Dopamine prediction error signaling in a unique nigrostriatal circuit is critical for associative fear learning

Abstract

Learning by experience that certain cues in the environment predict danger is crucial for survival. How dopamine (DA) circuits drive this form of associative learning is not fully understood. Here, in male mice, we demonstrate that DA neurons projecting to a unique subregion of the dorsal striatum, the posterior tail of the striatum (TS), encode a prediction error (PE) signal during associative fear learning. These DA neurons are necessary specifically during acquisition of fear learning, but not once the fear memory is formed, and are not required for forming cue-reward associations. Notably, temporally-precise inhibition or excitation of DA terminals in TS impairs or enhances fear learning, respectively. Furthermore, neuronal activity in TS is crucial for the acquisition of associative fear learning and learning-induced activity patterns in TS critically depend on DA input. Together, our results reveal that DA PE signaling in a non-canonical nigrostriatal circuit is important for driving associative fear learning.

Full text

Article https://doi.org/10.1038/s41467-025-58382-5
Dopamine prediction error signaling in a
unique nigrostriatal circuit is critical for
associative fear learning
Daphne Zafiri
1,2
, Ximena I. Salinas-Hernández
1,2
,EloahS.DeBiasi
1
,
Leonor Rebelo
1
&SevilDuvarci
1
Learning by experience that certain cues in the environment predict danger is
crucial for survival. How dopamine (DA) circuits drive this form of associative
learning is not fully understood. Here, in male mice, we demonstrate that DA
neurons projecting to a unique subregion of the dorsal striatum, the posterior
tail of the striatum (TS), encode a prediction error (PE) signal during associa-
tive fear learning. These DA neurons are necessary specifically during acqui-
sition of fear learning, but not once the fear memory is formed, and are not
required for forming cue-reward associations. Notably, temporally-precise
inhibition or excitation of DA terminals in TS impairs or enhances fear learning,
respectively. Furthermore, neuronal activity in TS is crucial for the acquisition
of associative fear learning and learning-induced activity patterns in TS criti-
cally depend on DA input. Together, our results reveal that DA PE signaling in a
non-canonical nigrostriatal circuit is important for driving associative fear
learning.
Associative fear learning (‘threat learning’1)─the ability to associate
stimuli with threats ─enables animals to predict and avoid danger and
hence is criticalfor survival. However,learned fear that is excessive can
also be maladaptive and havepathophysiological consequences. Much
evidence indicates that anxiety disorders,such as post-traumatic stress
disorder (PTSD), result from dysregulation of normal fear learning
mechanisms2–4. Therefore, elucidating the neural mechanisms under-
lying fear learning is critical for understanding the pathophysiology of
anxiety disorders and thus has high clinical significance. In the
laboratory, this kind of associative learning is commonly studied using
Pavlovian fear conditioning (FC), in which an initially neutral stimulus
(conditioned stimulus, CS), typically a tone is paired in time with an
aversive unconditioned stimulus (US), such as a mild electrical foot
shock. As the CS-US association is learned, the CS acquires the ability
to elicit fear responses that are associated with the US (such as beha-
vioral freezing) so that it can elicit conditioned fear responses when
later presented alone. Traditionally, the amygdala, particularly its lat-
eral nucleus (LA), has been recognized as the primary brain region for
acquiring CS-US associations during FC5–10. However, recent studies
suggest that plasticity in brain structures beyond the canonical
amygdala circuitry is also involved in acquisition of fear memories. Of
note, the posterior tail of the dorsal striatum (TS) has recently been
indicated in fear conditioning11,12.
TS is a unique subregion within the dorsal striatum that receives a
combination of inputs from structures such as the auditory, visual and
rhinal cortices as well as the amygdala, that sets it apart from other
dorsal striatal subregions13,14. Importantly, TS also receives dopami-
nergic innervation from a unique subpopulation of midbrain dopa-
mine (DA) neurons that are predominantly located in the substantia
nigra lateralis15 (SNL).These DA neurons exhibit a distinct input-output
organization compared to other midbrain DA neurons, project selec-
tively to TS, and do not overlap with the subpopulations that project to
the rest of the striatum, cortex and the amygdala15. Notably, recent
studies have shown that these DA neurons are particularly important
for novelty-induced threat avoidance16,17, and they have also been
shown to be involved in fear learning12. However, the precise role TS-
projecting DA neurons play during associative fear learning is incom-
pletely understood18.
