Non-verbal behaviours as predictors for treatment response in patients with depression or schizophrenia: a systematic...

Non-verbal behaviours (NVBs) reveal information about mood and emotions, potentially providing an objective means to measure and monitor early treatment responses. Previous research has examined NVB changes during treatment in patients with depression and psychosis but a systematic evaluation of the evidence is lacking. Furthermore, this review could inform the fast moving field of digital and precision psychiatry due to the use of AI-based technology that could transform the potential of NVBs as reliable biomarkers for treatment response.

Medline, Embase, PsycINFO and CINAHL were searched in June 2024. Included papers studied adults diagnosed with depression or schizophrenia and measured NVBs at least twice during separate clinical interviews. Outcomes of interest were changes in clinical symptoms and NVBs following treatment initiation. Treatment strategies included hospitalisation, pharmacological, psychological, neuromodulatory, other non-standardised interventions, or a combination of approaches. Two reviewers independently extracted data and assessed risk of bias. The protocol for the review was registered on PROSPERO.

20 papers were identified; 15 on depression, two on schizophrenia and three evaluating both conditions separately. Methodological variations across studies made comparisons challenging. NVBs consistently associated with improvement in depression symptoms included: increased smiling, facial expressivity, and amplified head and body movements. Results across studies were more consistent when considering general categories of behaviour, versus discrete facial behaviours. No commonalities were observed in NVB changes over time for patients with schizophrenia.

The existing evidence is presently insufficient to establish distinct behavioural profiles for clinical improvement depression or schizophrenia. Despite implicit challenges, there is considerable future scope in the evaluation of NVBs as predictors of clinical outcomes or change.

https://www.crd.york.ac.uk/PROSPERO, identifier CRD42022368599.