Neuron 51, 541–547, September 7, 2006 ª2006 Elsevier Inc.DOI 10.1016/j.neuron.2006.07.026
ReportDopamine Scales Performance
in the Absence of New Learning
Barbara Cagniard,1Jeff A. Beeler,1
Jonathan P. Britt,3Daniel S. McGehee,3
Michela Marinelli,2and Xiaoxi Zhuang1,*
1Department of Neurobiology, Pharmacology,
The University of Chicago
Chicago, Illinois 60637
2Department of Cellular and Molecular Pharmacology
Rosalind Franklin University of Medicine and Science
3333 Green Bay Road
North Chicago, Illinois 60064
3Department of Anesthesia and Critical Care
The University of Chicago
Chicago, Illinois 60637
Learning and motivation are integral in shaping an
organism’s adaptive behavior. The dopamine system
has been implicated in both processes; however, dis-
sociating the two, both experimentally and conceptu-
ally, has posed significant challenges. We have devel-
oped an animal model that dissociates expression or
scaling of a learned behavior from learning itself. An
inducible dopamine transporter (DAT) knockdown
mouse line has been generated, which exhibits signif-
icantly slower reuptake of released dopamine and
increased tonic firing of dopamine neurons without al-
tering phasic burst firing. Mice were trained in experi-
mental tasks prior to inducing a hyperdopaminergic
tone and then retested. Elevated dopamine enhanced
performance in goal-directed operant responses.
These data demonstrate that alterations in dopaminer-
gic tone can scale the performance of a previously
learned behavior in the absence of new learning.
Both learning and motivation shape an organism’s re-
sponse to its environment. Adaptive behavior requires
assigning significance to environmental stimuli and
associating appropriate behavioral responses. Equally
important, an organism must scale a learned response
in relation to its current needs. Although the dopamine
(DA) system is strongly implicated in each phenomena,
a primary difficulty is dissociating the two, both experi-
mentally and conceptually (Dickinson and Balleine,
There are two broad perspectives on DA function. In
the first, DA facilitates reinforcement learning by provid-
tions (Wise, 2004). In this context, Schultz et al. have
demonstrated that DA cells fire phasically in response
1994; Schultz et al., 1993; Schultz et al., 1997). This work
provided the empirical basis for the prediction-error
theory of DA (Montague et al., 1996), which posits that
DA facilitates learning by signaling discrepancies be-
tween predictions and actual events.
In the second perspective, DA’s root function is to
facilitate motivation. The incentive salience hypothesis,
for example, suggests that DA enhances the energizing
effect of reward or reward-predicting cues (Robinson
and Berridge, 1993; Berridge and Robinson, 2003). Sal-
amone et al. hypothesize that DA maintains behavior
when response costs are high (Cousins and Salamone,
1994; Salamone and Correa, 2002). Thus, DA modulates
the expression of learned behavior by effectively scaling
the response generated by previously established
We have designed a genetic approach to manipulate
DA signaling to address the question of whether DA
can directly scale performance of a learned task in the
absence of new learning. An inducible DA transporter
(DAT) knockdown mouse line was developed. When
without affecting phasic DA activity. The inducible
knockdown allows us isolate the putative performance-
scaling effects of DA from learning effects by training
our subjects prior to inducing the genetic alteration.
We report here the clearest evidence to date that DA
directly scales behavioral performance in the absence
of new learning.
Generation of Inducible DAT Knockdown
The tetracycline inducible system (Gossen and Bujard,
1992) was used (Figure 1A). A DAT-tTA line was gener-
ated by gene-targeting the tetracycline responsive
der the transcriptional control of the DAT promoter and
limiting expression to DA neurons. This line is a DAT
knockout. A second line, tetO-DAT, was generated by
DAT under the transcriptional control of tetO. This line is
wild-type expression levels and the same expression
pattern. The residual DAT expression is due to the mini-
mal promoter activity in tetO. We generated DAT-tTA/
tetO-DAT compound heterozygotes by crossing these
two lines. DAT protein expression in these mice is lower
than that in wild-type mice with the same expression
pattern (Figure 1B). This expression is mediated by tTA
binding to the tetO promoter, as doxycycline (Dox, a tet-
(Figure 1B). We used Western blotting to examine the
time course of Dox treatment effect. DAT protein was al-
most completely absent after 6 weeks of Dox treatment
(Figure 1C). We quantified DAT levels in each sample
and used the formula N(t) = N(0) 3 (1/2)t/tauto estimate
that the half-life (tau) of DAT is approximately 7 days.
cycline inducible system have been published (Bond
et al., 2000; Gross et al., 2002). The present study
represents a strategy first proposed by Hen et al.
(Lucas and Hen, 1995), which allows independent use
of two knockin lines.