Received: 11 December 2023
Accepted: 13 March 2025
Check for updates
1
Institute of Neurophysiology, Neuroscience Center, Goethe University, Frankfurt, Germany.
2
These authors contributed equally: Daphne Zafiri, Ximena I.
Salinas-Hernández. e-mail: duvarci@med.uni-frankfurt.de
Nature Communications | (2025) 16:3066 1
1234567890():,;
1234567890():,;
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Associative learning is driven by prediction errors (PE) that signal
the discrepancy between predicted and actual outcomes19;andnew
learning happens when outcomes donot match expectations.It is well-
established that ventral midbrain DA neurons, located in the ventral
tegmental area (VTA) and the substantia nigra (SN), encode reward
prediction errors (RPE) which act as teaching signals to drive reinfor-
cement learning20–24. Recent studies have further demonstrated that
ventral midbrain DA neurons encode positive PE signals not only for
rewards, but also for omission of aversive outcomes25–28 to drive
associative learning. Interestingly, dorsal tegmental DA neurons that
project to the amygdala have recently been shown to encode a PE
signal to mediate fear learning29. However, although ventral midbrain
DA neurons have been shown to play critical roles in aversion and fear
learning30–33, whether they contribute to PE signaling necessary for
driving fear learning has remained elusive. Importantly, whether DA
neurons projecting to brain structures outside the canonical amygdala
circuitry encode a PE signal that is required for driving associative fear
learning is largely unknown.
In this study, we investigate the precise role DA projections to TS
play during associative fear learning. By performing measurements of
DA terminal activity as well as DA release in TS, we found that DA
projections to TS encode a PE signal for the aversive outcome during
associative fear learning. Selective lesioning of TS-projecting DA neu-
rons demonstrated that these neurons are important specifically dur-
ing acquisition of fear learning, but not oncethe CS-US association was
learned. Notably, the activity of these DA neurons was required for
associating sensory cues with the aversive US, but not the context with
aversive US or the sensory cues with reward. Conversely, temporally-
precise optogenetic manipulation of DA terminals in TS during FC
demonstrated that this PE signal drives fear learning and boosting this
signal enhances associative fear learning. To gain further insights into
the functional role of TS, we performed Ca2+ recordings of TS activity
and found a PE-like activity pattern, as well as potentiation of CS
responses, during associative fear learning. Bidirectional manipula-
tions of TS activity further showed that the neuronal activity in TS is
required for fear learning. Finally, we demonstrated that DA input was
necessary for the fear learning-induced activity patterns in TS during
FC. Taken together, our results reveal a key role for DA PE signaling in a
unique nigrostriatal circuit for driving associative fear learning.
Results
DA neurons projecting to TS encode a PE signal during asso-
ciative fear learning
In order to investigate the activity of DA neurons projecting to TS in a
projection-specific manner, we used fiber photometry to measure
activity-dependent Ca2+ signals at the terminals of DA neurons in TS.
We injected a Cre-dependent adeno-associated virus (AAV) expressing
the genetically encoded Ca2+ indicator GCaMP in SN of transgenic mice
expressing Cre recombinase under the control of the dopamine
transporter (DAT) promoter (DAT-Cre mice; Fig. 1a). In these mice, the
expression of Cre is highly selective for DA neurons34.Inlinewiththis,
we observed a high degree of overlap between Cre-dependent
GCaMP6m expression and immunohistochemical staining against
tyrosine hydroxylase (TH; Fig. 1c) in the SN. An optical fiber implanted
in TS (Fig. 1a, b, Supplementary Fig. 1) enabled recording of Ca2+
transients in the axon terminals of DA neurons (Fig. 1d). In addition, to
test whether the observed changes in fluorescence reflect neuronal
activity we injected a Cre-dependent AAV expressing the control
fluorophore EYFP in a separate group of control mice. Transient fluc-
tuations in fluorescence were absent in mice expressing the control
fluorophore (n= 2, Supplementary Fig. 2a, b), consistent with our
previous results25,28.
