An Artificial Neural Network for Movement Pattern Analysis to Estimate Blood Alcohol Content Level

Fine-Grained Intoxicated Gait Classification Using a Bilinear CNN

Consuming excessive amounts of alcohol causes impaired mobility and judgment and driving accidents, resulting in more than 800 injuries and fatalities each day. Passive methods to detect intoxicated drivers beyond the safe driving limit can facilitate Just-In-Time alerts and reduce Driving Under the Influence (DUI) incidents. Popularly-owned smartphones are not only equipped with motion sensors (accelerometer and gyroscope) that can be employed for passively collecting gait (walk) data but also have the processing power to run computationally expensive machine learning models. In this paper, we advance the state-of-the-art by proposing a novel method that utilizes a Bi-linear Convolution Neural Network (BiCNN) for analyzing smartphone accelerometer and gyroscope data to determine whether a smartphone user is over the legal driving limit (0.08) from their gait. After segmenting the gait data into steps, we converted the smartphone motion sensor data to a Gramian Angular Field (GAF) image and then leveraged the BiCNN architecture for intoxication classification. Distinguishing GAF-encoded images of the gait of intoxicated vs. sober users is challenging as the differences between the classes (intoxicated vs. sober) are subtle, also known as a fine-grained image classification problem. The BiCNN neural network has previously produced state-of-the-art results on fine-grained image classification of natural images. To the best of our knowledge, our work is the first to innovatively utilize the BiCNN to classify GAF encoded images of smartphone gait data in order to detect intoxication. Prior work had explored using the BiCNN to classify natural images or explored other gait-related tasks but not intoxication Our complete intoxication classification pipeline consists of several important pre-processing steps carefully adapted to the BAC classification task, including step detection and segmentation, data normalization to account for inter-subject variability, data fusion, GAF image generation from time-series data, and a BiCNN classification model. In rigorous evaluation, our BiCNN model achieves an accuracy of 83.5%, outperforming the previous state-of-the-art and demonstrating the feasibility of our approach.

Leveraging Mobile Phone Sensors, Machine Learning and Explainable Artificial Intelligence to Predict Imminent Same-Day Binge Drinking Events to Support Just-In-Time Adaptive Interventions: A Feasibility Study

BACKGROUND

Digital Just-In-Time Adaptive Interventions (JITAIs) can reduce binge drinking events (BDEs: consuming 4+/5+ drinks per occasion for women/men) in young adults, but need to be optimized for timing and content. Delivering just-in-time support messages in the hours prior to BDEs could improve intervention impact.

OBJECTIVE

We determined the feasibility of developing a machine learning model to accurately predict future, that is, same-day, 1 to 6-hours prior BDEs using smartphone sensor data. We aimed to identify the most informative phone sensor features associated with BDEs on weekend and weekdays, respectively, to determine the key features that explain prediction model performance.

METHODS

We collected phone sensor data from 75 young adults (ages 21-25; mean =22.4, SD=1.9) with risky drinking behavior who reported drinking behavior over 14 weeks. Participants in this secondary analysis were enrolled in a clinical trial. We developed machine learning models testing different algorithms (e.g., XGBoost, decision tree) to predict same-day BDEs (versus low-risk drinking events and non-drinking periods) using smartphone sensor data (e.g., accelerometer, GPS). We tested various "prediction distance" time windows (more proximal: 1-hour; to distant: 6-hour) from drinking onset. We also tested various analysis time windows (i.e., amount of data to be analyzed), ranging from 1 to 12 hours prior to drinking onset, because this determines the amount of data that needs to be stored on the phone to compute the model. Explainable AI (XAI) was used to explore interactions between the most informative phone sensor features contributing to BDEs.

RESULTS

The XGBoost model performed best in predicting imminent same-day BDE, with 95.0% accuracy on weekends and 94.3% accuracy on weekdays (F1 score = 0.95 and 0.94, respectively). This XGBoost model needed 12- and 9-hours of phone sensor data at 3- and 6- hours prediction distance from the onset of drinking, on weekends and weekdays, respectively, prior to predicting same-day BDEs. The most informative phone sensor features for BDE prediction were time (e.g., time of day) and GPS-derived, such as radius of gyration (an indicator of travel). Interactions among key features (e.g., time of day, GPS-derived features) contributed to prediction of same-day BDE.

CONCLUSIONS

We demonstrated the feasibility and potential use of smartphone sensor data and machine learning to accurately predict imminent (same-day) BDEs in young adults. The prediction model provides "windows of opportunity" and with the adoption of XAI, we identified "key contributing features" to trigger JITAI prior to the onset of BDEs, with the potential to reduce the likelihood of BDEs in young adults.

