Data Mining in Clinical Practice for the Quantification of Motor Impairment in Parkinson's Disease

Advances in biomedical signal acquisition systems for motion analysis have led to lowcost and ubiquitous wearable sensors which can be used to record movement data in different settings. This implies the potential availability of large amounts of quantitative data. It is then crucial to identify and to extract the information of clinical relevance from the large amount of available data. This quantitative and objective information can be an important aid for clinical decision making. Data mining is the process of discovering such information in databases through data processing, selection of informative data, and identification of relevant patterns. The databases considered in this thesis store motion data from wearable sensors (specifically accelerometers) and clinical information (clinical data, scores, tests). The main goal of this thesis is to develop data mining tools which can provide quantitative information to the clinician in the field of movement disorders. This thesis will focus on motor impairment in Parkinson's disease (PD). Different databases related to Parkinson subjects in different stages of the disease were considered for this thesis. Each database is characterized by the data recorded during a specific motor task performed by different groups of subjects. The data mining techniques that were used in this thesis are feature selection (a technique which was used to find relevant information and to discard useless or redundant data), classification, clustering, and regression. The aims were to identify high risk subjects for PD, characterize the differences between early PD subjects and healthy ones, characterize PD subtypes and automatically assess the severity of symptoms in the home setting.

Analyzing the dependence structure of microarray data: a copula–based approach

The main aim of this Ph.D. dissertation is the study of clustering dependent data by means of copula functions with particular emphasis on microarray data. Copula functions are a popular multivariate modeling tool in each field where the multivariate dependence is of great interest and their use in clustering has not been still investigated. The first part of this work contains the review of the literature of clustering methods, copula functions and microarray experiments. The attention focuses on the K–means (Hartigan, 1975; Hartigan and Wong, 1979), the hierarchical (Everitt, 1974) and the model–based (Fraley and Raftery, 1998, 1999, 2000, 2007) clustering techniques because their performance is compared. Then, the probabilistic interpretation of the Sklar’s theorem (Sklar’s, 1959), the estimation methods for copulas like the Inference for Margins (Joe and Xu, 1996) and the Archimedean and Elliptical copula families are presented. In the end, applications of clustering methods and copulas to the genetic and microarray experiments are highlighted. The second part contains the original contribution proposed. A simulation study is performed in order to evaluate the performance of the K–means and the hierarchical bottom–up clustering methods in identifying clusters according to the dependence structure of the data generating process. Different simulations are performed by varying different conditions (e.g., the kind of margins (distinct, overlapping and nested) and the value of the dependence parameter ) and the results are evaluated by means of different measures of performance. In light of the simulation results and of the limits of the two investigated clustering methods, a new clustering algorithm based on copula functions (‘CoClust’ in brief) is proposed. The basic idea, the iterative procedure of the CoClust and the description of the written R functions with their output are given. The CoClust algorithm is tested on simulated data (by varying the number of clusters, the copula models, the dependence parameter value and the degree of overlap of margins) and is compared with the performance of model–based clustering by using different measures of performance, like the percentage of well–identified number of clusters and the not rejection percentage of H0 on . It is shown that the CoClust algorithm allows to overcome all observed limits of the other investigated clustering techniques and is able to identify clusters according to the dependence structure of the data independently of the degree of overlap of margins and the strength of the dependence. The CoClust uses a criterion based on the maximized log–likelihood function of the copula and can virtually account for any possible dependence relationship between observations. Many peculiar characteristics are shown for the CoClust, e.g. its capability of identifying the true number of clusters and the fact that it does not require a starting classification. Finally, the CoClust algorithm is applied to the real microarray data of Hedenfalk et al. (2001) both to the gene expressions observed in three different cancer samples and to the columns (tumor samples) of the whole data matrix.

