Semantic Categorization of Naming Responses Based on Prearticulatory Electrical Brain Activity.

PURPOSE: Object naming requires visual decoding, conceptualization, semantic categorization, and phonological encoding, all within 400 to 600 ms of stimulus presentation and before a word is spoken. In this study, we sought to predict semantic categories of naming responses based on prearticulatory brain activity recorded with scalp EEG in healthy individuals. METHODS: We assessed 19 healthy individuals who completed a naming task while undergoing EEG. The naming task consisted of 120 drawings of animate/inanimate objects or abstract drawings. We applied a one-dimensional, two-layer, neural network to predict the semantic categories of naming responses based on prearticulatory brain activity. RESULTS: Classifications of animate, inanimate, and abstract responses had an average accuracy of 80%, sensitivity of 72%, and specificity of 87% across participants. Across participants, time points with the highest average weights were between 470 and 490 milliseconds after stimulus presentation, and electrodes with the highest weights were located over the left and right frontal brain areas. CONCLUSIONS: Scalp EEG can be successfully used in predicting naming responses through prearticulatory brain activity. Interparticipant variability in feature weights suggests that individualized models are necessary for highest accuracy. Our findings may inform future applications of EEG in reconstructing speech for individuals with and without speech impairments.

Seizure forecasting using machine learning models trained by seizure diaries.

Objectives.People with refractory epilepsy are overwhelmed by the uncertainty of their next seizures. Accurate prediction of future seizures could greatly improve the quality of life for these patients. New evidence suggests that seizure occurrences can have cyclical patterns for some patients. Even though these cyclicalities are not intuitive, they can be identified by machine learning (ML), to identify patients with predictable vs unpredictable seizure patterns.Approach.Self-reported seizure logs of 153 patients from the Human Epilepsy Project with more than three reported seizures (totaling 8337 seizures) were used to obtain inter-seizure interval time-series for training and evaluation of the forecasting models. Two classes of prediction methods were studied: (1) statistical approaches using Bayesian fusion of population-wise and individual-wise seizure patterns; and (2) ML-based algorithms including least squares, least absolute shrinkage and selection operator, support vector machine (SVM) regression, and long short-term memory regression. Leave-one-person-out cross-validation was used for training and evaluation, by training on seizure diaries of all except one subject and testing on the left-out subject.Main results.The leading forecasting models were the SVM regression and a statistical model that combined the median of population-wise seizure time-intervals with a test subject's prior seizure intervals. SVM was able to forecast 50%, 70%, 81%, 84%, and 87% of seizures of unseen subjects within 0, 1, 2, 3 to 4 d of mean absolute forecasting error, respectively. The subject-wise performances show that patients with more frequent seizures were generally better predicted.Significance.ML models can leverage non-random patterns within self-reported seizure diaries to forecast future seizures. While diary-based seizure forecasting alone is only one of many aspects of clinical care of patients with epilepsy, studying the level of predictability across seizures and patients paves the path towards a better understanding of predictable vs unpredictable seizures on individualized and population-wise bases.

Radiological identification of temporal lobe epilepsy using artificial intelligence: a feasibility study.

Temporal lobe epilepsy is associated with MRI findings reflecting underlying mesial temporal sclerosis. Identifying these MRI features is critical for the diagnosis and management of temporal lobe epilepsy. To date, this process relies on visual assessment by highly trained human experts (e.g. neuroradiologists, epileptologists). Artificial intelligence is increasingly recognized as a promising aid in the radiological evaluation of neurological diseases, yet its applications in temporal lobe epilepsy have been limited. Here, we applied a convolutional neural network to assess the classification accuracy of temporal lobe epilepsy based on structural MRI. We demonstrate that convoluted neural networks can achieve high accuracy in the identification of unilateral temporal lobe epilepsy cases even when the MRI had been originally interpreted as normal by experts. We show that accuracy can be potentiated by employing smoothed grey matter maps and a direct acyclic graphs approach. We further discuss the foundations for the development of computer-aided tools to assist with the diagnosis of epilepsy.

Artificial intelligence for classification of temporal lobe epilepsy with ROI-level MRI data: A worldwide ENIGMA-Epilepsy study.

