Neuroimaging in Acute Stroke

Document Sample
Neuroimaging in Acute Stroke Powered By Docstoc
					Physiologic Basis of fMRI Signals
       Focus on Perfusion MRI
              John A. Detre, M.D.
     Center for Functional Neuroimaging
 Cognitive Rehabilitation Research Consortium
           University of Pennsylvania
          Moss Rehabilitation Institute
               Philadelphia, PA
Spatiotemporal Scales for Neuroscience Methods
                              adapted from Churchland

                 Brain
                            MEG+ERP                                             PET


                                                                                            Lesions

                 Map
   Log Size




                                  Optical          fMRI
                                   Dyes                                         2-deoxyglucose
              Column

                Layer                                                Microlesions

               Neuron             Single Unit
                                                                         Light
              Dendrite           Patch Clamp                          Microscopy



              Synapse

                         Millisecond            Second     Minute        Hour         Day     Week

                                                          Log Time
     Imaging Physiological Correlates of Neural
                     Function

                              metabolic response
                                                           FDG PET
                              - ATP tightly regulated
                              - glucose consumption
electrical activity           - oxygen consumption
  - excitatory                                            H215O PET
  - inhibitory
  - soma action potential
                              hemodynamic response
                              - blood flow              optical imaging
          electrophysiology   - blood volume
                              - blood oxygenation
               EEG
                                                            fMRI

               MEG
           Brain Mapping with fMRI
• Hemodynamic/metabolic response used as surrogate
  marker for neural activity
• BOLD fMRI signal represents a complex interaction
  between CBF, CBV, CMRO2:
  – CBF >> CMRO2  less deoxyhemoglobin with activation
  – CBF is monitored indirectly
  – “Tracer” is venous (though T2* effects can extend beyond vein)
• Both magnitude and temporal pattern are modeled in
  fMRI data analysis
• Potentially affected by normal physiology,
  pathophysiology, pharmacology
                        Activation-Flow Coupling
• Blood flow and metabolism changes accompany brain activation
• First described in late 1800’s by Mosso (Italy) and Sherrington (England)
• Physiological basis remains poorly understood today


                                                Mosso, Atti R. Accad. Lincei 1880
                                                                          (Italian)




Sherrington and Roy,
     J. Physiol. 1890


                 Sir Charles Sherrington
          Physiology of Functional Activation

                          neuron
                                              ???                arteriole
              receptors    energetics



membrane                           glucose
potential &                                              •  CBF with task activation
                 ion channels      oxygen    capillary
firing
                                   waste                 • Linear with metabolism
                                                            – vs. BOLD (see Hyder)




                                                                 venule
 Coupling of CBF, CMRGlu, and CMRO2
     during Functional Activation
• Uncoupling of CBF, CMRGlu, and CMRO2
  Fox and Raichle, PNAS 1996
  – CBF=CMRGlu>>CMRO2
  – Predicts reduction in deoxyhemoglobin with activation
• No increase in activated CBF with hypoglycemia
  Powers et al., Am. J. Physiol. 1996
  – Suggests CBF is not required to supply glucose substrate
• No increase in activated CBF with hypoxia
  Mintun et al., PNAS 2001
  – Suggests CBF is not required to supply O2 substrate
  Potential Physiological Influences on fMRI Signal
   cerebrovascular                                           structural lesions
       disease                                                (compression)

 medications          blood                         blood            autoregulation
                       flow                                          (vasodilation)
                                                   volume
   hypoxia
                                                                       volume status

 hypercarbia                   BOLD
                              contrast
anesthesia/sleep                  biophysical effects

