Standardized evaluation of algorithms for computer-aided diagnosis of dementia based on structural MRI: The CADDementia challenge

Algorithms for computer-aided diagnosis of dementia based on structural MRI have demonstrated high performance in the literature, but are difficult to compare as different data sets and methodology were used for evaluation. In addition, it is unclear how the algorithms would perform on previously unseen data, and thus, how they would perform in clinical practice when there is no real opportunity to adapt the algorithm to the data at hand. To address these comparability, generalizability and clinical applicability issues, we organized a grand challenge that aimed to objectively compare algorithms based on a clinically representative multi-center data set. Using clinical practice as the starting point, the goal was to reproduce the clinical diagnosis. Therefore, we evaluated algorithms for multi-class classification of three diagnostic groups: patients with probable Alzheimer's disease, patients with mild cognitive impairment and healthy controls. The diagnosis based on clinical criteria was used as reference standard, as it was the best available reference despite its known limitations. For evaluation, a previously unseen test set was used consisting of 354 T1-weighted MRI scans with the diagnoses blinded. Fifteen research teams participated with a total of 29 algorithms. The algorithms were trained on a small training set (n = 30) and optionally on data from other sources (e.g., the Alzheimer's Disease Neuroimaging Initiative, the Australian Imaging Biomarkers and Lifestyle flagship study of aging). The best performing algorithm yielded an accuracy of 63.0% and an area under the receiver-operating-characteristic curve (AUC) of 78.8%. In general, the best performances were achieved using feature extraction based on voxel-based morphometry or a combination of features that included volume, cortical thickness, shape and intensity. The challenge is open for new submissions via the web-based framework: http://caddementia.grand-challenge.org.

Dimensionality reduction assisted tensor clustering

This paper is concerned with tensor clustering with the assistance of dimensionality reduction approaches. A class of formulation for tensor clustering is introduced based on tensor Tucker decomposition models. In this formulation, an extra tensor mode is formed by a collection of tensors of the same dimensions and then used to assist a Tucker decomposition in order to achieve data dimensionality reduction. We design two types of clustering models for the tensors: PCA Tensor Clustering model and Non-negative Tensor Clustering model, by utilizing different regularizations. The tensor clustering can thus be solved by the optimization method based on the alternative coordinate scheme. Interestingly, our experiments show that the proposed models yield comparable or even better performance compared to most recent clustering algorithms based on matrix factorization.

Joint multiple dictionary learning for tensor sparse coding

Traditional dictionary learning algorithms are used for finding a sparse representation on high dimensional data by transforming samples into a one-dimensional (1D) vector. This 1D model loses the inherent spatial structure property of data. An alternative solution is to employ Tensor Decomposition for dictionary learning on their original structural form —a tensor— by learning multiple dictionaries along each mode and the corresponding sparse representation in respect to the Kronecker product of these dictionaries. To learn tensor dictionaries along each mode, all the existing methods update each dictionary iteratively in an alternating manner. Because atoms from each mode dictionary jointly make contributions to the sparsity of tensor, existing works ignore atoms correlations between different mode dictionaries by treating each mode dictionary independently. In this paper, we propose a joint multiple dictionary learning method for tensor sparse coding, which explores atom correlations for sparse representation and updates multiple atoms from each mode dictionary simultaneously. In this algorithm, the Frequent-Pattern Tree (FP-tree) mining algorithm is employed to exploit frequent atom patterns in the sparse representation. Inspired by the idea of K-SVD, we develop a new dictionary update method that jointly updates elements in each pattern. Experimental results demonstrate our method outperforms other tensor based dictionary learning algorithms.

Rapid preconditioning of data for accelerating convex hull algorithms

Given a dataset of two-dimensional points in the plane with integer coordinates, the method proposed reduces a set of n points down to a set of s points s ≤ n, such that the convex hull on the set of s points is the same as the convex hull of the original set of n points. The method is O(n). It helps any convex hull algorithm run faster. The empirical analysis of a practical case shows a percentage reduction in points of over 98%, that is reflected as a faster computation with a speedup factor of at least 4.

Decomposition of surface electromyographic signal using Hidden Markov Model

The detection of physiological signals from the motor system (electromyographic signals) is being utilized in the practice clinic to guide the therapist in a more precise and accurate diagnosis of motor disorders. In this context, the process of decomposition of EMG (electromyographic) signals that includes the identification and classification of MUAP (Motor Unit Action Potential) of a EMG signal, is very important to help the therapist in the evaluation of motor disorders. The EMG decomposition is a complex task due to EMG features depend on the electrode type (needle or surface), its placement related to the muscle, the contraction level and the health of the Neuromuscular System. To date, the majority of researches on EMG decomposition utilize EMG signals acquired by needle electrodes, due to their advantages in processing this type of signal. However, relatively few researches have been conducted using surface EMG signals. Thus, this article aims to contribute to the clinical practice by presenting a technique that permit the decomposition of surface EMG signal via the use of Hidden Markov Models. This process is supported by the use of differential evolution and spectral clustering techniques. The developed system presented coherent results in: (1) identification of the number of Motor Units actives in the EMG signal; (2) presentation of the morphological patterns of MUAPs in the EMG signal; (3) identification of the firing sequence of the Motor Units. The model proposed in this work is an advance in the research area of decomposition of surface EMG signals.

