Pharmacological MRI studies of brain network adaptivity and connectivity

1
Pharmacological MRI studies of brain network
adaptivity and connectivity

• Ed Bullmore
• Psychiatric Neuroimaging
• NATO Advanced Research Workshop
• Chiavari ITALY, 30th Sept 2002

2
Collaborators

• Garry Honey
• Fernando Zelaya
• Steve Williams
• Virginia Ng
• John Suckling

• Caroline Stephenson
• Paul Fletcher
• John Brown
• Carol Routledge
• Cinly Ooi

Acknowledgements Wellcome Trust, BMA,
GlaxoSmithKline, MRC Wolfson Brain Imaging Centre
Co-operative Group.
3
Pharmacological MRI what is it?Laurence Reed et
al (2002)
Locating drug effects on an experimentally
generated neurocognitive signal - locating PK
related drug effects on resting MR signal (more
usual in animal models)
4
Pharma MRI what is it good for?

• Drug fingerprinting
• Functional effects of drugs, including new
  compounds
• of particular value in neuropsychiatric drug
  development?
• Transmitter mechanics of systems
• Transmitter-specific mechanisms for dynamic
  large-scale neurocognitive network organisation
• adaptivity, e.g., local activation changes due
  to task difficulty or repetition
• univariate statistics, e.g. GLM model parameters
• connectivity, e.g., correlation between regional
  time series activity
• multivariate statistics, e.g. correlation or
  path coefficients

5
Neurodevelopmental changes in transmitter
mechanics can be measured by systems fMRI

• Can we measure brain activation with fMRI in
  healthy elderly people?
• (Yes)
• Can we see drug effects on brain activation in
  this age group?
• (Yes)
• Drugs paired within treatment groups to optimise
  power of instructive comparisions between
  dopamine ant/agonist drugs (A) and between
  different drug models of amnesia (B)

6
Design and analysis of a graded, blocked fMRI
activation paradigm
Object location learning trials presented at two
levels of difficulty, 4 s per trial, 6 trials per
24 s block or epoch.
Design allows efficient estimation of mutually
orthogonal components of variance mean
activation-baseline contrast, repetition- and
load-related
7
Systems adaptivity to load and repetition is
mediated by pharmacologically dissociable
mechanisms

• 24 elderly volunteers scanned during repeated
  performance of object-location learning at two
  levels of difficulty
• DA/ACh effects on load-response versus GABAergic
  effects on between-block repetition-suppression
• Bullmore et al (2002) Cerebral Cortex

8
Ant/agonist DA drugs have opposing effects on
atypical load-response of a spatial attention
network
Temporo-parietal cortex
Lateral premotor cortex frontal eye fields
Dorsal cingulate gyrus
Medial posterior parietal cortex
Methylphenidate significantly increases, and
sulpiride significantly decreases, load-response
in epicenters of a large scale neurocognitive
network for spatial attention. implying tonic
dopaminergic control on adaptivity of this system
to increased difficulty of object-location
learning?
9
GABAergic drugs specifically attenuate typical
repetition-suppression Caroline Stephenson et al
(2002)

• 12 healthy young volunteers, each scanned three
  times
• Placebo, flumazenil, lorazepam in
  counterbalanced order
• N-back verbal working memory paradigm graded
  block design
• 3T fMRI data acquisition at WBIC, Cambridge
• Data analysed for orthogonal block repetition
  and load effects

Frontal load-response
Occipito-temporal repetition-suppression
10
Ant/agonist GABAergic drugs have opposing effects
on atypical adaptivity of hippocampal systems
Flumazenil enhances hippocampal repetition
consolidation and fronto-hippocampal
load-response
Perhaps implying that these aspects of brain
adaptivity are tonically regulated by an
endogenous GABA-BZ receptor ligand?
11
Mechanisms for network adaptivity

• Adaptivity to difficulty and (between-block)
  repetition could be anatomically dissociated on
  two tasks
• difficultyfronto-striatal repetitionoccipito-te
  mporal
• Adaptivity to difficulty and repetition were also
  pharmacologically dissociable on two tasks
• difficultydopamine/cholinergic (Mattay et al
  2000) repetitionGABAergic (Thiel et al 2001)
• Ant/agonist drugs in combination can identify
  systems adaptivity regulated by levels of
  endogenous ligand
• spatial attentionDA hippocampalendo-BZ?
• Age-related differences in GABAergic effects are
  compatible with age-related changes in BZ
  receptor properties
• possible role for pharma MRI in mapping
  age-related changes in human transmitter systems
  (Zhang et al 2001)

12
Neoclassical Brain Network Modelling by Path
Analysis Fitting a
Wernicke-Lichtheim-like diagram to fMRI data
Bullmore et al (2000) NeuroImage Semantic
categorisation and subvocal rehearsal task
induces correlated activation of 5 main left
brain regions Causal interactions between these
regions are summarised in a path diagram of and
quantified by partial regression coefficients
13
The human brain is connectedso whats new?
anatomical connectivity
Wright et al (1999) Cerebral Cortex
physiological connectivity
Bullmore et al (2000) NeuroImage
14
Dopaminergic mechanisms modulate connectivity or
coherence in cortico-striato-thalamic systems
Multiple, anatomically parallel, functionally
segregated cortico-striato-thalamic circuits
after De Long et al (1986) Modulation of
inter-regional coherence by L-DOPA in patients
with Parkinsons disease Brown et al (2001) J
Neurosci
15
Path analysis shows enhanced nigro-striatal
connectivity following DA blockade
Caudate
Midbrain
DLPFC
Putamen
Thalamus
Prior knowledge
Activation map of data

• Mixed effects modeling identifies salient and
  specific effect of sulpiride in the context of
  considerable inter-subject variability

Path model for observed inter-regional covariances
16
Scale-invariant dopaminergic mechanisms for
dynamically (de)segregating anatomically parallel
CST circuits?

• LDOPA and MPTP related changes in inter-regional
  coherence measures from electrophysiological
  studies of PD patients and monkeys
• Mixed effects modeling of path (and correlation)
  coefficients shows analogous effects of D2
  blockade on nigrostriatal connectivity measured
  with fMRI in normal volunteers

17
http//www.psychiatry.cam.ac.uk/bmu