fMRIs

• Historical images of the brain

  • some still in use today
  • oftern interpretive, labour/skill intensive
  • phenology - mental abilities can be determined by examining shape & size of skull
  • elecrophysiological mapping - electrical measurements to understand function & organization of brain
• MRI

  Magnetic

  Resonance

  Imaging
• MRI - physics

  • strong magnetic field (Bo) & radio enery produces image
  • powerful magnet causes certian nuclei in body to align with magnetic field
  • Radio frequency pulses applied to aligned nuclei = temporarily deviate from aligned state
  • When turned off = return to original alignment releasing energy process
• Magnetic moment 


  nuclei that align have a magnetic moment = ability to interact with magnetic fields - due to presence of odd number of protons &/or neutrons in atom's nucleus
• How image acquired

  • nuclei spin around main magnetic field
  • radio frequency pulse tips nuclei out of alinemnt with Bo & sychnronise phase of spins
  • RF off = nuclei gradually return to original alignment & start to lose phase coherence
  • changes in alignment & phase coherence of nuclear spins = detected as MRI signal
• Blood Oxygen Level Dependent (BOLD) response - basis of fMRI


  1. After neural activity increases = immediate decrease in blood oxygenation - "initial dip" in hemodynamic response function
  2. Following initial dip = blood flow increases to compensate for heightened demand = often actual increase in regional blood oxygenation
  3. Blood flow peaks around 6s after initial dip, then gradually returns to baseline levels
  4. Sometimes, "post-stimulus undershoot" = blood flow briefly drops below baseline
• Blood Oxygen Level Dependent (BOLD) response - measures


  measures changes in relative levels of deoxyhemoglobin & oxyhemoglobin in response to regional cortical activity in brain
• BOLD

  • Deoxyhemoglobin & oxyhemoglobin = different magnetic properties = affected differently
  • neural activity increases = increase in oxygen demand = more oxygenated blood flows into area
  • differences = local magnetic field strength in brain changes = affects MRI signal in area
  • increased signal detected as "activity" in image
  • change from image = far removed from nerutal events - long chain of events
• fMRI - experimental logic

  • cognitive subtraction - comparing brain activity between 2 or more experimental conditions that differ in presence or absence of particular cognitive process
  • measuring time for specific cognitive process to occur by comparing reaction times between tasks with different components
• fMRI - experimental logic - method

  1. T1: Participants hit button when see light = baseline RT
  2. T2: Participants hit button when light is green not red - discrimination
  3. T3: Participants hit left button when light green & right when red = decision-making component

  • subtracting T1 from T2 = isolate time taken to discriminate between colors
  • subtracting T2 from T3 reveals time taken to make a decision
• fMRI - experimental logic - assumption?

  • assumption of pure insertion suggests - component process can be added to task without affecting other components
  • differences in task difficulty between conditions may impact attention = confounding variable
• Experimental design 1 - block design

  • stimuli or tasks presented in blocks or groups
  • blocks alternate with periods of rest or baseline conditions
  • active blocks = brain activity measured using fMRI
  • simple, more effective at identifying regions consistently activated, study sustained cognitive processes over extended periods
  • not be ideal for detecting rapid changes, lacks ecological validity
• Experimental design 2 - event-related fMRI

  • stimuli or events are presented individually & at irregular intervals
  • allows researchers to isolate neural responses associated with each individual event & examine how brain responds to different types
• Event-related fMRI - advantages

  • allow for greater flexibility in experimental design - manipulate timing & sequence of stimuli more precisely = more naturalistic
  • examine how brain activity changes over time in response to different events
• event-related fMRI

  • individual stimulus = less statistical power
  • more susceptible to issues like habituation or carryover effects
• Analysis - block model 


  • data analyzed by dividing experiment into blocks representing a period of time during which a particular condition or stimulus is presented
  • straightforward & robust
  • may overlook transient or rapidly evolving neural responses
• analysis - block model convolved with hemodynamic response function

  • account for delayed & prolonged hemodynamic response observed in brain following neural activity
  • generates a predicted BOLD signal for each experimental condition - compared to actual BOLD signal obtained
  • better capture temporal dynamics of neural activity & improve accuracy of analyses
• voxel-wise analysis


  • analysis done independently at every voxel
  • Contrasts = test for voxels where activation in 1 condition greater than another
  • Voxels with significant T statistics can then be colored in according to size of T
• Blobs


  • clusters of significant statistics for either a main effect or a contrast between 2 sets of regressors at each voxel
  • Shows areas where signal change significantly predicted by model (or where degree of prediction differed between contrasted conditions)
  • end result after much preprocessing & analysis
  • Change in signal due to regional hemodynamics

  = activations distantly related to underlying neurological events
• What has functional brain imaging told us

  • Identified functional areas
  • Corroborated  findings from other methods
  • Allowed localization of function from undamaged brains
  • Meta-analyses bring some order to flood of data - but are these any more useful than electrophysiology map
• New directions

  • Functional-connectivity analyses: calculate correlations between activations in different areas
  • Dynamic causal modelling: explicit models of  distributed networks tested to see which best fits observed data
  • Both techniques investigate distributed processing & overcome some limitations of lesion studies & earlier fMRI studies