Search

Objective: Parkinson’s Disease (PD) is a progressive neurodegenerative disorder, manifesting with subtle early signs, which, often hinder timely and early diagnosis and treatment. The development of accessible, technology-based methods for longitudinal PD symptoms tracking in daily living, offers the potential for transforming disease assessment and accelerating diagnosis. Methods: A privacy-aware method for classifying patients and healthy controls (HC), on the grounds of speech impairment present in PD, is proposed. Voice features from running speech signals were extracted from passively-captured recordings over voice calls. Language-aware training of multiple- and single-instance learning classifiers was employed to fuse and predict on voice features and demographic data from a multilingual cohort of 498 subjects (392/106 self-reported HC/PD patients). Results: By means of leave-one-subject-out cross-validation, the best-performing models yielded 0.69/0.68/0.63/0.83 area under the Receiver Operating Characteristic curve (AUC) for the binary classification of PD patient vs. HC in sub-cohorts of English/Greek/German/Portuguese-speaking subjects, respectively. Out-of sample testing of the best performing models was conducted in an additional dataset, generated by 63 clinically-assessed subjects (24/39 HC/early PD patients). Testing has resulted in 0.84/0.93/0.83 AUC for the English/Greek/German-speaking sub-cohorts, respectively. Conclusions: The proposed approach outperforms other methods proposed for language-aware PD detection considering the ecological validity of the voice data. Significance: This paper introduces for the first time a high-frequency, privacy-aware and unobtrusive PD screening tool based on analysis of voice samples captured during routine phone calls. [ABSTRACT FROM AUTHOR]

Seizure prediction for drug-refractory epilepsy patients can improve their quality of life, reduce their anxiety, and help them take the necessary precautions. Nowadays, numerous deep learning algorithms have been proposed to predict seizure onset and obtain better performance than traditional machine learning methods. However, these methods require a large set of parameters and large hardware resources; they also have high energy consumption. Therefore, these methods cannot be implemented on compact, low-power wearable, or implantable medical devices. The devices should operate on a real-time basis to continually inform the epileptic patients. In this paper, we describe energy-efficient and hardware-friendly methods to predict the epileptic seizures. A model of only 45 kB was obtained by the neural architecture search and was evaluated across three datasets. The overall sensitivity, false prediction rate, and area under receiver operating characteristic curve were 99.81%, 0.005/h, 1 and 93.48%, 0.063/h, 0.977 and 85.19%, 0.116/h, 0.933, respectively, for three public datasets. This model was further reduced with several methods for model compression. The performances for the model sizes less than 50 kB for scalp EEG data and less than 10 kB for intracranial EEG data outperformed all the other models of similar model sizes. In particular, the energy consumption estimation was less than 10 mJ per inference for scalp EEG signal and less than 0.5 mJ per inference for intracranial EEG signal, which meet the requirements for low-power wearable and implantable devices, respectively. [ABSTRACT FROM AUTHOR]

Objective: Pain assessment in children continues to challenge clinicians and researchers, as subjective experiences of pain require inference through observable behaviors, both involuntary and deliberate. The presented approach supplements the subjective self-report-based method by fusing electrodermal activity (EDA) recordings with video facial expressions to develop an objective pain assessment metric. Such an approach is specifically important for assessing pain in children who are not capable of providing accurate self-pain reports, requiring nonverbal pain assessment. We demonstrate the performance of our approach using data recorded from children in post-operative recovery following laparoscopic appendectomy. We examined separately and combined the usefulness of EDA and video facial expression data as predictors of children's self-reports of pain following surgery through recovery. Findings indicate that EDA and facial expression data independently provide above chance sensitivities and specificities, but their fusion for classifying clinically significant pain vs. clinically nonsignificant pain achieved substantial improvement, yielding 90.91% accuracy, with 100% sensitivity and 81.82% specificity. The multimodal measures capitalize upon different features of the complex pain response. Thus, this paper presents both evidence for the utility of a weighted maximum likelihood algorithm as a novel feature selection method for EDA and video facial expression data and an accurate and objective automated classification algorithm capable ofdiscriminating clinically significant pain from clinically nonsignificant pain in children. [ABSTRACT FROM AUTHOR]

