Automated Diagnosis of Early Alzheimer’s disease using Fuzzy Neural Network
01 Jan 2009-pp 1455-1458
TL;DR: The aim of this research is to develop computing algorithms that can partially or fully automate the extraction of features from MRI of neuroanatomical structures in MTL regions, which aid in diagnosis of AD.
Abstract: This paper is presented towards the development of an automated diagnosis of Alzheimer’s disease (AD) from Magnetic Resonance Images (MRI) using Fuzzy Neural Network (FNN) algorithm. AD is a chronic degenerative disease of the central nervous system. The diagnosis of AD at an early stage is a major concern due to the growing number of the elderly population affected, as well as the lack of a standard and effective diagnosis procedure available to the healthcare providers. Medial Temporal Lobe (MTL) structure in brain has been reported to be involved earliest and most extensively in the pathology of AD. The aim of this research is to develop computing algorithms that can partially or fully automate the extraction of features from MRI of neuroanatomical structures in MTL regions, which aid in diagnosis of AD. Hippocampus volume reductions and ventricular expansions are observed and play significant role in MTL region of brain to identify AD, various other features are also considered and measured. The extracted feature values may be uncertain and it introduces fuzziness in input given to the Artificial Neural Network (ANN). Input uncertainty distribution is effectively solved by designing FNN. The back-propagation neural network algorithm was applied to the analysis of regional patterns corresponding to AD. A trained network was able to successfully classify MRI scans of normal subjects from Mild Cognitive Impairment (MCI), which could be a valuable early indicator of AD. This automated diagnosis will help the neurologist to find the level of disorders and measure the development stage of atrophy in the brain.
Citations
More filters
TL;DR: Different segmentation and machine learning techniques used for the diagnosis of AD are reviewed including thresholding, supervised and unsupervised learning, probabilistic techniques, Atlas-based approaches, and fusion of different image modalities.
Abstract: Alzheimer's disease (AD) is a common health problem in elderly people. There has been considerable research toward the diagnosis and early detection of this disease in the past decade. The sensitivity of biomarkers and the accuracy of the detection techniques have been defined to be the key to an accurate diagnosis. This paper presents a state-of-the-art review of the research performed on the diagnosis of AD based on imaging and machine learning techniques. Different segmentation and machine learning techniques used for the diagnosis of AD are reviewed including thresholding, supervised and unsupervised learning, probabilistic techniques, Atlas-based approaches, and fusion of different image modalities. More recent and powerful classification techniques such as the enhanced probabilistic neural network of Ahmadlou and Adeli should be investigated with the goal of improving the diagnosis accuracy. A combination of different image modalities can help improve the diagnosis accuracy rate. Research is needed on the combination of modalities to discover multi-modal biomarkers.
91 citations
Cites background from "Automated Diagnosis of Early Alzhei..."
..., 2006), fuzzy neural network (Anand et al., 2009; Forero Mendoza et al., 2014), 3D convolutional neural network (Payan and Montana, 2015), and PNN (Sankari, 2011)....
[...]
References
More filters
TL;DR: The sensitivity of diagnosis for dementia of the Alzheimer type (DAT) without any other diagnosis was 87%, and the specificity was 78%, but the ischemic scale score did not discriminate well between patients with pure multi-infarct dementia and those with both DAT and multi- infarCT dementia.
Abstract: • Clinical and pathologic diagnoses are compared in 65 patients who had dementia and who had been studied longitudinally during life. The sensitivity of diagnosis for dementia of the Alzheimer type (DAT) without any other diagnosis was 87%, and the specificity was 78%. The ischemic scale score did not discriminate well between patients with pure multi-infarct dementia and those with both DAT and multi-infarct dementia. However, 35 of 38 cases of pure DAT had a score of 4 or less on the ischemic scale.
620 citations
TL;DR: H hippocampus and parahippocampal atrophy occurs at a similar rate regardless of diagnostic group, and those who develop dementia may have smaller hippocampi to begin with, but become symptomatic because of accelerated loss of temporal lobe volume.
Abstract: Objective To determine initial locus and rate of degeneration of temporal lobe structures (total lobe, hippocampus and parahippocampus) in preclinical dementia. Background Postmortem studies suggest that the earliest changes in Alzheimer9s disease are neurofibrillary tangle formation in hippocampus and adjacent cortex. MRI volume analysis of temporal lobe structures over time in subjects prior to developing dementia may allow the identification of when these processes begin, the rate they develop, and which areas are key to symptom development. Methods 30 nondemented (NOD), healthy, elderly individuals enrolled in a prospective study of healthy aging evaluated annually over a mean of 42 months. Twelve subjects with subsequent cognitive decline were assigned to the preclinical dementia group (PreD). All 120 annual MRI studies analyzed by volumetric techniques assessed group differences in temporal lobe volumes and rates of brain loss. Results NOD as well as PreD subjects had significant, time-dependent decreases in hippocampal and parahippocampal volume. Rates of volume loss between the groups did not significantly differ. PreD cases had significantly smaller hippocampi when asymptomatic. Parahippocampal volume did not differ between PreD and NOD cases. Significant time-dependent temporal lobe atrophy was present only in PreD. Conclusions Hippocampal and parahippocampal atrophy occurs at a similar rate regardless of diagnostic group. Those who develop dementia may have smaller hippocampi to begin with, but become symptomatic because of accelerated loss of temporal lobe volume. Temporal lobe volume loss may mark the beginning of the disease process within six years prior to dementia onset.
