Faculty Bibliography

Neuropsychiatric disorders pose a high societal cost, but their treatment is hindered by lack of objective outcomes and fidelity metrics. AI technologies and specifically Natural Language Processing (NLP) have emerged as tools to study mental health interventions (MHI) at the level of their constituent conversations. However, NLP's potential to address clinical and research challenges remains unclear. We therefore conducted a pre-registered systematic review of NLP-MHI studies using PRISMA guidelines (osf.io/s52jh) to evaluate their models, clinical applications, and to identify biases and gaps. Candidate studies (n = 19,756), including peer-reviewed AI conference manuscripts, were collected up to January 2023 through PubMed, PsycINFO, Scopus, Google Scholar, and ArXiv. A total of 102 articles were included to investigate their computational characteristics (NLP algorithms, audio features, machine learning pipelines, outcome metrics), clinical characteristics (clinical ground truths, study samples, clinical focus), and limitations. Results indicate a rapid growth of NLP MHI studies since 2019, characterized by increased sample sizes and use of large language models. Digital health platforms were the largest providers of MHI data. Ground truth for supervised learning models was based on clinician ratings (n = 31), patient self-report (n = 29) and annotations by raters (n = 26). Text-based features contributed more to model accuracy than audio markers. Patients' clinical presentation (n = 34), response to intervention (n = 11), intervention monitoring (n = 20), providers' characteristics (n = 12), relational dynamics (n = 14), and data preparation (n = 4) were commonly investigated clinical categories. Limitations of reviewed studies included lack of linguistic diversity, limited reproducibility, and population bias. A research framework is developed and validated (NLPxMHI) to assist computational and clinical researchers in addressing the remaining gaps in applying NLP to MHI, with the goal of improving clinical utility, data access, and fairness.

BACKGROUND:At-home Ketamine-assisted therapy (KAT) with psychosocial support and remote monitoring through telehealth platforms addresses access barriers, including the COVID-19 pandemic. Large-scale evaluation of this approach is needed for questions regarding safety and effectiveness for depression and anxiety. METHODS:In this prospective study, a large outpatient sample received KAT over four weeks through a telehealth provider. Symptoms were assessed using the Patient Health Questionnaire (PHQ-9) for depression, and the Generalized Anxiety Disorder scale (GAD-7) for anxiety. Demographics, adverse events, and patient-reported dissociation were also analyzed. Symptom trajectories were identified using Growth Mixture Modeling, along with outcome predictors. RESULTS:A sample of 1247 completed treatment with sufficient data, 62.8 % reported a 50 % or greater improvement on the PHQ-9, d = 1.61, and 62.9 % on the GAD-7, d = 1.56. Remission rates were 32.6 % for PHQ-9 and 31.3 % for GAD-7, with 0.9 % deteriorating on the PHQ-9, and 0.6 % on the GAD-7. Four patients left treatment early due to side effects or clinician disqualification, and two more due to adverse events. Three patient subpopulations emerged, characterized by Improvement (79.3 %), Chronic (11.4 %), and Delayed Improvement (9.3 %) for PHQ-9 and GAD-7. Endorsing side effects at Session 2 was associated with delayed symptom improvement, and Chronic patients were more likely than the other two groups to report dissociation at Session 4. CONCLUSION/CONCLUSIONS:At-home KAT response and remission rates indicated rapid and significant antidepressant and anxiolytic effects. Rates were consistent with laboratory- and clinic-administered ketamine treatment. Patient screening and remote monitoring maintained low levels of adverse events. Future research should assess durability of effects.

BACKGROUND:Somatic Symptom and Related Disorders (SSRD), including chronic pain, result in frequent primary care visits, depression and anxiety symptoms, and diminished quality of life. Treatment access remains limited due to structural barriers and functional impairment. Digital delivery offers to improve access and enables transcript analysis via Natural Language Processing (NLP) to inform treatment. Therefore, we investigated asynchronous message-delivered SSRD treatment, and used NLP methods to identify symptom reduction markers from emotional valence. METHODS:173 individuals diagnosed with SSRD received interventions from licensed therapists via messaging 5 days/week for 8 weeks. Depression and anxiety symptoms were measured with the PHQ-9 and GAD-7 from baseline every three weeks. Symptoms trajectories were identified using unsupervised random forest clustering. Emotional valence expressed and use of emotional words were extracted from patients' de-identified transcripts, respectively using VADER and NCR Lexicon. Valence differences were examined using logistic regression. RESULTS:Two subpopulations were identified showing symptoms Improvement (n = 72; 41.62 %) and non-response (n = 101; 58.38 %). Improvement patients expressed more positive valence in the first week of treatment (OR = 1.84, CI: 1.12-3.02; p = .015) and were less likely to express negative valence by the end of treatment (OR = 0.05; CI: 0.30-0.83; p = .008). Non-response patients used more negative valence words, including pain. LIMITATIONS/CONCLUSIONS:Findings were derived from observational data obtained during an ecological intervention, without the inclusion of a control group. CONCLUSIONS:NLP identified linguistic markers distinguishing changes in anxiety and depression symptoms over treatment. Digital interventions offer new forms of delivery and provide the opportunity to automatically collect data for linguistic analysis.

