureshi et al, 2019). It could be envisaged that ML and deep learning methods could lead beyond current clinical diagnosis by establishing, in an unsupervised fashion, groups of patients with similar MRI and clinical scores.…”

Section: Discussionmentioning

confidence: 99%

A Machine Learning Approach for the Differential Diagnosis of Alzheimer and Vascular Dementia Fed by MRI Selected Features

Castellazzi

Cuzzoni²,

Ramusino

et al. 2020

Front. Neuroinform.

View full text Add to dashboard Cite

Among dementia-like diseases, Alzheimer disease (AD) and vascular dementia (VD) are two of the most frequent. AD and VD may share multiple neurological symptoms that may lead to controversial diagnoses when using conventional clinical and MRI criteria. Therefore, other approaches are needed to overcome this issue. Machine learning (ML) combined with magnetic resonance imaging (MRI) has been shown to improve the diagnostic accuracy of several neurodegenerative diseases, including dementia. To this end, in this study, we investigated, first, whether different kinds of ML algorithms, combined with advanced MRI features, could be supportive in classifying VD from AD and, second, whether the developed approach might help in predicting the prevalent disease in subjects with an unclear profile of AD or VD. Three ML categories of algorithms were tested: artificial neural network (ANN), support vector machine (SVM), and adaptive neuro-fuzzy inference system (ANFIS). Multiple regional metrics from resting-state fMRI (rs-fMRI) and diffusion tensor imaging (DTI) of 60 subjects (33 AD, 27 VD) were used as input features to train the algorithms and find the best feature pattern to classify VD from AD. We then used the identified VD-AD discriminant feature pattern as input for the most performant ML algorithm to predict the disease prevalence in 15 dementia patients with a "mixed VD-AD dementia" (MXD) clinical profile using their baseline MRI data. ML predictions were compared with the diagnosis evidence from a 3-year clinical follow-up. ANFIS emerged as the most efficient algorithm in discriminating AD from VD, reaching a classification accuracy greater than 84% using a small feature pattern. Moreover, ANFIS showed improved classification accuracy when trained with a multimodal input feature data set (e.g., DTI + rs-fMRI metrics) rather than a unimodal feature data set. Castellazzi et al. Dementia Classification Using Machine Learning When applying the best discriminant pattern to the MXD group, ANFIS achieved a correct prediction rate of 77.33%. Overall, results showed that our approach has a high discriminant power to classify AD and VD profiles. Moreover, the same approach also showed potential in predicting earlier the prevalent underlying disease in dementia patients whose clinical profile is uncertain between AD and VD, therefore suggesting its usefulness in supporting physicians' diagnostic evaluations.

show abstract

Section: Discussionmentioning

confidence: 99%

A Machine Learning Approach for the Differential Diagnosis of Alzheimer and Vascular Dementia Fed by MRI Selected Features

Castellazzi

Cuzzoni²,

Ramusino

et al. 2020

Front. Neuroinform.

View full text Add to dashboard Cite

show abstract

“…In this study, we have exploited Gwangju Alzheimer Research Data (GARD) [34][35][36] and Alzheimer Neuroimaging Initiative Data. ADNI was exploited for comparison with stat-of-the-art methods while extensive analysis was done for GARD database.…”

Section: Data Setmentioning

confidence: 99%

“…The sMRI scans were acquired from the registered subjects at the NRCD during the time period of 2014 to March 2018. The subject selection, MRI acquisition and exclusion criteria are mentioned in [34][35][36].…”

Section: Gard Datasetmentioning

confidence: 99%

Ensemble of ROI-based convolutional neural network classifiers for staging the Alzheimer disease spectrum from magnetic resonance imaging

et al. 2020

Self Cite

View full text Add to dashboard Cite

Patches from three orthogonal views of selected cerebral regions can be utilized to learn convolutional neural network (CNN) models for staging the Alzheimer disease (AD) spectrum including preclinical AD, mild cognitive impairment due to AD, and dementia due to AD and normal controls. Hippocampi, amygdalae and insulae were selected from the volumetric analysis of structured magnetic resonance images (MRIs). Three-view patches (TVPs) from these regions were fed to the CNN for training. MRIs were classified with the SoftMax-normalized scores of individual model predictions on TVPs. The significance of each region of interest (ROI) for staging the AD spectrum was evaluated and reported. The results of the ensemble classifier are compared with state-of-the-art methods using the same evaluation metrics. Patch-based ROI ensembles provide comparable diagnostic performance for AD staging. In this work, TVP-based ROI analysis using a CNN provides informative landmarks in cerebral MRIs and may have significance in clinical studies and computer-aided diagnosis system design.

