A Comparison of Shallow and Deep Learning Methods for Predicting Cognitive Performance of Stroke Patients From MRI Lesion Images

Diagnosis of cognitive and motor disorders levels in stroke patients through explainable machine learning based on MRI

BACKGROUND

Globally, stroke is the third most significant cause of disability. A stroke may produce motor, sensory, perceptual, or cognitive disorders that result in disability and affect the likelihood of recovery, affecting a person's ability to function. Evaluation post-stroke is critical for optimal stroke care.

PURPOSE

Traditional methods for classifying the clinical disorders of cognitive and motor in stroke patients use assessment and interrogative measures, which are time-consuming, complex, and labor-intensive. In response to the current situation, this study develops an algorithm to automatically classify motor and cognitive disorders in stroke patients by 3D brain MRI to assist physicians in diagnosis.

METHODS

First, radiomics and fusion features are extracted from the OAx T2 Propeller of 3D brain MRI. Then, we use 14 machine learning models and one model ensemble method to predict Fugl-Meyer and MMSE levels of stroke patients. Next, we evaluate the models using accuracy, recall, f1-score, and area under the curve (AUC). Finally, we employ SHAP to explain the output of the model.

RESULTS

The best predictive models come from Random Forest (RF) Classifier with fusion features in cognitive classification and Linear Discriminant Analysis (LDA) with radiomics features in motor classification. The highest accuracies are 92.0 and 82.5% for cognitive and motor disorders.

CONCLUSIONS

MRI brain maps can classify the cognitive and motor disorders of stroke patients. Radiomics features demonstrate its merits. The proposed algorithms with MRI images can efficiently assist physicians in diagnosing the cognitive and motor disorders of stroke patients in clinical practice. Additionally, this lessens labor costs, improves diagnostic effectiveness, and avoids the subjective difference that comes with manual assessment.

NSF DARE-transforming modeling in neurorehabilitation: a patient-in-the-loop framework

In 2023, the National Science Foundation (NSF) and the National Institute of Health (NIH) brought together engineers, scientists, and clinicians by sponsoring a conference on computational modelling in neurorehabiilitation. To facilitate multidisciplinary collaborations and improve patient care, in this perspective piece we identify where and how computational modelling can support neurorehabilitation. To address the where, we developed a patient-in-the-loop framework that uses multiple and/or continual measurements to update diagnostic and treatment model parameters, treatment type, and treatment prescription, with the goal of maximizing clinically-relevant functional outcomes. This patient-in-the-loop framework has several key features: (i) it includes diagnostic and treatment models, (ii) it is clinically-grounded with the International Classification of Functioning, Disability and Health (ICF) and patient involvement, (iii) it uses multiple or continual data measurements over time, and (iv) it is applicable to a range of neurological and neurodevelopmental conditions. To address the how, we identify state-of-the-art and highlight promising avenues of future research across the realms of sensorimotor adaptation, neuroplasticity, musculoskeletal, and sensory & pain computational modelling. We also discuss both the importance of and how to perform model validation, as well as challenges to overcome when implementing computational models within a clinical setting. The patient-in-the-loop approach offers a unifying framework to guide multidisciplinary collaboration between computational and clinical stakeholders in the field of neurorehabilitation.

Scaling behaviours of deep learning and linear algorithms for the prediction of stroke severity

Deep learning has allowed for remarkable progress in many medical scenarios. Deep learning prediction models often require 105-107 examples. It is currently unknown whether deep learning can also enhance predictions of symptoms post-stroke in real-world samples of stroke patients that are often several magnitudes smaller. Such stroke outcome predictions however could be particularly instrumental in guiding acute clinical and rehabilitation care decisions. We here compared the capacities of classically used linear and novel deep learning algorithms in their prediction of stroke severity. Our analyses relied on a total of 1430 patients assembled from the MRI-Genetics Interface Exploration collaboration and a Massachusetts General Hospital-based study. The outcome of interest was National Institutes of Health Stroke Scale-based stroke severity in the acute phase after ischaemic stroke onset, which we predict by means of MRI-derived lesion location. We automatically derived lesion segmentations from diffusion-weighted clinical MRI scans, performed spatial normalization and included a principal component analysis step, retaining 95% of the variance of the original data. We then repeatedly separated a train, validation and test set to investigate the effects of sample size; we subsampled the train set to 100, 300 and 900 and trained the algorithms to predict the stroke severity score for each sample size with regularized linear regression and an eight-layered neural network. We selected hyperparameters on the validation set. We evaluated model performance based on the explained variance (R2) in the test set. While linear regression performed significantly better for a sample size of 100 patients, deep learning started to significantly outperform linear regression when trained on 900 patients. Average prediction performance improved by ∼20% when increasing the sample size 9× [maximum for 100 patients: 0.279 ± 0.005 (R2, 95% confidence interval), 900 patients: 0.337 ± 0.006]. In summary, for sample sizes of 900 patients, deep learning showed a higher prediction performance than typically employed linear methods. These findings suggest the existence of non-linear relationships between lesion location and stroke severity that can be utilized for an improved prediction performance for larger sample sizes.

