A computational model for identifying behavioral patterns in people with neuropsychiatric disorders

Neuropsychiatric disorders represent a great economic and social burden, whose estimates range from 14% to 32% of the global burden of diseases. One in four people suffer from mental disorders during their lifetime and the estimates of people suffering from depression and some type of neuropsychiatric disorder are about 322 million and 420 million, respectively. Therefore, this paper presents Eigenroutines, a model based on the hypothesis that the ubiquitous availability of mobile devices and the massive usage of these devices have created an opportunity of identifying behavior, supporting the management of neuropsychiatric disorders. The model is also capable of building psychiatric epidemiological profiles, using context histories. The model evaluation occurred in three steps, where the first confirmed that the model has the ability to identify and distinguish neuropsychiatric disorders, analyzing context histories from a group of 55 individuals, composed of 23 people with unipolar or bipolar depression and 32 healthy people. The model presented an accuracy of 78% for classifying the routine of an individual in depressed or healthy. The second step evaluated the extraction of epidemiological profiles based on context histories from 1066 individuals stored by the platform. Finally, the last step evaluated the limitations of the model regarding temporal complexity and scalability.