DA Release and Uptake Parameters in Mutant Mice
We characterized DA release and uptake in mutant mice
during stages of Dox treatment. In brain slices, stimula-
tioninthenucleus accumbens (NAc)shellelicited oxida-
tion currents that were identified as DA using fast-scan
cyclic voltammetry (FSCV), allowing us to employ am-
perometry to increase temporal resolution. The rise
time, peak amplitude, and decay parameters for the
evoked DA signals did not differ between wild-type
and mutant mice not fed Dox (Figure 2). Mutant mice
treated with Dox, however, exhibited a marked reduc-
tion in DA clearance rates, as illustrated by the longer
decay times of the DA oxidation currents [Figure 2B;
F(4,96) = 93.25, p < 0.0001]. The effect was time-depen-
dent, with the slowest decay after more than 8 weeks of
Dox exposure. The current decay was best fit by two ex-
ponentials. The first exponential showed considerable
variability that was not related to Dox treatment. The
second exponential was greatly prolonged with Dox
treatment (Figure 2B). The peak levels of DA release
were not significantly different between groups [not
shown; F(4,96) = 1.98, p = 0.07]. Interestingly, DAT ex-
pression was suppressed after 6 weeks, but functional
elimination of DAT took more than 6 weeks of treatment,
suggesting that minimal DAT expression can have a
large functional impact.
Action Potential Activity of DA Neurons
in Mutant Mice
We evaluated action potential activity of DA neurons in
the ventral tegmental area (VTA). Suppressing DAT ex-
pression by Dox produced an increase in overall firing
rate of DA neurons (Figure 3A, T = 2.2, p < 0.04). Mice
treated with Dox also exhibited a high prevalence of
neurons with firing rates >4 Hz (73%) compared with
mice that were not fed Dox (30%; Chi2= 7.44, p < 0.01;
We performed separate analyses on the bursting
and nonbursting components of neuronal activity. As
Figure 3D shows, nonbursting firing rates were higher
in mice treated with Dox versus those not exposed to
Dox (T = 22.38, p < 0.04). However, no differences
were observed in the quantity or characteristics of
bursts (Table 1). Thus, DAT knockdown mice treated
with Dox exhibit a chronically elevated DA tone without
changes in phasic DA activity.
Downregulation of D2 autoreceptors may explain ele-
vated tonic firing. We conducted radioligand binding
autoradiography using [3H]-spiperone. There was no
change in D2 autoreceptor expression levels (212 6 16
fmol/mg in untreated group and 203 6 21 fmol/mg in
>8 week Dox-treated group; n = 5, T = 0.38, p = 0.70).
Figure 1. Generation of the Inducible DAT Knockdown Transgenic
(A) The inducible DAT knockdown mice were generated by crossing
straining indicated that 4 weeks of Dox treatment decreased DAT
expression while preserving the expression pattern. (C) Western
blot analysis of striatal DAT protein indicated that DAT protein was
gradually decreased by Dox. Str: striatum, NAc: nucleus accum-
bens, SN: substantia nigra, VTA: ventral tegmental area.
Figure 2. Dopamine Release and Uptake Parameters
(A) Normalized representative amperometric traces of dopamine
efflux in the NAc shell. (B) Second order decay time constant (t2)
of evoked dopamine release. Decreasing DAT by Dox significantly
reduced dopamine reuptake (***p < 0.0001). Error bars = SEM.
Induced Knockdown of DAT Expression
Mice were trained and tested on progressive ratio (PR)
schedules, which measure the work that animals are
willing to do for a reward (Hodos, 1961). We tested all
mice on a PR7 schedule (the work requirement in-
creases by 7 lever presses after each reward), food de-
prived (1 week, w10% below baseline body weight) and
then not food deprived (1 week). No mice were on Dox
during this portion of the experiment. There was no per-
formance difference between groups in either food-
deprived [Figure 4A, F(2,25) = 1.4, p = 0.26] or not food-
deprived conditions [Figure 4B, F(2,25) = 0.31, p = 0.74].
We then treated half of the mutant and all the wild-
type mice with Dox for 8 weeks. We again tested all
mice on a PR7 schedule, food-deprived and then not
food-deprived (1week each). Inducible DAT knockdown
on Dox (Inducible: No Dox / Dox) mice displayed more
lever pressing than wild-type (WT: No Dox / Dox) mice
or DAT knockdown not on Dox (Inducible: No Dox / No
Dox) mice when food deprived [Figure 4A, F(2,25) = 5.7,
p = 0.009]. Increased response in the Dox-treated mice
cannot be attributed to nonspecific hyperactivity as
this elevated responding was not observed on the inac-
a fixed ratio schedule (Figure 4D, discussed below), or
when the mice were sated [Figure 4B, F(2,25) = 0.046,
p = 0.96].
We next tested all mice (under food deprivation) in the
concurrent choice task (Cousins and Salamone, 1994).
Mice had a choice between lever pressing (FR30) for a
preferred food (chocolate-flavored 20 mg pellet) or con-
suming a less preferred standard rodent chow that
was freely available on the floor of the operant box.