Arrhythmia detection and classification is a crucial step for diagnosing cardiovascular diseases. However, deep learning models that are commonly used and trained in end-to-end fashion are not able to provide good interpretability. In this paper, we address this deficiency by proposing the first novel interpretable arrhythmia classification approach based on a human-machine collaborative knowledge representation. Our approach first employs an AutoEncoder to encode electrocardiogram signals into two parts: hand-encoding knowledge and machine-encoding knowledge. A classifier then takes as input the encoded knowledge to classify arrhythmia heartbeats with or without human in the loop (HIL). Experiments and evaluation on the MIT-BIH Arrhythmia Database demonstrate that our new approach not only can effectively classify arrhythmia while offering interpretability, but also can improve the classification accuracy by adjusting the hand-encoding knowledge with our HIL mechanism. [ABSTRACT FROM AUTHOR]

Evaluating progress throughout a patient's rehabilitation episode is critical for determining the effectiveness of the selected treatments and is an essential ingredient in personalised and evidence-based rehabilitation practice. The evaluation process is complex due to the inherently large human variations in motor recovery and the limitations of commonly used clinical measurement tools. Information recorded during a robot-assisted rehabilitation process can provide an effective means to continuously quantitatively assess movement performance and rehabilitation progress. However, selecting appropriate motion features for rehabilitation evaluation has always been challenging. This paper exploits unsupervised feature learning techniques to reduce the complexity of building the evaluation model of patients’ progress. A new feature learning technique is developed to select the most significant features from a large amount of kinematic features measured from robotics, providing clinically useful information to health practitioners with reduction of modeling complexity. A novel indicator that uses monotonicity and trendability is proposed to evaluate kinematic features. The data used to develop the feature selection technique consist of kinematic data from robot-aided rehabilitation for a population of stroke patients. The selected kinematic features allow for human variations across a population of patients as well as over the sequence of rehabilitation sessions. The study is based on data records pertaining to 41 stroke patients using three different robot assisted exercises for upper limb rehabilitation. Consistent with the literature, the results indicate that features based on movement smoothness are the best measures among 17 kinematic features suitable to evaluate rehabilitation progress. [ABSTRACT FROM AUTHOR]

Objective: Detection of Atrial fibrillation (AF) from premature atrial contraction (PAC) and premature ventricular contraction (PVC) is difficult as frequent occurrences of these ectopic beats can mimic the typical irregular patterns of AF. In this paper, we present a novel density Poincaré plot-based machine learning method to detect AF from PAC/PVCs using electrocardiogram (ECG) recordings. Methods: First, we propose the generation of this new density Poincaré plot which is derived from the difference of the heart rate (DHR) and provides the overlapping phase-space trajectory information of the DHR. Next, from this density Poincaré plot, several image processing domain-based approaches including statistical central moments, template correlation, Zernike moment, discrete wavelet transform and Hough transform features are used to extract suitable features. Subsequently, the infinite latent feature selection algorithm is implemented to rank the features. Finally, classification of AF vs. PAC/PVC is performed using K-Nearest Neighbor, Support vector machine (SVM) and Random Forest (RF) classifiers. Our method is developed and validated using a subset of Medical Information Mart for Intensive Care (MIMIC) III database containing 10 AF and 10 PAC/PVC subjects. Results- During the segment-wise 10-fold cross-validation, SVM achieved the best performance with 98.99% sensitivity, 95.18% specificity and 97.45% accuracy with the extracted features. In subject-wise scenario, RF achieved the highest accuracy of 91.93%. Moreover, we further validated the proposed method using two other databases: wearable armband ECG data and the Physionet AFPDB. 100% PAC detection accuracy was obtained for both databases without any further training. Conclusion: Our proposed density Poincaré plot-based method showed superior performance when compared with four existing algorithms; thus showing the efficacy of the extracted image domain-based features. Significance: From intensive care unit's ECG to wearable armband ECGs, the proposed method is shown to discriminate PAC/PVCs from AF with high accuracy. [ABSTRACT FROM AUTHOR]