Background: Although network analysis studies of psychiatric syndromes have increased in recent years, most have emphasized centrality symptoms and robust edges. Broadening the focus to include bridge symptoms within a systematic review could help to elucidate symptoms having the strongest links in network models of psychiatric syndromes. We conducted this systematic review and statistical evaluation of network analyses on depressive and anxiety symptoms to identify the most central symptoms and bridge symptoms, as well as the most robust edge indices of networks. Methods: A systematic literature search was performed in PubMed, PsycINFO, Web of Science, and EMBASE databases from their inception to May 25, 2022. To determine the most influential symptoms and connections, we analyzed centrality and bridge centrality rankings and aggregated the most robust symptom connections into a summary network. After determining the most central symptoms and bridge symptoms across network models, heterogeneity across studies was examined using linear logistic regression. Results: Thirty-three studies with 78,721 participants were included in this systematic review. Seventeen studies with 23 cross-sectional networks based on the Patient Health Questionnaire (PHQ) and Generalized Anxiety Disorder (GAD-7) assessments of clinical and community samples were examined using centrality scores. Twelve cross-sectional networks based on the PHQ and GAD-7 assessments were examined using bridge centrality scores. We found substantial variability between study samples and network features. "˜Sad mood"™, "˜Uncontrollable worry"™, and "˜Worrying too much"™ were the most central symptoms, while "˜Sad mood"™, "˜Restlessness"™, and "˜Motor disturbance"™ were the most frequent bridge centrality symptoms. In addition, the connection between "˜Sleep"™ and "˜Fatigue"™ was the most frequent edge for the depressive and anxiety symptoms network model. Conclusion: Central symptoms, bridge symptoms and robust edges identified in this systematic review can be viewed as potential intervention targets. We also identified gaps in the literature and future directions for network analysis of comorbid depression and anxiety.

PURPOSE/OBJECTIVE:Breast cancer survivors may demonstrate elevated psychological distress, which can also hinder adherence to survivorship care plans. Our goal was to study heterogeneity of behavioral health and functioning in breast cancer survivors, and identify both risk and protective factors to improve targets for wellness interventions. METHODS:Breast cancer survivors (n = 187) consented to complete self-reported psychological measures and to access their medical records. Latent class analysis (LCA) was used to classify heterogeneous subpopulations based on levels of depression, post-traumatic stress, fear of cancer recurrence, cancer-related pain, and fatigue. Multinomial logistic regression and auxiliary analysis in a 3-step modeling conditional approach was used to identify characteristics of the group based on demographics, treatment history and characteristics, and current medication prescriptions. RESULTS:Three subpopulations of breast cancer survivors were identified from the LCA: a modal Resilient group (48.2%, n = 90), a Moderate Symptoms group (34%, n = 65), and an Elevated Symptoms group (n = 17%, n = 32) with clinically-relevant impairment. Results from the logistic regression indicated that individuals in the Elevated Symptoms group were less likely to have a family history of breast cancer; they were more likely to be closer to time of diagnosis and younger, have received chemotherapy and psychotropic prescriptions, and have higher BMI. Survivors in the Elevated Symptoms group were also less likely to be prescribed estrogen inhibitors than the Moderate Symptoms group. CONCLUSIONS:This study identified subgroups of breast cancer survivors based on behavioral, psychological, and treatment-related characteristics, with implications for targeted monitoring and survivorship care plans. IMPLICATIONS FOR CANCER SURVIVORS/CONCLUSIONS:Results showed the majority of cancer survivors were resilient, with minimal psychological distress. Results also suggest the importance of paying special attention to younger patients getting chemotherapy, especially those without a family history of breast cancer.