Combined metabolomics and tandem machine-learning models for wound age estimation: a novel analytical strategy

Wound age estimation is one of the most challenging and indispensable issues for forensic pathologists. Although many methods based on physical findings and biochemical tests can be used to estimate wound age, an objective and reliable method for inferring the time interval after injury remains difficult. In the present study, endogenous metabolites of contused skeletal muscle were investigated to estimate the time interval after injury. Animal model of skeletal muscle injury was established using Sprague-Dawley rat, and the contused muscles were sampled at 4, 8, 12, 16, 20, 24, 28, 32, 36, 40, 44, and 48 h postcontusion (n = 9). Then, the samples were analysed using ultraperformance liquid chromatography coupled with high-resolution mass spectrometry. A total of 43 differential metabolites in contused muscle were determined by metabolomics method. They were applied to construct a two-level tandem prediction model for wound age estimation based on multilayer perceptron algorithm. As a result, all muscle samples were eventually divided into the following subgroups: 4, 8, 12, 16-20, 24-32, 36-40, and 44-48 h. The tandem model exhibited a robust performance and achieved a prediction accuracy of 92.6%, which was much higher than that of the single model. In summary, the multilayer perceptron-multilayer perceptron tandem machine-learning model based on metabolomics data can be used as a novel strategy for wound age estimation in future forensic casework.

Key Points

The changes of metabolite profile were correlated with the time interval after injury in contused skeletal muscle.A panel of 43 endogenous metabolites screened by ultraperformance liquid chromatography coupled with high-resolution mass spectrometry could distinguish the wound ages.The multilayer perceptron (MLP) algorithm exhibited a robust performance in wound age estimation using metabolites.The combination of matabolomics and MLP-MLP tandem model could improve the accuracy of inferring the time interval after injury.

PiTLiD: Identification of Plant Disease From Leaf Images Based on Convolutional Neural Network

With the development of plant phenomics, the identification of plant diseases from leaf images has become an effective and economic approach in plant disease science. Among the methods of plant diseases identification, the convolutional neural network (CNN) is the most popular one for its superior performance. However, CNN's representation power is still a challenge in dealing with small datasets, which greatly affects its popularization. In this work, we propose a new method, namely PiTLiD, based on pretrained Inception-V3 convolutional neural network and transfer learning to identify plant leaf diseases from phenotype data of plant leaf with small sample size. To evaluate the robustness of the proposed method, the experiments on several datasets with small-scale samples were implemented. The results show that PiTLiD performs better than compared methods. This study provides a plant disease identification tool based on a deep learning algorithm for plant phenomics. All the source data and code are accessible at https://github.com/zhanglab-wbgcas/PiTLiD.

A shallow deep learning approach to classify skin cancer using down-scaling method to minimize time and space complexity

The complex feature characteristics and low contrast of cancer lesions, a high degree of inter-class resemblance between malignant and benign lesions, and the presence of various artifacts including hairs make automated melanoma recognition in dermoscopy images quite challenging. To date, various computer-aided solutions have been proposed to identify and classify skin cancer. In this paper, a deep learning model with a shallow architecture is proposed to classify the lesions into benign and malignant. To achieve effective training while limiting overfitting problems due to limited training data, image preprocessing and data augmentation processes are introduced. After this, the ‘box blur’ down-scaling method is employed, which adds efficiency to our study by reducing the overall training time and space complexity significantly. Our proposed shallow convolutional neural network (SCNN_12) model is trained and evaluated on the Kaggle skin cancer data ISIC archive which was augmented to 16485 images by implementing different augmentation techniques. The model was able to achieve an accuracy of 98.87% with optimizer Adam and a learning rate of 0.001. In this regard, parameter and hyper-parameters of the model are determined by performing ablation studies. To assert no occurrence of overfitting, experiments are carried out exploring k-fold cross-validation and different dataset split ratios. Furthermore, to affirm the robustness the model is evaluated on noisy data to examine the performance when the image quality gets corrupted.This research corroborates that effective training for medical image analysis, addressing training time and space complexity, is possible even with a lightweighted network using a limited amount of training data.