New approaches to open problems in gene expression microarray data

In the past decade, the advent of efficient genome sequencing tools and high-throughput experimental biotechnology has lead to enormous progress in the life science. Among the most important innovations is the microarray tecnology. It allows to quantify the expression for thousands of genes simultaneously by measurin the hybridization from a tissue of interest to probes on a small glass or plastic slide. The characteristics of these data include a fair amount of random noise, a predictor dimension in the thousand, and a sample noise in the dozens. One of the most exciting areas to which microarray technology has been applied is the challenge of deciphering complex disease such as cancer. In these studies, samples are taken from two or more groups of individuals with heterogeneous phenotypes, pathologies, or clinical outcomes. these samples are hybridized to microarrays in an effort to find a small number of genes which are strongly correlated with the group of individuals. Eventhough today methods to analyse the data are welle developed and close to reach a standard organization (through the effort of preposed International project like Microarray Gene Expression Data -MGED- Society [1]) it is not unfrequant to stumble in a clinician's question that do not have a compelling statistical method that could permit to answer it.The contribution of this dissertation in deciphering disease regards the development of new approaches aiming at handle open problems posed by clinicians in handle specific experimental designs. In Chapter 1 starting from a biological necessary introduction, we revise the microarray tecnologies and all the important steps that involve an experiment from the production of the array, to the quality controls ending with preprocessing steps that will be used into the data analysis in the rest of the dissertation. While in Chapter 2 a critical review of standard analysis methods are provided stressing most of problems that In Chapter 3 is introduced a method to adress the issue of unbalanced design of miacroarray experiments. In microarray experiments, experimental design is a crucial starting-point for obtaining reasonable results. In a two-class problem, an equal or similar number of samples it should be collected between the two classes. However in some cases, e.g. rare pathologies, the approach to be taken is less evident. We propose to address this issue by applying a modified version of SAM [2]. MultiSAM consists in a reiterated application of a SAM analysis, comparing the less populated class (LPC) with 1,000 random samplings of the same size from the more populated class (MPC) A list of the differentially expressed genes is generated for each SAM application. After 1,000 reiterations, each single probe given a "score" ranging from 0 to 1,000 based on its recurrence in the 1,000 lists as differentially expressed. The performance of MultiSAM was compared to the performance of SAM and LIMMA [3] over two simulated data sets via beta and exponential distribution. The results of all three algorithms over low- noise data sets seems acceptable However, on a real unbalanced two-channel data set reagardin Chronic Lymphocitic Leukemia, LIMMA finds no significant probe, SAM finds 23 significantly changed probes but cannot separate the two classes, while MultiSAM finds 122 probes with score >300 and separates the data into two clusters by hierarchical clustering. We also report extra-assay validation in terms of differentially expressed genes Although standard algorithms perform well over low-noise simulated data sets, multi-SAM seems to be the only one able to reveal subtle differences in gene expression profiles on real unbalanced data. In Chapter 4 a method to adress similarities evaluation in a three-class prblem by means of Relevance Vector Machine [4] is described. In fact, looking at microarray data in a prognostic and diagnostic clinical framework, not only differences could have a crucial role. In some cases similarities can give useful and, sometimes even more, important information. The goal, given three classes, could be to establish, with a certain level of confidence, if the third one is similar to the first or the second one. In this work we show that Relevance Vector Machine (RVM) [2] could be a possible solutions to the limitation of standard supervised classification. In fact, RVM offers many advantages compared, for example, with his well-known precursor (Support Vector Machine - SVM [3]). Among these advantages, the estimate of posterior probability of class membership represents a key feature to address the similarity issue. This is a highly important, but often overlooked, option of any practical pattern recognition system. We focused on Tumor-Grade-three-class problem, so we have 67 samples of grade I (G1), 54 samples of grade 3 (G3) and 100 samples of grade 2 (G2). The goal is to find a model able to separate G1 from G3, then evaluate the third class G2 as test-set to obtain the probability for samples of G2 to be member of class G1 or class G3. The analysis showed that breast cancer samples of grade II have a molecular profile more similar to breast cancer samples of grade I. Looking at the literature this result have been guessed, but no measure of significance was gived before.

Pattern recognition analysis on heavy ion reaction data

One of the problems in the analysis of nucleus-nucleus collisions is to get information on the value of the impact parameter b. This work consists in the application of pattern recognition techniques aimed at associating values of b to groups of events. To this end, a support vec- tor machine (SVM) classifier is adopted to analyze multifragmentation reactions. This method allows to backtracing the values of b through a particular multidimensional analysis. The SVM classification con- sists of two main phase. In the first one, known as training phase, the classifier learns to discriminate the events that are generated by two different model:Classical Molecular Dynamics (CMD) and Heavy- Ion Phase-Space Exploration (HIPSE) for the reaction: 58Ni +48 Ca at 25 AMeV. To check the classification of events in the second one, known as test phase, what has been learned is tested on new events generated by the same models. These new results have been com- pared to the ones obtained through others techniques of backtracing the impact parameter. Our tests show that, following this approach, the central collisions and peripheral collisions, for the CMD events, are always better classified with respect to the classification by the others techniques of backtracing. We have finally performed the SVM classification on the experimental data measured by NUCL-EX col- laboration with CHIMERA apparatus for the previous reaction.