show abstract

“…Prior to this saturation at the onset of the triggering signal, the incoming signals can be processed quickly even though the rate of initiation may be significantly more than that which is moved into and out of the neuroreceptor field when saturated. The relation of Equation (28) is used to characterize this saturation process, if X sat (t) > X max and t > t min then X sat (t) = X max (28) where X sat (t) is the neural response in saturated condition, X max is the maximum achievable rate of neural excitation given a saturated state, and t min is the time at which the post-synaptic neural membrane first achieves a saturated condition.…”

Section: Nonlinear Modelingmentioning

confidence: 99%

“…The applications for functional Magnetic Response Imaging (fMRI) in relation to brain function are numerous, profound, and range widely including studies on dementia [28], drug dependency [16,5], brain-computer interface [31,13], ADHD clinical testing [14], psychiatric disorders [12], stress [33], neurological disorders [8], parkinson's disease [27], aging [24], and cancer [26].…”

Section: Introductionmentioning

confidence: 99%

Analytic Model of fMRI BOLD Signals for Separable Metrics of Neural and Metabolic Activity

Schaper¹

2019

Preprint

View full text Add to dashboard Cite

The applications of fMRI (functional magnetic resonance imaging) are broad covering diagnostic and clinical extents of brain function, which involves the analysis of BOLD (blood oxygen level-dependent) contrast signals. The BOLD signals are sourced from both neural and metabolic functions, and thus to enable a detailed examination of fMRI studies, methods are sought to separate the neural and metabolic functions, such that the neural component, which is often the metric of interest, can be independently examined, especially in relation to neural connectivity. In this work, a modeling approach is developed that separates the neural and metabolic functions from the overall BOLD signal. The newly developed model is initially developed within a linear framework and demonstrates excellent comparison in data fit at 97.4% to the three Gamma function, which has been widely used to characterize fMRI BOLD experimental data. The neural component of the model formulation is comprised of a proper transfer function of two poles and two zeros, and characterizes the salient features of the BOLD signal, including the initial dip, peak, undershoot, and stabilization period. The linear model is extended to characterize nonlinear fMRI BOLD signal responses through the integration of saturation functions to both the leading and trailing edges of the signal. The nonlinear model representation not only explains the muted response in amplitude and oscillations, but also explains nuanced oscillations during the hold and settling phases of fMRI BOLD responses as exemplified in comparison to published data of sensorimotor responses. Further, the newly developed decomposition is derived within a framework for modeling neurovascular coupling environments, and thus lends credibility to the modeling framework. In developing the decomposition of the neural and metabolic transfer functions, it is a conclusion that the BOLD signal correlates very well with the fast dynamics associated with neural response to external stimuli. * charlesschaper@comcast.net Graphical AbstractThe normalized impulse response of the BOLD signal and the corresponding neural activity according to the newly developed model. There is a correspondence of the critical points for the oscillatory response of the neural function and metabolic reactivity, including the initial dip, peak and subsequent undershoot. Thus, the BOLD signal is a correlated representation of the underlying neural response.

show abstract

Evaluation of Functional Decline in Alzheimer’s Dementia Using 3D Deep Learning and Group ICA for rs-fMRI Measurements

Cited by 35 publications

References 56 publications

A Machine Learning Approach for the Differential Diagnosis of Alzheimer and Vascular Dementia Fed by MRI Selected Features

A Machine Learning Approach for the Differential Diagnosis of Alzheimer and Vascular Dementia Fed by MRI Selected Features

Ensemble of ROI-based convolutional neural network classifiers for staging the Alzheimer disease spectrum from magnetic resonance imaging

Analytic Model of fMRI BOLD Signals for Separable Metrics of Neural and Metabolic Activity

Contact Info

Product

Resources

About