Emotion processing in the minimally conscious state
- T Bekinschtein1,
- J Niklison1,
- L Sigman1,
- F Manes1,2,6,
- R Leiguarda2,
- J Armony3,
- A Owen4,
- S Carpintiero5,
- L Olmos6
- 1Cognitive Neurology Section, Institute for Neurological Research (FLENI), Buenos Aires, Argentina
- 2Department of Neurology, Institute for Neurological Research (FLENI), Buenos Aires, Argentina
- 3Department of Experimental Psychology, McGill University, Montreal, Canada
- 4MRC-CBU, Cambridge, UK
- 5Functional Neuroimaging Laboratory, Institute for Neurological Research (FLENI), Buenos Aires, Argentina
- 6Rehabilitation Institute (FLENI), Buenos Aires, Argentina
- Correspondence to: Dr F Manes Cognitive Neurology Section, Institute for Neurological Research (FLENI), Montañeses 2325 (C1428AQK). Buenos Aires, Argentina;
- Received 22 December 2003
- Accepted 13 February 2004
- Revised 12 February 2004
As a newly described condition distinct from coma or the vegetative state, minimally conscious state (MCS) is characterised by a threshold level of consciousness, and diagnostic criteria have recently been proposed.1 In MCS, cognitively mediated behaviour occurs inconsistently, but is reproducible or sustained enough to be differentiated from reflexive behaviour. It is clinically essential to distinguish this condition from persistent vegetative state (PVS), due to a potentially more favourable outcome.1 So far, whether patients in MCS can process emotion is unknown.
Cortical processing has been described in PVS using auditory and visual functional paradigms with positron emission tomography.2,3 However, to date hardly any functional imaging studies are available in patients in MCS.4 We used fMRI to assess brain activity induced by an emotional stimulus in a patient in MCS.
A 17 year old man was riding his bicycle when he was hit by a train. The accident resulted in head trauma and immediate coma, progressing to MCS over the course of 4 months, when he was admitted to our institution. This research protocol was approved by the Institutional Ethics Committee. At the time of the fMRI study, 5 months after the accident, the patient localised noxious stimuli, had spontaneous eye opening, detectable sleep/wake cycles, sustained visual fixation, and contingent smiling, thus meeting criteria for MCS. A structural MRI study showed mild cortical atrophy and dilated ventricles. Auditory evoked potentials showed decreased conduction velocities at brainstem level. The patient increased his level of awareness 2.5 months after the functional study was conducted. Auditory evoked potentials after recovery were within normal range, while MRI showed much less ventricle dilatation. Six months after recovering full consciousness, he was able to chat normally and feed himself. Currently we are retesting the patient with the same paradigm.
Non-familiar voice v silence and mother’s voice v non-familiar voice recognition were tested in an fMRI block design with 30 seconds per epoch. The patient listened to his mother reading a story, followed 30 seconds later by an age matched voice reading the same story, for 30 seconds with silence epochs in between. Blood oxygen level dependent images were acquired using a T2 weighted gradient echo planar sequence on a General Electric Signa CVI, 1.5T system with real time image processing of multislice and multiphase images during patient stimulation and rest periods. The Medx 3.4 Sensor System was used to carry out fMRI post-processing, including motion correction and Gaussian smoothing. An uncorrected significance threshold of P<0.001 was used because amygdala and insula activation was expected, owing to emotional voice processing. Activated clusters were localised following co-registration with an anatomical T1-IR volume.
Subtraction of the phrases read by the age matched voice from silence was the control experiment, showing a significant focus of activation in the transverse and superior temporal gyri, which spread to the planum temporale; more anterior activation was found in the superior (right) and inferior (left) insula (fig 1A). The subtraction of the mother’s phrases from the age matched voice disclosed a strong activation of the amygdala and insula spreading to the inferior frontal gyrus; there was also weaker activation of the transverse temporal gyrus, temporal operculum, and planum temporale (fig 1B, C). Activation was lower on the right hemisphere in both comparisons, non-familiar voice v silence and familiar voice v non-familiar.
To the best of our knowledge, our results provide for the first time anatomical evidence for the response of an MCS patient to a familiar voice, in which both amygdala and insula appear to play a major role.
The activation pattern of the control experiment agrees with previous studies.5 Our results showed that the mother’s voice activates the extended amygdala, an emotionally related structure, and a directly connected area such as the insula, perhaps acting jointly as limbic integration cortex. Although residual cerebral activity was unequivocal in our case, representing fragmentary cognitive processing, it should not be assumed that it depicts a fully integrated system required for normal levels of awareness; however, our findings highlight the legal and ethical implications of careless bedside chatter. Whether functional imaging represents a reliable method to evaluate neural processing in MCS patients, in whom cognitive output is extremely difficult to assess, remains to be seen.
This work was supported by the Institute for Neurological Research, FLENI.
Competing interests: none declared