Towards a safe and efficient clinical implementation of machine learning in radiation oncology by exploring model interpretability, explainability and data-model dependency

The interest in machine learning (ML) has grown tremendously in recent years, partly due to the performance leap that occurred with new techniques of deep learning, convolutional neural networks for images, increased computational power, and wider availability of large datasets. Most fields of medicine follow that popular trend and, notably, radiation oncology is one of those that are at the forefront, with already a long tradition in using digital images and fully computerized workflows. ML models are driven by data, and in contrast with many statistical or physical models, they can be very large and complex, with countless generic parameters. This inevitably raises two questions, namely, the tight dependence between the models and the datasets that feed them, and the interpretability of the models, which scales with its complexity. Any problems in the data used to train the model will be later reflected in their performance. This, together with the low interpretability of ML models, makes their implementation into the clinical workflow particularly difficult. Building tools for risk assessment and quality assurance of ML models must involve then two main points: interpretability and data-model dependency. After a joint introduction of both radiation oncology and ML, this paper reviews the main risks and current solutions when applying the latter to workflows in the former. Risks associated with data and models, as well as their interaction, are detailed. Next, the core concepts of interpretability, explainability, and data-model dependency are formally defined and illustrated with examples. Afterwards, a broad discussion goes through key applications of ML in workflows of radiation oncology as well as vendors' perspectives for the clinical implementation of ML.

FAST-IT: Find A Simple Test - In TIA (transient ischaemic attack): a prospective cohort study to develop a multivariable prediction model for diagnosis of TIA through proteomic discovery and candidate lipid mass spectrometry, neuroimaging and machine learning-study protocol

INTRODUCTION

Transient ischaemic attack (TIA) may be a warning sign of stroke and difficult to differentiate from minor stroke and TIA-mimics. Urgent evaluation and diagnosis is important as treating TIA early can prevent subsequent strokes. Recent improvements in mass spectrometer technology allow quantification of hundreds of plasma proteins and lipids, yielding large datasets that would benefit from different approaches including machine learning. Using plasma protein, lipid and radiological biomarkers, our study will develop predictive algorithms to distinguish TIA from minor stroke (positive control) and TIA-mimics (negative control). Analysis including machine learning employs more sophisticated modelling, allowing non-linear interactions, adapting to datasets and enabling development of multiple specialised test-panels for identification and differentiation.

METHODS AND ANALYSIS

Patients attending the Emergency Department, Stroke Ward or TIA Clinic at the Royal Adelaide Hospital with TIA, minor stroke or TIA-like symptoms will be recruited consecutively by staff-alert for this prospective cohort study. Advanced neuroimaging will be performed for each participant, with images assessed independently by up to three expert neurologists. Venous blood samples will be collected within 48 hours of symptom onset. Plasma proteomic and lipid analysis will use advanced mass spectrometry (MS) techniques. Principal component analysis and hierarchical cluster analysis will be performed using MS software. Output files will be analysed for relative biomarker quantitative differences between the three groups. Differences will be assessed by linear regression, one-way analysis of variance, Kruskal-Wallis H-test, χ² test or Fisher's exact test. Machine learning methods will also be applied including deep learning using neural networks.

ETHICS AND DISSEMINATION

Patients will provide written informed consent to participate in this grant-funded study. The Central Adelaide Local Health Network Human Research Ethics Committee approved this study (HREC/18/CALHN/384; R20180618). Findings will be disseminated through peer-reviewed publication and conferences; data will be managed according to our Data Management Plan (DMP2020-00062).

Identification of glomerulosclerosis using IBM Watson and shallow neural networks

Background

Advanced stages of different renal diseases feature glomerular sclerosis at a histological level which is observed by light microscopy on tissue samples obtained by performing a kidney biopsy. Computer-aided diagnosis (CAD) systems leverage the potential of artificial intelligence (AI) in healthcare to support physicians in the diagnostic process.

Methods

We propose a novel CAD system that processes histological images and discriminates between sclerotic and non-sclerotic glomeruli. To this goal, we designed, tested, and compared two artificial neural network (ANN) classifiers. The former implements a shallow ANN classifying hand-crafted features extracted from Regions of Interest (ROIs) by means of image-processing procedures. The latter, instead, employs the IBM Watson Visual Recognition System, which uses a deep artificial neural network making decisions taking the images as input, without the need to design any procedure for describing images with features. The input dataset consisted of 428 sclerotic glomeruli and 2344 non-sclerotic glomeruli derived from images of kidney biopsies scanned by the Aperio ScanScope System.

Results

Both AI approaches allowed to very accurately distinguish (mean MCC 0.95 and mean Accuracy 0.99) between sclerotic and non-sclerotic glomeruli. Although the systems may seem interchangeable, the approach based on feature extraction and classification would allow clinicians to gain information on the most discriminating features. In fact, further procedures could explain the classifier’s decision by analysing which subset of features impacted the most on the final decision.