Objective: in this paper we propose a system to detect a subject's sympathetic reaction, which is related to unexpected or challenging events during a car drive. Methods: we use the Electrocardiogram (ECG) signal and the Skin Potential Response (SPR) signal, which has several advantages with respect to other Electrodermal (EDA) signals. We record one SPR signal for each hand, and use an algorithm that, selecting the smoother signal, is able to remove motion artifacts. We extract statistical features from the ECG and SPR signals in order to classify signal segments and identify the presence or absence of emotional events via a Supervised Learning Algorithm. The experiments were carried out in a company which specializes in driving simulator equipment, using a motorized platform and a driving simulator. Different subjects were tested with this setup, with different challenging events happening on predetermined locations on the track. Results: we obtain an Accuracy as high as 79.10% for signal blocks and as high as 91.27% for events. Conclusion: results demonstrate the good performance of the presented system in detecting sympathetic reactions, and the effectiveness of the motion artifact removal procedure. Significance: our work demonstrates the possibility to classify the emotional state of the driver, using the ECG and EDA signals and a slightly invasive setup. In particular, the proposed use of SPR and of the motion artifact removal procedure are crucial for the effectiveness of the system. [ABSTRACT FROM AUTHOR]

Recognizing activities of daily living (ADL) provides vital contextual information that enhances the effectiveness of various mobile health and wellness applications. Development of wearable motion sensors along with machine learning algorithms offer a great opportunity for ADL recognition. However, the performance of the ADL recognition systems may significantly degrade when they are used by a new user due to inter-subject variability. This issue limits the usability of these systems. In this paper, we propose a deep learning assisted personalization framework for ADL recognition with the aim to maximize the personalization performance while minimizing solicitation of inputs or labels from the user to reduce user's burden. The proposed framework consists of unsupervised retraining of automatic feature extraction layers and supervised fine-tuning of classification layers through a novel active learning model based on a given model's uncertainty. We design a Bayesian deep convolutional neural network with stochastic latent variables that allows us to estimate both aleatoric (data-dependent) and epistemic (model-dependent) uncertainties in recognition task. In this study, for the first time, we show how distinguishing between the two aforementioned sources of uncertainty leads to more effective active learning. The experimental results show that our proposed method improves the accuracy of ADL recognition on a new user by 25% on average compared to the case of using a model for a new user with no personalization with an average final accuracy of 89.2%. Moreover, our method achieves higher personalization accuracy while significantly reducing user's burden in terms of soliciting inputs and labels compared to other methods. [ABSTRACT FROM AUTHOR]

Objective: This paper targets a major challenge in developing practical electroencephalogram (EEG)-based brain–computer interfaces (BCIs): how to cope with individual differences so that better learning performance can be obtained for a new subject, with minimum or even no subject-specific data? Methods: We propose a novel approach to align EEG trials from different subjects in the Euclidean space to make them more similar, and hence improve the learning performance for a new subject. Our approach has three desirable properties: first, it aligns the EEG trials directly in the Euclidean space, and any signal processing, feature extraction, and machine learning algorithms can then be applied to the aligned trials; second, its computational cost is very low; and third, it is unsupervised and does not need any label information from the new subject. Results: Both offline and simulated online experiments on motor imagery classification and event-related potential classification verified that our proposed approach outperformed a state-of-the-art Riemannian space data alignment approach, and several approaches without data alignment. Conclusion: The proposed Euclidean space EEG data alignment approach can greatly facilitate transfer learning in BCIs. Significance: Our proposed approach is effective, efficient, and easy to implement. It could be an essential pre-processing step for EEG-based BCIs. [ABSTRACT FROM AUTHOR]

In this paper, we present a deep learning framework “Rehab-Net” for effectively classifying three upper limb movements of the human arm, involving extension, flexion, and rotation of the forearm, which, over the time, could provide a measure of rehabilitation progress. The proposed framework, Rehab-Net is formulated with a personalized, light weight and low-complex, customized convolutional neural network (CNN) model, using two-layers of CNN, interleaved with pooling layers, followed by a fully connected layer that classifies the three movements from tri-axial acceleration input data collected from the wrist. The proposed Rehab-Net framework was validated on sensor data collected in two situations: 1) semi-naturalistic environment involving an archetypal activity of “making-tea” with four stroke survivors and 2) natural environment, where ten stroke survivors were free to perform any desired arm movement for the duration of 120 min. We achieved an overall accuracy of 97.89% on semi-naturalistic data and 88.87% on naturalistic data which exceeded state-of-the-art learning algorithms namely, linear discriminant analysis, support vector machines, and k-means clustering with an average accuracy of 48.89%, 44.14%, and 27.64%. Subsequently, a computational complexity analysis of the proposed model has been discussed with an eye toward hardware implementation. The clinical significance of this study is to accurately monitor the clinical progress of the rehabilitated subjects under the ambulatory settings. [ABSTRACT FROM AUTHOR]

