significantly between treatments (Fig. 3C;H3,194=
17.47,P< 0.0001). Bees with elevated dopa-
mine levels learned better than the two control
groups (Fig. 3C;P< 0.05) and the group treated
with flupentixol (Fig. 3C;P< 0.01). Treatment
with flupentixol tended to reduce discrimina-
tion at the end of training (fig. S3D), but the
corresponding score did not differ from that of
controls.
Finally, we determined whether manipulat-
ing dopamine brain levels would affect retrieval
of an appetitive memory evoked 1 hour after the
last conditioning trial. After conditioning for-
agers to discriminate the same previous two
odorants, individuals that responded to the
positive but not to the negative odorant in
the last conditioning trial (fig. S4) were cold-
anesthetized and assigned to different topicaltreatments (dopamine, flupentixol, or DMF).
An additional group remained untreated (sham)
(see SM for more details). Bees were then tested
with both odorants, and the percentage of bees
exhibiting specific memory [i.e., responding
to the positive but not to the negative odorant
( 30 )] was quantified for each group. This per-
centage varied significantly with the treat-
ment (Fig. 3D;c^2 = 12.47, df: 3,P< 0.005), as
the group topically exposed to dopamine had
a better retrieval performance than the other
groups, which exhibited no differences in re-
trieval (c^2 = 4.05, df: 2, not significant). Im-
portantly, altering serotonin levels affected
neither sucrose responsiveness (fig. S5A) nor
olfactory appetitive learning (fig. S5B). Thus,
the changes induced in appetitive responsive-
ness, learning, and memory are specific to
dopaminergic signaling and point to a driving
role for dopamine in an individual wanting
system.
Our results suggest that distinct food-related
motivational and communication states deter-
mine different levels of dopamine in the brain
of honey bee foragers, consistent with the
existence of a dopamine-based wanting sys-
tem activated by both colony and individual
appetitive needs. Foragers driven by an en-
hanced motivation for food exhibit higher
dopamine levels. Accordingly, blockade of
dopaminergic signaling results in a decrease
of foraging activities, consistent with the in-
hibition of a wanting system mediated by
dopaminergic signaling ( 1 , 4 , 5 ). Dopamine
levels also increase when foragers report dis-
tance and direction of a profitable food source
by means of the waggle dance and upon de-
parture toward the appetitive goal. This finding
suggests that dancers do not only report the
vectorial information leading to the food source
but also recollect transiently the appetitive
properties of the food source at the start of
dancing and when leaving the hive, thereby
elevating dopamine levels in their brain. The
waggle dance is thus more than an innate,
automatic behavior ( 7 ), as it sets the frame-
work for a dancer’s transient evocation of
food-source properties.
In addition to this socially embedded want-
ing system, individual hunger also leads to
an increase in the level of dopamine in the
forager brain. Accordingly, artificial elevation
of dopamine increases the incentive value of
sucrose solution and individual appetitive re-
sponsiveness and improves olfactory learning
and memory retrieval. These results are consist-
ent with alliesthesia modulation of incentive
value by relevant physiological state ( 31 ) as, for
instance, the improvement of memory retrieval
upon dopamine elevation and in the absence
of reward indicates that the incentive salience
of a learned odor can be enhanced through an
activation of the wanting system. Honey bees,
unlike some mammals, are strictly eusocialSCIENCEscience.org 29 APRIL 2022•VOL 376 ISSUE 6592 511
Fig. 3. An individual dopamine-dependent
wanting system drives both appetitive
responsiveness and appetitive learning and
memory in honey bees.(A) Increased starvation
enhances dopamine levels in the brain of
foragers. Dopamine (DA) and serotonin (5-HT)
measurements were performed in harnessed
23-day-old bees after immobilizing them and
subjecting them to starvation periods of 1
or 2 hours (DA1h:n= 42; DA2h:n= 41; 5-HT1h:
n= 39; 5-HT2h:n= 41). Increased hunger
translates into higher DA levels in the bee brain.
Lowercase letters (a, b, and c) above the bars
indicate significant differences between groups
(Tukey tests,P< 0.05). (B) Artificial increase
of DA levels in honey bee foragers enhances
their appetitive responsiveness to sucrose
solution. Honey bee foragers captured at a feeder
upon landing and before they started feeding
were fed 30ml of 1.0 M sucrose solution and
then topically exposed on the thorax to one of two
doses of either dopamine (DA 1 :n= 29; DA 2 :
n= 30) or 6,7-ADTN (ADTN 1 :n= 34; ADTN 2 :n= 38).
Control bees were treated with DMF (n= 47) or remained untreated (sham;n= 47). A sucrose responsiveness
score (SRS) was computed for each bee subjected to an increasing series of six sucrose concentrations
on the basis of its appetitive proboscis extension response (PER) in response to these stimulations. Higher SRSs
indicate higher appetitive motivation. The highest dose of both DA and 6,7-ADTN significantly increased SRSs,
thus revealing a higher appetitive motivation. For each treatment, the figure shows the median, 10th, 25th, 75th,
and 90th percentiles as box plots with error bars; outliers are shown as 5th and 95th percentiles; the blue line
within each box plot shows the mean. Lowercase letters (a and b) above the bars indicate significant differences
between groups (Kruskal-Wallis test, post hoc mean rank comparisons,P<0.05).(C) Artificial increase of
DA levels in honey bee foragers enhances their appetitive olfactory learning. Honey bee foragers captured at a
feeder upon landing and before they started feeding were harnessed and then topically exposed on the
thorax to either dopamine (n= 48) or flupentixol (n= 54). Control bees were treated with DMF (n= 44) or
remained untreated (sham;n= 48). Bees were trained along six trials to discriminate two odorants, one
rewarded (CS+) and another not rewarded (CS−), using the olfactory conditioning of the PER. An acquisition
score (ACQS) was computed for each bee on the basis of its responses to the conditioned odorants in the last two
CS+ and last two CS−trials. Higher positive ACQSs indicate better learning performances. Bees in which DA
was increased artificially demonstrated improved learning performance and exhibited significantly higher ACQSs.
For each treatment, the figure shows the median, 10th, 25th, 75th, and 90th percentiles as box plots with
error bars; outliers are shown as 5th and 95th percentiles; the blue line within each box plot shows the mean.
Lowercase letters (a and b) above the bars indicate significant differences between groups (Kruskal-Wallis test,
post hoc mean rank comparisons,P<0.05).(D) Artificial increase of DA levels in honey bee foragers
enhances their appetitive olfactory memory retrieval. Honey bee foragers captured at a feeder upon landing
but before they started feeding were harnessed and trained to discriminate two odorants, one rewarded (CS+)
and another not rewarded (CS−), using the olfactory conditioning of the PER. Bees that managed a successful
discrimination in the last conditioning trial (i.e., responding to the CS+ and not to the CS−) were then
topically exposed on the thorax to either dopamine (n= 33) or the flupentixol (n= 33). Control bees were treated
with DMF (n= 33) or remained untreated (sham;n= 32). Memory retrieval was tested 1 hour after the last
conditioning trial and 30 min after the topical exposure by presenting bees with the CS+ and the CS−in the
absence of reward. The percentage of bees exhibiting specific memory (i.e., responding to the CS+ and not to
the CS−) differed between treatments. It was significantly higher in the group treated with DA. Lowercase letters
(a and b) above the bars indicate significant differences between groups (c^2 contingency analysis,P< 0.005).
RESEARCH | REPORTS