Equipping computational pathology systems with artifact processing pipelines: a showcase for computation and performance trade-offs.

BACKGROUND: Histopathology is a gold standard for cancer diagnosis. It involves extracting tissue specimens from suspicious areas to prepare a glass slide for a microscopic examination. However, histological tissue processing procedures result in the introduction of artifacts, which are ultimately transferred to the digitized version of glass slides, known as whole slide images (WSIs). Artifacts are diagnostically irrelevant areas and may result in wrong predictions from deep learning (DL) algorithms. Therefore, detecting and excluding artifacts in the computational pathology (CPATH) system is essential for reliable automated diagnosis. METHODS: In this paper, we propose a mixture of experts (MoE) scheme for detecting five notable artifacts, including damaged tissue, blur, folded tissue, air bubbles, and histologically irrelevant blood from WSIs. First, we train independent binary DL models as experts to capture particular artifact morphology. Then, we ensemble their predictions using a fusion mechanism. We apply probabilistic thresholding over the final probability distribution to improve the sensitivity of the MoE. We developed four DL pipelines to evaluate computational and performance trade-offs. These include two MoEs and two multiclass models of state-of-the-art deep convolutional neural networks (DCNNs) and vision transformers (ViTs). These DL pipelines are quantitatively and qualitatively evaluated on external and out-of-distribution (OoD) data to assess generalizability and robustness for artifact detection application. RESULTS: We extensively evaluated the proposed MoE and multiclass models. DCNNs-based MoE and ViTs-based MoE schemes outperformed simpler multiclass models and were tested on datasets from different hospitals and cancer types, where MoE using (MobileNet) DCNNs yielded the best results. The proposed MoE yields 86.15 % F1 and 97.93% sensitivity scores on unseen data, retaining less computational cost for inference than MoE using ViTs. This best performance of MoEs comes with relatively higher computational trade-offs than multiclass models. Furthermore, we apply post-processing to create an artifact segmentation mask, a potential artifact-free RoI map, a quality report, and an artifact-refined WSI for further computational analysis. During the qualitative evaluation, field experts assessed the predictive performance of MoEs over OoD WSIs. They rated artifact detection and artifact-free area preservation, where the highest agreement translated to a Cohen Kappa of 0.82, indicating substantial agreement for the overall diagnostic usability of the DCNN-based MoE scheme. CONCLUSIONS: The proposed artifact detection pipeline will not only ensure reliable CPATH predictions but may also provide quality control. In this work, the best-performing pipeline for artifact detection is MoE with DCNNs. Our detailed experiments show that there is always a trade-off between performance and computational complexity, and no straightforward DL solution equally suits all types of data and applications. The code and HistoArtifacts dataset can be found online at Github and Zenodo , respectively.

A Gated Multiscale Multitask Learning Model Using Time-Frequency Representation for Health Assessment and Remaining Useful Life Prediction.

Health assessment and remaining useful life prediction are usually seen as separate tasks in industrial systems. Some multitask models use common features to handle these tasks synchronously, but they lack the usage of the representation in different scales and time-frequency domain. A lack of balance also exists among these scales. Therefore, a gated multiscale multitask learning model known as GMM-Net is proposed in this paper. By using the time-frequency representation, GMM-Net can obtain features of different scales via different kernels and compose the features by a gating network. A detailed loss function whose weight can be searched in a smaller scale is designed. The model is tested with different weights in the total loss function, and an optimal weight is found. Using this optimal weight, it is observed that the proposed method converges to a smaller loss and has a smaller model size than long short-term memory (LSTM) and gated recurrent unit (GRU) with less training time. The experiment results demonstrate the effectiveness of the proposed method.

Novel Meta-Learning Techniques for the Multiclass Image Classification Problem.