Artificial intelligence has recently gained popularity across different medical fields to aid in the detection of diseases based on pathology samples or medical imaging findings. Brain magnetic resonance imaging (MRI) is a key assessment tool for patients with temporal lobe epilepsy (TLE). The role of machine learning and artificial intelligence to increase detection of brain abnormalities in TLE remains inconclusive. We used support vector machine (SV) and deep learning (DL) models based on region of interest (ROI-based) structural (n = 336) and diffusion (n = 863) brain MRI data from patients with TLE with ("lesional") and without ("non-lesional") radiographic features suggestive of underlying hippocampal sclerosis from the multinational (multi-center) ENIGMA-Epilepsy consortium. Our data showed that models to identify TLE performed better or similar (68-75%) compared to models to lateralize the side of TLE (56-73%, except structural-based) based on diffusion data with the opposite pattern seen for structural data (67-75% to diagnose vs. 83% to lateralize). In other aspects, structural and diffusion-based models showed similar classification accuracies. Our classification models for patients with hippocampal sclerosis were more accurate (68-76%) than models that stratified non-lesional patients (53-62%). Overall, SV and DL models performed similarly with several instances in which SV mildly outperformed DL. We discuss the relative performance of these models with ROI-level data and the implications for future applications of machine learning and artificial intelligence in epilepsy care.

Temporal Lobe Epilepsy Surgical Outcomes Can Be Inferred Based on Structural Connectome Hubs: A Machine Learning Study.

OBJECTIVE: Medial temporal lobe epilepsy (TLE) is the most common form of medication-resistant focal epilepsy in adults. Despite removal of medial temporal structures, more than one-third of patients continue to have disabling seizures postoperatively. Seizure refractoriness implies that extramedial regions are capable of influencing the brain network and generating seizures. We tested whether abnormalities of structural network integration could be associated with surgical outcomes. METHODS: Presurgical magnetic resonance images from 121 patients with drug-resistant TLE across 3 independent epilepsy centers were used to train feed-forward neural network models based on tissue volume or graph-theory measures from whole-brain diffusion tensor imaging structural connectomes. An independent dataset of 47 patients with TLE from 3 other epilepsy centers was used to assess the predictive values of each model and regional anatomical contributions toward surgical treatment results. RESULTS: The receiver operating characteristic area under the curve based on regional betweenness centrality was 0.88, significantly higher than a random model or models based on gray matter volumes, degree, strength, and clustering coefficient. Nodes most strongly contributing to the predictive models involved the bilateral parahippocampal gyri, as well as the superior temporal gyri. INTERPRETATION: Network integration in the medial and lateral temporal regions was related to surgical outcomes. Patients with abnormally integrated structural network nodes were less likely to achieve seizure freedom. These findings are in line with previous observations related to network abnormalities in TLE and expand on the notion of underlying aberrant plasticity. Our findings provide additional information on the mechanisms of surgical refractoriness. ANN NEUROL 2020;88:970-983.

The ENIGMA-Epilepsy working group: Mapping disease from large data sets.

Epilepsy is a common and serious neurological disorder, with many different constituent conditions characterized by their electro clinical, imaging, and genetic features. MRI has been fundamental in advancing our understanding of brain processes in the epilepsies. Smaller-scale studies have identified many interesting imaging phenomena, with implications both for understanding pathophysiology and improving clinical care. Through the infrastructure and concepts now well-established by the ENIGMA Consortium, ENIGMA-Epilepsy was established to strengthen epilepsy neuroscience by greatly increasing sample sizes, leveraging ideas and methods established in other ENIGMA projects, and generating a body of collaborating scientists and clinicians to drive forward robust research. Here we review published, current, and future projects, that include structural MRI, diffusion tensor imaging (DTI), and resting state functional MRI (rsfMRI), and that employ advanced methods including structural covariance, and event-based modeling analysis. We explore age of onset- and duration-related features, as well as phenomena-specific work focusing on particular epilepsy syndromes or phenotypes, multimodal analyses focused on understanding the biology of disease progression, and deep learning approaches. We encourage groups who may be interested in participating to make contact to further grow and develop ENIGMA-Epilepsy.

Neural structures supporting spontaneous and assisted (entrained) speech fluency.

