Conversational assessment using artificial intelligence is as clinically useful as depression scales and preferred by users.

BACKGROUND: Depression is prevalent, chronic, and burdensome. Due to limited screening access, depression often remains undiagnosed. Artificial intelligence (AI) models based on spoken responses to interview questions may offer an effective, efficient alternative to other screening methods. OBJECTIVE: The primary aim was to use a demographically diverse sample to validate an AI model, previously trained on human-administered interviews, on novel bot-administered interviews, and to check for algorithmic biases related to age, sex, race, and ethnicity. METHODS: Using the Aiberry app, adults recruited via social media (N = 393) completed a brief bot-administered interview and a depression self-report form. An AI model was used to predict form scores based on interview responses alone. For all meaningful discrepancies between model inference and form score, clinicians performed a masked review to determine which one they preferred. RESULTS: There was strong concurrent validity between the model predictions and raw self-report scores (r = 0.73, MAE = 3.3). 90 % of AI predictions either agreed with self-report or with clinical expert opinion when AI contradicted self-report. There was no differential model performance across age, sex, race, or ethnicity. LIMITATIONS: Limitations include access restrictions (English-speaking ability and access to smartphone or computer with broadband internet) and potential self-selection of participants more favorably predisposed toward AI technology. CONCLUSION: The Aiberry model made accurate predictions of depression severity based on remotely collected spoken responses to a bot-administered interview. This study shows promising results for the use of AI as a mental health screening tool on par with self-report measures.

Inclusion of genetic variants in an ensemble of gradient boosting decision trees does not improve the prediction of citalopram treatment response.

Identifying in advance who is unlikely to respond to a specific antidepressant treatment is crucial to precision medicine efforts. The current work leverages genome-wide genetic variation and machine learning to predict response to the antidepressant citalopram using data from the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) trial (n = 1257 with both valid genomic and outcome data). A confirmatory approach selected 11 SNPs previously reported to predict response to escitalopram in a sample different from the current study. A novel exploratory approach selected SNPs from across the genome using nested cross-validation with elastic net logistic regression with a predominantly lasso penalty (alpha = 0.99). SNPs from each approach were combined with baseline clinical predictors and treatment response outcomes were predicted using a stacked ensemble of gradient boosting decision trees. Using pre-treatment clinical and symptom predictors only, out-of-fold prediction of a novel treatment response definition based on STAR*D treatment guidelines was acceptable, AUC = .659, 95% CI [0.629, 0.689]. The inclusion of SNPs using confirmatory or exploratory selection methods did not improve the out-of-fold prediction of treatment response (AUCs were .662, 95% CI [0.632, 0.692] and .655, 95% CI [0.625, 0.685], respectively). A similar pattern of results were observed for the secondary outcomes of the presence or absence of distressing side effects regardless of treatment response and achieving remission or satisfactory partial response, assuming medication tolerance. In the current study, incorporating SNP variation into prognostic models did not enhance the prediction of citalopram response in the STAR*D sample.

Mood-Reactive Self-Esteem and Depression Vulnerability: Person-Specific Symptom Dynamics via Smart Phone Assessment.

Cognitive theories of depression suggest that mood-reactive self-esteem, a pattern of cognitive reactivity where low self-esteem is temporally dependent on levels of sadness, represents vulnerability for depression. Few studies have directly tested this hypothesis, particularly using intensive data collection methods (i.e., experience sampling) required to capture the temporal dynamics of sadness and self-esteem as they unfold naturally, over time. In this study we used participants' smartphones to collect multiple daily ratings of sadness and self-esteem over three weeks, in the real world. We then applied dynamic factor modeling to explore theoretically driven hypotheses about the temporal dependency of self-esteem on sadness (i.e., mood-reactive self-esteem) and its relationship to indices of depression vulnerability both contemporaneously (e.g., rumination, sad mood persistence) and prospectively (e.g., future symptomatology). In sum, individuals who demonstrated mood-reactive self-esteem reported higher levels of rumination at baseline, more persistent sad mood over three weeks, and increased depression symptoms at the end of three weeks above and beyond a trait-like index of self-esteem. The integration of smartphone assessment and person-specific analytics employed in this study offers an exiting new avenue to advance the study and treatment of depression.