As many individuals experience potentially traumatic or stressful life events, understanding factors that are likely to promote resilience is imperative. Given the demonstrated efficacy of exercise for depression treatment, we examined if exercise buffers against the risk of developing psychiatric symptoms following life stressors. 1405 participants (61% female) from a longitudinal panel cohort experienced disability onset (43%), bereavement (26%), heart attack (20%), divorce (11%), and job loss (3%). They reported time spent exercising and depressive symptoms (Center for Epidemiologic Studies Depression scale) across three time points collected in two-year intervals: T0 (pre-stressor), T1 (acutely post-stressor), and T2 (post-stressor). Participants were classified in previously identified heterogeneous depression trajectories pre-to post-life stressor: resilient (69%), emerging (11.5%), chronic (10%), and improving (9.5%). Multinomial logistic regression found that more T0 exercise predicted likelihood of classification in resilient versus other groups (all p <.02). Controlling for covariates, only the higher likelihood of classification in resilient versus improving remained (p =.03). Follow-up repeated measures general linear model (GLM) assessed whether trajectory was associated with exercise at each time, controlling for covariates. GLM indicated significant within-subjects effects for time (p =.016, partial η² = 0.003) and time*trajectory (p =.020, partial η² = 0.005) on exercise and significant between-subjects effects of trajectory (p <.001, partial η² = 0.016) and all covariates. The resilient group showed consistent high exercise levels. The improving group had consistent moderate exercise. The emerging and chronic groups were associated with lower exercise post-stressor. Pre-stressor exercise may buffer against depression and ongoing exercise may be associated with lower depression levels following a major life stressor.

BACKGROUND:Evidence-based pharmacological treatments for posttraumatic stress disorder (PTSD) are few and of limited efficacy. Previous work suggests that angiotensin type 1 receptor inhibition facilitates fear inhibition and extinction, important for recovery from PTSD. This study tests the efficacy of the angiotensin type 1 receptor antagonist losartan, an antihypertensive drug, repurposed for the treatment of PTSD. METHODS:A randomized controlled trial was conducted for 10 weeks in 149 men and women meeting DSM-5 PTSD criteria. Losartan (vs. placebo) was flexibly titrated from 25 to 100 mg/day by week 6 and held at highest tolerated dose until week 10. Primary outcome was the Clinician-Administered PTSD Scale for DSM-5 (CAPS-5) change score at 10 weeks from baseline. A key secondary outcome was change in CAPS-5 associated with a single nucleotide polymorphism of the ACE gene. Additional secondary outcomes included changes in the PTSD Checklist for DSM-5 and the Patient Health Questionnaire-9, and proportion of responders with a Clinical Global Impressions-Improvement scale of "much improved" or "very much improved." RESULTS:Both groups had robust improvement in PTSD symptoms, but there was no significant difference on the primary end point, CAPS-5 measured as week 10 change from baseline, between losartan and placebo (mean change difference, 0.9, 95% confidence interval, -3.2 to 5.0). There was no significant difference in the proportion of Clinical Global Impressions-Improvement scale responders for losartan (58.6%) versus placebo (57.9%), no significant differences in changes in PTSD Checklist for DSM-5 or Patient Health Questionnaire-9, and no association between ACE genotype and CAPS-5 improvement on losartan. CONCLUSIONS:At these doses and durations, there was no significant benefit of losartan compared with placebo for the treatment of PTSD. We discuss implications for failure to determine the benefit of a repurposed drug with strong a priori expectations of success based on preclinical and epidemiological data.

The loss of a loved one is a potentially traumatic event that can result in disparate outcomes and symptom patterns. Machine learning methods offer computational tools to probe this heterogeneity and understand grief psychopathology in its complexity. In this article, we examine the latest contributions to the scientific study of bereavement reactions garnered through the use of computational methods. We focus on findings originating from trajectory modeling studies, as well as the recent insights originating from the network analysis of prolonged grief symptoms. We also discuss applications of artificial intelligence for the accurate identification of major depression and post-traumatic stress, as examples for their potential applications to the study of loss reactions.

PURPOSE/OBJECTIVE:Breast cancer survivors may be at risk for increased rates of emotional distress and poorer quality of life. Survivorship care plans (SCPs) promoting wellness activities may support well-being; however, survivors may not receive or engage in their SCPs. This study aimed to assess receipt and participation in SCP activities as well as barriers to engagement amongst breast cancer survivors. METHODS:Breast cancer survivors (n = 187; 99% female, Mean age = 57.7) consented and completed self-reported assessments of SCP recommendations, engagement and interest in wellness activities, and potential barriers to engagement. RESULTS:A minority of participants recalled receiving an SCP (21%). The most physician recommended (62%) and completed (53%) activity was exercise. Interest in adding other wellness activities to the SCP was high, with reported interest levels of approximately 50% for several activities (e.g., mind body, nutrition, psychotherapy interventions). Fully half reported that having a physician-designed plan would influence participation in activities. The most common reported barriers to SCP activity engagement were lack of time (82%), work/school (65%), and lack of information (65%). CONCLUSION/CONCLUSIONS:Few survivors recalled receiving a formal SCP, and lack of information about wellness activities was a commonly reported barrier to participation. Interest in wellness activities was generally high and may indicate the need for more formal prescription or motivation enhancement techniques to promote SCP engagement. There may be a clinical need to emphasize SCP recommendations to enhance recall and increase engagement in wellness activities that may reduce psychological distress and improve quality of life.