Abstract
Major Depressive Disorder (MDD) is one of the most common psychological disorders. The multiplicity of its clinical patterns, the varieties of symptoms and the different types of clinical evolution generate many diagnostic difficulties. Currently, MDD assessment is performed through the use of assessment scales and interviews with the patient. This highlights the need to adopt methods that can make an objective, rapid and effective diagnosis. Mental disorders and embodiment can alter the brain processes that are related to cognition and therefore to the motor system. As a result, movement analysis by means of wearable sensors is attracting the attention of clinicians as it represents a solution that can be quickly translated into ecological environments. In this study we explored the potential of an instrumented movement assessment, targeting the long-term goal of self-administered assessment in ecological settings. Using 4 inertial measurement units (IMUs), we tested parameters that could be predictive of pathology during a timed up and go test. By means of age/sex adjusted logistic regression we identified instrumental parameters significantly discriminating MDD patients and controls. Building on earlier findings in literature for pathology recognition in movement, a particular attention was devoted to parameters concerning movement complexity evaluated by multiscale entropy analysis of signals. Our tests confirmed that complexity parameters can significantly discriminate between patients and controls (Nagelkerke’s R2 = 0.523).