Distributed parallelization of greedy Mobile Network Optimization algorithms

The Mobile Network Optimization (MNO) technologies have advanced at a tremendous pace in recent years. And the Dynamic Network Optimization (DNO) concept emerged years ago, aimed to continuously optimize the network in response to variations in network traffic and conditions. Yet, DNO development is still at its infancy, mainly hindered by a significant bottleneck of the lengthy optimization runtime. This paper identifies parallelism in greedy MNO algorithms and presents an advanced distributed parallel solution. The solution is designed, implemented and applied to real-life projects whose results yield a significant, highly scalable and nearly linear speedup up to 6.9 and 14.5 on distributed 8-core and 16-core systems respectively. Meanwhile, optimization outputs exhibit self-consistency and high precision compared to their sequential counterpart. This is a milestone in realizing the DNO. Further, the techniques may be applied to similar greedy optimization algorithm based applications.

Asynchronous distributed parallelization of mobile network optimization algorithms

It has been years since the introduction of the Dynamic Network Optimization (DNO) concept, yet the DNO development is still at its infant stage, largely due to a lack of breakthrough in minimizing the lengthy optimization runtime. Our previous work, a distributed parallel solution, has achieved a significant speed gain. To cater for the increased optimization complexity pressed by the uptake of smartphones and tablets, however, this paper examines the potential areas for further improvement and presents a novel asynchronous distributed parallel design that minimizes the inter-process communications. The new approach is implemented and applied to real-life projects whose results demonstrate an augmented acceleration of 7.5 times on a 16-core distributed system compared to 6.1 of our previous solution. Moreover, there is no degradation in the optimization outcome. This is a solid sprint towards the realization of DNO.

Extracting planetary waves from geomagnetic time series using Empirical Mode Decomposition

Empirical Mode Decomposition is presented as an alternative to traditional analysis methods to decompose geomagnetic time series into spectral components. Important comments on the algorithm and its variations will be given. Using this technique, planetary wave modes of 5-, 10-, and 16-day mean periods can be extracted from magnetic field components of three different stations in Germany. In a second step, the amplitude modulation functions of these wave modes can be shown to contain significant contribution from solar cycle variation through correlation with smoothed sunspot numbers. Additionally, the data indicate connections with geomagnetic jerk occurrences, supported by a second set of data providing reconstructed near-Earth magnetic field for 150 years. Usually attributed to internal dynamo processes within the Earth's outer core, the question of who is impacting whom will be briefly discussed here.

Cadaver decomposition in terrestrial ecosystems

A dead mammal (i.e. cadaver) is a high quality resource (narrow carbon:nitrogen ratio, high water content) that releases an intense, localised pulse of carbon and nutrients into the soil upon decomposition. Despite the fact that as much as 5,000 kg of cadaver can be introduced to a square kilometre of terrestrial ecosystem each year, cadaver decomposition remains a neglected microsere. Here we review the processes associated with the introduction of cadaver-derived carbon and nutrients into soil from forensic and ecological settings to show that cadaver decomposition can have a greater, albeit localised, effect on belowground ecology than plant and faecal resources. Cadaveric materials are rapidly introduced to belowground floral and faunal communities, which results in the formation of a highly concentrated island of fertility, or cadaver decomposition island (CDI). CDIs are associated with increased soil microbial biomass, microbial activity (C mineralisation) and nematode abundance. Each CDI is an ephemeral natural disturbance that, in addition to releasing energy and nutrients to the wider ecosystem, acts as a hub by receiving these materials in the form of dead insects, exuvia and puparia, faecal matter (from scavengers, grazers and predators) and feathers (from avian scavengers and predators). As such, CDIs contribute to landscape heterogeneity. Furthermore, CDIs are a specialised habitat for a number of flies, beetles and pioneer vegetation, which enhances biodiversity in terrestrial ecosystems.

