What are some of the issues in choosing bolus vs bolus+infusion?
Which one predicts equilibrium coefficients?
Which one achieves equilibrium?
Which one is more difficult, experimentally?
Which is more difficult to analyze – mathematically?
Which would be good for a slow tracer?
Which would be good for a fast tracer?
Which one is cheaper?
Which one is good for a studying the effect of horse-riding on striatal dopamine?
Which one …. for studying distribution of receptors in equestrians?
Which one …. for an irreversible tracer?
what was the general
argument here?
what were the assumptions?
model similar to prev.
like Morris paper, they allow
for time-varying dopamine
but here, the function is
preset and the assumption is
that DA is always in equilib
between free and bound.
DA(t) is an exponential that
peaks instantaneously at
start of stimulus.
Is this realistic?
involved model of noise in DBP
does it make sense?
Noise goes down with more counts
(NEC) in either the Reference region or
the target region.
Noise goes down with pixel size of the
regions
Noise goes down as the ratio of the
target to reference region goes up.
want to optimize the sampling
of the pre-stimulus equilib
value and the post-stimulus
equilib value
given lots of simulations based
on differing input functions, we
can ask, what are the best
overall windows for pre and
post
Trade-offs
longer values means more
counts, better statistics – but
perhaps more bias too.
later windows mean closer to
equilib – less bias but worse
statistics
Why doesn’t the post-stim
curve stay flat? Does it ever?
major finding of Watabe 2000:
optimal sampling of B/I data pre and
post stim gives better significance,
larger effect size, fewer subjects
required.
where is most of the gain – why?
H is the transfer function for a B/I
study. Ie., the tissue response to an
arbitrary input.
It is made up of the response to a
bolus (d) (aka the impulse
response) and the response to an
infusion (step) (aka the step
response)
response
to :
= response
to :
+
-
+
=
+
Kbol tells us how much bolus response and how much infusion response
we want – usually to get to equilib in tissue fastest… but not all tissues
behave the same way, so it’s a compromise.
Kbol is in units of minutes because a bolus is characterized by a dose but
an infusion is characterized by a dose/time
Estimated Kbol from bolus tissue responses
--- same (female) subject
12000
Bolus response
16000
caudate
cerebellum
cerebellum_wm
cingulum_ant
cingulum_post
frontal
hippocampus
occipital
pallidum
parietal
putamen
temporal
thalamus
caudate
cerebellum
12000
cerebellum_wm
cingulum_ant
10000
cingulum_post
frontal
8000
hippocampus
occipital
pallidum
6000
parietal
Concentration (Bq/ml)
10000
14000
Time (min)
JJ986 - Kbol=180
8000
6000
4000
putamen
4000
2000
temporal
thalamus
2000
0
0
0
50
100
150
200
250
300
350
0
400
50
100
JJ986 - Kbol=190
200
250
300
350
JJ986 - Kbol=195
12000
12000
caudate
cerebellum
cerebellum_wm
6000
cingulum_ant
cingulum_post
frontal
hippocampus
4000
occipital
pallidum
parietal
8000
putamen
temporal
thalamus
2000
caudate
cerebellum
cerebellum_wm
10000
Concentration (Bq/ml)
10000
Concentration (Bq/ml)
150
Time (mins)
Concentration
8000
cingulum_ant
cingulum_post
frontal
6000
hippocampus
occipital
pallidum
4000
parietal
putamen
temporal
thalamus
2000
0
0
0
50
100
150
200
Time (min)
250
300
350
-40
10
60
110
160
Time (mins)
210
260
310
360
Issues:
B vs B/I
anesthesia
dependence
self-admin
SA is it high or low?
hint:
Early or Late
(monkey 11C-raclopride images)
BP image
Fill in the boxes to explain the paradigms
?
?
plot the curves to show the dynamic
11C-raclopride data in dorsal anterior
caudate
?
?
why correct for vascular
contrib?
why didn’t they do it in bolus
case?
where is the activation?
does this make sense?
why do they call this
exploratory?
what are the general
steps to get to this
point?
what correction has to
be made?
why do we need to
normalize all the data
‘spatially’ first?
what is a t value?