Multiclass image classification is a complex task that has been thoroughly investigated in the past. Decomposition-based strategies are commonly employed to address it. Typically, these methods divide the original problem into smaller, potentially simpler problems, allowing the application of numerous well-established learning algorithms that may not apply directly to the original task. This work focuses on the efficiency of decomposition-based methods and proposes several improvements to the meta-learning level. In this paper, four methods for optimizing the ensemble phase of multiclass classification are introduced. The first demonstrates that employing a mixture of experts scheme can drastically reduce the number of operations in the training phase by eliminating redundant learning processes in decomposition-based techniques for multiclass problems. The second technique for combining learner-based outcomes relies on Bayes' theorem. Combining the Bayes rule with arbitrary decompositions reduces training complexity relative to the number of classifiers even further. Two additional methods are also proposed for increasing the final classification accuracy by decomposing the initial task into smaller ones and ensembling the output of the base learners along with that of a multiclass classifier. Finally, the proposed novel meta-learning techniques are evaluated on four distinct datasets of varying classification difficulty. In every case, the proposed methods present a substantial accuracy improvement over existing traditional image classification techniques.

AC2: An Efficient Protein Sequence Compression Tool Using Artificial Neural Networks and Cache-Hash Models.

Recently, the scientific community has witnessed a substantial increase in the generation of protein sequence data, triggering emergent challenges of increasing importance, namely efficient storage and improved data analysis. For both applications, data compression is a straightforward solution. However, in the literature, the number of specific protein sequence compressors is relatively low. Moreover, these specialized compressors marginally improve the compression ratio over the best general-purpose compressors. In this paper, we present AC2, a new lossless data compressor for protein (or amino acid) sequences. AC2 uses a neural network to mix experts with a stacked generalization approach and individual cache-hash memory models to the highest-context orders. Compared to the previous compressor (AC), we show gains of 2-9% and 6-7% in reference-free and reference-based modes, respectively. These gains come at the cost of three times slower computations. AC2 also improves memory usage against AC, with requirements about seven times lower, without being affected by the sequences' input size. As an analysis application, we use AC2 to measure the similarity between each SARS-CoV-2 protein sequence with each viral protein sequence from the whole UniProt database. The results consistently show higher similarity to the pangolin coronavirus, followed by the bat and human coronaviruses, contributing with critical results to a current controversial subject. AC2 is available for free download under GPLv3 license.

Diagnosis of Autism Spectrum Disorders in Young Children Based on Resting-State Functional Magnetic Resonance Imaging Data Using Convolutional Neural Networks.

Statistics show that the risk of autism spectrum disorder (ASD) is increasing in the world. Early diagnosis is most important factor in treatment of ASD. Thus far, the childhood diagnosis of ASD has been done based on clinical interviews and behavioral observations. There is a significant need to reduce the use of traditional diagnostic techniques and to diagnose this disorder in the right time and before the manifestation of behavioral symptoms. The purpose of this study is to present the intelligent model to diagnose ASD in young children based on resting-state functional magnetic resonance imaging (rs-fMRI) data using convolutional neural networks (CNNs). CNNs, which are by far one of the most powerful deep learning algorithms, are mainly trained using datasets with large numbers of samples. However, obtaining comprehensive datasets such as ImageNet and achieving acceptable results in medical imaging domain have become challenges. In order to overcome these two challenges, the two methods of "combining classifiers," both dynamic (mixture of experts) and static (simple |Bayes) approaches, and "transfer learning" were used in this analysis. In addition, since diagnosis of ASD will be much more effective at an early age, samples ranging in age from 5 to 10 years from global Autism Brain Imaging Data Exchange I and II (ABIDE I and ABIDE II) datasets were used in this research. The accuracy, sensitivity, and specificity of presented model outperform the results of previous studies conducted on ABIDE I dataset (the best results obtained from Adamax optimization technique: accuracy = 0.7273, sensitivity = 0.712, specificity = 0.7348). Furthermore, acceptable classification results were obtained from ABIDE II dataset (the best results obtained from Adamax optimization technique: accuracy = 0.7, sensitivity = 0.582, specificity = 0.804) and the combination of ABIDE I and ABIDE II datasets (the best results obtained from Adam optimization technique: accuracy = 0.7045, sensitivity = 0.679, specificity = 0.7421). We can conclude that the proposed architecture can be considered as an efficient tool for diagnosis of ASD in young children. From another perspective, this proposed method can be applied to analyzing rs-fMRI data related to brain dysfunctions.

Mixture of Experts with Entropic Regularization for Data Classification.