Goal: In this paper, we propose methods for (1) automatic feature extraction and classification for acetic acid and Lugol's iodine cervigrams and (2) methods for combining features/diagnosis of different contrasts in cervigrams for improved performance. Methods: We developed algorithms to pre-process pathology-labeled cervigrams and extract simple but powerful color and textural-based features. The features were used to train a support vector machine model to classify cervigrams based on corresponding pathology for visual inspection with acetic acid, visual inspection with Lugol's iodine, and a combination of the two contrasts. Results: The proposed framework achieved a sensitivity, specificity, and accuracy of 81.3%, 78.6%, and 80.0%, respectively, when used to distinguish cervical intraepithelial neoplasia (CIN+) relative to normal and benign tissues. This is superior to the average values achieved by three expert physicians on the same data set for discriminating normal/benign cases from CIN+ (77% sensitivity, 51% specificity, and 63% accuracy). Conclusion: The results suggest that utilizing simple color- and textural-based features from visual inspection with acetic acid and visual inspection with Lugol's iodine images may provide unbiased automation of cervigrams. Significance: This would enable automated, expert-level diagnosis of cervical pre-cancer at the point of care. [ABSTRACT FROM AUTHOR]

Objective: Chinese women are seriously threatened by breast cancer with high morbidity and mortality. The lack of robust prognosis models results in difficulty for doctors to prepare an appropriate treatment plan that may prolong patient survival time. An alternative prognosis model framework to predict invasive disease-free survival (iDFS) for early stage breast cancer patients, called MP4Ei, is proposed. MP4Ei framework gives an excellent performance to predict the relapse or metastasis breast cancer of Chinese patients in five years. Methods: MP4Ei is built based on statistical theory and gradient boosting decision tree framework. 5246 patients, derived from the clinical research center for breast in West China Hospital of Sichuan University, with early-stage (stage I–III) breast cancer are eligible for inclusion. Stratified feature selection, including statistical and ensemble methods, is adopted to select 23 out of the 89 patient features about the patient’ demographics, diagnosis, pathology, and therapy. Then, 23 selected features as the input variables are imported into the XGBoost algorithm, with Bayesian parameter tuning and cross validation, to find out the optimum simplified model for five-year iDFS prediction. Results: For eligible data, with 4196 patients (80%) for training, and with 1050 patients (20%) for testing, MP4Ei achieves comparable accuracy with AUC 0.8451, which has a significant advantage (p < 0.05). Conclusion: This work demonstrates the complete iDFS prognosis model with very competitive performance. Significance: The proposed method in this paper could be used in clinical practice to predict patients’ prognosis and future surviving state, which may help doctors make treatment plan. [ABSTRACT FROM AUTHOR]

Objective: The objectives of this paper are to develop and test the ability of a wearable physiological sensors system, based on ECG, EDA, and EEG, to capture human stress and to assess whether the detected changes in physiological signals correlate with changes in salivary cortisol level, which is a reliable, objective biomarker of stress. Methods: 15 healthy participants, eight males and seven females, mean age 40.8 ± 9.5 years, wore a set of three commercial sensors to record physiological signals during the Maastricht Acute Stress Test, an experimental protocol known to elicit robust physical and mental stress in humans. Salivary samples were collected throughout the different phases of the test. Statistical analysis was performed using a support vector machine (SVM) classification algorithm. A correlation analysis between extracted physiological features and salivary cortisol levels was also performed. Results: 15 features extracted from heart rate variability, electrodermal, and electroencephalography signals showed a high degree of significance in disentangling stress from a relaxed state. The classification algorithm, based on significant features, provided satisfactory outcomes with 86% accuracy. Furthermore, correlation analysis showed that the observed changes in physiological features were consistent with the trend of salivary cortisol levels (R2 = 0.714). Conclusion: The tested set of wearable sensors was able to successfully capture human stress and quantify stress level. Significance: The results of this pilot study may be useful in designing portable and remote control systems, such as medical devices used to turn on interventions and prevent stress consequences. [ABSTRACT FROM AUTHOR]