Conclusions

We developed a customizable support system that can facilitate the work of renal pathologists both in clinical and research settings.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s40620-021-01200-0.

Classifying Social Determinants of Health from Unstructured Electronic Health Records Using Deep Learning-based Natural Language Processing

OBJECTIVE

Social determinants of health (SDOH) are non-medical factors that can profoundly impact patient health outcomes. However, SDOH are rarely available in structured electronic health record (EHR) data such as diagnosis codes, and more commonly found in unstructured narrative clinical notes. Hence, identifying social context from unstructured EHR data has become increasingly important. Yet, previous work on using natural language processing to automate extraction of SDOH from text (a) usually focuses on an ad hoc selection of SDOH, and (b) does not use the latest advances in deep learning. Our objective was to advance automatic extraction of SDOH from clinical text by (a) systematically creating a set of SDOH based on standard biomedical and psychiatric ontologies, and (b) training state-of-the-art deep neural networks to extract mentions of these SDOH from clinical notes.

DESIGN

A retrospective cohort study.

SETTING AND PARTICIPANTS

Data were extracted from the Medical Information Mart for Intensive Care (MIMIC-III) database. The corpus comprised 3,504 social related sentences from 2,670 clinical notes.

METHODS

We developed a framework for automated classification of multiple SDOH categories. Our dataset comprised narrative clinical notes under the "Social Work" category in the MIMIC-III Clinical Database. Using standard terminologies, SNOMED-CT and DSM-IV, we systematically curated a set of 13 SDOH categories and created annotation guidelines for these. After manually annotating the 3,504 sentences, we developed and tested three deep neural network (DNN) architectures - convolutional neural network (CNN), long short-term memory (LSTM) network, and the Bidirectional Encoder Representations from Transformers (BERT) - for automated detection of eight SDOH categories. We also compared these DNNs to three baselines models: (1) cTAKES, as well as (2) L2-regularized logistic regression and (3) random forests on bags-of-words. Model evaluation metrics included micro- and macro- F1, and area under the receiver operating characteristic curve (AUC).

RESULTS

All three DNN models accurately classified all SDOH categories (minimum micro-F1 =.632, minimum macro-AUC=.854). Compared to the CNN and LSTM, BERT performed best in most key metrics (micro-F1 = 0.690, macro-AUC=0.907). The BERT model most effectively identified the "occupational" category (F1=.774, AUC=.965) and least effectively identified the "non-SDOH" category (F=.491, AUC=.788). BERT outperformed cTAKES in distinguishing social vs non-social sentences (BERT F1 = .87 vs. cTAKES F1=.06), and outperformed logistic regression (micro-F1=0.649, macro-AUC=0.696) and random forest (micro-F1=0.502, macro-AUC=0.523) trained on bag-of-words.

CONCLUSIONS

Our study framework with DNN models demonstrated improved performance for efficiently identifying a systematic range of SDOH categories from clinical notes in the EHR. Improved identification of patient SDOH may further improve healthcare outcomes.

Bridging structural MRI with cognitive function for individual level classification of early psychosis via deep learning

INTRODUCTION

Recent efforts have been made to apply machine learning and deep learning approaches to the automated classification of schizophrenia using structural magnetic resonance imaging (sMRI) at the individual level. However, these approaches are less accurate on early psychosis (EP) since there are mild structural brain changes at early stage. As cognitive impairments is one main feature in psychosis, in this study we apply a multi-task deep learning framework using sMRI with inclusion of cognitive assessment to facilitate the classification of patients with EP from healthy individuals.

METHOD

Unlike previous studies, we used sMRI as the direct input to perform EP classifications and cognitive estimations. The proposed deep learning model does not require time-consuming volumetric or surface based analysis and can provide additionally cognition predictions. Experiments were conducted on an in-house data set with 77 subjects and a public ABCD HCP-EP data set with 164 subjects.

RESULTS

We achieved 74.9 ± 4.3% five-fold cross-validated accuracy and an area under the curve of 71.1 ± 4.1% on EP classification with the inclusion of cognitive estimations.

DISCUSSION

We reveal the feasibility of automated cognitive estimation using sMRI by deep learning models, and also demonstrate the implicit adoption of cognitive measures as additional information to facilitate EP classifications from healthy controls.