This setup (‘‘choice’’ condition) was used on days 1, 3,
and 5 of each week. On days 2 and 4, only FR30 was
available (‘‘no choice’’ condition). Before Dox treatment,
there was no difference in performance between groups
either under the choice [Figure 4C, F(2,25) = 0.53, p =
0.60] or no choice [Figure 4D, F(2,25) = 0.78, p = 0.47]
conditions. After Dox treatment, the Dox group lever-
pressed more and earned more pellets in the choice
[Figure 4C, F(2,25) = 7.0, p = 0.004] but not in the no
choice conditions [Figure 4D, F(2,25) = 3.3, p = 0.052].
Moreover, in the choice condition, the Dox group
showed a greater preference than wild-type controls
for lever pressing for pellets [Figure 4E, F(2,25) = 6.3,
p = 0.006]. There was no difference in total food con-
sumed [not shown, F(2,25) = 1.2, p = 0.32]. Thus, DAT
knockdown mice showed enhanced motivation for the
preferred reward, but not an enhanced appetite for
food in general.
To ensure that the above differences were not due to
contamination by learning during the testing week, we
analyzed day 1 results from the PR7 food-restricted
condition separately (since this was the first session
after Dox). It is clear that the Dox group displayed
more lever pressing compared with all other groups
[Figure 4F, F(2,25) = 3.6, p = 0.04].
Induced Knockdown of DAT Expression
Does Not Affect Learning
The above studies allowed us to isolate performance-
scaling effects of DA from learning effects by training
our subjects prior to inducing changes in dopaminergic
tone. To directly test the effect of elevated DA on learn-
ing, we used Pavlovian conditioning with the above
three groups of mice (inducible DAT knockdown treated
with Dox, inducible DAT knockdown not treated with
Dox, and wild-type treated with Dox). Mice were pre-
sented with a cue light that lasted for 12 s followed by
food pellet delivery (unconditioned stimulus; US) at the
offset of the light. The auditory cue from the pellet drop-
ping and the light together represent a compound con-
ditioned stimulus (CS). We assessed the acquisition of
Pavlovian association between the CS and US with the
conditioned response (CR), which was head entry in the
feeder. Figure 4G represents head entries during the pre-
sentation of the light (0 to 12 s, 2 s bins) and after the
presentation of the light (12 to 24 s, 2 s bins) in the last
Figure 3. Action Potential Activity of VTA
(A) Mutant mice fed with Dox (five mice, 19
cells) showed higher dopamine neuron firing
rates when compared with untreated mice
(No Dox, five mice, 20 cells). (B) Representa-
tive 10 s traces. (C) Mutant mice fed with Dox
exhibited a majority of cells with fast (>4 Hz)
firing rates compared with untreated mice.
(D) Mutant mice fed with Dox show enhanced
nonbursting activity of dopamine neurons
compared with untreated mice (*p < 0.05).
Error bars = SEM.
Table 1. Induced Knockdown of DAT Does Not Affect the Quantity or Characteristics of Bursts
Firing activity Amount of burstingCharacteristics of bursts
frequency (Hz)Spikes/burst Burst duration (ms)
3.06 6 0.41
4.20 6 0.31*
3.07 6 0.43
4.35 6 0.31*
17.45 6 5.45
20.99 6 4.71
4.17 6 1.99
4.35 6 1.37
0.28 6 0.10
0.37 6 0.08
2.43 6 0.16
2.47 6 0.10
92.06 6 15.20
100.69 6 9.24
18.09 6 1.24
17.06 6 0.75
For firing activity and amount of bursting: n = 20 (No Dox) and n = 19 (Dox). For characteristics of bursts, only cells showing bursting activity were
considered: n = 15 (No Dox) and n = 17 (Dox) (*p < 0.04).
Tonic Dopamine Scales Performance
session. We found very few head entries during the 12 s
cue light. All groups showed high levels of head entries
at the occurrence of the auditory cue, indicating that
they learned to associate the CS with the US. Notably,
mutant mice on Dox showed significantly more head en-
tries throughout sessions [F(2,25) = 4.2, p = 0.026], con-
sistent with elevated motivation. However, the pattern
and timing of CR were identical between the three
groups; that is, all groups discriminated between CS
and non-CS. To examine learning, we focused on those
head entries occurring in the 2 s bin following the CS (as
percentage of head entries during 12 s cue light and 12 s
post-cue) across training sessions. Although the audi-
ing earlier operant tasks was associated with reward,
low initial rates of CR and the clear acquisition curve in-
dicate that prior experience in operant tasks transferred
little to this new task. There was no genotype difference
in the acquisition curve [Figure 4H, group X session in-
teraction F(20,210) = 0.926, p = 0.55], indicating that all
animals learned the task equally.
Here we report a genetically altered mouse line in which
DAT activity can be virtually eliminated by the adminis-
tration of Dox, resulting in reduced DA reuptake, in-
creased tonic firing of DA neurons, and unaltered phasic
firing. By training mice prior to inducing the above
dopaminergic tone on the expression of an already
learned behavior. Thus, we show that alterations in DA
transmission can scale the performance of a previously
learned behavior in the absence of new learning.