Today, there is growing interest in the automatic classification of a variety of tasks, such as weather forecasting, product recommendations, intrusion detection, and people recognition. "Mixture-of-experts" is a well-known classification technique; it is a probabilistic model consisting of local expert classifiers weighted by a gate network that is typically based on softmax functions, combined with learnable complex patterns in data. In this scheme, one data point is influenced by only one expert; as a result, the training process can be misguided in real datasets for which complex data need to be explained by multiple experts. In this work, we propose a variant of the regular mixture-of-experts model. In the proposed model, the cost classification is penalized by the Shannon entropy of the gating network in order to avoid a "winner-takes-all" output for the gating network. Experiments show the advantage of our approach using several real datasets, with improvements in mean accuracy of 3-6% in some datasets. In future work, we plan to embed feature selection into this model.

Capturing heterogeneous group differences using mixture-of-experts: Application to a study of aging.

In MRI studies, linear multi-variate methods are often employed to identify regions or connections that are affected due to disease or normal aging. Such linear models inherently assume that there is a single, homogeneous abnormality pattern that is present in all affected individuals. While kernel-based methods can implicitly model a non-linear effect, and therefore the heterogeneity in the affected group, extracting and interpreting information about affected regions is difficult. In this paper, we present a method that explicitly models and captures heterogeneous patterns of change in the affected group relative to a reference group of controls. For this purpose, we use the Mixture-of-Experts (MOE) framework, which combines unsupervised modeling of mixtures of distributions with supervised learning of classifiers. MOE approximates the non-linear boundary between the two groups with a piece-wise linear boundary, thus allowing discovery of multiple patterns of group differences. In the case of patient/control comparisons, each such pattern aims to capture a different dimension of a disease, and hence to identify patient subgroups. We validated our model using multiple simulation scenarios and performance measures. We applied this method to resting state functional MRI data from the Baltimore Longitudinal Study of Aging, to investigate heterogeneous effects of aging on brain function in cognitively normal older adults (>85years) relative to a reference group of normal young to middle-aged adults (<60years). We found strong evidence for the presence of two subgroups of older adults, with similar age distributions in each subgroup, but different connectivity patterns associated with aging. While both older subgroups showed reduced functional connectivity in the Default Mode Network (DMN), increases in functional connectivity within the pre-frontal cortex as well as the bilateral insula were observed only for one of the two subgroups. Interestingly, the subgroup showing this increased connectivity (unlike the other subgroup) was, cognitively similar at baseline to the young and middle-aged subjects in two of seven cognitive domains, and had a faster rate of cognitive decline in one of seven domains. These results suggest that older individuals whose baseline cognitive performance is comparable to that of younger individuals recruit their "cognitive reserve" later in life, to compensate for reduced connectivity in other brain regions.

EMPT: a sparsity Transformer for EEG-based motor imagery recognition.

Introduction: Transformer network is widely emphasized and studied relying on its excellent performance. The self-attention mechanism finds a good solution for feature coding among multiple channels of electroencephalography (EEG) signals. However, using the self-attention mechanism to construct models on EEG data suffers from the problem of the large amount of data required and the complexity of the algorithm. Methods: We propose a Transformer neural network combined with the addition of Mixture of Experts (MoE) layer and ProbSparse Self-attention mechanism for decoding the time-frequency-spatial domain features from motor imagery (MI) EEG of spinal cord injury patients. The model is named as EEG MoE-Prob-Transformer (EMPT). The common spatial pattern and the modified s-transform method are employed for achieving the time-frequency-spatial features, which are used as feature embeddings to input the improved transformer neural network for feature reconstruction, and then rely on the expert model in the MoE layer for sparsity mapping, and finally output the results through the fully connected layer. Results: EMPT achieves an accuracy of 95.24% on the MI EEG dataset for patients with spinal cord injury. EMPT has also achieved excellent results in comparative experiments with other state-of-the-art methods. Discussion: The MoE layer and ProbSparse Self-attention inside the EMPT are subjected to visualisation experiments. The experiments prove that sparsity can be introduced to the Transformer neural network by introducing MoE and kullback-leibler divergence attention pooling mechanism, thereby enhancing its applicability on EEG datasets. A novel deep learning approach is presented for decoding EEG data based on MI.

Mixture-of-experts and semantic-guided network for brain tumor segmentation with missing MRI modalities.