Human versus animal: contrasting decomposition dynamics of mammalian analogues in experimental taphonomy

Taphonomic studies regularly employ animal analogues for human decomposition due to ethical restrictions relating to the use of human tissue. However, the validity of using animal analogues in soil decomposition studies is still questioned. This study compared the decomposition of skeletal muscle tissues (SMTs) from human (Homo sapiens), pork (Sus scrofa), beef (Bos taurus), and lamb (Ovis aries) interred in soil microcosms. Fixed interval samples were collected from the SMT for microbial activity and mass tissue loss determination; samples were also taken from the underlying soil for pH, electrical conductivity, and nutrient (potassium, phosphate, ammonium, and nitrate) analysis. The overall patterns of nutrient fluxes and chemical changes in nonhuman SMT and the underlying soil followed that of human SMT. Ovine tissue was the most similar to human tissue in many of the measured parameters. Although no single analogue was a precise predictor of human decomposition in soil, all models offered close approximations in decomposition dynamics.

Moisture can be the dominant environmental parameter governing cadaver decomposition in soil

Forensic taphonomy involves the use of decomposition to estimate postmortem interval (PMI) or locate clandestine graves. Yet, cadaver decomposition remains poorly understood, particularly following burial in soil. Presently, we do not know how most edaphic and environmental parameters, including soil moisture, influence the breakdown of cadavers following burial and alter the processes that are used to estimate PMI and locate clandestine graves. To address this, we buried juvenile rat (Rattus rattus) cadavers (∼18 g wet weight) in three contrasting soils from tropical savanna ecosystems located in Pallarenda (sand), Wambiana (medium clay), or Yabulu (loamy sand), Queensland, Australia. These soils were sieved (2 mm), weighed (500 g dry weight), calibrated to a matric potential of -0.01 megapascals (MPa), -0.05 MPa, or -0.3 MPa (wettest to driest) and incubated at 22 °C. Measurements of cadaver decomposition included cadaver mass loss, carbon dioxide-carbon (CO2-C) evolution, microbial biomass carbon (MBC), protease activity, phosphodiesterase activity, ninhydrin-reactive nitrogen (NRN) and soil pH. Cadaver burial resulted in a significant increase in CO2-C evolution, MBC, enzyme activities, NRN and soil pH. Cadaver decomposition in loamy sand and sandy soil was greater at lower matric potentials (wetter soil). However, optimal matric potential for cadaver decomposition in medium clay was exceeded, which resulted in a slower rate of cadaver decomposition in the wettest soil. Slower cadaver decomposition was also observed at high matric potential (-0.3 MPa). Furthermore, wet sandy soil was associated with greater cadaver decomposition than wet fine-textured soil. We conclude that gravesoil moisture content can modify the relationship between temperature and cadaver decomposition and that soil microorganisms can play a significant role in cadaver breakdown. We also conclude that soil NRN is a more reliable indicator of gravesoil than soil pH.

Freezing skeletal muscle tissue does not affect its decomposition in soil: Evidence from temporal changes in tissue mass, microbial activity and soil chemistry based on excised samples

The study of decaying organisms and death assemblages is referred to as forensic taphonomy, or more simply the study of graves. This field is dominated by the fields of entomology, anthropology and archaeology. Forensic taphonomy also includes the study of the ecology and chemistry of the burial environment. Studies in forensic taphonomy often require the use of analogues for human cadavers or their component parts. These might include animal cadavers or skeletal muscle tissue. However, sufficient supplies of cadavers or analogues may require periodic freezing of test material prior to experimental inhumation in the soil. This study was carried out to ascertain the effect of freezing on skeletal muscle tissue prior to inhumation and decomposition in a soil environment under controlled laboratory conditions. Changes in soil chemistry were also measured. In order to test the impact of freezing, skeletal muscle tissue (Sus scrofa) was frozen (−20 °C) or refrigerated (4 °C). Portions of skeletal muscle tissue (∼1.5 g) were interred in microcosms (72 mm diameter × 120 mm height) containing sieved (2 mm) soil (sand) adjusted to 50% water holding capacity. The experiment had three treatments: control with no skeletal muscle tissue, microcosms containing frozen skeletal muscle tissue and those containing refrigerated tissue. The microcosms were destructively harvested at sequential periods of 2, 4, 6, 8, 12, 16, 23, 30 and 37 days after interment of skeletal muscle tissue. These harvests were replicated 6 times for each treatment. Microbial activity (carbon dioxide respiration) was monitored throughout the experiment. At harvest the skeletal muscle tissue was removed and the detritosphere soil was sampled for chemical analysis. Freezing was found to have no significant impact on decomposition or soil chemistry compared to unfrozen samples in the current study using skeletal muscle tissue. However, the interment of skeletal muscle tissue had a significant impact on the microbial activity (carbon dioxide respiration) and chemistry of the surrounding soil including: pH, electroconductivity, ammonium, nitrate, phosphate and potassium. This is the first laboratory controlled study to measure changes in inorganic chemistry in soil associated with the decomposition of skeletal muscle tissue in combination with microbial activity.