The performance-scaling effects of DA were food
deprivation dependent. Increased response due to ele-
vated DA was only observed in deprived, not sated,
animals. Thus, elevated DA had no impact on behavior
in the absence of pronounced motivational drive. This
highlights the importance of considering DA in the con-
incentive motivation model(Bindra, 1974)andthe incen-
tive saliencemodelofBerridge andRobinson (Robinson
and Berridge, 1993; Berridge and Robinson, 2003),
physiological state interacts with conditioning to pro-
duce incentive motivation, with the deprivation state
essentially multiplying the incentive value of relevant
reward cues. Our data indicate that dopaminergic tone
does not simply modulate sensitivity to food deprivation
or appetite. Rather, it modulates the performance/de-
ployment of learned behaviors to satisfy motivational
The choices made in a PR schedule (to continue to le-
ver-press) and in the concurrent choice task (to work for
behaviors. Our data do not address the role of DA in
scaling habitual behavior. It has been demonstrated
that habitual behaviors are less sensitive to motivational
state (Balleine and Dickinson, 1998; Dickinson and Bal-
leine, 2002; Yin et al., 2004). Interestingly, using an ex-
tinction-reinstatement paradigm with rats overtrained
to traverse a runway, Ettenberg et al. (Ettenberg and
Horvitz, 1990; Horvitz and Ettenberg, 1988) report that
DA blockade has no immediate effect on performance.
The insensitivity of habits to motivational state may ren-
der themless amenable to scaling by DA, which was fur-
ther demonstrated in a recent study (Choi et al., 2005).
This is consistent with the present findings that in the
absence of motivational drive (e.g., sated animals), DA
does not scale learned behavioral responses.
Figure 4. Induced Hyperdopaminergic Tone Scales Performance in
the Absence of Learning
Wild-type and mutant mice were tested in a PR7 schedule first be-
fore, then after, Dox treatment (WT: No Dox / Dox and Inducible:
No Dox / Dox, respectively). A third mutant group was never
treated with Dox (Inducible: No Dox / No Dox). Before Dox treat-
ment, there was no difference between any groups in PR7 perfor-
mance either under food deprivation (A) or no food deprivation (B).
Dox treatment of mutant mice significantly enhanced PR7 perfor-
mance under food deprivation ([A], p = 0.009) but not under no
food deprivation (B).In the concurrent choicetask, before Dox treat-
under the choice (C) or no choice (D) conditions. Dox treatment of
mutant mice resulted in more lever-presses under the choice ([C],
p=0.004),butnot underthe no choice, condition ([D], p= 0.052). Un-
der the choice condition, Dox-treated mutant mice also showed
olate-flavored pellets ([E], p = 0.006). The above difference was not
due to contamination by learning during the testing week as indi-
cated byday1results (F).Dox-treated mutant micedisplayedhigher
duced hyperdopaminergic tone affects reinforcement learning, the
acquisition of Pavlovian association between a CS (the light cue
and the auditory cue from the food pellet drop) and the US was as-
sessed with the conditioned response (head entry in the feeder).
High levels of head entries at pellet delivery indicate that all mice
learned the task (G). Analysis of head entries occurring in the 2 s
bin following the CS across training sessions indicates that there
was no genotype difference in the acquisition curve (H) (*p < 0.05;
**p < 0.01). Error bars = SEM.
Although tonic DA exhibits performance-scaling ef-
fects without altering reinforcement learning, DA may
contribute to reinforcement learning through phasic ac-
tivity, which is unaltered in our study. Phasic DA release
provides the temporal resolution necessary to represent
the contingencies in reinforcement learning (Schultz,
2002). Our data are consistent with the differential as-
signment of learning and performance-scaling functions
to phasic and tonic activity, respectively. Alternatively,
recent studies by the Palmiter group (Cannon and Pal-
miter, 2003; Hnasko et al., 2005) suggest that DA may
not be as critical to reinforcement learning as generally
Acute inhibition of DAT and the resulting elevation in
extracellular DA may activate DA autoreceptors and de-
crease DA neuron firing (Lacey et al., 1987); however,
chronically reduced DAT and elevated DA may have
very different effects. Elevated DA neuron firing arising
from reduced DAT has been observed previously in
the constitutive DAT knockout mice, and it was specu-
lated that this might be due to the lack of D2 autorecep-
tor function in these mice (Gainetdinov et al., 1998;
Jones et al., 1999). However, we did not find any change
in D2 autoreceptor expression with induced DAT knock-
down, suggesting that downregulation of D2 autorecep-
tors is unlikely to mediate the increased tonic activity
observed in the present study. Recently, chronic activa-
tion of D2 autoreceptors was implicated in increasing
tonic DA activity, with downregulation of A-type K+
channels as a potential mechanism (Hahn et al., 2006).
Alternatively, changes in afferent control of DA activ-
ity may underlie the observed increased tonic firing
(Marinelli et al., 2006; Floresco et al., 2003).
