Research methods

We focus on identifying digital biomarkers that provide insights into the clinical diagnosis of emotional disorders. Using data from passive monitoring (e.g., smartphone usage, wearables) and active assessments (e.g., self-reports, mood tracking apps), we combine machine learning with clinical data to uncover patterns that reflect the underlying mental health conditions.
This approach allows for a more accurate, timely, and non-invasive diagnosis of disorders such as depression, anxiety, and PTSD in older adults.

In addition to stand-alone interventions, we explore blended interventions that combine digital tools with traditional face-to-face therapy. Blended care models allow clinicians to monitor patients remotely, provide ongoing feedback, and deliver supplementary therapeutic exercises between sessions.
This approach bridges the gap between digital innovation and human interaction, improving engagement and outcomes in older adults who may need both digital and personal support.

Our work on prognostic biomarkers aims to predict the course of emotional disorders over time. By analyzing longitudinal data, we aim to anticipate symptom severity, relapse likelihood, and treatment response. Predictive modeling using advanced machine learning helps to identify individuals at higher risk for worsening symptoms or those who might benefit from specific interventions, allowing for earlier and more personalized therapeutic strategies.

We are developing stand-alone digital interventions that can be delivered without direct therapist involvement. These interventions are designed to be self-guided and accessible via smartphones or tablets, allowing older adults to engage with cognitive behavioral therapy (CBT), mindfulness, and mood regulation exercises independently. Our interventions are adapted to the cognitive and emotional needs of the elderly, focusing on user-friendliness and clinical efficacy.