The pressures faced by college students frequently lead to heightened levels of stress. Wearable devices, which collect sensor data in a non-intrusive manner, present an opportunity for early detection of stress. Nonetheless, there is a lack of diversity in current research concerning psychological assessments, physiological metrics, and time series features. In this work, we utilize a Fitbit dataset and evaluate its use in predicting stress through machine learning. Our results demonstrate that physiological data such as calories burned and sleep hold promise for stress screening, with F1 scores reaching up to 0.81. These findings illustrate the potential of wearable technology for continuous stress monitoring and emphasize the need for selecting appropriate data aggregation levels and physiological modalities for effective screening.

Major depressive disorder (MDD) and post-traumatic stress disorder (PTSD) are mental disorders that reduce quality of life. As they are challenging to detect in a timely manner, recent studies explore the mental illness screening potential of language models on modalities such as transcripts. Such datasets suffer from a limited number of participants. To overcome these challenges, we take a two-pronged approach: (1) we leverage multi-task learning to model multiple illnesses concurrently namely MDD and PTSD screening, and (2) we plug in pre-trained language models as a backbone, namely, Bidirectional Encoder Representations from Transformers (BERT) variants – which are capable of learning the linguistic content of clinical interview transcripts. In particular, we experiment with three multi-task weighting strategies and five BERT variants, applying them to 15 transcript sets extracted from the Distress Analysis Interview Corpus. Our results indicate that leveraging multi-task learning, especially with meta-weighting, increases the screening performance compared to single-task learning. Our multi-task learning model strategies improved the F1 scores for all 15 datasets for both MDD and PTSD screening. Notably, multitask learning improved MDD screening ability by 20 percent with transcripts regarding ‘regret’ and achieved F1 of 0.89 and 0.82 respectively for MDD and PTSD screening with transcripts regarding ‘medical history’. Our findings may help researchers develop more effective mental illness screening models.

Major depressive disorder (MDD) and post-traumatic stress disorder (PTSD) are prevalent mental health conditions with severe physical and social impacts. They are expensive and detection is difficult, requiring substantial time from trained mental professionals. To alleviate this issue, recent studies explore the diagnostic potential of deep learning models trained on modalities extracted from clinical interview videos, conducted by a virtual agent. However, deep learning models are challenging to train because of the long sequences and small number of participants common in the mental health community. To combat these challenges, we leverage multi-task learning, using temporal facial features as input, to screen for MDD and PTSD. The multi-task architecture is based on a bidirectional GRU model with self-attention. We evaluate our multi-task model on temporal facial features extracted from responses to 15 clinical interview questions conducted by a virtual agent. The results suggest that multi-task learning increases the generalization performance compared to single-task learning. For MDD screening, multi-task learning improved the balanced accuracy over single-task learning for 11 of the 15 datasets. In fact, our multi-learning model increased the MDD screening ability by 25 percent to a balanced accuracy of 0.87 in some scenarios. This work provides valuable findings for the future of mental screening applications leveraging temporal facial features.

Corpuses of unstructured textual data, such as text messages between individuals, are often predictive of medical issues such as depression. The text data usually used in healthcare applications has high value and great variety, but is typically small in volume. Generating labeled unstructured text data is important to improve models by augmenting these small datasets, as well as to facilitate anonymization. While methods for labeled data generation exist, not all of them generalize well to small datasets. In this work, we thus perform a much needed systematic comparison of conditional text generation models that are promising for small datasets due to their unified architectures. We identify and implement a family of nine conditional sequence generative adversarial networks for text generation, which we collectively refer to as cSeqGAN models. These models are characterized along two orthogonal design dimensions: weighting strategies and feedback mechanisms. We conduct a comparative study evaluating the generation ability of the nine cSeqGAN models on three diverse text datasets with depression and sentiment labels. To assess the quality and realism of the generated text, we use standard machine learning metrics as well as human assessment via a user study. While the unconditioned models produced predictive text, the cSeqGAN models produced more realistic text. Our comparative study lays a solid foundation and provides important insights for further text generation research, particularly for the small datasets common within the healthcare domain.

Depression is a leading cause of disability with tremendous socioeconomic costs. In spite of early detection being crucial to improving prognosis, this mental illness remains largely undiagnosed. Depression classification from voice holds the promise to revolutionize diagnosis by ubiquitously integrating this screening capability into virtual assistants and smartphone technologies. Unfortunately, due to privacy concerns, audio datasets with depression labels have a small number of participants, causing current classification models to suffer from low performance. To tackle this challenge, we introduce Audio-Assisted BERT (AudiBERT), a novel deep learning framework that leverages the multimodal nature of human voice. To alleviate the small data problem, AudiBERT integrates pretrained audio and text representation models for the respective modalities augmented by a dual self-attention mechanism into a deep learning architecture. AudiBERT applied to depression classification consistently achieves promising performance with an increase in F1 scores between 6% and 30% compared to state-of-the-art audio and text models for 15 thematic question datasets. Using answers from medically targeted and general wellness questions, our framework achieves F1 scores of up to 0.92 and 0.86, respectively, demonstrating the feasibility of depression screening from informal dialogue using voice-enabled technologies.