Our results highlight the power of genetic approaches
for independently manipulating phasic and tonic DA
function. Traditional pharmacological approaches have
laid a critical foundation, demonstrating that DA block-
ade or depletion decreases animals’ willingness to
work for food reward (Cousins and Salamone, 1994; Sal-
amone and Correa, 2002). However, similar approaches
have also supported the competing reinforcement
learning hypothesis. When DA function is impaired or
sition of learned responses (Wise and Schwartz, 1981)
and maintenance of reinforced behaviors (McFarland
and Ettenberg, 1995; Wise et al., 1978).
Pharmacological manipulations will necessarily alter
both tonic and phasic DA transmission. If DA is impor-
tant for both learning and motivation, drug manipula-
tions will induce alterations in both. Moreover, an acute
drug challenge initiates a cascade of acute, dynamic
physiological events, compounding the difficulty of as-
signing any observed behavioral effect to alterations in
DA per se. Most drugs are also promiscuous in their
molecular targets; e.g., amphetamine acts on the sero-
tonin and norepinephrine transporters in addition to the
DA transporter (Wall et al., 1995). Behaviorally, many
drugs that target the monoamine systems are them-
selves potent stimuli, which may become a confounding
factor in Pavlovian or instrumental conditioning studies
based on stimulus control of conditioned responses.
The inducible DAT knockdown mice, in contrast, are
in a sustained and stable state at the time of behavioral
The controversy between the reinforcement learning
and motivational hypotheses of DA, as well as conflict-
ical approaches outlined above. The predominant agent
for elevating DA is amphetamine, which has been vari-
ously reported to decrease PR responding (Caul and
decrease response at low doses but increase it at high
doses (Mobini et al., 2000), or increase response at low
doses but decrease it at high doses (Mayorga et al.,
2000). More precisely targeted genetic manipulations
will complement pharmacological approaches and
help arbitrate competing interpretations and conflicting
In the present study, using animals with a sustained
alteration in tonic, but not phasic, DA activity, we show
that DA can scale the performance of a previously
learned behavior in the absence of new learning. If DA
also plays a critical role in learning, it may be that the
and learning processes, providing a mechanism to inte-
recent theoretical work, it was suggested that tonic DA
may underlie a learned average reward value for a given
environment or context (Niv et al., 2005; see also
McClure et al., 2003). This expected average reward, in
turn,reflects anopportunity costfor inaction andconse-
quently serves to establish response vigor. The present
study provides an initial step toward exploring empiri-
cally how DA may serve to integrate motivation and
Generation of Inducible DAT Knockdown
Construction of DAT genomic DNA for gene targeting is described
elsewhere (Zhuang et al., 2001). The genetic strategy is described
in the results section and Figure 1A. 129/SvJ ES cells (Specialty Me-
dia) were used. Male chimeras were mated with C57BL6/J females.
PGK-neo (floxed) in DAT-tTA was deleted by the germline deleter
EIIa-cre. DAT-tTA/+ and tetO-DAT/+ mice were mated to obtain
DAT-tTA/tetO-DAT (inducible DAT knockdown) and wild-type litter-
mates. For electrophysiology and electrochemistry experiments,
knockdown of DAT expression was achieved by replacing regular
rodent chow with rodent chow mixed with 200 mg Dox per kg
food (Bio-Serv). For the behavior experiments, knockdown of DAT
expression was achieved by replacing water with Dox water (0.4
mg/ml in 5% sucrose solution and 5% sucrose solution for control).
food were different and could affect behavior experiments using
food reward. All animal procedures were approved by the Institu-
tional Animal Care and Use Committee at The University of Chicago.
Mouse brains werelysed. Fifty micrograms of protein was subjected
to SDS-PAGE and transferred to polyvinylidene fluoride (PVDF)
membrane. Nonspecific sites were blocked with 5% nonfat dry
milk. Membranes were incubated with rat anti-mouse DAT antibody
(1:1000, Chemicon) in TBS with 2% nonfat dry milk. Signals were de-
(ICN) and enhanced chemiluminescence (Pierce). b-actin antibody
was used to confirm equal sample loading.
Animals were transcardially perfused with 4% paraformaldehyde.
Brains were equilibrated in 30% sucrose. Frozen sections (50 mm)
were blocked with normal goat serum and incubated with rat anti-
mouse DAT antibody (1:10,000, Chemicon) over two nights at 4?C.
Tonic Dopamine Scales Performance
Immunoreactivity was visualized with the ABC method (Vector) and
We used 50 pM [3H]-spiperone (NEN) and 100 nM ketanserin in
50 mM Tris buffer containing (in mM) 120 NaCl, 5 KCl, 2 CaCl2, and
1 MgCl2(pH 7.4). Nonspecific binding was determined by adding
1 mM spiperone. After 6 week exposure to BioMax MS film, optical
density was quantified using NIH Image and compared to 3H stan-
dards (American Radiolabeled Chemicals).