CLINICALTRIAL

User-centered preferences for a gait-informed alcohol intoxication app

BACKGROUND

mHealth technology can be used as a potential intervention for alcohol-related consequences. Applications designed to monitor alcohol use and relay information to the user may help to reduce risky behavior. Acceptability of such applications needs to be assessed.

METHODS

Survey data from 139 participants (29.8 years on average, 58% female) completing a single-session study for developing an application to detect blood alcohol concentration (BAC) from gait was analyzed to examine user preferences. Participants reported on their interest in an application for monitoring BAC from gait. Participants also reported on their preference for controlling features of the application. Acceptability and feasibility data were collected. Data were examined for the entire sample as well as differences in preference by age and gender were examined.

RESULTS

The majority of the sample indicated that they were interested in using an mHealth application to infer BAC from their gait. Users were interested in being able to control features of the application, such as monitoring BAC and reporting information to other individuals. Adults, as compared to emerging adults, preferred the ability to turn off the BAC-monitoring feature of the app. Females reported a preference for an app that does not allow the user to turn off notifications for BAC as well as safety features of the app.

CONCLUSIONS

Results of the survey data indicate general interest in mHealth technology that monitors BAC from passive input. These results suggest that such an app may be accepted and used as an intervention for monitoring alcohol levels, which could mediate drinking and alcohol-related consequences.

Estimation of Blood Alcohol Concentration From Smartphone Gait Data Using Neural Networks

Driving is a dynamic activity, which requires quick reflexes and decision making in order to respond to sudden changes in traffic conditions. Alcohol consumption impairs motor and cognitive skills, and causes many driving-related accidents annually. Passive methods of proactively detecting drivers who are too drunk to drive in order to notify them and prevent accidents, have recently been proposed. The effects of alcohol on a drinker's gait (walk) is a reliable indicator of their intoxication level. In this paper, we investigate detecting drinkers' intoxication levels from their gait by using neural networks to analyze sensor data gathered from their smartphone. Using data gathered from a large controlled alcohol study, we perform regression analysis using a Bi-directional Long Short Term Memory (Bi-LSTM) and Convolutional Neural Network (CNN) architectures to predict a person's Blood Alcohol Concentration (BAC) from their smartphone's accelerometer and gyroscope data. We innovatively proposed a comprehensive suite of pre-processing techniques and model-specific extensions to vanilla CNN and bi-LSTM models, which are well thought out and adapted specifically for BAC estimation. Our Bi-LSTM architecture achieves an RMSE of 0.0167 and the CNN architecture achieves an RMSE of 0.0168, outperforming state-of-the-art intoxication detection models using Bayesian Regularized Multilayer Perceptrons (MLP) (RMSE of 0.017) and the Random Forest (RF), with hand-crafted features. Moreover, our models learn features from raw sensor data, obviating the need for hand-crafted features, which is time consuming. Moreover, they achieve lower variance across folds and are hence more generalizable.

Alcohol Use Disorder in the Age of Technology: A Review of Wearable Biosensors in Alcohol Use Disorder Treatment

Biosensors enable observation and understanding of latent physiological occurrences otherwise unknown or invasively detected. Wearable biosensors monitoring physiological constructs across a wide variety of mental and physical health conditions have become an important trend in innovative research methodologies. Within substance use research, explorations of biosensor technology commonly focus on identifying physiological indicators of intoxication to increase understanding of addiction etiology and to inform treatment recommendations. In this review, we examine the state of research in this area as it pertains to treatment of alcohol use disorders specifically highlighting the gaps in our current knowledge with recommendations for future research. Annually, alcohol use disorders affect approximately 15 million individuals. A primary focus of existing wearable technology-based research among people with alcohol use disorders is identifying alcohol intoxication. A large benefit of wearable biosensors for this purpose is they provide continuous readings in a passive manner compared with the gold standard measure of blood alcohol content (BAC) traditionally measured intermittently by breathalyzer or blood draw. There are two primary means of measuring intoxication with biosensors: gait and sweat. Gait changes have been measured via smart sensors placed on the wrist, in the shoe, and mobile device sensors in smart phones. Sweat measured by transdermal biosensors detects the presence of alcohol in the blood stream correlating to BAC. Transdermal biosensors have been designed in tattoos/skin patches, shirts, and most commonly, devices worn on the ankle or wrist. Transdermal devices were initially developed to help monitor court-ordered sobriety among offenders with alcohol use disorder. These devices now prove most useful in continuously tracking consumption throughout clinical trials for behavioral treatment modalities. More recent research has started exploring the uses for physical activity trackers and physiological arousal sensors to guide behavioral interventions for relapse prevention. While research has begun to demonstrate wearable devices' utility in reducing alcohol consumption among individuals aiming to cutdown on their drinking, monitoring sustained abstinence in studies exploring contingency management for alcohol use disorders, and facilitating engagement in activity-based treatment interventions, their full potential to further aid in understanding of, and treatment for, alcohol use disorders has yet to be explored.

