Search

Research on automated mental health assessment tools has been growing in recent years, often aiming to address the subjectivity and bias that existed in the current clinical practice of the psychiatric evaluation process. Despite the substantial health and economic ramifications, the potential unfairness of those automated tools was understudied and required more attention. In this work, we systematically evaluated the fairness level in a multimodal remote mental health dataset and an assessment system, where we compared the fairness level in race, gender, education level, and age. Demographic parity ratio (DPR) and equalized odds ratio (EOR) of classifiers using different modalities were compared, along with the F1 scores in different demographic groups. Post-training classifier threshold optimization was employed to mitigate the unfairness. No statistically significant unfairness was found in the composition of the dataset. Varying degrees of unfairness were identified among modalities, with no single modality consistently demonstrating better fairness across all demographic variables. Post-training mitigation effectively improved both DPR and EOR metrics at the expense of a decrease in F1 scores. Addressing and mitigating unfairness in these automated tools are essential steps in fostering trust among clinicians, gaining deeper insights into their use cases, and facilitating their appropriate utilization.

Abstract A comprehensive understanding of collocated social interactions can help campuses and organizations better support their community. Universities could determine new ways to conduct classes and design programs by studying how students have collocated in the past. However, this needs data that describe large groups over a long period. Harnessing user devices to infer collocation, while tempting, is challenged by privacy concerns, power consumption, and maintenance issues. Alternatively, embedding new sensors across the entire campus is expensive. Instead, we investigate an easily accessible data source that can retroactively depict multiple users on campus over a semester, a managed WiFi network. Despite the coarse approximations of collocation provided by WiFi network logs, we demonstrate that leveraging such data can express meaningful outcomes of collocated social interaction. Since a known outcome of collocating with peers is improved performance, we inspected if automatically–inferred collocation behaviors can indicate the individual performance of project group members on a campus. We studied 163 students (in 54 project groups) over 14 weeks. After describing how we determine collocation with the WiFi logs, we present a study to analyze how collocation within groups relates to a student’s final score. We found that modeling collocation behaviors showed a significant correlation (Pearson’s r = 0.24 $r =0.24$ ) with performance (better than models of peer feedback or individual behaviors). These findings emphasize that it is feasible and valuable to characterize collocated social interactions with archived WiFi network logs. We conclude the paper with a discussion of applications for repurposing WiFi logs to describe collocation, along with privacy considerations, and directions for future work.

Tremor, defined as an “involuntary, rhythmic, oscillatory movement of a body part”, is a key feature of many neurological conditions including Parkinson’s disease and essential tremor. Clinical assessment continues to be performed by visual observation with quantification on clinical scales. Methodologies for objectively quantifying tremor are promising but remain non-standardized across centers. Our center performs full-body behavioral testing with 3D motion capture for clinical and research purposes in patients with Parkinson’s disease, essential tremor, and other conditions. The objective of this study was to assess the ability of several candidate processing pipelines to identify the presence or absence of tremor in kinematic data from patients with confirmed movement disorders and compare them to expert ratings from movement disorders specialists. We curated a database of 2272 separate kinematic data recordings from our center, each of which was contemporaneously annotated as tremor present or absent by a movement physician. We compared the ability of six separate processing pipelines to recreate clinician ratings based on F1 score, in addition to accuracy, precision, and recall. The performance across algorithms was generally comparable. The average F1 score was 0.84±0.02 (mean ± SD; range 0.81–0.87). The second highest performing algorithm (cross-validated F1=0.87) was a hybrid that used engineered features adapted from an algorithm in longstanding clinical use with a modern Support Vector Machine classifier. Taken together, our results suggest the potential to update legacy clinical decision support systems to incorporate modern machine learning classifiers to create better-performing tools.

Spatial navigation patterns in indoor space usage can reveal important cues about the cognitive health of participants. In this work, we present a low-cost, scalable, open-source edge computing system using Bluetooth low energy (BLE) beacons for tracking indoor movements in a large, 1700 m2 facility used to carry out therapeutic activities for participants with mild cognitive impairment (MCI). The facility is instrumented with 39 edge computing systems, along with an on-premise fog server. The participants carry a BLE beacon, in which BLE signals are received and analyzed by the edge computing systems. Edge computing systems are sparsely distributed in the wide, complex indoor space, challenging the standard trilateration technique for localizing subjects, which assumes a dense installation of BLE beacons. We propose a graph trilateration approach that considers the temporal density of hits from the BLE beacon to surrounding edge devices to handle the inconsistent coverage of edge devices. This proposed method helps us tackle the varying signal strength, which leads to intermittent detection of beacons. The proposed method can pinpoint the positions of multiple participants with an average error of 4.4 m and over 85% accuracy in region-level localization across the entire study area. Our experimental results, evaluated in a clinical environment, suggest that an ordinary medical facility can be transformed into a smart space that enables automatic assessment of individuals’ movements, which may reflect health status or response to treatment.