Kernel Methods for Tree Structured Data

Machine learning comprises a series of techniques for automatic extraction of meaningful information from large collections of noisy data. In many real world applications, data is naturally represented in structured form. Since traditional methods in machine learning deal with vectorial information, they require an a priori form of preprocessing. Among all the learning techniques for dealing with structured data, kernel methods are recognized to have a strong theoretical background and to be effective approaches. They do not require an explicit vectorial representation of the data in terms of features, but rely on a measure of similarity between any pair of objects of a domain, the kernel function. Designing fast and good kernel functions is a challenging problem. In the case of tree structured data two issues become relevant: kernel for trees should not be sparse and should be fast to compute. The sparsity problem arises when, given a dataset and a kernel function, most structures of the dataset are completely dissimilar to one another. In those cases the classifier has too few information for making correct predictions on unseen data. In fact, it tends to produce a discriminating function behaving as the nearest neighbour rule. Sparsity is likely to arise for some standard tree kernel functions, such as the subtree and subset tree kernel, when they are applied to datasets with node labels belonging to a large domain. A second drawback of using tree kernels is the time complexity required both in learning and classification phases. Such a complexity can sometimes prevents the kernel application in scenarios involving large amount of data. This thesis proposes three contributions for resolving the above issues of kernel for trees. A first contribution aims at creating kernel functions which adapt to the statistical properties of the dataset, thus reducing its sparsity with respect to traditional tree kernel functions. Specifically, we propose to encode the input trees by an algorithm able to project the data onto a lower dimensional space with the property that similar structures are mapped similarly. By building kernel functions on the lower dimensional representation, we are able to perform inexact matchings between different inputs in the original space. A second contribution is the proposal of a novel kernel function based on the convolution kernel framework. Convolution kernel measures the similarity of two objects in terms of the similarities of their subparts. Most convolution kernels are based on counting the number of shared substructures, partially discarding information about their position in the original structure. The kernel function we propose is, instead, especially focused on this aspect. A third contribution is devoted at reducing the computational burden related to the calculation of a kernel function between a tree and a forest of trees, which is a typical operation in the classification phase and, for some algorithms, also in the learning phase. We propose a general methodology applicable to convolution kernels. Moreover, we show an instantiation of our technique when kernels such as the subtree and subset tree kernels are employed. In those cases, Direct Acyclic Graphs can be used to compactly represent shared substructures in different trees, thus reducing the computational burden and storage requirements.

Development of a new galaxy classification technique and its application to the zCOSMOS survey

The purpose of this Thesis is to develop a robust and powerful method to classify galaxies from large surveys, in order to establish and confirm the connections between the principal observational parameters of the galaxies (spectral features, colours, morphological indices), and help unveil the evolution of these parameters from $z \sim 1$ to the local Universe. Within the framework of zCOSMOS-bright survey, and making use of its large database of objects ($\sim 10\,000$ galaxies in the redshift range $0 < z \lesssim 1.2$) and its great reliability in redshift and spectral properties determinations, first we adopt and extend the \emph{classification cube method}, as developed by Mignoli et al. (2009), to exploit the bimodal properties of galaxies (spectral, photometric and morphologic) separately, and then combining together these three subclassifications. We use this classification method as a test for a newly devised statistical classification, based on Principal Component Analysis and Unsupervised Fuzzy Partition clustering method (PCA+UFP), which is able to define the galaxy population exploiting their natural global bimodality, considering simultaneously up to 8 different properties. The PCA+UFP analysis is a very powerful and robust tool to probe the nature and the evolution of galaxies in a survey. It allows to define with less uncertainties the classification of galaxies, adding the flexibility to be adapted to different parameters: being a fuzzy classification it avoids the problems due to a hard classification, such as the classification cube presented in the first part of the article. The PCA+UFP method can be easily applied to different datasets: it does not rely on the nature of the data and for this reason it can be successfully employed with others observables (magnitudes, colours) or derived properties (masses, luminosities, SFRs, etc.). The agreement between the two classification cluster definitions is very high. ``Early'' and ``late'' type galaxies are well defined by the spectral, photometric and morphological properties, both considering them in a separate way and then combining the classifications (classification cube) and treating them as a whole (PCA+UFP cluster analysis). Differences arise in the definition of outliers: the classification cube is much more sensitive to single measurement errors or misclassifications in one property than the PCA+UFP cluster analysis, in which errors are ``averaged out'' during the process. This method allowed us to behold the \emph{downsizing} effect taking place in the PC spaces: the migration between the blue cloud towards the red clump happens at higher redshifts for galaxies of larger mass. The determination of $M_{\mathrm{cross}}$ the transition mass is in significant agreement with others values in literature.