Accurate brain tumor segmentation with multi-modal MRI images is crucial, but missing modalities in clinical practice often reduce accuracy. The aim of this study is to propose a mixture-of-experts and semantic-guided network to tackle the issue of missing modalities in brain tumor segmentation. We introduce a transformer-based encoder with novel mixture-of-experts blocks. In each block, four modality experts aim for modality-specific feature learning. Learnable modality embeddings are employed to alleviate the negative effect of missing modalities. We also introduce a decoder guided by semantic information, designed to pay higher attention to various tumor regions. Finally, we conduct extensive comparison experiments with other models as well as ablation experiments to validate the performance of the proposed model on the BraTS2018 dataset. The proposed model can accurately segment brain tumor sub-regions even with missing modalities. It achieves an average Dice score of 0.81 for the whole tumor, 0.66 for the tumor core, and 0.52 for the enhanced tumor across the 15 modality combinations, achieving top or near-top results in most cases, while also exhibiting a lower computational cost. Our mixture-of-experts and sematic-guided network achieves accurate and reliable brain tumor segmentation results with missing modalities, indicating its significant potential for clinical applications. Our source code is already available at https://github.com/MaggieLSY/MESG-Net .

Definition and analysis of gray matter atrophy subtypes in mild cognitive impairment based on data-driven methods.

Introduction: Mild cognitive impairment (MCI) is an important stage in Alzheimer's disease (AD) research, focusing on early pathogenic factors and mechanisms. Examining MCI patient subtypes and identifying their cognitive and neuropathological patterns as the disease progresses can enhance our understanding of the heterogeneous disease progression in the early stages of AD. However, few studies have thoroughly analyzed the subtypes of MCI, such as the cortical atrophy, and disease development characteristics of each subtype. Methods: In this study, 396 individuals with MCI, 228 cognitive normal (CN) participants, and 192 AD patients were selected from ADNI database, and a semi-supervised mixture expert algorithm (MOE) with multiple classification boundaries was constructed to define AD subtypes. Moreover, the subtypes of MCI were obtained by using the multivariate linear boundary mapping of support vector machine (SVM). Then, the gray matter atrophy regions and severity of each MCI subtype were analyzed and the features of each subtype in demography, pathology, cognition, and disease progression were explored combining the longitudinal data collected for 2 years and analyzed important factors that cause conversion of MCI were analyzed. Results: Three MCI subtypes were defined by MOE algorithm, and the three subtypes exhibited their own features in cortical atrophy. Nearly one-third of patients diagnosed with MCI have almost no significant difference in cerebral cortex from the normal aging population, and their conversion rate to AD are the lowest. The subtype characterized by severe atrophy in temporal lobe and frontal lobe have a faster decline rate in many cognitive manifestations than the subtype featured with diffuse atrophy in the whole cortex. APOE ε4 is an important factor that cause the conversion of MCI to AD. Conclusion: It was proved through the data-driven method that MCI collected by ADNI baseline presented different subtype features. The characteristics and disease development trajectories among subtypes can help to improve the prediction of clinical progress in the future and also provide necessary clues to solve the classification accuracy of MCI.

Multimodal Gated Mixture of Experts Using Whole Slide Image and Flow Cytometry for Multiple Instance Learning Classification of Lymphoma.

In this study, we present a deep-learning-based multimodal classification method for lymphoma diagnosis in digital pathology, which utilizes a whole slide image (WSI) as the primary image data and flow cytometry (FCM) data as auxiliary information. In pathological diagnosis of malignant lymphoma, FCM serves as valuable auxiliary information during the diagnosis process, offering useful insights into predicting the major class (superclass) of subtypes. By incorporating both images and FCM data into the classification process, we can develop a method that mimics the diagnostic process of pathologists, enhancing the explainability. In order to incorporate the hierarchical structure between superclasses and their subclasses, the proposed method utilizes a network structure that effectively combines the mixture of experts (MoE) and multiple instance learning (MIL) techniques, where MIL is widely recognized for its effectiveness in handling WSIs in digital pathology. The MoE network in the proposed method consists of a gating network for superclass classification and multiple expert networks for (sub)class classification, specialized for each superclass. To evaluate the effectiveness of our method, we conducted experiments involving a six-class classification task using 600 lymphoma cases. The proposed method achieved a classification accuracy of 72.3%, surpassing the 69.5% obtained through the straightforward combination of FCM and images, as well as the 70.2% achieved by the method using only images. Moreover, the combination of multiple weights in the MoE and MIL allows for the visualization of specific cellular and tumor regions, resulting in a highly explanatory model that cannot be attained with conventional methods. It is anticipated that by targeting a larger number of classes and increasing the number of expert networks, the proposed method could be effectively applied to the real problem of lymphoma diagnosis.