Deep Learning in Neuroimaging: Overcoming Challenges With Emerging Approaches

Deep learning (DL) is of great interest in psychiatry due its potential yet largely untapped ability to utilize multidimensional datasets (such as fMRI data) to predict clinical outcomes. Typical DL methods, however, have strong assumptions, such as large datasets and underlying model opaqueness, that are suitable for natural image prediction problems but not medical imaging. Here we describe three relatively novel DL approaches that may help accelerate its incorporation into mainstream psychiatry research and ultimately bring it into the clinic as a prognostic tool. We first introduce two methods that can reduce the amount of training data required to develop accurate models. These may prove invaluable for fMRI-based DL given the time and monetary expense required to acquire neuroimaging data. These methods are (1) transfer learning - the ability of deep learners to incorporate knowledge learned from one data source (e.g., fMRI data from one site) and apply it toward learning from a second data source (e.g., data from another site), and (2) data augmentation (via Mixup) - a self-supervised learning technique in which "virtual" instances are created. We then discuss explainable artificial intelligence (XAI), i.e., tools that reveal what features (and in what combinations) deep learners use to make decisions. XAI can be used to solve the "black box" criticism common in DL and reveal mechanisms that ultimately produce clinical outcomes. We expect these techniques to greatly enhance the applicability of DL in psychiatric research and help reveal novel mechanisms and potential pathways for therapeutic intervention in mental illness.

Deep learning-based landslide susceptibility mapping

Landslides are considered as one of the most devastating natural hazards in Iran, causing extensive damage and loss of life. Landslide susceptibility maps for landslide prone areas can be used to plan for and mitigate the consequences of catastrophic landsliding events. Here, we developed a deep convolutional neural network (CNN–DNN) for mapping landslide susceptibility, and evaluated it on the Isfahan province, Iran, which has not previously been assessed on such a scale. The proposed model was trained and validated using training (80%) and testing (20%) datasets, each containing relevant data on historical landslides, field records and remote sensing images, and a range of geomorphological, geological, environmental and human activity factors as covariates. The CNN–DNN model prediction accuracy was tested using a wide range of statistics from the confusion matrix and error indices from the receiver operating characteristic (ROC) curve. The CNN–DNN model was evaluated comprehensively by comparing it to several state-of-the-art benchmark machine learning techniques including the support vector machine (SVM), logistic regression (LR), Gaussian naïve Bayes (GNB), multilayer perceptron (MLP), Bernoulli Naïve Bayes (BNB) and decision tree (DT) classifiers. The CNN–DNN model for landslide susceptibility mapping was found to predict more accurately than the benchmark algorithms, with an AUC = 90.9%, IRs = 84.8%, MSE = 0.17, RMSE = 0.40, and MAPE = 0.42. The map provided by the CNN–DNN clearly revealed a high-susceptibility area in the west and southwest, related to the main Zagros trend in the province. These findings can be of great utility for landslide risk management and land use planning in the Isfahan province.

RealVS: Toward Enhancing the Precision of Top Hits in Ligand-Based Virtual Screening of Drug Leads from Large Compound Databases

Accurate modeling of compound bioactivities is essential for the virtual screening of drug leads. In real-world scenarios, pharmacists tend to choose from the top-k hit compounds ranked by predicted bioactivities from a large database with interest to continue wet experiments for drug discovery. Significant improvement of the precision of the top hits in ligand-based virtual screening of drug leads is more valuable than conventional schemes for accurately predicting the bioactivities of all compounds from a large database. Here, we proposed a new method, RealVS, to significantly improve the top hits' precision and learn interpretable key substructures associated with compound bioactivities. The features of RealVS involve the following points. (1) Abundant transferable information from the source domain was introduced for alleviating the insufficiency of inactive ligands associated with drug targets. (2) The adversarial domain alignment was adopted to fit the distribution of generated features of compounds from the training data set and that from the screening database for greater model generalization ability. (3) A novel objective function was proposed to simultaneously optimize the classification loss, regression loss, and adversarial loss, where most inactive ligands tend to be screened out before activity regression prediction. (4) Graph attention networks were adopted for learning key substructures associated with ligand bioactivities for better model interpretability. The results on a large number of benchmark data sets show that our method has significantly improved the precision of top hits under various k values in ligand-based virtual screening of drug leads from large compound databases, which is of great value in real-world scenarios. The web server of RealVS is freely available at noveldelta.com/RealVS for academic purposes, where virtual screening of hits from large compound databases is accessible.

