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.

Correlates and predictors of the severity of suicidal ideation in adolescence: an examination of brain connectomics and psychosocial characteristics

BACKGROUND

Suicidal ideation (SI) typically emerges during adolescence but is challenging to predict. Given the potentially lethal consequences of SI, it is important to identify neurobiological and psychosocial variables explaining the severity of SI in adolescents.

METHODS

In 106 participants (59 female) recruited from the community, we assessed psychosocial characteristics and obtained resting-state fMRI data in early adolescence (baseline: aged 9-13 years). Across 250 brain regions, we assessed local graph theory-based properties of interconnectedness: local efficiency, eigenvector centrality, nodal degree, within-module z-score, and participation coefficient. Four years later (follow-up: ages 13-19 years), participants self-reported their SI severity. We used least absolute shrinkage and selection operator (LASSO) regressions to identify a linear combination of psychosocial and brain-based variables that best explain the severity of SI symptoms at follow-up. Nested-cross-validation yielded model performance statistics for all LASSO models.

RESULTS

A combination of psychosocial and brain-based variables explained subsequent severity of SI (R² = .55); the strongest was internalizing and externalizing symptom severity at follow-up. Follow-up LASSO regressions of psychosocial-only and brain-based-only variables indicated that psychosocial-only variables explained 55% of the variance in SI severity; in contrast, brain-based-only variables performed worse than the null model.

CONCLUSIONS

A linear combination of baseline and follow-up psychosocial variables best explained the severity of SI. Follow-up analyses indicated that graph theory resting-state metrics did not increase the prediction of the severity of SI in adolescents. Attending to internalizing and externalizing symptoms is important in early adolescence; resting-state connectivity properties other than local graph theory metrics might yield a stronger prediction of the severity of SI.

Study Protocol for Teen Inflammation Glutamate Emotion Research (TIGER)

This article provides an overview of the study protocol for the Teen Inflammation Glutamate Emotion Research (TIGER) project, a longitudinal study in which we plan to recruit 60 depressed adolescents (ages 13–18 years) and 30 psychiatrically healthy controls in order to examine the inflammatory and glutamatergic pathways that contribute to the recurrence of depression in adolescents. TIGER is the first study to examine the effects of peripheral inflammation on neurodevelopmental trajectories by assessing changes in cortical glutamate in depressed adolescents. Here, we describe the scientific rationale, design, and methods for the TIGER project. This article is intended to serve as an introduction to this project and to provide details for investigators who may be seeking to replicate or extend these methods for other related research endeavors.