Comparing Different Approaches for Clustering Categorical Data

There are different ways to do cluster analysis of categorical data in the literature and the choice among them is strongly related to the aim of the researcher, if we do not take into account time and economical constraints. Main approaches for clustering are usually distinguished into model-based and distance-based methods: the former assume that objects belonging to the same class are similar in the sense that their observed values come from the same probability distribution, whose parameters are unknown and need to be estimated; the latter evaluate distances among objects by a defined dissimilarity measure and, basing on it, allocate units to the closest group. In clustering, one may be interested in the classification of similar objects into groups, and one may be interested in finding observations that come from the same true homogeneous distribution. But do both of these aims lead to the same clustering? And how good are clustering methods designed to fulfil one of these aims in terms of the other? In order to answer, two approaches, namely a latent class model (mixture of multinomial distributions) and a partition around medoids one, are evaluated and compared by Adjusted Rand Index, Average Silhouette Width and Pearson-Gamma indexes in a fairly wide simulation study. Simulation outcomes are plotted in bi-dimensional graphs via Multidimensional Scaling; size of points is proportional to the number of points that overlap and different colours are used according to the cluster membership.

Lesion detection and classification in breast cancer: evaluation of approaches based on morphological features, tracer kinetic modelling and semi-quantitative parameters in MR functional imaging (DCE-MRI)

The diagnosis, grading and classification of tumours has benefited considerably from the development of DCE-MRI which is now essential to the adequate clinical management of many tumour types due to its capability in detecting active angiogenesis. Several strategies have been proposed for DCE-MRI evaluation. Visual inspection of contrast agent concentration curves vs time is a very simple yet operator dependent procedure, therefore more objective approaches have been developed in order to facilitate comparison between studies. In so called model free approaches, descriptive or heuristic information extracted from time series raw data have been used for tissue classification. The main issue concerning these schemes is that they have not a direct interpretation in terms of physiological properties of the tissues. On the other hand, model based investigations typically involve compartmental tracer kinetic modelling and pixel-by-pixel estimation of kinetic parameters via non-linear regression applied on region of interests opportunely selected by the physician. This approach has the advantage to provide parameters directly related to the pathophysiological properties of the tissue such as vessel permeability, local regional blood flow, extraction fraction, concentration gradient between plasma and extravascular-extracellular space. Anyway, nonlinear modelling is computational demanding and the accuracy of the estimates can be affected by the signal-to-noise ratio and by the initial solutions. The principal aim of this thesis is investigate the use of semi-quantitative and quantitative parameters for segmentation and classification of breast lesion. The objectives can be subdivided as follow: describe the principal techniques to evaluate time intensity curve in DCE-MRI with focus on kinetic model proposed in literature; to evaluate the influence in parametrization choice for a classic bi-compartmental kinetic models; to evaluate the performance of a method for simultaneous tracer kinetic modelling and pixel classification; to evaluate performance of machine learning techniques training for segmentation and classification of breast lesion.

Learning with Kernels on Graphs: DAG-based kernels, data streams and RNA function prediction.

In many application domains data can be naturally represented as graphs. When the application of analytical solutions for a given problem is unfeasible, machine learning techniques could be a viable way to solve the problem. Classical machine learning techniques are defined for data represented in a vectorial form. Recently some of them have been extended to deal directly with structured data. Among those techniques, kernel methods have shown promising results both from the computational complexity and the predictive performance point of view. Kernel methods allow to avoid an explicit mapping in a vectorial form relying on kernel functions, which informally are functions calculating a similarity measure between two entities. However, the definition of good kernels for graphs is a challenging problem because of the difficulty to find a good tradeoff between computational complexity and expressiveness. Another problem we face is learning on data streams, where a potentially unbounded sequence of data is generated by some sources. There are three main contributions in this thesis. The first contribution is the definition of a new family of kernels for graphs based on Directed Acyclic Graphs (DAGs). We analyzed two kernels from this family, achieving state-of-the-art results from both the computational and the classification point of view on real-world datasets. The second contribution consists in making the application of learning algorithms for streams of graphs feasible. Moreover,we defined a principled way for the memory management. The third contribution is the application of machine learning techniques for structured data to non-coding RNA function prediction. In this setting, the secondary structure is thought to carry relevant information. However, existing methods considering the secondary structure have prohibitively high computational complexity. We propose to apply kernel methods on this domain, obtaining state-of-the-art results.