A Clinical Decision Web to Predict ICU Admission or Death for Patients Hospitalised with COVID-19 Using Machine Learning Algorithms

The purpose of the study was to build a predictive model for estimating the risk of ICU admission or mortality among patients hospitalized with COVID-19 and provide a user-friendly tool to assist clinicians in the decision-making process. The study cohort comprised 3623 patients with confirmed COVID-19 who were hospitalized in the SALUD hospital network of Aragon (Spain), which includes 23 hospitals, between February 2020 and January 2021, a period that includes several pandemic waves. Up to 165 variables were analysed, including demographics, comorbidity, chronic drugs, vital signs, and laboratory data. To build the predictive models, different techniques and machine learning (ML) algorithms were explored: multilayer perceptron, random forest, and extreme gradient boosting (XGBoost). A reduction dimensionality procedure was used to minimize the features to 20, ensuring feasible use of the tool in practice. Our model was validated both internally and externally. We also assessed its calibration and provide an analysis of the optimal cut-off points depending on the metric to be optimized. The best performing algorithm was XGBoost. The final model achieved good discrimination for the external validation set (AUC = 0.821, 95% CI 0.787-0.854) and accurate calibration (slope = 1, intercept = -0.12). A cut-off of 0.4 provides a sensitivity and specificity of 0.71 and 0.78, respectively. In conclusion, we built a risk prediction model from a large amount of data from several pandemic waves, which had good calibration and discrimination ability. We also created a user-friendly web application that can aid rapid decision-making in clinical practice.

Improved machine learning performances with transfer learning to predicting need for hospitalization in arboviral infections against the small dataset

The prediction of hospital patients and outpatients with suspected arboviral infection individuals in research-limited settings of the urban areas is defined as a challenging process for clinicians. Dengue, Chikungunya, and Zika arboviruses have gained attention in recent years because of the high prevalence in the society and financial burden of major global health systems. In this study, we proposed a machine learning algorithm based prediction model over retrospective medical records, which are named as SISA (the Severity Index for Suspected Arbovirus) and SISAL (the Severity Index for Suspected Arbovirus with Laboratory) datasets. Therefore, we aim to inform the clinicians about the use of machine learning with transfer learning success for diagnosis and comprehensive comparison of the classification performances over the SISA/SISAL datasets in the resource-limited settings that may cause to the small datasets of arboviral infection. In this study, Convolutional Neural Network and Long Short-Term Memory have achieved 100% accuracy and 1 of area under the curve (AUC) score, Fully Connected Deep Network has provided 92.86% accuracy and 0.969 AUC score in the SISAL dataset with transfer learning. Moreover, 98.73% accuracy and 0.988 AUC score were obtained by Convolutional Neural Network and Long Short-Term Memory for the SISA dataset. Furthermore, Linear Discriminant Analysis (shallow algorithm) has provided reaching up to 96.43% accuracy. Notably, deep learning based models have achieved improved performances compared to the previously reported study.

A comparison of feature extraction methods for prediction of neuropsychological scores from functional connectivity data of stroke patients

Multivariate prediction of human behavior from resting state data is gaining increasing popularity in the neuroimaging community, with far-reaching translational implications in neurology and psychiatry. However, the high dimensionality of neuroimaging data increases the risk of overfitting, calling for the use of dimensionality reduction methods to build robust predictive models. In this work, we assess the ability of four well-known dimensionality reduction techniques to extract relevant features from resting state functional connectivity matrices of stroke patients, which are then used to build a predictive model of the associated deficits based on cross-validated regularized regression. In particular, we investigated the prediction ability over different neuropsychological scores referring to language, verbal memory, and spatial memory domains. Principal Component Analysis (PCA) and Independent Component Analysis (ICA) were the two best methods at extracting representative features, followed by Dictionary Learning (DL) and Non-Negative Matrix Factorization (NNMF). Consistent with these results, features extracted by PCA and ICA were found to be the best predictors of the neuropsychological scores across all the considered cognitive domains. For each feature extraction method, we also examined the impact of the regularization method, model complexity (in terms of number of features that entered in the model) and quality of the maps that display predictive edges in the resting state networks. We conclude that PCA-based models, especially when combined with L1 (LASSO) regularization, provide optimal balance between prediction accuracy, model complexity, and interpretability.

Deep learning in systems medicine

Systems medicine (SM) has emerged as a powerful tool for studying the human body at the systems level with the aim of improving our understanding, prevention and treatment of complex diseases. Being able to automatically extract relevant features needed for a given task from high-dimensional, heterogeneous data, deep learning (DL) holds great promise in this endeavour. This review paper addresses the main developments of DL algorithms and a set of general topics where DL is decisive, namely, within the SM landscape. It discusses how DL can be applied to SM with an emphasis on the applications to predictive, preventive and precision medicine. Several key challenges have been highlighted including delivering clinical impact and improving interpretability. We used some prototypical examples to highlight the relevance and significance of the adoption of DL in SM, one of them is involving the creation of a model for personalized Parkinson's disease. The review offers valuable insights and informs the research in DL and SM.