Characterizing motor subtypes of Parkinson’s disease (PD) is an important aspect of clinical care that is useful for prognosis and medical management. Although all PD cases involve the loss of dopaminergic neurons in the brain, individual cases may present with different combinations of motor signs, which may indicate differences in underlying pathology and potential response to treatment. However, the conventional method for distinguishing PD motor subtypes involves resource-intensive physical examination by a movement disorders specialist. Moreover, the standardized rating scales for PD rely on subjective observation, which requires specialized training and unavoidable inter-rater variability. In this work, we propose a system that uses machine learning models to automatically and objectively identify some PD motor subtypes, specifically Tremor-Dominant (TD) and Postural Instability and Gait Difficulty (PIGD), from 3D kinematic data recorded during walking tasks for patients with PD (MDS-UPDRS-III Score, 34.7 ± 10.5, average disease duration 7.5 ± 4.5 years). This study demonstrates a machine learning model utilizing kinematic data that identifies PD motor subtypes with a 79.6% F1 score (N = 55 patients with parkinsonism). This significantly outperformed a comparison model using classification based on gait features (19.8% F1 score). Variants of our model trained to individual patients achieved a 95.4% F1 score. This analysis revealed that both temporal, spectral, and statistical features from lower body movements are helpful in distinguishing motor subtypes. Automatically assessing PD motor subtypes simply from walking may reduce the time and resources required from specialists, thereby improving patient care for PD treatments. Furthermore, this system can provide objective assessments to track the changes in PD motor subtypes over time to implement and modify appropriate treatment plans for individual patients as needed.

Freezing of gait (FOG) is a poorly understood heterogeneous gait disorder seen in patients with parkinsonism which contributes to significant morbidity and social isolation. FOG is currently measured with scales that are typically performed by movement disorders specialists (i.e., MDS-UPDRS), or through patient completed questionnaires (N-FOG-Q) both of which are inadequate in addressing the heterogeneous nature of the disorder and are unsuitable for use in clinical trials The purpose of this study was to devise a method to measure FOG objectively, hence improving our ability to identify it and accurately evaluate new therapies. A major innovation of our study is that it is the first study of its kind that uses the largest sample size (>30 h, N = 57) in order to apply explainable, multi-task deep learning models for quantifying FOG over the course of the medication cycle and at varying levels of parkinsonism severity. We trained interpretable deep learning models with multi-task learning to simultaneously score FOG (cross-validated F1 score 97.6%), identify medication state (OFF vs. ON levodopa; cross-validated F1 score 96.8%), and measure total PD severity (MDS-UPDRS-III score prediction error ≤ 2.7 points) using kinematic data of a well-characterized sample of N = 57 patients during levodopa challenge tests. The proposed model was able to explain how kinematic movements are associated with each FOG severity level that were highly consistent with the features, in which movement disorders specialists are trained to identify as characteristics of freezing. Overall, we demonstrate that deep learning models’ capability to capture complex movement patterns in kinematic data can automatically and objectively score FOG with high accuracy. These models have the potential to discover novel kinematic biomarkers for FOG that can be used for hypothesis generation and potentially as clinical trial outcome measures.

Supervised training of human activity recognition (HAR) systems based on body-worn inertial measurement units (IMUs) is often constrained by the typically rather small amounts of labeled sample data. Systems like IMUTube have been introduced that employ cross-modality transfer approaches to convert videos of activities of interest into virtual IMU data. We demonstrate for the first time how such large-scale virtual IMU datasets can be used to train HAR systems that are substantially more complex than the state-of-the-art. Complexity is thereby represented by the number of model parameters that can be trained robustly. Our models contain components that are dedicated to capture the essentials of IMU data as they are of relevance for activity recognition, which increased the number of trainable parameters by a factor of 1100 compared to state-of-the-art model architectures. We evaluate the new model architecture on the challenging task of analyzing free-weight gym exercises, specifically on classifying 13 dumbbell execises. We have collected around 41 h of virtual IMU data using IMUTube from exercise videos available from YouTube. The proposed model is trained with the large amount of virtual IMU data and calibrated with a mere 36 min of real IMU data. The trained model was evaluated on a real IMU dataset and we demonstrate the substantial performance improvements of 20% absolute F1 score compared to the state-of-the-art convolutional models in HAR.