Mice were decapitated; brains were removed into cold, sucrose-ar-
tificial cerebrospinal fluid (ACSF) containing the following (in mM):
200 sucrose, 25 NaHCO3, 20 glucose, 10 ascorbic acid, 2.5 KCl,
2.5 CaCl2, 1 MgCl2, and 1 NaH2PO4(pH 7.4), saturated with 95%
O2and 5% CO2. Coronal slices 250 mm thick were prepared with a
vibratome (VT100S, Leica). Slices were incubated for 1 hr in bath cir-
NaHCO3, 20 glucose, 2.5 KCl, 2.5 CaCl2, 1 MgCl2, 1 NaHCO3, and 1
ascorbic acid (pH 7.4), saturated with 95% O2and 5% CO2. For re-
cording, slices were perfused (2 ml/min) with this same 32?C ACSF
without ascorbic acid.
Carbon fiber recording electrodes were placed in the NAc shell
w150 mm from a bipolar stimulating electrode with a 250 mm tip
separation. DA release was evoked by single-pulse stimulations
(400 mA, 1 ms) delivered every 2 min. Currents were recorded using
an Axopatch 200B amplifier with a DigiData 1200 interface and
pCLAMP 8 software (Axon Instruments). For FSCV, the electrode
voltage was ramped from 2400 mV to +1000 mV and then back at
200 V/s at 100 ms intervals. Current was filtered at 10 kHz and digi-
tionofthe oxidized substance andforcalibrationwith5mMDAatthe
end of the experiment. For amperometry, a constant voltage of +400
mV was applied. Amperometric traces were filtered at 1 kHz, digi-
tized at 2.5 kHz, and digitally filtered at 100 Hz.
Extracellular Single-Unit Recordings of DA Neurons
All mice were naı ¨ve to behavioral testing. Dox exposure was >8
weeks for the Dox group. Mice were anesthetized with choloral hy-
drate and neurons were recorded in the VTA (0.4–1.0 AP, 0.2–0.6
L, and 4.5–5.5 V mm from brain surface) as previously described
(Mathon et al., 2005; White and Wang, 1984).
Signals were recorded using a Fintronics amplifier with a DigiData
1200 interface and axoscope software (Axon Instruments). DA cells
were identified according to standard physiological criteria (Grace
and Bunney, 1983, 1984b). This included a triphasic (+/2/+) wave
form with >2.5 ms duration from start to end (at 400 Hz to 0.5 kHz,
White and Wang, 1984) and >1.1 ms from start to trough of negative
peak (at 300 Hz to 0.8 kHz, Ungless et al., 2004).
Data were analyzed as follows. Firing rate: total number of spikes
over time. Bursts: clusters of spikes occurring at high frequencies,
with interspike interval <80 ms at start and >160 ms at end of burst
(Grace and Bunney, 1984a). Tonic firing rate: firing activity with
bursts subtracted (Mathon et al., 2005).
All experiments were carried out during the light period (06:00–
18:00). The same mice were used in the following three behavioral
tests in the order described. All tests were conducted in mouse op-
erant conditioning chambers that have two retractable levers,
poke hole on the back wall, and a feeder with photobeam (Med
Associates). Twenty milligram chocolate flavored pellets (Bio-Serv)
were used as reinforcers. For behavioral effects of DAT knockdown,
mice were treated for at least 8 weeks with Dox water.
In the PR operant task, mice were first trained under a fixed ratio 1
mice reached a criterion of 30 lever presses in less than 45 min on
two consecutive days, they were shifted to a PR7 schedule. Two
parameters were recorded: the breakpoint and number of lever
presses on the active and inactive levers. For breakpoint, we used
two commonly employed criteria: (1) no active lever press for
5 min (built into the operant program) and (2) no reinforcement for
5 min (post hoc analysis). In the 45 min session time, many mutant
mice treated with Dox did not reach breakpoint with either criterion;
consequently, we focused our analysis on total lever presses.
In the concurrent choice task, under food deprivation, mice had
the choice between lever pressing (FR30 in 30 min session) for
a more preferred food (chocolate-flavored 20 mg pellet) or consum-
ing a less preferred, standard rodent chow that was concurrently
and freely available on the floor of the operant box. This choice con-
dition schedule was used on day 1, 3, and 5 of each week; and on
days 2 and 4, only the no choice condition FR30 was available. Test-
ing lasted for 3 weeks.
In Pavlovian associative learning, mice were habituated to the
conditioning chambers for 15 min on two consecutive days, during
which four food pellets were placed in the feeders. All animals ate
the pellets by the end of the second session. Mice were then trained
for 11 days with 20 daily trials (180 s variable intertrial interval). In
each trial, mice were exposed to a 12 s illumination of the signal light
on the back wall, followed by a single 20 mg food pellet. Interrup-
tions of a feeder photobeam indicated magazine entries.
This work was supported in part by NIMH MH66216 (X.Z.), NIDA
DA015918 (D.M.), Tourette Syndrome Association Postdoctoral
Fellowship (B.C.), NIDA F32 DA020427-01 (J.A.B.), and NIDA T32
DA07255-13 (J.P.B.). We thank Peter Balsam, Kent Berridge, Natha-
niel Daw, Peter Dayan, Jon Horvitz, and Yael Niv for critical reading
of the manuscript. We thank Lindsay Cotterly, Khalid Fakhro, Ali
Hussain, and Wanhao Chi for technical assistance.