A deep learning application to approximate the geometric orifice and coaptation areas of the polymeric heart valves under time - varying transvalvular pressure

Machine learning and deep learning frameworks have been presented as a substitute for lengthy computational analysis, such as finite element analysis, computational fluid dynamics, and fluid-structure interaction. In this study, our objective was to apply a deep learning framework to predict the geometric orifice (GOA) and the coaptation areas (CA) of the polymeric heart valves under the time-varying transvalvular pressure. 377 different valve geometries were generated by changing the control coordinates of the attachment and the belly curve. The GOA and the CA values were obtained at the maximum and the minimum transvalvular pressure, respectively. The results showed that the applied framework can accurately predict the GOA and the CA despite being trained with a relatively smaller data set. The presented framework can reduce the required time of the lengthy FE frameworks.

Using U-Net-Like Deep Convolutional Neural Networks for Precise Tree Recognition in Very High Resolution RGB (Red, Green, Blue) Satellite Images

Very high resolution satellite imageries provide an excellent foundation for precise mapping of plant communities and even single plants. We aim to perform individual tree recognition on the basis of very high resolution RGB (red, green, blue) satellite images using deep learning approaches for northern temperate mixed forests in the Primorsky Region of the Russian Far East. We used a pansharpened satellite RGB image by GeoEye-1 with a spatial resolution of 0.46 m/pixel, obtained in late April 2019. We parametrized the standard U-Net convolutional neural network (CNN) and trained it in manually delineated satellite images to solve the satellite image segmentation problem. For comparison purposes, we also applied standard pixel-based classification algorithms, such as random forest, k-nearest neighbor classifier, naive Bayes classifier, and quadratic discrimination. Pattern-specific features based on grey level co-occurrence matrices (GLCM) were computed to improve the recognition ability of standard machine learning methods. The U-Net-like CNN allowed us to obtain precise recognition of Mongolian poplar (Populus suaveolens Fisch. ex Loudon s.l.) and evergreen coniferous trees (Abies holophylla Maxim., Pinus koraiensis Siebold & Zucc.). We were able to distinguish species belonging to either poplar or coniferous groups but were unable to separate species within the same group (i.e. A. holophylla and P. koraiensis were not distinguishable). The accuracy of recognition was estimated by several metrics and exceeded values obtained for standard machine learning approaches. In contrast to pixel-based recognition algorithms, the U-Net-like CNN does not lead to an increase in false-positive decisions when facing green-colored objects that are similar to trees. By means of U-Net-like CNN, we obtained a mean accuracy score of up to 0.96 in our computational experiments. The U-Net-like CNN recognizes tree crowns not as a set of pixels with known RGB intensities but as spatial objects with a specific geometry and pattern. This CNN’s specific feature excludes misclassifications related to objects of similar colors as objects of interest. We highlight that utilization of satellite images obtained within the suitable phenological season is of high importance for successful tree recognition. The suitability of the phenological season is conceptualized as a group of conditions providing highlighting objects of interest over other components of vegetation cover. In our case, the use of satellite images captured in mid-spring allowed us to recognize evergreen fir and pine trees as the first class of objects (“conifers”) and poplars as the second class, which were in a leafless state among other deciduous tree species.

Digital Biomarkers for the Early Detection of Mild Cognitive Impairment: Artificial Intelligence Meets Virtual Reality

Elderly people affected by Mild Cognitive Impairment (MCI) usually report a perceived decline in cognitive functions that deeply impacts their quality of life. This subtle waning, although it cannot be diagnosable as dementia, is noted by caregivers on the basis of their relative’s behaviors. Crucially, if this condition is also not detected in time by clinicians, it can easily turn into dementia. Thus, early detection of MCI is strongly needed. Classical neuropsychological measures – underlying a categorical model of diagnosis - could be integrated with a dimensional assessment approach involving Virtual Reality (VR) and Artificial Intelligence (AI). VR can be used to create highly ecologically controlled simulations resembling the daily life contexts in which patients’ daily instrumental activities (IADL) usually take place. Clinicians can record patients’ kinematics, particularly gait, while performing IADL (Digital Biomarkers). Then, Artificial Intelligence employs Machine Learning (ML) to analyze them in combination with clinical and neuropsychological data. This integrated computational approach would enable the creation of a predictive model to identify specific patterns of cognitive and motor impairment in MCI. Therefore, this new dimensional cognitive-behavioral assessment would reveal elderly people’s neural alterations and impaired cognitive functions, typical of MCI and dementia, even in early stages for more time-sensitive interventions.