Non-fluent speech is one of the most common impairments in post-stroke aphasia. The rehabilitation of non-fluent speech in aphasia is particularly challenging as patients are rarely able to produce and practice fluent speech production. Speech entrainment is a behavioural technique that enables patients with non-fluent aphasia to speak fluently. However, its mechanisms are not well understood and the level of improved fluency with speech entrainment varies among individuals with non-fluent aphasia. In this study, we evaluated the behavioural and neuroanatomical factors associated with better speech fluency with the aid of speech entrainment during the training phase of speech entrainment. We used a lesion-symptom mapping approach to define the relationship between chronic stroke location on MRI and the number of different words per second produced during speech entrainment versus picture description spontaneous speech. The behavioural variable of interest was the speech entrainment/picture description ratio, which, if ≥1, indicated an increase in speech output during speech entrainment compared to picture description. We used machine learning (shallow neural network) to assess the statistical significance and out-of-sample predictive accuracy of the neuroanatomical model, and its regional contributors. We observed that better assisted speech (higher speech entrainment/picture description ratio) was achieved by individuals who had preservation of the posterior middle temporal gyrus, inferior fronto-occipital fasciculus and uncinate fasciculus, while exhibiting lesions in areas typically associated with non-fluent aphasia, such as the superior longitudinal fasciculus, precentral, inferior frontal, supramarginal and insular cortices. Our findings suggest that individuals with dorsal stream damage but preservation of ventral stream structures are more likely to achieve more fluent speech with the aid of speech entrainment compared to spontaneous speech. This observation provides insight into the mechanisms of non-fluent speech in aphasia and has potential implications for future research using speech entrainment for rehabilitation of non-fluent aphasia.

White matter connectomes at birth accurately predict cognitive abilities at age 2.

Cognitive ability is an important predictor of mental health outcomes that is influenced by neurodevelopment. Evidence suggests that the foundational wiring of the human brain is in place by birth, and that the white matter (WM) connectome supports developing brain function. It is unknown, however, how the WM connectome at birth supports emergent cognition. In this study, a deep learning model was trained using cross-validation to classify full-term infants (n = 75) as scoring above or below the median at age 2 using WM connectomes generated from diffusion weighted magnetic resonance images at birth. Results from this model were used to predict individual cognitive scores. We additionally identified WM connections important for classification. The model was also evaluated in a separate set of preterm infants (n = 37) scanned at term-age equivalent. Findings revealed that WM connectomes at birth predicted 2-year cognitive score group with high accuracy in both full-term (89.5%) and preterm (83.8%) infants. Scores predicted by the model were strongly correlated with actual scores (r = 0.98 for full-term and r = 0.96 for preterm). Connections within the frontal lobe, and between the frontal lobe and other brain areas were found to be important for classification. This work suggests that WM connectomes at birth can accurately predict a child's 2-year cognitive group and individual score in full-term and preterm infants. The WM connectome at birth appears to be a useful neuroimaging biomarker of subsequent cognitive development that deserves further study.

Identifying disease-related subnetwork connectome biomarkers by sparse hypergraph learning.

The functional brain network has gained increased attention in the neuroscience community because of its ability to reveal the underlying architecture of human brain. In general, majority work of functional network connectivity is built based on the correlations between discrete-time-series signals that link only two different brain regions. However, these simple region-to-region connectivity models do not capture complex connectivity patterns between three or more brain regions that form a connectivity subnetwork, or subnetwork for short. To overcome this current limitation, a hypergraph learning-based method is proposed to identify subnetwork differences between two different cohorts. To achieve our goal, a hypergraph is constructed, where each vertex represents a subject and also a hyperedge encodes a subnetwork with similar functional connectivity patterns between different subjects. Unlike previous learning-based methods, our approach is designed to jointly optimize the weights for all hyperedges such that the learned representation is in consensus with the distribution of phenotype data, i.e. clinical labels. In order to suppress the spurious subnetwork biomarkers, we further enforce a sparsity constraint on the hyperedge weights, where a larger hyperedge weight indicates the subnetwork with the capability of identifying the disorder condition. We apply our hypergraph learning-based method to identify subnetwork biomarkers in Autism Spectrum Disorder (ASD) and Attention Deficit Hyperactivity Disorder (ADHD). A comprehensive quantitative and qualitative analysis is performed, and the results show that our approach can correctly classify ASD and ADHD subjects from normal controls with 87.65 and 65.08% accuracies, respectively.

Deep learning applied to whole-brain connectome to determine seizure control after epilepsy surgery.

