ConvNets for automatic detection of polyglutamine SCAs from brain MRIs: state of the art applications.

Polyglutamine spinocerebellar ataxias (polyQ SCAs) are a group of neurodegenerative diseases, clinically and genetically heterogeneous, characterized by loss of balance and motor coordination due to dysfunction of the cerebellum and its connections. The diagnosis of each type of polyQ SCA, alongside with genetic tests, includes medical images analysis, and its automation may help specialists to distinguish between each type. Convolutional neural networks (ConvNets or CNNs) have been recently used for medical image processing, with outstanding results. In this work, we present the main clinical and imaging features of polyglutamine SCAs, and the basics of CNNs. Finally, we review studies that have used this approach to automatically process brain medical images and may be applied to SCAs detection. We conclude by discussing the possible limitations and opportunities of using ConvNets for SCAs diagnose in the future.

Convolutional Neural Networks for Segmenting Cerebellar Fissures from Magnetic Resonance Imaging.

The human cerebellum plays an important role in coordination tasks. Diseases such as spinocerebellar ataxias tend to cause severe damage to the cerebellum, leading patients to a progressive loss of motor coordination. The detection of such damages can help specialists to approximate the state of the disease, as well as to perform statistical analysis, in order to propose treatment therapies for the patients. Manual segmentation of such patterns from magnetic resonance imaging is a very difficult and time-consuming task, and is not a viable solution if the number of images to process is relatively large. In recent years, deep learning techniques such as convolutional neural networks (CNNs or convnets) have experienced an increased development, and many researchers have used them to automatically segment medical images. In this research, we propose the use of convolutional neural networks for automatically segmenting the cerebellar fissures from brain magnetic resonance imaging. Three models are presented, based on the same CNN architecture, for obtaining three different binary masks: fissures, cerebellum with fissures, and cerebellum without fissures. The models perform well in terms of precision and efficiency. Evaluation results show that convnets can be trained for such purposes, and could be considered as additional tools in the diagnosis and characterization of neurodegenerative diseases.

Neurophysiological features in spinocerebellar ataxia type 2: Prospects for novel biomarkers.

Electrophysiological biomarkers are useful to assess the degeneration and progression of the nervous system in pre-ataxic and ataxic stages of the Spinocerebellar Ataxia Type 2 (SCA2). These biomarkers are essentially defined by their clinical significance, discriminating patients and/or preclinical subjects from healthy controls in cross-sectional studies, their significant changes over time in longitudinal studies, and their correlation with the cytosine-guanine-adenine (CAG) repeat expansion and/or clinical ataxia scores, time of evolution and time to ataxia onset. We classified electrophysiological biomarkers into three main types: (1) preclinical, (2) disease progression and (3) genetic damage. We review the data that identify sural nerve potential amplitude, maximum saccadic velocity, sleep efficiency, rapid eye movement (REM) sleep percentage, K-complex density, REM sleep without atonia percentage, corticomuscular coherence, central motor conduction time, visual P300 latency, and antisaccadic error correction latency as reliable preclinical, progression and/or genetic damage biomarkers of SCA2. These electrophysiological biomarkers will facilitate the conduction of clinical trials that test the efficacy of emerging treatments in SCA2.

Touchscreen games to detect cognitive impairment in senior adults. A user-interaction pilot study.