Post-stroke deficit prediction from lesion and indirect structural and functional disconnection

Behavioural deficits in stroke reflect both structural damage at the site of injury, and widespread network dysfunction caused by structural, functional, and metabolic disconnection. Two recent methods allow for the estimation of structural and functional disconnection from clinical structural imaging. This is achieved by embedding a patient's lesion into an atlas of functional and structural connections in healthy subjects, and deriving the ensemble of structural and functional connections that pass through the lesion, thus indirectly estimating its impact on the whole brain connectome. This indirect assessment of network dysfunction is more readily available than direct measures of functional and structural connectivity obtained with functional and diffusion MRI, respectively, and it is in theory applicable to a wide variety of disorders. To validate the clinical relevance of these methods, we quantified the prediction of behavioural deficits in a prospective cohort of 132 first-time stroke patients studied at 2 weeks post-injury (mean age 52.8 years, range 22-77; 63 females; 64 right hemispheres). Specifically, we used multivariate ridge regression to relate deficits in multiple functional domains (left and right visual, left and right motor, language, spatial attention, spatial and verbal memory) with the pattern of lesion and indirect structural or functional disconnection. In a subgroup of patients, we also measured direct alterations of functional connectivity with resting-state functional MRI. Both lesion and indirect structural disconnection maps were predictive of behavioural impairment in all domains (0.16 < R2 < 0.58) except for verbal memory (0.05 < R2 < 0.06). Prediction from indirect functional disconnection was scarce or negligible (0.01 < R2 < 0.18) except for the right visual field deficits (R2 = 0.38), even though multivariate maps were anatomically plausible in all domains. Prediction from direct measures of functional MRI functional connectivity in a subset of patients was clearly superior to indirect functional disconnection. In conclusion, the indirect estimation of structural connectivity damage successfully predicted behavioural deficits post-stroke to a level comparable to lesion information. However, indirect estimation of functional disconnection did not predict behavioural deficits, nor was a substitute for direct functional connectivity measurements, especially for cognitive disorders.

Digitizing the Pharma Neurons - A Technological Operation in Progress!

BACKGROUND

Digitization and automation are the buzzwords in clinical research and pharma companies are investigating heavily here. Right from drug discovery to personalized medicine, digital patients and patient engagement, there is great consideration of technology at each step.

METHODS

The published data and online information available is reviewed to give an overview of digitization in pharma, across the drug development cycle, industry collaborations and innovations. The regulatory guidelines, innovative collaborations across industry, academics and thought leadership are presented. Also included are some ideas, suggestions, way forwards while digitizing the pharma neurons, the regulatory stand, benefits and challenges.

RESULTS

The innovations range from discovering personalized medicine to conducting virtual clinical trials, and maximizing data collection from the real-world experience. To address the increasing demand for the real-world data and the needs of tech-savvy patients, the innovations are shaping up accordingly. Pharma companies are collaborating with academics and they are co-innovating the technology for example Massachusetts Institute of Technology's program. This focuses on the modernization of clinical trials, strategic use of artificial intelligence and machine learning using real-world evidence, assess the risk-benefit ratio of deploying digital analytics in medicine, and proactively identifying the solutions.

CONCLUSION

With unfolding data on the impact of science and technology amalgamation, we need shared mindset between data scientists and medical professionals to maximize the utility of enormous health and medical data. To tackle this efficiently, there is a need of cross-collaboration and education, and align with ethical and regulatory requirements. A perfect blend of industry, regulatory, and academia will ensure successful digitization of pharma neurons.

DeepSnap-Deep Learning Approach Predicts Progesterone Receptor Antagonist Activity With High Performance

The progesterone receptor (PR) is important therapeutic target for many malignancies and endocrine disorders due to its role in controlling ovulation and pregnancy via the reproductive cycle. Therefore, the modulation of PR activity using its agonists and antagonists is receiving increasing interest as novel treatment strategy. However, clinical trials using the PR modulators have not yet been found conclusive evidences. Recently, increasing evidence from several fields shows that the classification of chemical compounds, including agonists and antagonists, can be done with recent improvements in deep learning (DL) using deep neural network. Therefore, we recently proposed a novel DL-based quantitative structure-activity relationship (QSAR) strategy using transfer learning to build prediction models for agonists and antagonists. By employing this novel approach, referred as DeepSnap-DL method, which uses images captured from 3-dimension (3D) chemical structure with multiple angles as input data into the DL classification, we constructed prediction models of the PR antagonists in this study. Here, the DeepSnap-DL method showed a high performance prediction of the PR antagonists by optimization of some parameters and image adjustment from 3D-structures. Furthermore, comparison of the prediction models from this approach with conventional machine learnings (MLs) indicated the DeepSnap-DL method outperformed these MLs. Therefore, the models predicted by DeepSnap-DL would be powerful tool for not only QSAR field in predicting physiological and agonist/antagonist activities, toxicity, and molecular bindings; but also for identifying biological or pathological phenomena.