Genetic and hormonal sensitivity to threat: testing a serotonin transporter genotype × testosterone interaction.

BACKGROUND: Striking parallels are observed when comparing the literature on the 5-HTTLPR of the serotonin transporter gene (SLC6A4) to the testosterone (T) literature on measures of stress reactivity and neural activity. Short (S) allele carriers and individuals higher in testosterone levels show exaggerated stress responses, amygdala hyperactivity, and reduction of amygdala-prefrontal cortex coupling when exposed to threat. METHODS: Three studies tested the hypothesis that higher T, S carriers would show increased cortisol responses to threat. RESULTS: Supporting the hypothesis, a T × 5-HTTLPR interaction was obtained across all studies. Threats to status via social exclusion (Study 1), cognitive/perceptual failure (Study 2), and physical competence (Study 3) all produced elevated cortisol levels in S carriers with higher T levels. An unexpected result was that 5-HTTLPR long (L) allele homozygotes with higher T showed lower cortisol levels in response to threat-a pattern of response that closely parallels that reported for psychopathic individuals. Finally, combining effect sizes across studies showed that the likelihood that these effects were due to Type 1 errors was quite low. CONCLUSIONS: What emerges from these studies is a novel yet reliable, and synergistic relationship between 5-HTTLPR genotype and testosterone on stress reactivity, possibly conferring vulnerability for multiple neuropsychiatric disorders.

Major and minor depression in female adolescents: onset, course, symptom presentation, and demographic associations.

We examined the epidemiology and phenomenology of major depressive disorder (MDD) and minor depression among a community sample of 496 female adolescents. Diagnostic interviews were conducted annually for 7 years, allowing us to examine onset, course, and symptom presentation among participants 12 through 20 years old. Approximately 1 of 6 girls experienced MDD. MDD episodes had a mean duration of 5.3 months (SD=4.2). One-year prevalence for MDD peaked at age 16 (5.3%). White racial status and younger age were associated with greater worthlessness and suicidality during an MDD episode. One of 5 girls met criteria for minor depression. Adolescents from racial/ethnic minority groups were at especially high risk for minor depression. Adolescence is a high-risk period for depression in young women, although its prevalence and phenomenology vary as a function of age and race/ethnicity.

Family functioning is associated with depressive symptoms in caregivers of acute stroke survivors.

OBJECTIVE: To determine whether family functioning is uniquely associated with caregiver depressive symptoms in the immediate aftermath of stroke. DESIGN: Cross-sectional data from the baseline assessment of an intervention study for stroke survivors and their families. SETTING: Neurology inpatient service of a large urban hospital. PARTICIPANTS: Stroke survivors (n=192), each with a primary caregiver. The mean age of stroke survivors was 66 years, and most, 57%, were men (n=110). The mean age of caregivers was 57 years, and 73% (n=140) of the caregivers were women. Eighty-five percent of caregivers were white. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: Measures were chosen to assess caregivers' depressive symptoms (Centers for Epidemiologic Studies Depression Scale), family functioning (Family Assessment Device), and additional factors such as health status (Medical Outcomes Study 36-Item Short-Form Health Survey) and stroke survivors' cognitive abilities (modified Mini-Mental State Examination) and functional impairments (FIM and Frenchay Activities Index). RESULTS: Depressive symptoms were mild to moderate in 14% and severe in 27% of caregivers. Family functioning was assessed as unhealthy in 34% of caregiver-patient dyads. In statistical regression models, caregiver depression was associated with patients' sex, caregivers' general health, and family functioning. CONCLUSIONS: Forty-one percent of caregivers experienced prominent depressive symptoms after their family member's stroke. Higher depression severity in caregivers was associated with caring for a man, and having worse health and poor family functioning. After stroke, the assessment of caregivers' health and family functioning may help determine which caregivers are most at risk for a depressive syndrome.