Learning Methods and Algorithms for Semantic Text Classification across Multiple Domains

Information is nowadays a key resource: machine learning and data mining techniques have been developed to extract high-level information from great amounts of data. As most data comes in form of unstructured text in natural languages, research on text mining is currently very active and dealing with practical problems. Among these, text categorization deals with the automatic organization of large quantities of documents in priorly defined taxonomies of topic categories, possibly arranged in large hierarchies. In commonly proposed machine learning approaches, classifiers are automatically trained from pre-labeled documents: they can perform very accurate classification, but often require a consistent training set and notable computational effort. Methods for cross-domain text categorization have been proposed, allowing to leverage a set of labeled documents of one domain to classify those of another one. Most methods use advanced statistical techniques, usually involving tuning of parameters. A first contribution presented here is a method based on nearest centroid classification, where profiles of categories are generated from the known domain and then iteratively adapted to the unknown one. Despite being conceptually simple and having easily tuned parameters, this method achieves state-of-the-art accuracy in most benchmark datasets with fast running times. A second, deeper contribution involves the design of a domain-independent model to distinguish the degree and type of relatedness between arbitrary documents and topics, inferred from the different types of semantic relationships between respective representative words, identified by specific search algorithms. The application of this model is tested on both flat and hierarchical text categorization, where it potentially allows the efficient addition of new categories during classification. Results show that classification accuracy still requires improvements, but models generated from one domain are shown to be effectively able to be reused in a different one.

Classification algorithms for Intelligent Transport Systems

Intelligent Transport Systems (ITS) consists in the application of ICT to transport to offer new and improved services to the mobility of people and freights. While using ITS, travellers produce large quantities of data that can be collected and analysed to study their behaviour and to provide information to decision makers and planners. The thesis proposes innovative deployments of classification algorithms for Intelligent Transport System with the aim to support the decisions on traffic rerouting, bus transport demand and behaviour of two wheelers vehicles. The first part of this work provides an overview and a classification of a selection of clustering algorithms that can be implemented for the analysis of ITS data. The first contribution of this thesis is an innovative use of the agglomerative hierarchical clustering algorithm to classify similar travels in terms of their origin and destination, together with the proposal for a methodology to analyse drivers’ route choice behaviour using GPS coordinates and optimal alternatives. The clusters of repetitive travels made by a sample of drivers are then analysed to compare observed route choices to the modelled alternatives. The results of the analysis show that drivers select routes that are more reliable but that are more expensive in terms of travel time. Successively, different types of users of a service that provides information on the real time arrivals of bus at stop are classified using Support Vector Machines. The results shows that the results of the classification of different types of bus transport users can be used to update or complement the census on bus transport flows. Finally, the problem of the classification of accidents made by two wheelers vehicles is presented together with possible future application of clustering methodologies aimed at identifying and classifying the different types of accidents.

Tensor-Train decomposition for image classification problems

In these last years a great effort has been put in the development of new techniques for automatic object classification, also due to the consequences in many applications such as medical imaging or driverless cars. To this end, several mathematical models have been developed from logistic regression to neural networks. A crucial aspect of these so called classification algorithms is the use of algebraic tools to represent and approximate the input data. In this thesis, we examine two different models for image classification based on a particular tensor decomposition named Tensor-Train (TT) decomposition. The use of tensor approaches preserves the multidimensional structure of the data and the neighboring relations among pixels. Furthermore the Tensor-Train, differently from other tensor decompositions, does not suffer from the curse of dimensionality making it an extremely powerful strategy when dealing with high-dimensional data. It also allows data compression when combined with truncation strategies that reduce memory requirements without spoiling classification performance. The first model we propose is based on a direct decomposition of the database by means of the TT decomposition to find basis vectors used to classify a new object. The second model is a tensor dictionary learning model, based on the TT decomposition where the terms of the decomposition are estimated using a proximal alternating linearized minimization algorithm with a spectral stepsize.

