Title : Synapticare: Integrating sleep data and tau biomarkers to assess depression severity
Abstract:
After a year of treatment, only 11% of Major Depressive Disorder (MDD) patients achieved remission. This low recovery rate is not simply due to a lack of effective medications, as there are over 20 FDA approved, effective treatments for MDD. The challenge lies in identifying the best treatment for each individual patient. According to the American Psychiatric Association, a "watchful waiting" approach is recommended to assess whether a particular medication will work for a patient. Currently, depression treatment involves a "trial and error" approach, in which various medications or therapies are tried without a clear understanding of which will work best for the patient. This research focuses on the relationship between various sleep metrics, tau protein deposits, and depressive symptoms. The study aimed to develop a predictive model for the change in depression scores based on tau protein levels, sleep patterns, and depression severity, as measured by PHQ-9 scores. Neuroimaging data were used to measure tau protein levels across brain regions implicated in mood regulation and neurodegeneration. Depression severity was assessed using PHQ-9 scores, while sleep metrics were obtained through actigraphy over a 10-week period. For model development, changes in PHQ-9 scores served as the target variable. Features included tau deposition, baseline depression severity, treatment type, dosage, treatment duration, and sleep metrics. Performance was evaluated using R² and mean squared error (MSE). The Gradient-Boosting model demonstrated strong predictive performance. Results indicate that both tau pathology and sleep quality significantly influence changes in depression severity. These findings suggest that integrating tau- and sleep-based biomarkers into predictive frameworks may improve treatment personalization and clinical outcomes in MDD.
