Estimation of the Composition of the Resting State fMRI Networks in Subjects with Mild Depression and Healthy Volunteers

Depressive disorders can be associated with changes in not only interaction between neural networks, but also in their composition. Resting state fMRI scanning was performed for 4 min twice for each subject and the results of patients with mild depression (N=15) and healthy subjects (N=19) were analyzed. The fMRI signal was reduced into the independent components and the contrasts between the groups and between the first and second records were constructed for each component. During the first scanning, the auditory network of individuals with depression involved greater volume in the left insular region and lower volume in the right hemisphere. In record 2, depression patients were characterized by expansion of the executive network in the left hemisphere in the region of the middle and inferior frontal cortex. In healthy people, from record 1 to record 2, representation of the dorsal default mode network (DMN) increased in the left medial prefrontal area, the precuneus network expanded in the left hemisphere, and presentation of the ventral DMN in the right precuneus decreased. In the depression group, the auditory network lost some part of the left temporo-insular cortex; the sensorimotor network expanded in the left hemisphere to the cerebellum or to the central parietal region depending on the evaluation method, and the visuospatial network included or excluded a cluster in the left parietal lobe (in different points). Our findings indicate that connection of the auditory network with the left insular cortex could be a possible depression marker and also demonstrate a possibility of evaluating the composition of cerebral networks in intergroup comparisons and in dynamics without interventions.