NEMoE: a nutrition aware regularized mixture of experts model to identify heterogeneous diet-microbiome-host health interactions.

BACKGROUND: Unrevealing the interplay between diet, the microbiome, and the health state could enable the design of personalized intervention strategies and improve the health and well-being of individuals. A common approach to this is to divide the study population into smaller cohorts based on dietary preferences in the hope of identifying specific microbial signatures. However, classification of patients based solely on diet is unlikely to reflect the microbiome-host health relationship or the taxonomic microbiome makeup. RESULTS: We present a novel approach, the Nutrition-Ecotype Mixture of Experts (NEMoE) model, for establishing associations between gut microbiota and health state that accounts for diet-specific cohort variability using a regularized mixture of experts model framework with an integrated parameter sharing strategy to ensure data-driven diet-cohort identification consistency across taxonomic levels. The success of our approach was demonstrated through a series of simulation studies, in which NEMoE showed robustness with regard to parameter selection and varying degrees of data heterogeneity. Further application to real-world microbiome data from a Parkinson's disease cohort revealed that NEMoE is capable of not only improving predictive performance for Parkinson's Disease but also for identifying diet-specific microbial signatures of disease. CONCLUSION: In summary, NEMoE can be used to uncover diet-specific relationships between nutritional-ecotype and patient health and to contextualize precision nutrition for different diseases. Video Abstract.

Clinical Phenotyping with an Outcomes-driven Mixture of Experts for Patient Matching and Risk Estimation.

Observational medical data present unique opportunities for analysis of medical outcomes and treatment decision making. However, because these datasets do not contain the strict pairing of randomized control trials, matching techniques are to draw comparisons among patients. A key limitation to such techniques is verification that the variables used to model treatment decision making are also relevant in identifying the risk of major adverse events. This article explores a deep mixture of experts approach to jointly learn how to match patients and model the risk of major adverse events in patients. Although trained with information regarding treatment and outcomes, after training, the proposed model is decomposable into a network that clusters patients into phenotypes from information available before treatment. This model is validated on a dataset of patients with acute myocardial infarction complicated by cardiogenic shock. The mixture of experts approach can predict the outcome of mortality with an area under the receiver operating characteristic curve of 0.85 ± 0.01 while jointly discovering five potential phenotypes of interest. The technique and interpretation allow for identifying clinically relevant phenotypes that may be used both for outcomes modeling as well as potentially evaluating individualized treatment effects.

Implicit Anatomical Rendering for Medical Image Segmentation with Stochastic Experts.

Integrating high-level semantically correlated contents and low-level anatomical features is of central importance in medical image segmentation. Towards this end, recent deep learning-based medical segmentation methods have shown great promise in better modeling such information. However, convolution operators for medical segmentation typically operate on regular grids, which inherently blur the high-frequency regions, i.e., boundary regions. In this work, we propose MORSE, a generic implicit neural rendering framework designed at an anatomical level to assist learning in medical image segmentation. Our method is motivated by the fact that implicit neural representation has been shown to be more effective in fitting complex signals and solving computer graphics problems than discrete grid-based representation. The core of our approach is to formulate medical image segmentation as a rendering problem in an end-to-end manner. Specifically, we continuously align the coarse segmentation prediction with the ambiguous coordinate-based point representations and aggregate these features to adaptively refine the boundary region. To parallelly optimize multi-scale pixel-level features, we leverage the idea from Mixture-of-Expert (MoE) to design and train our MORSE with a stochastic gating mechanism. Our experiments demonstrate that MORSE can work well with different medical segmentation backbones, consistently achieving competitive performance improvements in both 2D and 3D supervised medical segmentation methods. We also theoretically analyze the superiority of MORSE.