             anemia


                                                           smoking
                               oxygen
                              utilization
                                                        degenerative
                                                          disease
      Brain Mapping with fMRI - Definitions
• Activation-Flow Coupling
  – Hemodynamic responses used as surrogate marker for neural
    activity in functional neuroimaging
• Blood Oxygenation Level Dependent (BOLD) fMRI
  –   represents a complex interaction between CBF, CBV, CMRO2
  –   “Tracer” is venous (though T2* effects can extend beyond vein)
  –   CBF is monitored indirectly
  –   Qualitative: only differences between conditions can be measured
• Arterial spin labeling (ASL) Perfusion MRI
  –   ASL = endogenous flow tracer (analogous to 15O-H2O in PET)
  –   Quantitative: provides CBF in ml/100g/min (classical units)
  –   Allows both resting CBF and CBF changes to be measured
  –   Harder to implement and has lower SNR compared to BOLD
 Direct MRI Measurement of Cerebral Blood Flow
         with Arterial Spin Labeling (ASL)
                                                                          arterial   imaging    control
                                                                          labeling     slice   labeling
          PET or SPECT                 M RI PERFUSION
      Ste ady State Me thod          Ste ady State Me thod

                         de cay
                                                         T1 re laxation




                                 arte rial
                              spin labe ling

15
  O infusion
or inhalation




                • Uses electromagnetically labeled arterial blood water as
                  an endogenous flow tracer
                • Provides quantifiable CBF in classical units
                • Effects of ASL are measured by interleaved subtractive
                  comparison with control labeling
                                                                                                          Detre et al., 1992 and ff
                      Physiological Basis of fMRI

        behavior                         neural function             disease


                biophysics***
                                          metabolism

   BOLD fMRI                                                       ASL CBF MRI

***BOLD contrast includes
contributions from biophysical effects     blood flow      ASL fMRI measures changes in CBF
                                                           directly, and hence measured signal
such as magnetic field strength                            changes may be more directly coupled
homogeneity and orientation of                             to neural activity
vascular structures.
              Brain Activation Analysis
   ASL or
T2*-weighted Average                     Statistical Thresholded Overlay on
 Snapshot Difference                    Significance Statistical T1 Anatomic
   Image      Image                        Image        Image       Image




TIME SERIES                        ON
               TASK




                             OFF
               fMRI SIGNAL
         fMRI with BOLD Contrast
                   task activation
   Photic Stimulation                Verbal Fluency Task




calcarine cortex          Broca’s area    Wernicke’s area
 Perfusion MRI with Arterial Spin Labeling
                                                              Control - Label
                                          Control Inversion
B Field Gradient




                                          Plane


                                          Imaging Slice


                                          Arterial Tagging
                                          Plane




                   Continuous Adiabatic                          Single Slice
                    Inversion Geometry                        Perfusion Image
                                                              (about 1% effect)
               Key Improvements in ASL MRI
• Transit time correction
   – (Alsop and Detre, 1998)
• Multislice
   – (Alsop and Detre, 1998)
• Background suppression
   – (Ye et al., 2000)
• High Field
   – (Wang et al., 2002)
• Multicoil/Parallel Imaging
   – (Wang et al, 2005)
• Snapshot 3D Imaging
   – (Duhamel and Alsop, 2004)
   – (Fernandez-Seara et al., 2005)
• Improved Labeling
   – (Garcia et al., 2005)                       Data from David Alsop, BIDC


      SNR gains exceeding approaching 1000% over the past decade
             Perfusion fMRI using ASL
• Observe CBF changes directly
   – CBF changes are more linearly coupled to neural activity than BOLD effects

• Resting and activated CBF in absolute units (ml/g/min)
   – Pathological conditions may affect resting CBF

• Despite reduced sensitivity vs. BOLD, advantages for:
   – Spatial resolution (localizes to brain rather than vein)
   – Low frequency designs (behavioral states)
   – Group analyses (? reduced intersubject variability)
   – Regions of high static susceptibility gradient (non-GE EPI)
   – Statistical advantages (white noise)
Localization of Functional Contrast

           Perfusion

                              Perfusion Activation




                                                     draining
                               BOLD Activation         vein
BOLD*
        *1.5T/Gradient Echo
Temporal Characteristics of Perfusion fMRI
  • Control/Label pair typically every 4-8 sec
    – “Turbo” ASL (Wong) can increase resolution by ~50%
    – Qualitative CBF (no control) in ~2 sec
    – S:N much lower than BOLD for event-related fMRI


