A novel approach to the study of spatio-temporal brain dynamics using change-based complexity

Abstract Brain complexity and neural oscillations are of vital significance for understanding brain dynamics. Although widely used in the quantification of complex behavior of the brain, standard complexity measures are hampered by some theoretical and implementation issues which limit their applicability to the study of nonlinear brain signals. Further, they have been used for spatial analysis but are insensitive to the temporal structure of data and not applicable to short EEG segments. We demonstrated the ability of a change-based complexity measure (Aksentijevic-Gibson complexity) to analyze neural activity in both spatial and temporal domains. We examined EEG recordings from a group of adolescents with schizophrenia and an age-matched group of healthy controls. AG revealed that the spatial complexity was significantly lower in the patient group across all electrodes with the lowest complexity at the frontal area of the brain. By focusing on outer frontal regions and temporal data averaged over schizophrenia adolescents, we observed periodic drops in complexity and low-delta oscillations at the left electrode F7 and numerous irregularly placed drops at the right side (F8). Periodicity was caused by zero complexity troughs in several individuals associated with large voltage declensions and connected with hypofrontality supporting left frontal pathology. Both averaged spatial and temporal complexities of healthy and schizophrenia adolescents were significantly different from the complexities of random sequences confirming that neural activity is more regular than a true random process. As a final step, we performed the analysis of short EEG segments for the two subjects (one control, one patient) in order to search for differences in neural oscillations. The examination revealed a clear 33 Hz gamma rhythm in a healthy participant and only a fast beta rhythm of 27 Hz in a patient as well as the presence of a full range of rhythms in a healthy EEG and only theta oscillations in that of a patient. The results support the large body of work which associates schizophrenia with lowered complexity, low-frequency rhythms and the scarcity of high-frequency oscillations.