INTRODUCTION: Alzheimer's disease is a degenerative brain disease and the most common cause of dementia. Today, 47 million people live with dementia worldwide. This number is projected to increase to more than 131 million by 2050, as populations age. Therefore, the World Health Organization considers serious cognitive deterioration a public health priority. OBJECTIVES: Advanced cognitive evaluation mechanisms are needed to help make an early diagnosis. These new mechanisms should overcome the limitations of current neuropsychological tests, including delayed detection; being perceived as intrusive; being non-ecological; being dependent on confounding factors; or their administration being expensive, among others. A promising novel approach consists of the introduction of serious games based on virtual reality and machine learning able to assess cognitive traits relevant to the diagnosis of mild cognitive impairment and Alzheimer's disease. METHODS: As a result of a preliminary pilot experiment, promising evidence was obtained about the predictive power of this solution. However, for these new serious games to be effective, evidence has to be gathered on the player experience by senior adults, avoiding the limitations of traditional tests at the same time. This study addresses these aspects with the participation of 74 senior users and 15 test administrators. RESULTS: Main findings confirm the usability and playability of Panoramix, a game battery designed according to the principles discussed above, its technological acceptability and its accessibility. For example, in relation to acceptability, the game battery was scored 4.39 in a 5-point scale, while its average usability score was 4.45 regardless of socio-cultural level or previous experience with digital technologies. In addition, health professionals confirm both, usability and playability, levels with an average score of 6.5 in a 7-point scale. Participants' willingness of using this kind of systems for cognitive evaluation was also confirmed. CONCLUSION: Promising results obtained pave the way for additional work to confirm the diagnostic validity according to clinical standards of these new cognitive assessment tools.

Learning to Detect Cognitive Impairment through Digital Games and Machine Learning Techniques.

OBJECTIVE: Alzheimer's disease (AD) is one of the most prevalent diseases among the adult population. The early detection of Mild Cognitive Impairment (MCI), which may trigger AD, is essential to slow down the cognitive decline process. METHODS: This paper presents a suit of serious games that aims at detecting AD and MCI overcoming the limitations of traditional tests, as they are time-consuming, affected by confounding factors that distort the result and usually administered when symptoms are evident and it is too late for preventive measures. The battery, named Panoramix, assesses the main early cognitive markers (i.e., memory, executive functions, attention and gnosias). Regarding its validation, it has been tested with a cohort study of 16 seniors, including AD, MCI and healthy individuals. RESULTS: This first pilot study offered initial evidence about psychometric validity, and more specifically about construct, criterion and external validity. After an analysis using machine learning techniques, findings show a promising 100% rate of success in classification abilities using a subset of three games in the battery. Thus, results are encouraging as all healthy subjects were correctly discriminated from those already suffering AD or MCI. CONCLUSIONS: The solid potential of digital serious games and machine learning for the early detection of dementia processes is demonstrated. Such a promising performance encourages further research to eventually introduce this technique for the clinical diagnosis of cognitive impairment.

Episodix: a serious game to detect cognitive impairment in senior adults. A psychometric study.

INTRODUCTION: Assessment of episodic memory is traditionally used to evaluate potential cognitive impairments in senior adults. The present article discusses the capabilities of Episodix, a game to assess the aforementioned cognitive area, as a valid tool to discriminate among mild cognitive impairment (MCI), Alzheimer's disease (AD) and healthy individuals (HC); that is, it studies the game's psychometric validity study to assess cognitive impairment. MATERIALS AND METHODS: After a preliminary study, a new pilot study, statistically significant for the Galician population, was carried out from a cross-sectional sample of senior adults as target users. A total of 64 individuals (28 HC, 16 MCI, 20 AD) completed the experiment from an initial sample of 74. Participants were administered a collection of classical pen-and-paper tests and interacted with the games developed. A total of six machine learning classification techniques were applied and four relevant performance metrics were computed to assess the classification power of the tool according to participants' cognitive status. RESULTS: According to the classification performance metrics computed, the best classification result is obtained using the Extra Trees Classifier (F1 = 0.97 and Cohen's kappa coefficient = 0.97). Precision and recall values are also high, above 0.9 for all cognitive groups. Moreover, according to the standard interpretation of Cohen's kappa index, classification is almost perfect (i.e., 0.81-1.00) for the complete dataset for all algorithms. LIMITATIONS: Weaknesses (e.g., accessibility, sample size or speed of stimuli) detected during the preliminary study were addressed and solved. Nevertheless, additional research is needed to improve the resolution of the game for the identification of specific cognitive impairments, as well as to achieve a complete validation of the psychometric properties of the digital game. CONCLUSION: Promising results obtained about psychometric validity of Episodix, represent a relevant step ahead towards the introduction of serious games and machine learning in regular clinical practice for detecting MCI or AD. However, more research is needed to explore the introduction of item response theory in this game and to obtain the required normative data for clinical validity.