  • Control/Label pair eliminates drift effects
    – White noise (instead of 1/f)
    – Stable over long durations (learning, behavioral state
      changes, pharmacological challenge etc.)
    – Sinc subtraction eliminates BOLD derivative
Only perfusion fMRI can detect activation with
   task and control separated by 24 hours
                                                 Wang et al., MRM 2002
                                                                         Perfusion vs. BOLD: Very Low Task Frequency
     ASL Perfusion fMRI vs. BOLD
Improved Intersubject Variability vs. BOLD
          Aguirre et al., NeuroImage 2002




    Single Subject            Group (Random Effects)
             Utility of ASL Perfusion fMRI
                   in Clinical Research
• Quantify CBF in cerebrovascular disorders
   – Perfusion imaging may reveal “functional” deficits without a structural
     correlate
   – Baseline CBF is a critical determinant of the capacity for activation-flow
     coupling with a task

• Correlate “resting” CBF with cognitive deficits in cohort
   – Allows functional localization of affected regions
   – Avoids confound of impaired task performance during fMRI
   – Avoids need for cognitively impaired subject to perform during fMRI

• CBF should be a stable biomarker across space and time
   – Ideal for multisite or longitudinal studies
   – This advantage has yet to be formally proven in a clinical study
     Perfusion MRI in Cerebrovascular Disease
 Intracranial Stenosis with Bilateral Cognitive Deficits
              Jefferson et al., AJNR 2006




T2-weighted structural MRI:   Preop Perfusion MRI:                Postop Perfusion MRI:
R>L ischemic changes          Bilateral (R>L) hypoperfusion       Bilateral increase in perfusion

                              L Hemisphere CBF= 27 ml/100g/min*   L Hemisphere CBF= 56 ml/100g/min
                              R Hemisphere CBF= 20 ml/100g/min    R Hemisphere CBF= 53 ml/100g/min
                                    Normal CBF= 50 ml/100g/min

       Perfusion MRI correlated better with cognitive deficits than structural MRI
   Cognitive Correlations using Resting Perfusion MRI
                 in Alzeimer’s Dementia
• 17 Patients with clinical Alzheimer’s Disease
• Noninvasive perfusion MRI in 5 mm slices
• Correlation of CBF with cognitive performance on:
   –   Semantic Category Membership Judgment
   –   Confrontation Naming
   –   Semantically-guided Category Naming Fluency
   –   Sentence Comprehension                                            RAW PERFUSION DATA




  SEMANTIC CATEGORY           CONFRONTATION NAMING     SEMANTICALLY-GUIDED           SENTENCE
MEMEBERSHIP JUDGEMENT                                CATEGORY NAMING FLUENCY       COMPREHENSION
 This approach to localization of dysfunction also avoids having patient perform a task during MRI
Dissociation of Activation-Flow Coupling
  Patient with Left Intracranial Carotid Stenosis

            BOLD fMRI                                    Perfusion MRI




                                                         Markedly decreased
 Lack of left hemispheric activation during motor task
                                                            left hemispheric
  most likely reflects low resting CBF rather than any      perfusion at rest
           reorganization of neural function
  Considerations in Task-Activation fMRI
               (Summary)

                                                diagnostic fMRI
    neural function             pathology

                                                pathophysiology



        behavior &    hemodynamic change
                                                      clinical fMRI
       performance


                                    motion

Neuroscience fMRI     RR & HR                         fMRI signal
                                    physics
                                              noise
                          Conclusions
• FMRI is measures neural activity indirectly
  – BOLD qualitatively reflects CBF and metabolism
  – ASL quantitatively reflect CBF
• Clinical FMRI poses special challenges
  – Task performance effects must be considered
  – Underlying pathophysiology may alter coupling of activation and flow
• FMRI identifies putative regions supporting task function
  – Does not establish necessity
  – Correlation with outcome, lesions, or TMS lesions can disambiguate
• Many fMRI applications in neurorehabilitation
  –   Mechanisms of neuroplasticity
  –   Biomarker for therapy
  –   Prediction of outcome
  –   Bionic interfaces

				
DOCUMENT INFO
Shared By:
Categories:
Stats:
views:12
posted:6/24/2011
language:English
pages:26