Background: Lack of quantifiable biomarkers is a major obstacle in diagnosing and treating depression. In adolescents, increasing suicidality during antidepressant treatment further complicates the problem. Objective: We sought to evaluate digital biomarkers for the diagnosis and treatment response of depression in adolescents through a newly developed smartphone app. Methods: We developed the Smart Healthcare System for Teens At Risk for Depression and Suicide app for Android-based smartphones. This app passively collected data reflecting the social and behavioral activities of adolescents, such as their smartphone usage time, physical movement distance, and the number of phone calls and text messages during the study period. Our study consisted of 24 adolescents (mean age 15.4 [SD 1.4] years, 17 girls) with major depressive disorder (MDD) diagnosed with Kiddie Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version and 10 healthy controls (mean age 13.8 [SD 0.6] years, 5 girls). After 1 week’s baseline data collection, adolescents with MDD were treated with escitalopram in an 8-week, open-label trial. Participants were monitored for 5 weeks, including the baseline data collection period. Their psychiatric status was measured every week. Depression severity was measured using the Children’s Depression Rating Scale-Revised and Clinical Global Impressions-Severity. The Columbia Suicide Severity Rating Scale was administered in order to assess suicide severity. We applied the deep learning approach for the analysis of the data. Deep neural network was employed for diagnosis classification, and neural network with weighted fuzzy membership functions was used for feature selection. Results: We could predict the diagnosis of depression with training accuracy of 96.3% and 3-fold validation accuracy of 77%. Of the 24 adolescents with MDD, 10 responded to antidepressant treatments. We predicted the treatment response of adolescents with MDD with training accuracy of 94.2% and 3-fold validation accuracy of 76%. Adolescents with MDD tended to move longer distances and use smartphones for longer periods of time compared to controls. The deep learning analysis showed that smartphone usage time was the most important feature in distinguishing adolescents with MDD from controls. Prominent differences were not observed in the pattern of each feature between the treatment responders and nonresponders. The deep learning analysis revealed that the total length of calls received as the most important feature predicting antidepressant response in adolescents with MDD. Conclusions: Our smartphone app demonstrated preliminary evidence of predicting diagnosis and treatment response in depressed adolescents. This is the first study to predict the treatment response of adolescents with MDD by examining smartphone-based objective data with deep learning approaches.

In this study, we propose a method to automatically find features from a dataset that are effective for classification or prediction, using a new method called multi-agent reinforcement learning and a guide agent. Each feature of the dataset has one of the main and guide agents, and these agents decide whether to select a feature. Main agents select the optimal features, and guide agents present the criteria for judging the main agents’ actions. After obtaining the main and guide rewards for the features selected by the agents, the main agent that behaves differently from the guide agent updates their Q-values by calculating the learning reward delivered to the main agents. The behavior comparison helps the main agent decide whether its own behavior is correct, without using other algorithms. After performing this process for each episode, the features are finally selected. The feature selection method proposed in this study uses multiple agents, reducing the number of actions each agent can perform and finding optimal features effectively and quickly. Finally, comparative experimental results on multiple datasets show that the proposed method can select effective features for classification and increase classification accuracy.

Deep neural networks have been successfully applied to generate predictive patterns from medical and diagnostic data. This paper presents an approach for assessing persons with Alzheimer’s disease (AD) mild cognitive impairment (MCI), compared with normal control (NC) persons, using the zoom-in neural network (ZNN) deep-learning algorithm. ZNN stacks a set of zoom-in learning units (ZLUs) in a feedforward hierarchy without backpropagation. The resting-state fMRI (rs-fMRI) dataset for AD assessments was obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). The Automated Anatomical Labeling (AAL-90) atlas, which provides 90 neuroanatomical functional regions, was used to assess and detect the implicated regions in the course of AD. The features of the ZNN are extracted from the 140-time series rs-fMRI voxel values in a region of the brain. ZNN yields the three classification accuracies of AD versus MCI and NC, NC versus AD and MCI, and MCI versus AD and NC of 97.7%, 84.8%, and 72.7%, respectively, with the seven discriminative regions of interest (ROIs) in the AAL-90.