A data-fusion approach to identifying developmental dyslexia from multi-omics datasets.

This exploratory study tested and validated the use of data fusion and machine learning techniques to probe high-throughput omics and clinical data with a goal of exploring the etiology of developmental dyslexia. Developmental dyslexia is the leading learning disability in school aged children affecting roughly 5-10% of the US population. The complex biological and neurological phenotype of this life altering disability complicates its diagnosis. Phenome, exome, and metabolome data was collected allowing us to fully explore this system from a behavioral, cellular, and molecular point of view. This study provides a proof of concept showing that data fusion and ensemble learning techniques can outperform traditional machine learning techniques when provided small and complex multi-omics and clinical datasets. Heterogenous stacking classifiers consisting of single-omic experts/models achieved an accuracy of 86%, F1 score of 0.89, and AUC value of 0.83. Ensemble methods also provided a ranked list of important features that suggests exome single nucleotide polymorphisms found in the thalamus and cerebellum could be potential biomarkers for developmental dyslexia and heavily influenced the classification of DD within our machine learning models.

Identifying multiple soil pollutions of potentially contaminated sites based on multi-gate mixture-of-experts network.

With the rapid increase in the amount and sources of big data, using big data and machine learning methods to identify site soil pollution has become a research hotspot. However, previous studies that used basic information of sites as pollution identification indexes mainly have problems of low accuracy and efficiency when conducting complex model predictions for multiple soil pollution types. In this study, we collected the environmental data of 199 sites in 6 typical industries involving heavy metal and organic pollution. After feature fusion and selection, 10 indexes based on pollution sources and pathways were used to establish the soil pollution identification index system. The Multi-gate Mixture-of-Experts network (MMoE) were constructed to carry out the multi-tasks of soil heavy metals, VOCs and SVOCs pollution identification simultaneously. The SHAP framework was used to reveal the importance of pollution identification indexes on the multiple outputs of MMoE and obtain their driving factors. The results showed that the accuracies of MMoE model were 0.600, 0.783 and 0.850 for soil heavy metals, VOCs and SVOCs pollution identifications, respectively, which were 0-20 % higher than their accuracies of BP neural networks of single tasks. The indexes of raw material containing organic compounds, enterprise scale, soil pollution traces and industry types have the different significant importance on site soil pollutions. This study proposed a more efficient and accurate method to identify site soil pollutions and their driving factors, which offers a step towards realizing intelligent identification and risk control of site soil pollution globally.

MoËT: Mixture of Expert Trees and its application to verifiable reinforcement learning.

Rapid advancements in deep learning have led to many recent breakthroughs. While deep learning models achieve superior performance, often statistically better than humans, their adoption into safety-critical settings, such as healthcare or self-driving cars is hindered by their inability to provide safety guarantees or to expose the inner workings of the model in a human understandable form. We present MoËT, a novel model based on Mixture of Experts, consisting of decision tree experts and a generalized linear model gating function. Thanks to such gating function the model is more expressive than the standard decision tree. To support non-differentiable decision trees as experts, we formulate a novel training procedure. In addition, we introduce a hard thresholding version, MoËTh, in which predictions are made solely by a single expert chosen via the gating function. Thanks to that property, MoËTh allows each prediction to be easily decomposed into a set of logical rules in a form which can be easily verified. While MoËT is a general use model, we illustrate its power in the reinforcement learning setting. By training MoËT models using an imitation learning procedure on deep RL agents we outperform the previous state-of-the-art technique based on decision trees while preserving the verifiability of the models. Moreover, we show that MoËT can also be used in real-world supervised problems on which it outperforms other verifiable machine learning models.

An adaptive closed-loop ECoG decoder for long-term and stable bimanual control of an exoskeleton by a tetraplegic.

