fMRI in Impaired Consciousness
7th May 2010
Christian SchwarzbauerAberdeen Biomedical Imaging Centre
Future diagnostic opportunities
and ethical challenges
Levels of consciousness
Adapted from Laureys (2007), Sci American
Psychosis ?
Hallucinations Delusions
Thought disorder
Example 1: fMRI in Schizophrenia
fMRI in SchizophreniaHallucination (Hearing voices) Acoustic stimulation
T. Dierks et al., Neuron 22, 615 1999.
S = Speech; R = Reversed speech; T = Pulsed tone
Ethical issuesThis study shows that hallucinations are ‘real’ from
the patients’ point of view (activation of primary auditory cortex)
Differentiation between hallucinations and real auditory input is:(i) not possible based on the actual perception(ii) may be possible by means of logical reasoning(e.g. ‘There is nobody in the room, so it must be a hallucination’)
Would it be ethically acceptable to use fMRI of hallucinations as evidence in support of mitigating circumstances in a court case?Assumption: fMRI reliably detects hallucinationsProblem: Detecting hallucinations in an fMRI experiment does not imply that the patient suffered from hallucinations at the time of the offence
Example 2: fMRI in Vegetative State
fMRI – Detecting awareness in the vegetative state
A. M. Owen et al., Science 313, 1402 (2006)
Communication scans
Is your father’s name Alexander?Yes response with the useof motor imagery
Vegetative state patient
Healthy control
Do you have any brothers?Yes response with the useof motor imagery
Monti et al, New England Journal of Medicine 362, 579-589 (2010)
Communication scans
Is your father’s name Thomas?No response with the useof spatial imagery
Vegetative state patient
Monti et al, New England Journal of Medicine 362, 579-589 (2010)
Healthy control
Do you have any sisters?No response with the useof spatial imagery
Ethical issues ‘The differential diagnosis of disorders of
consciousness is challenging. The rate of misdiagnosis is approximately 40%, and new methods are required to complement bedside testing...’
Communication scans: What if a patient expresses the wish to die?
Communication scans: fMRI response is a sufficient but not a necessary condition of consciousness (mathematical logic):fMRI response patient is conscious [true statement]fMRI response patient is conscious [not necessarily true]So the problem is: the fact that we see no fMRI response does not imply that the patient is unconscious
Monti et al, New England Journal of Medicine 362, 579-589 (2010)
Future diagnosticopportunities
Patient pathways following TBI
Brain death
Coma
Vegetative state
Minimally conscious state
Conscious state
Time
Level of consciousness
TBI
Diagnostic challenges Detection of level
of consciousnessPrediction of outcome
The Cambridge/Liege studyThe new england journal of medicine
10.1056/nejmoa0905370 nejm.org4
Table1.Characteristicsof thePatients.*
PatientNo. Location Age Sex
Diagnosison Admission
Causeof Disorder
IntervalsinceIctus
ResponseonMotor Imagery
Task
ResponseonSpatial Imagery
Taskyr mo
1 Cambridge 58 Male VS TBI 6.0 No No2 Cambridge 43 Female VS Anoxicbraininjury 50.0 No No3 Cambridge 41 Female VS TBI 10.0 No NA4 Cambridge 23 Female VS TBI 6.0 Yes Yes5 Cambridge 42 Male VS Anoxicbraininjury 8.0 No No6 Cambridge 46 Male VS TBI 2.0 Yes No7 Cambridge 52 Female VS Anoxicbraininjury,
encephalitis8.0 No NA
8 Cambridge 23 Male VS TBI 19.0 No No9 Cambridge 48 Female VS Anoxicbraininjury 18.0 No No10 Cambridge 34 Male VS TBI 13.0 No No11 Cambridge 35 Male VS Anoxicbrain injury 10.0 No No12 Cambridge 29 Male VS TBI 11.0 No No13 Cambridge 67 Male VS TBI 14.0 No No14 Cambridge 21 Male VS TBI 6.0 No No15 Cambridge 49 Male VS TBI 3.0 No NA16 Cambridge 56 Female VS Anoxicbrain injury 9.0 No No17 Liege 87 Male VS CVA <1.0 No No18 Liege 62 Male VS CVA 1.0 No No19 Liege 15 Male VS Anoxicbraininjury,
TBI20.5 No No
20 Liege 70 Female VS Meningitis 2.5 No No21 Liege 47 Male VS Anoxicbraininjury 18.8 No No22 Liege 22 Female VS TBI 30.2 Yes Yes23† Liege 22 Male VS TBI 60.8 Yes Yes24 Cambridge 23 Male MCS TBI 11.0 No No25 Cambridge 38 Female MCS TBI 3.0 No NA26 Cambridge 18 Male MCS TBI 8.0 No No27 Cambridge 26 Male MCS TBI 11.0 No NA28 Cambridge 64 Male MCS TBI 6.0 No No29 Cambridge 54 Female MCS Brain-stemstroke 5.0 No No30 Cambridge 29 Female MCS TBI 2.0 No NA31 Cambridge 19 Female MCS TBI 1.0 No No32 Cambridge 34 Male MCS TBI 52.0 No NA33 Cambridge 17 Male MCS TBI 7.0 No NA34 Cambridge 56 Male MCS Anoxicbrain injury 6.0 No No35 Cambridge 21 Male MCS TBI 51.0 No No36 Cambridge 53 Female MCS Anoxicbrain injury 13.0 No No37 Cambridge 36 Male MCS TBI 30.0 No NA38 Cambridge 25 Male MCS TBI 8.0 No No
Copyright © 2010 Massachusetts Medical Society. All rights reserved.Downloaded from www.nejm.org at TRIAL - UNIVERSITY OF ABERDEEN on February 9, 2010
Monti et al, New England Journal of Medicine 362, 579-589 (2010)
17%
83%
fMRI response
no fMRI response
Functional connectivity in the Default Network during resting state is preserved in a vegetative but not in a brain dead patient
Boly et al (2009), Human Brain Mapping 30:2393–2400
Functional connectivity: Default Network
Differences between minimally conscious patients and unconscious patients (vegetative state & coma):
Vanhaudenhuyse et al (2010), Brain 133; 161–171
Functional connectivity: Default Network
minimally conscious > unconscious
Main differences in posterior cingulate cortex / precuneus
Graph theoretical analysis of connectivity
Bullmore et al (2009), Neuroimage 47:1125
MRI Data
Parcellation
Signal decomposition
Pairwise association(e.g. correlation)
Thresholding
Display
Example of graph theoretical analysisAnaesthesia (unconscious) Control (conscious)
Schwarzbauer et al (2010), unpublished data
Future research and ethical implicationsResearch questionsWhat are the neuronal correlates of
consciousness? Is there a suitable biomarker?
(i) to identify level of consciousness(ii) to predict clinical outcome
Ethical implications It will be important to distinguish between direct
measures (e.g. fMRI communication scan) and indirect measures (e.g. network connectivity) of consciousness
Ethical decisions based on indirect measures will be considerably more challenging, but necessary (over 80% of vegetative state patients unresponsive to fMRI)
Thank [email protected]