OBJECTIVE: We evaluated whether deep learning applied to whole-brain presurgical structural connectomes could be used to predict postoperative seizure outcome more accurately than inference from clinical variables in patients with mesial temporal lobe epilepsy (TLE). METHODS: Fifty patients with unilateral TLE were classified either as having persistent disabling seizures (SZ) or becoming seizure-free (SZF) at least 1 year after epilepsy surgery. Their presurgical structural connectomes were reconstructed from whole-brain diffusion tensor imaging. A deep network was trained based on connectome data to classify seizure outcome using 5-fold cross-validation. RESULTS: Classification accuracy of our trained neural network showed positive predictive value (PPV; seizure freedom) of 88 ± 7% and mean negative predictive value (NPV; seizure refractoriness) of 79 ± 8%. Conversely, a classification model based on clinical variables alone yielded <50% accuracy. The specific features that contributed to high accuracy classification of the neural network were located not only in the ipsilateral temporal and extratemporal regions, but also in the contralateral hemisphere. SIGNIFICANCE: Deep learning demonstrated to be a powerful statistical approach capable of isolating abnormal individualized patterns from complex datasets to provide a highly accurate prediction of seizure outcomes after surgery. Features involved in this predictive model were both ipsilateral and contralateral to the clinical foci and spanned across limbic and extralimbic networks.

Multi-modal classification of neurodegenerative disease by progressive graph-based transductive learning.

Graph-based transductive learning (GTL) is a powerful machine learning technique that is used when sufficient training data is not available. In particular, conventional GTL approaches first construct a fixed inter-subject relation graph that is based on similarities in voxel intensity values in the feature domain, which can then be used to propagate the known phenotype data (i.e., clinical scores and labels) from the training data to the testing data in the label domain. However, this type of graph is exclusively learned in the feature domain, and primarily due to outliers in the observed features, may not be optimal for label propagation in the label domain. To address this limitation, a progressive GTL (pGTL) method is proposed that gradually finds an intrinsic data representation that more accurately aligns imaging features with the phenotype data. In general, optimal feature-to-phenotype alignment is achieved using an iterative approach that: (1) refines inter-subject relationships observed in the feature domain by using the learned intrinsic data representation in the label domain, (2) updates the intrinsic data representation from the refined inter-subject relationships, and (3) verifies the intrinsic data representation on the training data to guarantee an optimal classification when applied to testing data. Additionally, the iterative approach is extended to multi-modal imaging data to further improve pGTL classification accuracy. Using Alzheimer's disease and Parkinson's disease study data, the classification accuracy of the proposed pGTL method is compared to several state-of-the-art classification methods, and the results show pGTL can more accurately identify subjects, even at different progression stages, in these two study data sets.

Early brain development in infants at high risk for autism spectrum disorder.

Brain enlargement has been observed in children with autism spectrum disorder (ASD), but the timing of this phenomenon, and the relationship between ASD and the appearance of behavioural symptoms, are unknown. Retrospective head circumference and longitudinal brain volume studies of two-year olds followed up at four years of age have provided evidence that increased brain volume may emerge early in development. Studies of infants at high familial risk of autism can provide insight into the early development of autism and have shown that characteristic social deficits in ASD emerge during the latter part of the first and in the second year of life. These observations suggest that prospective brain-imaging studies of infants at high familial risk of ASD might identify early postnatal changes in brain volume that occur before an ASD diagnosis. In this prospective neuroimaging study of 106 infants at high familial risk of ASD and 42 low-risk infants, we show that hyperexpansion of the cortical surface area between 6 and 12 months of age precedes brain volume overgrowth observed between 12 and 24 months in 15 high-risk infants who were diagnosed with autism at 24 months. Brain volume overgrowth was linked to the emergence and severity of autistic social deficits. A deep-learning algorithm that primarily uses surface area information from magnetic resonance imaging of the brain of 6-12-month-old individuals predicted the diagnosis of autism in individual high-risk children at 24 months (with a positive predictive value of 81% and a sensitivity of 88%). These findings demonstrate that early brain changes occur during the period in which autistic behaviours are first emerging.

Detecting Anatomical Landmarks for Fast Alzheimer's Disease Diagnosis.

