**11. Conclusions**

196 Neuroimaging – Cognitive and Clinical Neuroscience

A recent meta-analysis by Northoff et al. (2006) of 27 PET and fMRI studies comparing hemodynamic brain responses obtained during active paradigms comparing processing of stimuli related to the self with those of non-self-referential stimuli identified activation in cortical midline structures in all studies and occurring across all functional domains (e.g. verbal, spatial, emotional, and facial). Cluster and factor analyses indicated functional specialization into ventral, dorsal, and posterior cortical midline areas. The latter encompasses the posterior cingulate cortex and adjacent precuneus and is considered to be involved in self-integration – that is linkage of self-referential stimuli to the personal context (Northoff et al., 2004). Neuroimaging studies during tasks involving self-processing (i.e. selfreflection, self-perspective and free thoughts) have also reported the activation of the medial prefrontal areas. Gusnard, Akbudak, Shulman , and Raichle (2001) for example showed medial frontal activation when subjects had to make two judgments in response to pleasant *vs* unpleasant pictures (i.e. self-referential) as compared to indoors *vs* outdoors pictures (i.e. not self-referential). The same area was also shown to be engaged when subjects had to make self-referential judgments about trait adjectives (i.e. self-referential processing) as compared to when they had to make case judgments (Kelley et al., 2002) and when subjects responded to statements requiring knowledge of, and reflection on, their own abilities, traits and attitudes , i.e. self-reflective thought (Johnson et al., 2002). Taking a self– perspective (i.e. being the agent of an history) also activated medial prefrontal/anterior cingulate cortices (Vogeley et al., 2001). Finally, activation of the mesiofrontal areas was describes in studies dealing with the conscious resting state, i.e. free thought (Mazoyer et al., 2001), a brain state which "instantiates functions that are integral to the self". The recovery of consciousness of one VS patient has previously been linked to an increase in the functional connectivity within fronto-parietal network, (Laureys et al, 1999) encompassing the areas known to be most active in resting–state conditions (Gusnard et al., 2001). A growing body of evidence from Positron emission tomography (PET) and fMRI studies of healthy volunteers in a variety of altered states of consciousness has emphasized the role of this "default-mode" network in the genesis of awareness. In keeping with this, functional impairments to this network have been observed during sleepwalking, absence of seizures, deep sleep and anesthesia (Bassetti et al., 2000). fMRI has also proved its utility in identifying a number of cognitive functions which may be preserved in DOC patients, the results of which have, in some cases, proved prognostic of positive outcomes (Owen et al., 2008). In one such fMRI study investigating language processing. Coleman et. al found evidence of speech processing in three out of seven behaviorally non-communicative VS patients (Coleman et al., 2007). Six months after the scan, each of these patients had made a marked behavioral recovery relative to those patients who did not demonstrate comparable activations. Similar findings have also been reported for the neural responses observed when patients hear their own name (Di et al., 2007). Multimodal imaging approach can provide a powerful tool for assessing the mechanisms involved in the recovery of consciousness in DOC patients. Further longitudinal studies with large cohorts will prove useful in assessing its full value in predicting outcome. Such insights may then provide guidance for decisions relating to rehabilitation programs by those orientating these towards the effective stimulation of those functions that appear preserved, in order to

**10. Neuroimaging of self-consciousness and recovery** 

maintain their integrity.

Patients with severe brain damage who are unable to communicate present several ethical concerns. Foremost is the concern that diagnostic and prognostic accuracy is certain, as treatment decisions typically include the possibility of withdrawal of life-support. Although imaging techniques have the potential to improve both diagnostic and prognostic accuracy, careful and repeated neurological assessment by a trained examiner remains best practice. Accurate clinical assessments of patients in these conditions must be obtained before they undergo neuroimaging. Moreover, in reports of neuroimaging studies, all relevant clinical details must be available for comparisons between studies.

Ethical concerns are commonly raised about the participation of patients with severe brain damage in neuroimaging studies. By definition, unconscious or minimally conscious patients cannot give informed consent to participate in clinical research and written approval must typically be obtained from family or legal representatives depending on governmental and hospital guidelines. Nonetheless, researchers studying these patients have been refused grants, ethics committee approval, and research publication; these decisions tend to be made on the basis that studies of patients who cannot provide consent are unethical. We prefer an ethical framework that balances access to research with medical advances alongside protection for defenseless patients. Severe brain damage represents an immense social and economic problem that warrants further research. Unconscious, minimally conscious, and locked-in patients deserve special procedural protections. However, it is important to stress that they are also at risk of being denied therapy that may be life-saving if clinical research cannot be done on these patient groups.

Patients who are in coma, VS, MCS, or locked-in syndrome present unique problems for diagnosis, prognosis, treatment, and everyday management. At the patient's bedside, assessment of cognitive function is difficult because voluntary movements may be very small, inconsistent, and easily exhausted. Functional neuroimaging will never replace the clinical assessment of patients with altered states of consciousness. Nevertheless, using population norms it can provide an objective measure of the regional distribution of cerebral activity at rest and under various conditions of stimulation. The quantification of brain activity differentiates patients who sometimes only differ by a brief and small movement of a finger. In our opinion, PET, MRS and fMRI will increase substantially our understanding of patients with severe brain damage.