To examine DA terminal activity during associative fear learning,
mice (n= 13) were trained in a FC paradigm (Fig. 1e) where a tone (CS)
was paired with an aversive foot shock (US) on day 2, following a tone
habituation session (Hab) on day 1.On day 3, mice received a fear recall
session consisting of CS presentations in the absence of the aversive
US. During the course of FC, freezing to the CS gradually increased
(Fig. 1f), indicating that the mice learned the association between the
CS and the US. Mice showed significantly higher freezing levels to the
CS at the end of FC session, as well as during fear recall, compared to
Hab (Fig. 1f). Notably, the activity of DA terminals in TS appeared to
resemble a PE signal during associative fear learning. In the beginning
of FC, DA terminals in TS showed strong excitation to the aversive US
which decreased during the course of conditioning (Fig. 1g–i), as the
CS-US association was learned and the CS came to predict the occur-
rence of the US. Indeed, there was a significant decrease in US
responses from the first to the last US (Fig. 1i). Conversely, while CS
responses were absent at the beginning of FC, they gradually increased
through the course of conditioning (Fig. 1g–i), mirroring the increase
in freezing to the CS (Fig. 1f). In line with the behavioral results,
responses to the CS were significantly larger during fear recall com-
pared to Hab session (Fig. 1j), indicating that the CS responses were
potentiated as the animalslearned the CS-US association. In almost all
animals, responses to the CS were larger during fear recall compared
to Hab session (Fig. 1k). Notably, mice expressing a control fluor-
ophore (EYFP) did not reveal any changes in fluorescence throughout
the FC protocol (Supplementary Fig. 2e). Together, these results
demonstrated that DA terminals in TS exhibited a PE-like activity pat-
tern during associative fear learning.
Because the presentation of the aversive US itself could result in a
nonspecific enhancement of CS responses, we next asked whether
potentiation of CS responses during FC was indeed a result of asso-
ciative learning and depended on the temporally contingent pre-
sentations of the CS and the US. To address this question, mice
underwent an unpaired training paradigm where they received the
same number of CS and US presentations as in FC, but the CS and the
US were explicitly unpaired in time during the training session onday 2
(Fig. 2a). In mice undergoing the unpaired training, wedid not observe
an increase in freezing to the CS between Hab, training and testing
sessions (Fig. 2b). Mirroring thesebehavior results, we also did not find
asignificant difference in the CS response between the Hab and the
testing session(n=7;p= 0.31, signed-rank test Fig. 2c–e). These results
indicate that potentiation of CS responses as well as the increased
behavioral fear responses to the CS during FC required the association
of the CS with the aversive US.
If DA terminals in TS encode a PE signal for the aversive US, we
expect that responses to unpredicted USs should be larger in magni-
tude compared to responses to predicted USs. To test this, we exam-
ined DA terminal activity while previously well-trained mice received
presentations of predicted (CS-US pairings) and unpredicted (US only)
foot shocks (Fig. 2f). We indeed found stronger responses to unpre-
dicted USs compared to CS-predicted ones (n=9;p=0.0078,signed-
rank test; Fig. 2g, h), consistent with the decrease in US responses that
we observed during the course of FC. Together, these results indicate
that TS-projecting DA neurons signal a PE for aversive outcomes.
Our results show that TS-projecting DA neurons exhibit a positive
prediction error for aversive stimuli. One question is whether these DA
neurons exhibit negative prediction errors and respond to omission of
aversive outcomes. To address this, we performed a partial con-
ditioning task inwhich the aversive US was omitted randomly in half of
the trials (Supplementary Fig. 3a). We found that DA terminalsin TS did
not exhibit a significant response to random omissions of the aversive
US (Supplementary Fig. 3b, c), consistent with previous reports35.
These results indicated a lack of inhibition by negative prediction
errors in TS-projecting DA neurons.
Our results so far demonstrate that TS-projecting DA neurons are
strongly activated by aversive USs. An important question is whether
these DA neurons are activated also strongly by rewards. Since DA
neurons are known to respond to rewards particularly when they are
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 2
Content courtesy of Springer Nature, terms of use apply. Rights reserved

unpredicted20–23,36, we used a reward taskin which mice weretrained to
enter a reward port to receive rewards randomly with 50% probability
(Fig. 2i), making reward delivery unpredicted. Interestingly, we found
that DA terminals in TS exhibited weak responses to rewards (n=19;
Fig. 2j, k), consistent with previous findings16,35. In all animals tested,
responses to unpredicted rewards were much smaller compared to
unpredicted footshock USs (Fig. 2l). Furthermore, increasing the value
of reward did not have a significant effect on the magnitude of reward
responses (Supplementary Fig. 4), suggesting that TS-projecting DA
neurons do not encode the value of rewards, in line with previous
reports16. Together, our results indicated that while DA terminals in TS
were strongly activated particularly by painful and aversive stimuli,
they responded only weakly to rewards, and did not respond to
omission of aversive outcomes.
