Klinische Neurophysiologie 2008; 39 - A31
DOI: 10.1055/s-2008-1072833

fMRI reveals cognitive and emotional processing in a long-term comatose patient

S Eickhoff 1, 2, M Dafotakis 1, C Grefkes 1, 3, T Stöcker 1, 4, JN Shah 1, 4, K Zilles 1, 2, 4, M Siebler 5
  • 1Forschungszentrum Jülich, Institut für Neurowissenschaften und Biophysik, Jülich
  • 2Heinrich-Heine-Universität, C. & O. Vogt Institut für Hirnforschung, Düsseldorf
  • 3Max-Planck-Institut für neurologische Forschung, Köln
  • 4Brain Imaging Center West (BICW), Jülich
  • 5Heinrich-Heine-Universität, Neurologische Klinik, Düsseldorf

History: At the age of 38, an otherwise healthy mother of two children was admitted to hospital with a Glasgow coma scale (GCS) of 4 following closed head injury. Diagnostics revealed traumatic subarachnoid haemorrhage and right sided space-occupying parenchymal haematoma; treated by craniotomy and haematoma resection. Without ever regaining consciousness, opening her eyes or showing other than stereotypic left arm movements, she was transferred to a nursing home. 35 months later the patient was presented for further evaluation using functional magnetic resonance imaging (fMRI).

Clinical status: Examination revealed a GCS of 5 and a grade IV coma on the Grady Coma Scale. When the permanently closed eyelids were lifted, the eyes showed horizontal deconjugation with variable abduction of either eye. The right pupil was wider and stood slightly higher (skew deviation). Vestibulo-ocular reflex testing and caloric irrigation indicated a bilateral internuclear ophthalmoplegia. Stereotypical masticatory movements, non-habituating snout, palmo-mental and (strong) grasp reflexes were present. Spastic tetraparesis was more pronounced on the lower extremity and on the right side with hypertonic contractures. Particularly when the patient was manipulated, stereotype movements of the left arm and hand were visible.

fMRI Methods: Three sensory paradigms were tested twice two months apart [Visual stimulation by full-field flicker light; auditory stimulation by non-emotional words; brushing the right or left forearm] using a block design. During the second examination we furthermore investigated responses to speech of varying familiarity (the patient's children, longstanding close friends, unknown person) and content (addressing the patient directly, stating both the speaker's and patient's name, or uttering random non-emotional words). All six conditions (3 speakers x 2 contents) were presented 20 times each in a pseudorandomised order using a block design.

fMRI Results: SPM5 analysis revealed robust activation of auditory, visual and somatosensory cortices by sensory stimulation (P<0.05, FWE corrected). Additionally, presentation of auditory words also activated Broca's area (left BA 44). Stronger response during speech directed to the patient as compared to neutral phrases, and a speaker-dependent modulation thereof, was detected in the left amygdala and the right anterior superior temporal sulcus. In both regions, children's voices elicited the strongest activity followed by her friends, while the response to the unknown voice was significantly smaller. Regardless of the speaker, addressing the patient resulted in significantly larger activations in these regions.

Implications: This case highlights the potential of neuroimaging to contribute to the evaluation of comatose patients by revealing cognitive and emotional capabilities not previously accessible. The detection of differentiated cortical processing, even in a patient with such poor clinical status, should bring about a reconsideration of the way these patients are cared for and relative's claims of „communication“ Moreover, assessment of residual cortical function by fMRI may help identifying patients who have the highest prospective to benefit from interventional treatment (e.g., deep brain stimulation). Finally, there might be a long-term potential for communication via brain-computer interfaces arising from this study and preceding observations in PVS patients.

Fig. 1

Fig. 2