Objective.The article aims at addressing 2 challenges to step motor brain-computer interface (BCI) out of laboratories: asynchronous control of complex bimanual effectors with large numbers of degrees of freedom, using chronic and safe recorders, and the decoding performance stability over time without frequent decoder recalibration.Approach.Closed-loop adaptive/incremental decoder training is one strategy to create a model stable over time. Adaptive decoders update their parameters with new incoming data, optimizing the model parameters in real time. It allows cross-session training with multiple recording conditions during closed loop BCI experiments. In the article, an adaptive tensor-based recursive exponentially weighted Markov-switching multi-linear model (REW-MSLM) decoder is proposed. REW-MSLM uses a mixture of expert (ME) architecture, mixing or switching independent decoders (experts) according to the probability estimated by a 'gating' model. A Hidden Markov model approach is employed as gating model to improve the decoding robustness and to provide strong idle state support. The ME architecture fits the multi-limb paradigm associating an expert to a particular limb or action.Main results.Asynchronous control of an exoskeleton by a tetraplegic patient using a chronically implanted epidural electrocorticography (EpiCoG) recorder is reported. The stable over a period of six months (without decoder recalibration) eight-dimensional alternative bimanual control of the exoskeleton and its virtual avatar is demonstrated.Significance.Based on the long-term (>36 months) chronic bilateral EpiCoG recordings in a tetraplegic (ClinicalTrials.gov, NCT02550522), we addressed the poorly explored field of asynchronous bimanual BCI. The new decoder was designed to meet to several challenges: the high-dimensional control of a complex effector in experiments closer to real-world behavior (point-to-point pursuit versus conventional center-out tasks), with the ability of the BCI system to act as a stand-alone device switching between idle and control states, and a stable performance over a long period of time without decoder recalibration.

A Sparse Mixture-of-Experts Model With Screening of Genetic Associations to Guide Disease Subtyping.

Motivation: Identifying new genetic associations in non-Mendelian complex diseases is an increasingly difficult challenge. These diseases sometimes appear to have a significant component of heritability requiring explanation, and this missing heritability may be due to the existence of subtypes involving different genetic factors. Taking genetic information into account in clinical trials might potentially have a role in guiding the process of subtyping a complex disease. Most methods dealing with multiple sources of information rely on data transformation, and in disease subtyping, the two main strategies used are 1) the clustering of clinical data followed by posterior genetic analysis and 2) the concomitant clustering of clinical and genetic variables. Both of these strategies have limitations that we propose to address. Contribution: This work proposes an original method for disease subtyping on the basis of both longitudinal clinical variables and high-dimensional genetic markers via a sparse mixture-of-regressions model. The added value of our approach lies in its interpretability in relation to two aspects. First, our model links both clinical and genetic data with regard to their initial nature (i.e., without transformation) and does not require post-processing where the original information is accessed a second time to interpret the subtypes. Second, it can address large-scale problems because of a variable selection step that is used to discard genetic variables that may not be relevant for subtyping. Results: The proposed method was validated on simulations. A dataset from a cohort of Parkinson's disease patients was also analyzed. Several subtypes of the disease and genetic variants that potentially have a role in this typology were identified. Software availability: The R code for the proposed method, named DiSuGen, and a tutorial are available for download (see the references).

A machine learning perspective on the emotional content of Parkinsonian speech.

Patients with Parkinson's disease (PD) have distinctive voice patterns, often perceived as expressing sad emotion. While this characteristic of Parkinsonian speech has been supported through the perspective of listeners, where both PD and healthy control (HC) subjects repeat the same speaking tasks, it has never been explored through a machine learning modelling approach. Our work provides an objective evaluation of this characteristic of the PD speech, by building a transfer learning system to assess how the PD pathology affects the sadness perception. To do so we introduce a Mixture-of-Experts (MoE) architecture for speech emotion recognition designed to be transferable across datasets. Firstly, by relying on publicly available emotional speech corpora, we train the MoE model and then we use it to quantify perceived sadness in never seen before PD and matched HC speech recordings. To build our models (experts), we extracted spectral features of the voicing parts of speech and we trained a gradient boosting decision trees model in each corpus to predict happiness vs. sadness. MoE predictions are created by weighting each expert's prediction according to the distance between the new sample and the expert-specific training samples. The MoE approach systematically infers more negative emotional characteristics in PD speech than in HC. Crucially, these judgments are related to the disease severity and the severity of speech impairment in the PD patients: the more impairment, the more likely the speech is to be judged as sad. Our findings pave the way towards a better understanding of the characteristics of PD speech and show how publicly available datasets can be used to train models that provide interesting insights on clinical data.