CS US
CS1
CS2
CS3
CS4
CS5
FC
a b
AAV5-FL EX-
GCaMP6m
DAT-Cre
Optical fiber
c
e
5 CS-US
FC Fear Recall
Hab
10 CS
10 CS
US
CS CS
CS
1s
CS
US
10 s 10 s 10 s
g
h
j
TS
SNDA
SNDA
TS d
Hab FC
Fear Recall
CS
CS
Time (s)
CS US
Time (s)
CS
dF/F
Time (s)
k
-0.05
0.1
0.25
-0.05 0.1 0.25
Average dF/F
(Hab)
Average dF/F
(Fear Recall)
i
% Freezing
80
60
40
20
0
DAY1 DAY2 DAY3
Hab FC Fear
Recall
CS CS CS
VTA
SNc
SNL
GCaMP TH DAPI
GCaMP TH
Merged
TS
BLA
CEA
GP
GCaMP TH DAPI
f
ASt
ASt
**
-0.1
0
0.1
0.2
0.3
Average dF/F
CS1 CS5
FC
0
0.5
1
1.5
US1 US5
FC
*
-0.05
0.05
0.15
0.25
Average dF/F
Hab Fear
Recall
***
US
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 3
Content courtesy of Springer Nature, terms of use apply. Rights reserved

DA release in TS signals a PE during associative fear learning
DA neurons have been shown to co-release glutamate and GABA in the
striatum37,38. It is therefore possible that the DA terminal activity might
not reflect DA release during associative fear learning. To address this,
we next examined the dynamics of DA release in TS during associative
fear learning by performing optical recordings of the genetically
encoded DA biosensor dLight39 using fiber photometry. To this end, an
AAV expressing dLight (AAV5-CAG-dLight1.1) was injected and optical
fibers were implanted in the TS (Fig. 3a, b, Supplementary Fig. 5). Mice
(n= 8) underwent thesame fear conditioning protocolas in the GCaMP
recording experiment (Fig. 3c), and showed increased freezing to the
CS following FC (Fig. 3d), indicating successful fear learning.
In line with Ca2+ recordings in DA terminals, we found strong
dLight responses to the US at the beginning of FC which significantly
decreased through the course of conditioning (Fig. 3e, f), as occur-
rence of the US became predictable. Conversely, we observed a sig-
nificant increase in the dLight activity in response to the CS when the
first and last CSs of FC were compared (Fig. 3e, f), indicating poten-
tiation of CS responses as the CS-US association was learned. Con-
sistent with increased behavioral freezing, CS responses exhibited a
significant increase from Hab to fear recall (Fig. 3e, g), and in all
recording sites (n= 10) CS responses during fear recall were larger
compared to Hab (Fig. 3g). Furthermore, we again observed small
responses to unpredicted rewards (Fig. 3h–j), and in all dLight
recording sites unpredicted footshock US responses (responses to the
first US of FC) were larger compared to unpredicted reward responses
(Fig. 3k). Taken together, these results indicate that DA release in TS
underlies PE signaling during associative fear learning.
TS-projecting DA neurons are required selectively for acquisi-
tion of cued associative fear learning
PE signals are thought to drive associative learning19.IfDAneurons
projecting to TS encode a PE signal and this signal is critical for driving
associative fear learning, we expect that lesioning these DA neurons
should impair particularly the acquisition of fear conditioning. To
address this, we performed projection-specificablationofTS-
projecting DA neurons using a DA neuron selective neurotoxin,
6-hydroxydopamine (6-OHDA; Fig. 4a). Importantly, the 6-OHDA
injections in TS caused reduction of DA axons specifically in TS, and
not in the neighboring structures such as the amygdala and the
amygdalostriatal transition area (ASt; Fig. 4b). Following 6-OHDA
lesions, mice were trained using an auditory FC protocol (Fig. 4c)
consisting of 5 CS-US pairings. Twenty-four hours after conditioning,
mice underwent a fear recall test consisting of 5 CS presentations. We
found that the lesioned mice (n=10)frozesignificantly less to the CS
throughout the conditioning session compared to saline-injected
control mice (n= 12; Fig. 4d), suggesting impaired fear learning. Fur-
thermore, impaired fear acquisition resulted in a weaker fear memory
when tested the next day (Fig. 4d). 6-OHDA lesioned mice froze sig-
nificantly less at the end of FC and the beginning of fear recall sessions
compared to control mice (Fig. 4e). These results demonstrated that
TS-projecting DA neurons are required for the acquisition of the CS-US
association.