Received: February 20, 2006
Revised: June 22, 2006
Accepted: July 27, 2006
Published: September 6, 2006
Balleine, B.W., and Dickinson, A. (1998). Goal-directed instrumental
action: contingency and incentive learning and their cortical sub-
strates. Neuropharmacology 37, 407–419.
Berridge, K.C., and Robinson, T.E. (2003). Parsing reward. Trends
Neurosci. 26, 507–513.
Bindra, D. (1974). A motivational view of learning, performance, and
behavior modification. Psychol. Rev. 81, 199–213.
Bond, C.T., Sprengel, R., Bissonnette, J.M., Kaufmann, W.A., Prib-
now, D., Neelands, T., Storck, T., Baetscher, M., Jerecic, J., Maylie,
J., et al. (2000). Respiration and parturition affected by conditional
overexpression of the Ca2+-activated K+ channel subunit, SK3.
Science 289, 1942–1946.
Cannon, C.M., and Palmiter, R.D. (2003). Reward without dopamine.
J. Neurosci. 23, 10827–10831.
Caul, W.F., and Brindle, N.A. (2001). Schedule-dependent effects of
haloperidol and amphetamine: multiple-schedule task shows
within-subject effects. Pharmacol. Biochem. Behav. 68, 53–63.
Choi, W.Y., Balsam, P.D., and Horvitz, J.C. (2005). Extended habit
training reduces dopamine mediation of appetitive response ex-
pression. J. Neurosci. 25, 6729–6733.
Cousins, M.S., and Salamone, J.D. (1994). Nucleus accumbens do-
pamine depletions in rats affect relative response allocation in
a novel cost/benefit procedure. Pharmacol. Biochem. Behav. 49,
eration of Motivational Systems. In Steven’s Handbook of Experi-
mental Psychology, Third Edition (New York: John Wiley & Sons),
Ettenberg, A., and Horvitz, J.C. (1990). Pimozide prevents the re-
sponse-reinstating effects of water reinforcement in rats. Pharma-
col. Biochem. Behav. 37, 465–469.
Floresco, S.B., West, A.R., Ash, B., Moore, H., and Grace, A.A.
(2003). Afferent modulation of dopamine neuron firing differentially
regulates tonic and phasic dopamine transmission. Nat. Neurosci.
Gainetdinov, R.R., Jones, S.R., Fumagalli, F., Wightman, R.M., and Download full-text
Caron, M.G. (1998). Re-evaluation of the role of the dopamine trans-
porter in dopamine system homeostasis. Brain Res. Brain Res. Rev.
Gossen, M., and Bujard, H. (1992). Tight control of gene expression
in mammalian cells by tetracycline-responsive promoters. Proc.
Natl. Acad. Sci. USA 89, 5547–5551.
Grace, A.A., and Bunney, B.S. (1983). Intracellular and extracellular
electrophysiology of nigral dopaminergic neurons–1. Identification
and characterization. Neuroscience 10, 301–315.
Grace,A.A., and Bunney, B.S. (1984a). The control of firing pattern in
nigral dopamine neurons: burst firing. J. Neurosci. 4, 2877–2890.
Grace,A.A.,and Bunney, B.S. (1984b). The control of firing pattern in
nigral dopamine neurons: single spike firing. J. Neurosci. 4, 2866–
Gross, C., Zhuang, X., Stark, K., Ramboz, S., Oosting, R., Kirby, L.,
Santarelli, L., Beck, S., and Hen, R. (2002). Serotonin1A receptor
acts during development to establish normal anxiety-like behaviour
in the adult. Nature 416, 396–400.
Hahn, J., Kullmann, P.H., Horn, J.P., and Levitan, E.S. (2006). D2 au-
toreceptors chronically enhance dopamine neuron pacemaker ac-
tivity. J. Neurosci. 26, 5240–5247.
Hnasko, T.S., Sotak, B.N., and Palmiter, R.D. (2005). Morphine re-
ward in dopamine-deficient mice. Nature 438, 854–857.
Science 134, 943–944.
Horvitz, J.C., and Ettenberg, A. (1988). Haloperidol blocks the re-
sponse-reinstating effects of food reward: a methodology for sepa-
rating neuroleptic effects on reinforcement and motor processes.
Pharmacol. Biochem. Behav. 31, 861–865.
Jones, S.R., Gainetdinov, R.R., Hu, X.T., Cooper, D.C., Wightman,
R.M.,White, F.J.,andCaron,M.G.(1999).Loss ofautoreceptorfunc-
tions in mice lacking the dopamine transporter. Nat. Neurosci. 2,
D2 receptors to increase potassium conductance in neurones of the
rat substantia nigra zona compacta. J. Physiol. 392, 397–416.