Leveraging Wikipedia knowledge to classify multilingual biomedical documents.

This article presents a classifier that leverages Wikipedia knowledge to represent documents as vectors of concepts weights, and analyses its suitability for classifying biomedical documents written in any language when it is trained only with English documents. We propose the cross-language concept matching technique, which relies on Wikipedia interlanguage links to convert concept vectors between languages. The performance of the classifier is compared to a classifier based on machine translation, and two classifiers based on MetaMap. To perform the experiments, we created two multilingual corpus. The first one, Multi-Lingual UVigoMED (ML-UVigoMED) is composed of 23,647 Wikipedia documents about biomedical topics written in English, German, French, Spanish, Italian, Galician, Romanian, and Icelandic. The second one, English-French-Spanish-German UVigoMED (EFSG-UVigoMED) is composed of 19,210 biomedical abstract extracted from MEDLINE written in English, French, Spanish, and German. The performance of the approach proposed is superior to any of the state-of-the art classifier in the benchmark. We conclude that leveraging Wikipedia knowledge is of great advantage in tasks of multilingual classification of biomedical documents.

A Bag of Concepts Approach for Biomedical Document Classification Using Wikipedia Knowledge*. Spanish-English Cross-language Case Study.

OBJECTIVES: The ability to efficiently review the existing literature is essential for the rapid progress of research. This paper describes a classifier of text documents, represented as vectors in spaces of Wikipedia concepts, and analyses its suitability for classification of Spanish biomedical documents when only English documents are available for training. We propose the cross-language concept matching (CLCM) technique, which relies on Wikipedia interlanguage links to convert concept vectors from the Spanish to the English space. METHODS: The performance of the classifier is compared to several baselines: a classifier based on machine translation, a classifier that represents documents after performing Explicit Semantic Analysis (ESA), and a classifier that uses a domain-specific semantic annotator (MetaMap). The corpus used for the experiments (Cross-Language UVigoMED) was purpose-built for this study, and it is composed of 12,832 English and 2,184 Spanish MEDLINE abstracts. RESULTS: The performance of our approach is superior to any other state-of-the art classifier in the benchmark, with performance increases up to: 124% over classical machine translation, 332% over MetaMap, and 60 times over the classifier based on ESA. The results have statistical significance, showing p-values < 0.0001. CONCLUSION: Using knowledge mined from Wikipedia to represent documents as vectors in a space of Wikipedia concepts and translating vectors between language-specific concept spaces, a cross-language classifier can be built, and it performs better than several state-of-the-art classifiers.

Design process and preliminary psychometric study of a video game to detect cognitive impairment in senior adults.

INTRODUCTION: Assessment of episodic memory has been traditionally used to evaluate potential cognitive impairments in senior adults. Typically, episodic memory evaluation is based on personal interviews and pen-and-paper tests. This article presents the design, development and a preliminary validation of a novel digital game to assess episodic memory intended to overcome the limitations of traditional methods, such as the cost of its administration, its intrusive character, the lack of early detection capabilities, the lack of ecological validity, the learning effect and the existence of confounding factors. MATERIALS AND METHODS: Our proposal is based on the gamification of the California Verbal Learning Test (CVLT) and it has been designed to comply with the psychometric characteristics of reliability and validity. Two qualitative focus groups and a first pilot experiment were carried out to validate the proposal. RESULTS: A more ecological, non-intrusive and better administrable tool to perform cognitive assessment was developed. Initial evidence from the focus groups and pilot experiment confirmed the developed game's usability and offered promising results insofar its psychometric validity is concerned. Moreover, the potential of this game for the cognitive classification of senior adults was confirmed, and administration time is dramatically reduced with respect to pen-and-paper tests. LIMITATIONS: Additional research is needed to improve the resolution of the game for the identification of specific cognitive impairments, as well as to achieve a complete validation of the psychometric properties of the digital game. CONCLUSION: Initial evidence show that serious games can be used as an instrument to assess the cognitive status of senior adults, and even to predict the onset of mild cognitive impairments or Alzheimer's disease.

