Theoretical approaches to the diagnosis of altered states of consciousness

Assessing the level of consciousness of noncommunicative brain-damaged patients is difficult, as one has to make inferences based on the patients' behavior. However, behavioral responses of brain-damaged patients are usually limited not only by their cognitive dysfunctions, but also by their frequent motor impairment. For these reasons, it is essential to resort to para-clinical markers of the level of consciousness. In recent years, a number of studies compared brain activity in comatose and vegetative state patients to that in healthy volunteers, and in other conditions of reduced consciousness such as sleep, anesthesia, or epileptic seizures. Despite the increasing amount of experimental results, no consensus on the brain mechanisms generating consciousness has yet been reached. Here, we discuss the need to combine a theoretical approach with current experimental procedures to obtain a coherent, parsimonious explanation for the loss of consciousness in several different conditions, such as coma, vegetative state, sleep, anesthesia, and epileptic seizures. In our view, without a theoretical account of how conscious experience is generated by the brain, it will remain difficult to understand the mechanisms underlying the generation of consciousness, and to predict reliably its presence or absence in noncommunicative brain-damaged patients. In this context, we review current theoretical approaches to consciousness, and how well they fit with current evidence on the neural correlates of experience. Specifically, we emphasize the principled approach provided by the Integrated Information Theory of Consciousness (IITC). We describe the different conditions where the theory predicts markedly reduced states of consciousness, and discuss several technical and conceptual issues limiting its applicability to measuring the level of consciousness of individual patients. Nevertheless, we argue that some of the predictions of the theory are potentially testable using available imaging techniques.