Background and objective: Knowing whether a subject is conscious or not is a current challenge with a deep potential clinical impact. Recent theoretical considerations suggest that consciousness is linked to the complexity of distributed interactions within the corticothalamic system. The fractal dimension (FD) is a quantitative parameter that has been extensively used to analyse the complexity of structural and functional patterns of the human brain. In this study we investigate FD to assess whether it can discriminate between consciousness and different states of unconsciousness in healthy individuals. Methods: We study 69 high-density electroencephalogram (hd-EEG) measurements after transcranial magnetic stimulation (TMS) in 18 healthy subjects progressing from wakefulness to non-rapid eye movement (NREM) sleep and sedation induced by different anaesthetic agents (xenon and propofol). We quantify the integration of thalamocortical networks by calculating the FD of a spatiotemporal voxelization obtained from the locations of all sources that are significantly activated by the perturbation (4DFD). Moreover, we study the temporal evolution of the evoked spatial distributions and compute a measure of the differentiation of the response by means of the Higuchi FD (HFD). Finally, a Fractal Dimension Index (FDI) of perturbational complexity is computed as the product of both quantities: integration FD (4DFD) and differentiation FD (HFD). Results: We found that FDI is significantly lower in sleep and sedation when compared to wakefulness and provides an almost perfect intra-subject discrimination between conscious and unconscious states. Conclusions: These results support the combination of FD measures of cortical integration and cortical differentiation as a novel paradigm of tracking complex spatiotemporal dynamics in the brain that could provide further insights into the link between complexity and the brain's capacity to sustain consciousness.
- Fractal dimension
- Transcranial magnetic stimulation
ASJC Scopus subject areas
- Computer Science Applications
- Health Informatics