Trends on the application of serious games to neuropsychological evaluation: A scoping review.

BACKGROUND: The dramatic technological advances witnessed in recent years have resulted in a great opportunity for changing the way neuropsychological evaluations may be performed in clinical practice. Particularly, serious games have been posed as the cornerstone of this still incipient paradigm-shift, as they have characteristics that make them especially advantageous in trying to overcome limitations associated with traditional pen-and-paper based neuropsychological tests: they can be easily administered and they can feature complex environments for the evaluation of neuropsychological constructs that are difficult to evaluate through traditional tests. The objective of this study was to conduct a scoping literature review in order to map rapidly the key concepts underpinning this research area during the last 25years on the use of serious games for neuropsychological evaluation. METHODS: MEDLINE, PsycINFO, Scopus and IEEE Xplore databases were systematically searched. The main eligibility criteria were to select studies published in a peer-reviewed journal; written in English; published in the last 25years; focused on the human population, and classified in the neuropsychological field. Moreover, to avoid risk of bias, studies were selected by consensus of experts, focusing primarily in psychometric properties. Therefore, selected studies were analyzed in accordance with a set of dimensions of analysis commonly used for evaluating neuropsychological tests. RESULTS: After applying the selected search strategy, 57 studies -including 54 serious games- met our selection criteria. The selected studies deal with visuospatial capabilities, memory, attention, executive functions, and complex neuropsychological constructs such as Mild Cognitive Impairment (MCI). Results show that the implementation of serious games for neuropsychological evaluation is tackled in several different ways in the selected studies, and that studies have so far been mainly exploratory, just aiming at testing the feasibility of the proposed approaches. DISCUSSION: It may be argued that the limited number of databases used might compromise this study. However, we think that the finally included sample is representative, in spite of how difficult is to achieve an optimum and maximum scope. Indeed, this review identifies other research issues related to the development of serious games beyond their reliability and validity. The main conclusion of this review is that there is a great interest in the research community in the use of serious games for neuropsychological evaluation. This scoping review is pertinent, in accordance with the increasing number of studies published in the last three years, they demonstrate its potential as a serious alternative to classic neuropsychological tests. Nevertheless, more research is needed in order to implement serious games that are reliable, valid, and ready to be used in the everyday clinical practice.

Biomedical literature classification using encyclopedic knowledge: a Wikipedia-based bag-of-concepts approach.

Automatic classification of text documents into a set of categories has a lot of applications. Among those applications, the automatic classification of biomedical literature stands out as an important application for automatic document classification strategies. Biomedical staff and researchers have to deal with a lot of literature in their daily activities, so it would be useful a system that allows for accessing to documents of interest in a simple and effective way; thus, it is necessary that these documents are sorted based on some criteria-that is to say, they have to be classified. Documents to classify are usually represented following the bag-of-words (BoW) paradigm. Features are words in the text-thus suffering from synonymy and polysemy-and their weights are just based on their frequency of occurrence. This paper presents an empirical study of the efficiency of a classifier that leverages encyclopedic background knowledge-concretely Wikipedia-in order to create bag-of-concepts (BoC) representations of documents, understanding concept as "unit of meaning", and thus tackling synonymy and polysemy. Besides, the weighting of concepts is based on their semantic relevance in the text. For the evaluation of the proposal, empirical experiments have been conducted with one of the commonly used corpora for evaluating classification and retrieval of biomedical information, OHSUMED, and also with a purpose-built corpus of MEDLINE biomedical abstracts, UVigoMED. Results obtained show that the Wikipedia-based bag-of-concepts representation outperforms the classical bag-of-words representation up to 157% in the single-label classification problem and up to 100% in the multi-label problem for OHSUMED corpus, and up to 122% in the single-label classification problem and up to 155% in the multi-label problem for UVigoMED corpus.