Big data analytics and modeling for Human Mobility

The fast development of Information Communication Technologies (ICT) offers new opportunities to realize future smart cities. To understand, manage and forecast the city's behavior, it is necessary the analysis of different kinds of data from the most varied dataset acquisition systems. The aim of this research activity in the framework of Data Science and Complex Systems Physics is to provide stakeholders with new knowledge tools to improve the sustainability of mobility demand in future cities. Under this perspective, the governance of mobility demand generated by large tourist flows is becoming a vital issue for the quality of life in Italian cities' historical centers, which will worsen in the next future due to the continuous globalization process. Another critical theme is sustainable mobility, which aims to reduce private transportation means in the cities and improve multimodal mobility. We analyze the statistical properties of urban mobility of Venice, Rimini, and Bologna by using different datasets provided by companies and local authorities. We develop algorithms and tools for cartography extraction, trips reconstruction, multimodality classification, and mobility simulation. We show the existence of characteristic mobility paths and statistical properties depending on transport means and user's kinds. Finally, we use our results to model and simulate the overall behavior of the cars moving in the Emilia Romagna Region and the pedestrians moving in Venice with software able to replicate in silico the demand for mobility and its dynamic.

Data-stream driven Fuzzy-granular approaches for system maintenance

Intelligent systems are currently inherent to the society, supporting a synergistic human-machine collaboration. Beyond economical and climate factors, energy consumption is strongly affected by the performance of computing systems. The quality of software functioning may invalidate any improvement attempt. In addition, data-driven machine learning algorithms are the basis for human-centered applications, being their interpretability one of the most important features of computational systems. Software maintenance is a critical discipline to support automatic and life-long system operation. As most software registers its inner events by means of logs, log analysis is an approach to keep system operation. Logs are characterized as Big data assembled in large-flow streams, being unstructured, heterogeneous, imprecise, and uncertain. This thesis addresses fuzzy and neuro-granular methods to provide maintenance solutions applied to anomaly detection (AD) and log parsing (LP), dealing with data uncertainty, identifying ideal time periods for detailed software analyses. LP provides deeper semantics interpretation of the anomalous occurrences. The solutions evolve over time and are general-purpose, being highly applicable, scalable, and maintainable. Granular classification models, namely, Fuzzy set-Based evolving Model (FBeM), evolving Granular Neural Network (eGNN), and evolving Gaussian Fuzzy Classifier (eGFC), are compared considering the AD problem. The evolving Log Parsing (eLP) method is proposed to approach the automatic parsing applied to system logs. All the methods perform recursive mechanisms to create, update, merge, and delete information granules according with the data behavior. For the first time in the evolving intelligent systems literature, the proposed method, eLP, is able to process streams of words and sentences. Essentially, regarding to AD accuracy, FBeM achieved (85.64+-3.69)%; eGNN reached (96.17+-0.78)%; eGFC obtained (92.48+-1.21)%; and eLP reached (96.05+-1.04)%. Besides being competitive, eLP particularly generates a log grammar, and presents a higher level of model interpretability.

Development and characterization of deep learning techniques for neuroimaging data

Deep learning methods are extremely promising machine learning tools to analyze neuroimaging data. However, their potential use in clinical settings is limited because of the existing challenges of applying these methods to neuroimaging data. In this study, first a data leakage type caused by slice-level data split that is introduced during training and validation of a 2D CNN is surveyed and a quantitative assessment of the model’s performance overestimation is presented. Second, an interpretable, leakage-fee deep learning software written in a python language with a wide range of options has been developed to conduct both classification and regression analysis. The software was applied to the study of mild cognitive impairment (MCI) in patients with small vessel disease (SVD) using multi-parametric MRI data where the cognitive performance of 58 patients measured by five neuropsychological tests is predicted using a multi-input CNN model taking brain image and demographic data. Each of the cognitive test scores was predicted using different MRI-derived features. As MCI due to SVD has been hypothesized to be the effect of white matter damage, DTI-derived features MD and FA produced the best prediction outcome of the TMT-A score which is consistent with the existing literature. In a second study, an interpretable deep learning system aimed at 1) classifying Alzheimer disease and healthy subjects 2) examining the neural correlates of the disease that causes a cognitive decline in AD patients using CNN visualization tools and 3) highlighting the potential of interpretability techniques to capture a biased deep learning model is developed. Structural magnetic resonance imaging (MRI) data of 200 subjects was used by the proposed CNN model which was trained using a transfer learning-based approach producing a balanced accuracy of 71.6%. Brain regions in the frontal and parietal lobe showing the cerebral cortex atrophy were highlighted by the visualization tools.