394 citations
TL;DR: These results verify that system-wide limbic degeneration occurs in patients with AD, and atrophy in selected limbic structures was used to distinguish patients withAD from normal elderly individuals with over 90% accuracy in this select clinical sample.
Abstract: Objective: To examine volumetric changes in limbic structures in patients with probable AD using planimetric measures on MRI. Method: Limbic structures (i.e., hippocampus, amygdala, anterior thalamus, hypothalamus, mamillary bodies, basal forebrain, septal area, fornix, and cingulate, orbitofrontal, and parahippocampal cortices) were traced on 3D T1-weighted MR images of 40 patients with mild to moderate AD and 40 age-, sex-, and education-matched normal control subjects. Limbic volumes were compared between groups and the predictive ability was assessed. Results: Overall, limbic structures showed significant atrophy in AD patients compared with normal control subjects. Differences ( p Conclusions: These results verify that system-wide limbic degeneration occurs in patients with AD. In addition, atrophy in selected limbic structures was used to distinguish patients with AD from normal elderly individuals with over 90% accuracy in this select clinical sample. The measures require further exploration in samples more representative of those seen by primary care physicians before their utility can be accurately assessed.
283 citations
TL;DR: The presence of rated atrophy in selected temporal structures makes the diagnosis of Alzheimer's disease more likely, but the absence does not rule out the possibility of early Alzheimer’s disease.
Abstract: • Objective. —To evaluate the use of simple ratings and linear measures of atrophy in the temporal lobe structures obtained with magnetic resonance imaging coronal scans in the diagnosis of early Alzheimer's disease. Design. —Prospective series. The National Institute for Neurological Disorders and Stroke—Alzheimer's Disease and Related Disorders Association criteria for probable Alzheimer's disease. Blinded assessment. Setting. —Dementia study in a university hospital. Subjects. —Patients with Alzheimer's disease (n=34), scoring 150 or more on the Extended Scale for Dementia, and age-matched healthy community volunteers (n=39) who had both magnetic resonance imaging coronal scans and a psychometric assessment using the Extended Scale for Dementia within 6 months were included. Measures. —Main measures: T1-weighted magnetic resonance imaging coronal scans, a 1.5-T system. The degree of atrophy rated (0 to 4) in both sides of the temporal neocortex, entorhinal cortex, hippocampal formation, temporal horns, third ventricle, lateral ventricles, and frontal and parietal cortex. Linear measures: the area of hippocampus and the maximal transverse width of temporal horns. Results. —Differentiation between patients with Alzheimer's disease and controls was limited by considerable variations in sensitivity and specificity. Receiver operating characteristics analysis revealed a clear order of discrimination, the entorhinal cortex and the temporal neocortex being the two best, followed by the temporal horns and hippocampal formation. For a given specificity of 90%, the corresponding sensitivity for the entorhinal cortex, temporal neocortex, temporal horns, and hippocampal formation was 95%, 63%, 56%, and 41 %, respectively. Linear measures differed significantly but showed considerable overlap. Conclusion. —The presence of rated atrophy in selected temporal structures makes the diagnosis of Alzheimer's disease more likely, but the absence does not rule out the possibility of early Alzheimer's disease.
151 citations
TL;DR: Multilayered perceptron (MLP) networks are shown to be a particular implementation of hierarchical sets of fuzzy threshold logic rules based on sigmoidal MFs, equivalent to crisp logical networks applied to input data with uncertainty.
Abstract: Probability that a crisp logical rule applied to imprecise input data is true may be computed using fuzzy membership function (MF). All reasonable assumptions about input uncertainty distributions lead to MFs of sigmoidal shape. Convolution of several inputs with uniform uncertainty leads to bell-shaped Gaussian-like uncertainty functions. Relations between input uncertainties and fuzzy rules are systematically explored and several new types of MFs discovered. Multilayered perceptron (MLP) networks are shown to be a particular implementation of hierarchical sets of fuzzy threshold logic rules based on sigmoidal MFs. They are equivalent to crisp logical networks applied to input data with uncertainty. Leaving fuzziness on the input side makes the networks or the rule systems easier to understand. Practical applications of these ideas are presented for analysis of questionnaire data and gene expression data.
58 citations