Economic forecasting is crucial since it benefits many different parties, such as governments, businesses, investors, and the general public. This paper presents a novel methodology for forecasting business cycles using an autoregressive integrated moving average (ARIMA), a popular linear model in time series forecasting, and a neural network with weighted fuzzy membership functions (NEWFM) as a forecasting model generator. The study used a dataset that included seven components of the leading composite index, which is used to predict positive or negative trends in several economic sectors before the GDP is compiled. The preprocessed time series data comprising the leading composite index using ARIMA were used as input vectors for the NEWFM to predict comprehensive business fluctuations. The prediction capability significantly improved through the duplicated refining process of the dataset using ARIMA and NEWFM. The combined ARIMA and NEWFM techniques exceeded ARIMA in both classification and prediction, yielding an accuracy of 91.61%.

Background: Speech abnormalities are recognized as early indicators of Alzheimer's disease (AD) and mild cognitive impairment (MCI). Objective: To determine whether deep-learning models trained on mel-spectrograms of brief speech tasks can (i) discriminate individuals with MCI and AD from cognitively normal controls (NC) and (ii) estimate cognitive status with clinically useful accuracy. Methods: Speech from 594 participants (185 NC, 231 MCI, 178 AD) was recorded through a mobile application that included 11 cognitive-linguistic tasks. Audio was converted into mel-spectrogram images and processed using a VGG16-based deep-learning model with transfer learning and fine-tuning of block 5. Task-specific feature vectors were extracted, concatenated, and used to train a deep neural network. The dataset was split into training, validation, and test sets (3:1:1), and five-split cross-validation was performed. Results: The model demonstrated an overall accuracy of 72.4% in classifying NC from the abnormal group (MCI and AD), with sensitivity and specificity of 72.5% and 72.2%, respectively, a balanced accuracy of 72.4%, and an AUC of 0.997. In binary classifications, the model achieved 82.9% accuracy (balanced accuracy 82.9%, AUC 0.992) for NC versus AD, 70.7% accuracy (balanced accuracy 70.3%, AUC 0.956) for NC versus MCI, and 77.5% accuracy (balanced accuracy 78.9%, AUC 0.889) for MCI versus AD. Tasks such as serial subtraction, storytelling, and picture description contributed most to classification performance, indicating their effectiveness in capturing cognitive deficits. Conclusions: Mel-spectrogram-based deep-learning analysis of speech shows promise as a rapid, non-invasive, and language-independent screening tool for early cognitive impairment, with potential advantages over traditional assessments such as the Mini-Mental State Examination.

The Clock Drawing Test (CDT) is a professional examination that can detect cognitive impairments, such as Parkinson’s and Alzheimer’s diseases, based on scoring criteria. The pooling layers of a convolutional neural network (CNN) compress features by reducing dimensionality, which tends to focus on a single dominant element. This can be detrimental to compressing information in CDT images, where all elements are significant features. Therefore, in this study, we developed a model that utilizes features obtained from multiple channels to focus on all the elements within an image using channel and spatial attention. We utilized supervised contrastive learning to classify patient and control groups solely from CDT images. The features obtained from the multiple channels of the MCC-net were used to compute contrastive loss and learn representations of the data. Subsequently, a classifier was trained to learn the decision boundaries between the data. When the MCC-net was trained for binary classification, the accuracy, sensitivity, specificity, and area under the curve reached their maximum values of 0.9718, 0.8358, 0.9789, and 0.9700, respectively. As far as our knowledge extends, this study represents the first instance of utilizing supervised contrastive learning, acquiring features from multiple channels, for classifying CDT images, and we achieved superior performance compared to other models. Furthermore, the model visualized the attention clock elements to provide evidence for the inference results and presents the potential of utilizing artificial intelligence to classify CDT images.