Lucas, J.J., and Hen, R. (1995). New players in the 5-HT receptor
field: genes and knockouts. Trends Pharmacol. Sci. 16, 246–252.
Marinelli, M., Rudick, C.N., Hu, X.T., and White, F.J. (2006). Excitabil-
ity of dopamine neurons: modulation and physiological conse-
quences. CNS Neurol. Disord. Drug Targets 5, 79–97.
Mathon, D.S., Ramakers, G.M., Pintar, J.E., and Marinelli, M. (2005).
Decreased firing frequency of midbrain dopamine neurons in mice
lacking mu opioid receptors. Eur. J. Neurosci. 21, 2883–2886.
Mayorga, A.J., Popke, E.J., Fogle, C.M., and Paule, M.G. (2000).
Similar effects of amphetamine and methylphenidate on the perfor-
mance of complex operant tasks in rats. Behav. Brain Res. 109, 59–
McClure, S.M., Daw, N.D., and Montague, P.R. (2003). A computa-
tional substrate for incentive salience. Trends Neurosci. 26, 423–
McFarland, K., and Ettenberg, A. (1995). Haloperidol differentially af-
fects reinforcement and motivational processes in rat running an al-
ley for intrvenous heroin. Psychopharmacology (Berl.) 122, 346–350.
Mirenowicz, J., and Schultz, W. (1994). Importance of unpredictabil-
ity for rewardresponsesinprimate dopamineneurons.J.Neurophy-
siol. 72, 1024–1027.
Mobini, S., Chiang, T.J., Ho, M.Y., Bradshaw, C.M., and Szabadi, E.
(2000). Comparison of the effects of clozapine, haloperidol, chlor-
promazine and d-amphetamine on performance on a time-con-
strained progressive ratio schedule and on locomotor behaviour in
the rat. Psychopharmacology (Berl.) 152, 47–54.
Montague, P.R., Dayan, P., and Sejnowski, T.J. (1996). A framework
for mesencephalic dopamine systems based on predictive Hebbian
learning. J. Neurosci. 16, 1936–1947.
Niv,Y., Daw,N.D., andDayan, P.(2005).How fast towork: Response
vigor, motivation and tonic dopamine. In NIPS 18, Y. Weiss,
B. Scho ¨lkopf, and J. Platt, eds. (Cambridge, MA: MIT Press),
Robinson, T.E., and Berridge, K.C. (1993). The neural basis of drug
craving: an incentive-sensitization theory of addiction. Brain Res.
Brain Res. Rev. 18, 247–291.
Salamone, J., and Correa, M. (2002). Motivationial views of rein-
forcement: implications for understanding the behavioral functions
of nucleus accumbens dopamine. Behav. Brain Res. 137, 3–25.
Schultz, W. (2002). Getting formal with dopamine and reward. Neu-
ron 36, 241–263.
Schultz, W., Apicella, P., and Ljungberg, T. (1993). Responses of
successive steps of learning a delayed response task. J. Neurosci.
Schultz, W., Dayan, P., and Montague, P.R. (1997). A neural sub-
strate of prediction and reward. Science 275, 1593–1599.
of dopamine neurons in the ventral tegmental area by aversive stim-
uli. Science 303, 2040–2042.
Wall, S.C., Gu, H., and Rudnick, G. (1995). Biogenic amine flux medi-
ated by cloned transporters stably expressed in cultured cell lines:
amphetamine specificity for inhibition and efflux. Mol. Pharmacol.
White, F.J., and Wang, R.Y. (1984). A10 dopamine neurons: role of
autoreceptors in determining firing rate and sensitivity to dopamine
agonists. Life Sci. 34, 1161–1170.
Wiley, J.L., and Compton, A.D. (2004). Progressive ratio perfor-
mance following challenge with antipsychotics, amphetamine, or
NMDA antagonists in adult rats treated perinatally with phencycli-
dine. Psychopharmacology (Berl.) 177, 170–177.
Wise, R.A. (2004). Dopamine, learning and motivation. Nat. Rev.
Neurosci. 5, 483–494.
Wise, R.A., and Schwartz, H.V. (1981). Pimozide attenuates acquisi-
tion of lever-pressing for food in rats. Pharmacol. Biochem. Behav.
Wise, R.A., Spindler, J., and Legault, L. (1978). Major attenuation of
food reward with performance-sparing doses of pimozide in the rat.
Can. J. Psychol. 32, 77–85.
Yin, H.H., Knowlton, B.J., and Balleine, B.W. (2004). Lesions of dor-
solateral striatum preserve outcome expectancy but disrupt habit
formation in instrumental learning. Eur. J. Neurosci. 19, 181–189.
Zhuang, X., Oosting, R.S., Jones, S.R., Gainetdinov, R.R., Miller,
G.W., Caron, M.G., and Hen, R. (2001). Hyperactivity and impaired
response habituation in hyperdopaminergic mice. Proc. Natl.
Acad. Sci. USA 98, 1982–1987.
Tonic Dopamine Scales Performance