APEsignalisexpectedspecifically to be critical for initiating and
driving new associative fear learning, but not the retrieval of fear
memories. We therefore hypothesized that TS-projecting DA neurons
might be criticalselectively during fearacquisitionon the conditioning
day, but not later once the CS-US association was learned. To test this,
we performed 6-OHDA ablation of TS-projecting DA neurons after fear
memory was formed (Fig. 4f). In support of our hypothesis, we found
that the control (n= 8) and the lesioned mice (n= 8) exhibited com-
parable levels of freezing to the CS during the fear recall test per-
formed 3 weeks after 6-OHDA lesions (Fig. 4g, h), suggesting that the
activity of these DA neurons was not necessary for retrieval and
expression of fear memories. These results indicated that TS-
projecting DA neurons were indeed important selectively for the
acquisition of fear conditioning but were no longer required once the
CS-US association was learned.
Given that TS is a sensory striatal region receiving sensory inputs
such as auditory and visual13, we hypothesized that DA projections to
TS might be important for associating specifically discrete sensory
cues with aversive outcomes. It is well-established that cued versus
contextual fear conditioning involves distinct neural circuits40.We
therefore next investigated whether TS-projecting DA neurons were
necessary for contextual fear learning. To this end, we performed a
contextual FC paradigm consisting of 5 US presentations in context A.
Mice received 6-OHDA or saline injections in TS as described above
and 3 weeks later underwent contextual FC (Fig. 4i). The animals were
tested for contextual fear memory the next day (Fig. 4j). Lesioned
(n= 7) and control (n= 7) mice showed comparable levels of freezing
to the context during contextual testing (Fig. 4k), indicating that DA
projections to TS were not required for learning the association
between the context and the US. Together, these results suggest that
TS-projecting DA neurons are critical for cued but not contextual fear
learning. However, whether they are required for associating stimuli
from sensory modalities other than auditory remained an important
question. To address this, we conducted a visual FC paradigm using a
discrete light cue as the CS. Mice received 6-OHDA or saline injections
in TS as described above and 3 weeks later underwent visual FC
(Supplementary Fig. 6a, b). We found that the lesioned mice (n=7)
froze significantly less to the CS throughout the conditioning session
compared to saline-injected control mice (n= 6; Supplementary
Fig. 6c, d), suggesting impaired visual fear learning. Furthermore,
impaired fear acquisition resulted in a weaker fear memory when
tested the next day (Supplementary Fig. 6c, d). These results demon-
strated that TS-projecting DA neurons are necessary for cued asso-
ciative fear learning.
Since we observed only small responses to rewards, we hypo-
thesized that TS-projecting DA neurons might not be required for
learning the association between a cue and reward. To test this, we
Fig. 1 | DA terminals in TS encode a PE signal during associative fear learning.
aSchematic of the surgical procedure. bExample histological image showing
expression of GCaMP (green), tyrosine hydroxylase (TH, red) at DA terminals and
DAPI (blue) staining in TS. The white vertical track indicates the optical fiber pla-
cement in TS. ASt amygdalostriatal transition area, BLA basolateral amygdala, CEA
central nucleus of the amygdala, GP globus pallidus. Scale bar: 0.5mm. cTop:
example histological image. SNc substantia nigra pars compacta, SNL substantia
nigra lateralis, VTA ventral tegmental area. Scale bar: 0.25 mm. Bottom: confocal
imagesshowing expression of GCaMPand TH stainingand the merged image. Scale
bar: 20 μm. dExample of c hange in fluorescence (dF/F) over time. Scale bar: 5 s,
0.5 dF/F. eTop: schematic of the behavioral protocol. Hab: tone habituation, FC:
fear conditioning. Bottom: schematic of CS and US presentations. Schematic rep-
rinted from ref. 28, copyright (2023), with permission from Elsevier. fFreezing to
the CS (n= 13 mice) during Hab, FC and Fear Recall.gTop: average activity around
each CS during Hab, FCand Fear Recallacross all mice (n= 13). The heat map shows
response amplitudes (dF/F). Bottom: average change in fluorescence around the
time of CS (gray area). The red area during FC represents the US presentation.
hChange in fluorescence around each CS and US during FC in an example animal.