Structural magnetic resonance imaging (MRI) is a very popular and effective technique used to diagnose Alzheimer's disease (AD). The success of computer-aided diagnosis methods using structural MRI data is largely dependent on the two time-consuming steps: 1) nonlinear registration across subjects, and 2) brain tissue segmentation. To overcome this limitation, we propose a landmark-based feature extraction method that does not require nonlinear registration and tissue segmentation. In the training stage, in order to distinguish AD subjects from healthy controls (HCs), group comparisons, based on local morphological features, are first performed to identify brain regions that have significant group differences. In general, the centers of the identified regions become landmark locations (or AD landmarks for short) capable of differentiating AD subjects from HCs. In the testing stage, using the learned AD landmarks, the corresponding landmarks are detected in a testing image using an efficient technique based on a shape-constrained regression-forest algorithm. To improve detection accuracy, an additional set of salient and consistent landmarks are also identified to guide the AD landmark detection. Based on the identified AD landmarks, morphological features are extracted to train a support vector machine (SVM) classifier that is capable of predicting the AD condition. In the experiments, our method is evaluated on landmark detection and AD classification sequentially. Specifically, the landmark detection error (manually annotated versus automatically detected) of the proposed landmark detector is 2.41 mm , and our landmark-based AD classification accuracy is 83.7%. Lastly, the AD classification performance of our method is comparable to, or even better than, that achieved by existing region-based and voxel-based methods, while the proposed method is approximately 50 times faster.

Scalable High-Performance Image Registration Framework by Unsupervised Deep Feature Representations Learning.

Feature selection is a critical step in deformable image registration. In particular, selecting the most discriminative features that accurately and concisely describe complex morphological patterns in image patches improves correspondence detection, which in turn improves image registration accuracy. Furthermore, since more and more imaging modalities are being invented to better identify morphological changes in medical imaging data, the development of deformable image registration method that scales well to new image modalities or new image applications with little to no human intervention would have a significant impact on the medical image analysis community. To address these concerns, a learning-based image registration framework is proposed that uses deep learning to discover compact and highly discriminative features upon observed imaging data. Specifically, the proposed feature selection method uses a convolutional stacked autoencoder to identify intrinsic deep feature representations in image patches. Since deep learning is an unsupervised learning method, no ground truth label knowledge is required. This makes the proposed feature selection method more flexible to new imaging modalities since feature representations can be directly learned from the observed imaging data in a very short amount of time. Using the LONI and ADNI imaging datasets, image registration performance was compared to two existing state-of-the-art deformable image registration methods that use handcrafted features. To demonstrate the scalability of the proposed image registration framework, image registration experiments were conducted on 7.0-T brain MR images. In all experiments, the results showed that the new image registration framework consistently demonstrated more accurate registration results when compared to state of the art.

Identifying Relationships in Functional and Structural Connectome Data Using a Hypergraph Learning Method.

The brain connectome provides an unprecedented degree of information about the organization of neuronal network architecture, both at a regional level, as well as regarding the entire brain network. Over the last several years the neuroimaging community has made tremendous advancements in the analysis of structural connectomes derived from white matter fiber tractography or functional connectomes derived from time-series blood oxygen level signals. However, computational techniques that combine structural and functional connectome data to discover complex relationships between fiber density and signal synchronization, including the relationship with health and disease, has not been consistently performed. To overcome this shortcoming, a novel connectome feature selection technique is proposed that uses hypergraphs to identify connectivity relationships when structural and functional connectome data is combined. Using publicly available connectome data from the UMCD database, experiments are provided that show SVM classifiers trained with structural and functional connectome features selected by our method are able to correctly identify autism subjects with 88 % accuracy. These results suggest our combined connectome feature selection approach may improve outcome forecasting in the context of autism.

Identifying High Order Brain Connectome Biomarkers via Learning on Hypergraph.

The functional connectome has gained increased attention in the neuroscience community. In general, most network connectivity models are based on correlations between discrete-time series signals that only connect two different brain regions. However, these bivariate region-to-region models do not involve three or more brain regions that form a subnetwork. Here we propose a learning-based method to explore subnetwork biomarkers that are significantly distinguishable between two clinical cohorts. Learning on hypergraph is employed in our work. Specifically, we construct a hypergraph by exhaustively inspecting all possible subnetworks for all subjects, where each hyperedge connects a group of subjects demonstrating highly correlated functional connectivity behavior throughout the underlying subnetwork. The objective function of hypergraph learning is to jointly optimize the weights for all hyperedges which make the separation of two groups by the learned data representation be in the best consensus with the observed clinical labels. We deploy our method to find high order childhood autism biomarkers from rs-fMRI images. Promising results have been obtained from comprehensive evaluation on the discriminative power and generality in diagnosis of Autism.