Today, medical image analysis papers require solid experiments to prove the usefulness of proposed methods. However, experiments are often performed on data selected by the researchers, which may come from different institutions, scanners, and populations. Different evaluation measures may be used, making it difficult to compare the methods. In this paper, we introduce a dataset of 7909 breast cancer histopathology images acquired on 82 patients, which is now publicly available from http://web.inf.ufpr.br/vri/breast-cancer-database. The dataset includes both benign and malignant images. The task associated with this dataset is the automated classification of these images in two classes, which would be a valuable computer-aided diagnosis tool for the clinician. In order to assess the difficulty of this task, we show some preliminary results obtained with state-of-the-art image classification systems. The accuracy ranges from 80% to 85%, showing room for improvement is left. By providing this dataset and a standardized evaluation protocol to the scientific community, we hope to gather researchers in both the medical and the machine learning field to advance toward this clinical application. [ABSTRACT FROM AUTHOR]

Objective: Gait analysis of animal disease models can provide valuable insights into in vivo compound effects and thus help in preclinical drug development. The purpose of this paper is to establish a computational gait analysis approach for the Noldus Catwalk system, in which footprints are automatically captured and stored. Methods: We present a - to our knowledge - first machine learning based approach for the Catwalk system, which comprises a step decomposition, definition and extraction of meaningful features, multivariate step sequence alignment, feature selection, and training of different classifiers (gradient boosting machine, random forest, and elastic net). Results: Using animal-wise leave-one-out cross validation we demonstrate that with our method we can reliable separate movement patterns of a putative Parkinson's disease animal model and several control groups. Furthermore, we show that we can predict the time point after and the type of different brain lesions and can even forecast the brain region, where the intervention was applied. We provide an in-depth analysis of the features involved into our classifiers via statistical techniques for model interpretation. Conclusion: A machine learning method for automated analysis of data from the Noldus Catwalk system was established. Significance: Our works shows the ability of machine learning to discriminate pharmacologically relevant animal groups based on their walking behavior in a multivariate manner. Further interesting aspects of the approach include the ability to learn from past experiments, improve with more data arriving and to make predictions for single animals in future studies. [ABSTRACT FROM PUBLISHER]

Objective: Walking task prediction in powered leg prostheses is an important problem in the development of biomimetic prosthesis controllers. This paper proposes a novel method to predict upcoming walking tasks by estimating the translational motion of leg joints using an integrated inertial measurement unit. Methods: We asked six subjects with unilateral transtibial amputations to traverse flat ground, ramps, and stairs using a powered prosthesis while inertial signals were collected. We then performed an offline analysis in which we simulated a real-time motion tracking algorithm on the inertial signals to estimate knee and ankle joint translations, and then used pattern recognition separately on the inertial and translational signal sets to predict the target walking tasks of individual strides. Results: Our analysis showed that using inertial signals to derive translational signals enabled a prediction error reduction of 6.8% compared to that attained using the original inertial signals. This result was similar to that seen by addition of surface electromyography sensors to integrated sensors in previous work, but was effected without adding any extra sensors. Finally, we reduced the size of the translational set to that of the inertial set and showed that the former still enabled a composite error reduction of 5.8%. Conclusion and Significance: These results indicate that translational motion tracking can be used to substantially enhance walking task prediction in leg prostheses without adding external sensing modalities. Our proposed algorithm can thus be used as a part of a task-adaptive and fully integrated prosthesis controller. [ABSTRACT FROM PUBLISHER]

Poor sleep is increasingly being recognized as an important prognostic parameter of health. For those with suspected sleep disorders, patients are referred to sleep clinics, which guide treatment. However, sleep clinics are not always a viable option due to their high cost, a lack of experienced practitioners, lengthy waiting lists, and an unrepresentative sleeping environment. A home-based noncontact sleep/wake monitoring system may be used as a guide for treatment potentially stratifying patients by clinical need or highlighting longitudinal changes in sleep and nocturnal patterns. This paper presents the evaluation of an undermattress sleep monitoring system for noncontact sleep/wake discrimination. A large dataset of sensor data with concomitant sleep/wake state was collected from both younger and older adults participating in a circadian sleep study. A thorough training/testing/validation procedure was configured and optimized feature extraction and sleep/wake discrimination algorithms evaluated both within and across the two cohorts. An accuracy, sensitivity, and specificity of 74.3%, 95.5%, and 53.2% is reported over all subjects using an external validation dataset (71.9%, 87.9%, and 56% and 77.5%, 98%, and 57% is reported for younger and older subjects, respectively). These results compare favorably with similar research, however this system provides an ambient alternative suitable for long-term continuous sleep monitoring, particularly among vulnerable populations. [ABSTRACT FROM PUBLISHER]