Modeling of a Novel Coaxial Ducted Fan Aerial Robot Combined with Corner Environment by Using Artificial Neural Network

A novel coaxial ducted fan aerial robot with a manipulator is proposed which can achieve some hover operation tasks in a corner environment, such as switching on and off a wall-attached button on the corner. In order to study the aerodynamic interference between the prototype and the environment when the aerial robot is hovering in the corner environment, a method for the comprehensive modeling of the prototype and corner environment based on the artificial neural network is presented. By using the CFD simulation software, the flow field of the prototype at different positions with the corner effect is analyzed. After determining the input, output and structure of the neural network model, the Adam and gradient descent algorithms are selected as the neural network training algorithms, respectively. In addition, to optimize the initial weights and biases of the neural network model, the genetic algorithm is precisely used. The three-dimensional prediction surfaces generated by the three methods of the neural network, kriging surface and the polynomial fitting are compared. The results show that the neural network has high prediction accuracy, and can be applied to the comprehensive modeling of the prototype and the corner environment.

Improving Mental Health Services: A 50-Year Journey from Randomized Experiments to Artificial Intelligence and Precision Mental Health

This conceptual paper describes the current state of mental health services, identifies critical problems, and suggests how to solve them. I focus on the potential contributions of artificial intelligence and precision mental health to improving mental health services. Toward that end, I draw upon my own research, which has changed over the last half century, to highlight the need to transform the way we conduct mental health services research. I identify exemplars from the emerging literature on artificial intelligence and precision approaches to treatment in which there is an attempt to personalize or fit the treatment to the client in order to produce more effective interventions.

Mobile Sensing in Substance Use Research: A Scoping Review

Background: Addictive disorders and substance use are significant health challenges worldwide, and relapse is a core component of addictive disorders. The dynamics surrounding relapse and especially the immediate period before it occurs is only partly understood, much due to difficulties collecting reliable and sufficient data from this narrow period. Mobile sensing has been an important way to improve data quality and enhance predictive capabilities for symptom worsening within physical and mental health care, but is less developed within substance use research. Methodology: This scoping review aimed to reviewing the currently available research on mobile sensing of substance use and relapse in substance use disorders. The search was conducted in January 2019 using PubMed and Web of Science. Results: Six articles were identified, all concerning subjects using alcohol. In the studies a range of mobile sensors and derived aggregated features were employed. Data collected through mobile sensing were predominantly used to make dichotomous inference on ongoing substance use or not and in some cases on the quantity of substance intake. Only one of the identified studies predicted later substance use. A range of statistical machine learning techniques was employed. Conclusions: The research on mobile sensing in this field remains scarce. The issues requiring further attention include more research on clinical populations in naturalistic settings, use of a priori knowledge in statistical modeling, focus on prediction of substance use rather than purely identification, and finally research on other substances than alcohol.

On Smartphone Sensability of Bi-Phasic User Intoxication Levels from Diverse Walk Types in Standardized Field Sobriety Tests

Intoxicated driving causes 10,000 deaths annually. Smartphone sensing of user gait (walk) to identify intoxicated users in order to prevent drunk driving, have recently emerged. Such systems gather motion sensor (accelerometer and gyroscope) data from the users' smartphone as they walk and classify them using machine or deep learning. Standard Field Sobriety Tests (SFSTs) involve various types of walks designed to cause an intoxicated person to lose their balance. However, SFSTs were designed to make intoxication apparent to a trained law enforcement officer who manually proctors them. No prior work has explored which types of walk yields the most accurate results when assessed autonomously by a smartphone intoxicated gait assessment system. In this paper, we compare how accurately Long Short Term Memory (LSTM), Convolution Neural Network (CNN), Random Forest, Gradient Boosted Machines (GBM) and neural network classifiers are able to detect intoxication levels of drunk subjects who performed normal, walk-and-turn and standing on one foot SFST walks. We also compared the accuracy of intoxication detection on the ascending (increasing intoxication) vs descending (decreasing intoxication) limbs of drinking sessions (bi-phasic). We found smartphone intoxication sensing more accurate on the descending limb of the drinking episode and that intoxication detection on the normal walks of subjects were just as accurate as the SFSTs.