Hierarchical and symmetric infant image registration by robust longitudinal-example-guided correspondence detection.

PURPOSE: To investigate anatomical differences across individual subjects, or longitudinal changes in early brain development, it is important to perform accurate image registration. However, due to fast brain development and dynamic tissue appearance changes, it is very difficult to align infant brain images acquired from birth to 1-yr-old. METHODS: To solve this challenging problem, a novel image registration method is proposed to align two infant brain images, regardless of age at acquisition. The main idea is to utilize the growth trajectories, or spatial-temporal correspondences, learned from a set of longitudinal training images, for guiding the registration of two different time-point images with different image appearances. Specifically, in the training stage, an intrinsic growth trajectory is first estimated for each training subject using the longitudinal images. To register two new infant images with potentially a large age gap, the corresponding images patches between each new image and its respective training images with similar age are identified. Finally, the registration between the two new images can be assisted by the learned growth trajectories from one time point to another time point that have been established in the training stage. To further improve registration accuracy, the proposed method is combined with a hierarchical and symmetric registration framework that can iteratively add new key points in both images to steer the estimation of the deformation between the two infant brain images under registration. RESULTS: To evaluate image registration accuracy, the proposed method is used to align 24 infant subjects at five different time points (2-week-old, 3-month-old, 6-month-old, 9-month-old, and 12-month-old). Compared to the state-of-the-art methods, the proposed method demonstrated superior registration performance. CONCLUSIONS: The proposed method addresses the difficulties in the infant brain registration and produces better results compared to existing state-of-the-art registration methods.

Evaluation of machine learning algorithms for treatment outcome prediction in patients with epilepsy based on structural connectome data.

The objective of this study is to evaluate machine learning algorithms aimed at predicting surgical treatment outcomes in groups of patients with temporal lobe epilepsy (TLE) using only the structural brain connectome. Specifically, the brain connectome is reconstructed using white matter fiber tracts from presurgical diffusion tensor imaging. To achieve our objective, a two-stage connectome-based prediction framework is developed that gradually selects a small number of abnormal network connections that contribute to the surgical treatment outcome, and in each stage a linear kernel operation is used to further improve the accuracy of the learned classifier. Using a 10-fold cross validation strategy, the first stage in the connectome-based framework is able to separate patients with TLE from normal controls with 80% accuracy, and second stage in the connectome-based framework is able to correctly predict the surgical treatment outcome of patients with TLE with 70% accuracy. Compared to existing state-of-the-art methods that use VBM data, the proposed two-stage connectome-based prediction framework is a suitable alternative with comparable prediction performance. Our results additionally show that machine learning algorithms that exclusively use structural connectome data can predict treatment outcomes in epilepsy with similar accuracy compared with "expert-based" clinical decision. In summary, using the unprecedented information provided in the brain connectome, machine learning algorithms may uncover pathological changes in brain network organization and improve outcome forecasting in the context of epilepsy.

Domain Transfer Learning for MCI Conversion Prediction.

Machine learning methods have successfully been used to predict the conversion of mild cognitive impairment (MCI) to Alzheimer's disease (AD), by classifying MCI converters (MCI-C) from MCI nonconverters (MCI-NC). However, most existing methods construct classifiers using data from one particular target domain (e.g., MCI), and ignore data in other related domains (e.g., AD and normal control (NC)) that may provide valuable information to improve MCI conversion prediction performance. To address is limitation, we develop a novel domain transfer learning method for MCI conversion prediction, which can use data from both the target domain (i.e., MCI) and auxiliary domains (i.e., AD and NC). Specifically, the proposed method consists of three key components: 1) a domain transfer feature selection component that selects the most informative feature-subset from both target domain and auxiliary domains from different imaging modalities; 2) a domain transfer sample selection component that selects the most informative sample-subset from the same target and auxiliary domains from different data modalities; and 3) a domain transfer support vector machine classification component that fuses the selected features and samples to separate MCI-C and MCI-NC patients. We evaluate our method on 202 subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) that have MRI, FDG-PET, and CSF data. The experimental results show the proposed method can classify MCI-C patients from MCI-NC patients with an accuracy of 79.4%, with the aid of additional domain knowledge learned from AD and NC.