Objective: The objective of this paper is to describe and evaluate an algorithm to reduce power usage and increase battery lifetime for wearable health-monitoring devices. Methods: We describe a novel dynamic computation offloading scheme for real-time wearable health monitoring devices that adjusts the partitioning of data processing between the wearable device and mobile application as a function of desired classification accuracy. Results: By making the correct offloading decision based on current system parameters, we show that we are able to reduce system power by as much as 20%. Conclusion: We demonstrate that computation offloading can be applied to real-time monitoring systems, and yields significant power savings. Significance: Making correct offloading decisions for health monitoring devices can extend battery life and improve adherence. [ABSTRACT FROM AUTHOR]

In this paper, we propose a method for screening for the presence of type 2 diabetes by means of the signal obtained from a pulse oximeter. The screening system consists of two parts: the first analyzes the signal obtained from the pulse oximeter, and the second consists of a machine-learning module. The system consists of a front end that extracts a set of features form the pulse oximeter signal. These features are based on physiological considerations. The set of features were the input of a machine-learning algorithm that determined the class of the input sample, i.e., whether the subject had diabetes or not. The machine-learning algorithms were random forests, gradient boosting, and linear discriminant analysis as benchmark. The system was tested on a database of \text1157 subjects (two samples per subject) collected from five community health centers. The mean receiver operating characteristic area found was \text69.4% (median value \text71.9% and range [\text75.4-\text61.1]%), with a specificity =  \text64% for a threshold that gave a sensitivity = \text65%. We present a screening method for detecting diabetes that has a performance comparable to the glycated haemoglobin (haemoglobin A1c HbA1c) test, does not require blood extraction, and yields results in less than 5 min. [ABSTRACT FROM AUTHOR]

In this paper, a multiscale convolutional network (MSCN) and graph-partitioning-based method is proposed for accurate segmentation of cervical cytoplasm and nuclei. Specifically, deep learning via the MSCN is explored to extract scale invariant features, and then, segment regions centered at each pixel. The coarse segmentation is refined by an automated graph partitioning method based on the pretrained feature. The texture, shape, and contextual information of the target objects are learned to localize the appearance of distinctive boundary, which is also explored to generate markers to split the touching nuclei. For further refinement of the segmentation, a coarse-to-fine nucleus segmentation framework is developed. The computational complexity of the segmentation is reduced by using superpixel instead of raw pixels. Extensive experimental results demonstrate that the proposed cervical nucleus cell segmentation delivers promising results and outperforms existing methods. [ABSTRACT FROM PUBLISHER]

Objective: This paper presents the results of a new approach for selection of RF time series features based on joint independent component analysis for in vivo characterization of prostate cancer. Methods: We project three sets of RF time series features extracted from the spectrum, fractal dimension, and the wavelet transform of the ultrasound RF data on a space spanned by five joint independent components. Then, we demonstrate that the obtained mixing coefficients from a group of patients can be used to train a classifier, which can be applied to characterize cancerous regions of a test patient. Results: In a leave-one-patient-out cross validation, an area under receiver operating characteristic curve of 0.93 and classification accuracy of 84% are achieved. Conclusion: Ultrasound RF time series can be used to accurately characterize prostate cancer, in vivo without the need for exhaustive search in the feature space. Significance: We use joint independent component analysis for systematic fusion of multiple sets of RF time series features, within a machine learning framework, to characterize PCa in an in vivo study. [ABSTRACT FROM PUBLISHER]

The current diagnosis process of dementia is resulting in a high percentage of cases with delayed detection. To address this problem, in this paper, we explore the feasibility of autonomously detecting mild cognitive impairment (MCI) in the older adult population. We implement a signal processing approach equipped with a machine learning paradigm to process and analyze real-world data acquired using home-based unobtrusive sensing technologies. Using the sensor and clinical data pertaining to 97 subjects, acquired over an average period of three years, a number of measures associated with the subjects' walking speed and general activity in the home were calculated. Different time spans of these measures were used to generate feature vectors to train and test two machine learning algorithms namely support vector machines and random forests. We were able to autonomously detect MCI in older adults with an area under the ROC curve of 0.97 and an area under the precision-recall curve of 0.93 using a time window of 24 weeks. This study is of great significance since it can potentially assist in the early detection of cognitive impairment in older adults. [ABSTRACT FROM PUBLISHER]