Looseness Monitoring of Bolted Spherical Joint Connection Using Electro-Mechanical Impedance Technique and BP Neural Networks

The bolted spherical joint (BSJ) has wide applications in various space grid structures. The bar and the bolted sphere are connected by the high-strength bolt inside the joint. High-strength bolt is invisible outside the joint, which causes the difficulty in monitoring the bolt looseness. Moreover, the bolt looseness leads to the reduction of the local stiffness and bearing capacity for the structure. In this regard, this study used the electro-mechanical impedance (EMI) technique and back propagation neural networks (BPNNs) to monitor the bolt looseness inside the BSJ. Therefore, a space grid specimen having bolted spherical joints and tubular bars was considered for experimental evaluation. Different torques levels were applied on the sleeve to represent different looseness degrees of joint connection. As the torque levels increased, the looseness degrees of joint connection increased correspondingly. The lead zirconate titanate (PZT) patch was used and integrated with the tubular bar due to its strong piezoelectric effect. The root-mean-square deviation (RMSD) of the conductance signatures for the PZT patch were used as the looseness-monitoring indexes. Taking RMSD values of sub-frequency bands and the looseness degrees as inputs and outputs respectively, the BPNNs were trained and tested in twenty repeated experiments. The experimental results show that the formation of the bolt looseness can be detected according to the changes of looseness-monitoring indexes, and the degree of bolt looseness by the trained BPNNs. Overall, this research demonstrates that the proposed structural health monitoring (SHM) technique is feasible for monitoring the looseness of bolted spherical connection in space grid structures.

Assessment of Alcohol Use in the Natural Environment

The current article critically reviews 3 methodological options for assessing drinking episodes in the natural environment. Ecological momentary assessment (EMA) typically involves using mobile devices to collect self-report data from participants in daily life. This technique is now widely used in alcohol research, but investigators have implemented diverse assessment strategies. This article focuses on "high-resolution" EMA protocols that oversample experiences and behaviors within individual drinking episodes. A number of approaches have been used to accomplish this, including using signaled follow-ups tied to drinking initiation, asking participants to log entries before and after individual drinks or drinking episodes, and delivering frequent signaled assessments during periods of the day when alcohol use is most common. Transdermal alcohol sensors (TAS) are devices that are worn continuously and are capable of detecting alcohol eliminated through the skin. These methods are appealing because they do not rely upon drinkers' self-report. Studies using TAS have been appearing with greater frequency over the past several years. New methods are making the use of TAS more tractable by permitting back-translation of transdermal alcohol concentration data to more familiar estimates of blood alcohol concentration or breath alcohol concentration. However, the current generation of devices can have problems with missing data and tend to be relatively insensitive to low-level drinking. An emerging area of research investigates the possibility of using mobile device data and machine learning to passively detect the user's drinking, with promising early findings. EMA, TAS, and sensor-based approaches are all valid, and tend to produce convergent information when used in conjunction with one another. Each has a unique profile of advantages, disadvantages, and threats to validity. Therefore, the nature of the underlying research question must dictate the method(s) investigators select.

Methods to reduce the incidence of false negative trial results in substance use treatment research

Treatment development and evaluation for substance use disorders are hindered when randomized controlled trials fail to show a treatment effect when one exists. This manuscript provides an overview of addressable methodological factors that may contribute to incorrect trial results. The collection of remote, naturalistic, real-time adherence and substance use data through ambulatory assessment methods in everyday life is presented as a partial solution. Other recommendations related to participant recruitment and selection, ensuring adequate consistency/fidelity and dose of treatment, and rigorously assessing clinical outcomes are discussed. With implementation of eligibility criteria verification, treatment adherence monitoring, and remote assessment of substance use and biomarkers, ambulatory assessment may help improve clinical trial success rates by improving precision, increasing reproducibility, and reducing the impact of methodological issues that may lead to inaccurate trial results.

Levenberg-Marquardt Neural Network Algorithm for Degree of Arteriovenous Fistula Stenosis Classification Using a Dual Optical Photoplethysmography Sensor

This paper proposes a noninvasive dual optical photoplethysmography (PPG) sensor to classify the degree of arteriovenous fistula (AVF) stenosis in hemodialysis (HD) patients. Dual PPG measurement node (DPMN) becomes the primary tool in this work for detecting abnormal narrowing vessel simultaneously in multi-beds monitoring patients. The mean and variance of Rising Slope (RS) and Falling Slope (FS) values between before and after HD treatment was used as the major features to classify AVF stenosis. Multilayer perceptron neural networks (MLPN) training algorithms are implemented for this analysis, which are the Levenberg-Marquardt, Scaled Conjugate Gradient, and Resilient Back-propagation, to identify the degree of HD patient stenosis. Eleven patients were recruited with mean age of 77 ± 10.8 years for analysis. The experimental results indicated that the variance of RS in the HD hand between before and after treatment was significant difference statistically to stenosis (p < 0.05). Levenberg-Marquardt algorithm (LMA) was significantly outperforms the other training algorithm. The classification accuracy and precision reached 94.82% and 92.22% respectively, thus this technique has a potential contribution to the early identification of stenosis for a medical diagnostic support system.