Scale bar: 2.5 s, 0.5 dF/F. iLeft: comparison of average change in fluorescence
during CS1 and CS5 of FC (n=13,**P= 0.0017, two-sided signed-rank test). Right:
comparison of average change in fluorescence during US1 and US5 of FC (n=13,
*P= 0.013, two -sided signed-r ank test). jAverag e change in fluorescence during the
CS for Hab and Fear Recall (n= 13, ***P= 0.0007, two-sided signed-rank test).
kScatter plot showing CS responses of each recording site during Hab and Fear
Recall. The dashed line indicates the unity line. Data points below the unity line
represent larger CS responses during Fear Recall. Shaded regions and error bars
represent mean ± s.e.m. Source data are provided as a Source Data file.
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 4
Content courtesy of Springer Nature, terms of use apply. Rights reserved

-0.05
0.1
0.25
-0.05 0.1 0.25
a b
10 CS 5 CS / 5US 10 CS
Hab Unpaired
Training Testing
DAY 1 DAY 2 DAY 3
c d
f Predicted US
(5 CS-US)
Unpredicted US
(5 US)
g
i
50% Reward
Probability
Reward Task
j k
1s
CS
US
10 s 10 s
10 s
h
dF/F
CS CS
Hab Testing
-0.01
0
0.01
-0.01 0 0.01
Average dF/F
(Hab)
Average dF/F
(Testing)
Time (s) Time (s)
% Freezing
80
60
40
20
0
DAY1
Hab Unpaired
Training Testing
DAY2 DAY3
0 10 20
Time (s)
US
-0.1
0.1
0.2
0.3
0
dF/F
0 10 20
Time (s)
CS US
Average dF/F
(Predicted US)
Average dF/F
(Unpredicted US)
0.1
dF/F
0.05
0
0
Time (s)
-2 2 4
NE
-0.05
0.05
0.15
Unpredicted
Reward
Average dF/F
-0.05
0.1
0.25
-0.05 0.1 0.25
Average dF/F
(Unpredicted US)
Average dF/F
(Unpred. Reward)
e
l
-0.01
0
0.01
Average dF/F
Hab Testing
0
0.15
0.3
Average dF/F
Pred.
US
Unpred.
US
**
Fig. 2 | Activity of DA terminals in TS in response to CSs, aversive USs, and
rewards that are unpredicted. a Top schematic of the behavioral protocol for
unpaired training. Hab tone habituation. Bottom: Schematic of CS and US pre-
sentations during Hab, Unpaired Training, and Testing sessions. bBehavioral
freezing to the CS (n= 6 mice) during Hab (average of 10 CSs), Unpaired Training
and Testing sessions (average of 10 CSs). cAverage change in fluorescence from
recording sites in TS (n= 7) around the time of CS (gray area) during Hab, and
Testing. dComparison of average change in fluorescence in the 5 s after CS onset
during Hab and Testing (n= 7, P = 0.31, two-sided signed-rank test). eScatter plot
showing CS responses of each recording site during Hab and Testing. The dashed
line indicates the unity line. Data points below the unity line represent larger CS
responses during Testing. fAver age change in fluorescence from recording sites in
TS (n = 9)around the time of CS (gray area) and US (red area) during Predicted and
Unpredicted US presentations. gComparison of average change in fluorescence
during US presentation for Predicted and Unpredicted USs (n=9,**P=0.0078,
two-sided signed-rank test).hScatter plot showingUS responses of each recording
site during Predicted and Unpredicted US presentations. The dashed line indicates
the unity line. Data points below the unity line represent larger responses for the
Unpredicted US. iSchematic of the reward task. Animals receivedreward randomly
50% of the time after entering the noseport. Schematic reprinted from ref. 28,
copyright (2023), with permission from Elsevier. jAverage change i n fluorescence
during rewarded noseport entries (NE) from all recording sites (n=19). kAverage
change in fluorescence in the 3 s after noseport entry during rewarded NE (n=19).
lScatter plot showing responses from each recording site (n= 19) during Unpre-
dicted US and reward presentations. The dashed line indicates the unity line. Data
points below the unity line represent larger responses for Unpredicted US. Shaded
regions and error bars represent mean ± s.e.m. Source data are provided as a
Source Data file.
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 5
Content courtesy of Springer Nature, terms of use apply. Rights reserved

performed a reward learning task (Fig. 4l, m) in which a tone CS was
paired with reward. The animals underwent the reward task following
6-OHDA lesioning of TS-projecting DA neurons (Fig. 4l). Notably, both
lesioned and control groups showed similar learning rates during the
reward task. The two groups did not differ in the number of rewarded
CSs (Fig. 4n, o), latency to enter the port (Supplementary Fig. 7c, d),
time they spent in the port during the CS (Supplementary Fig. 7e, f) as
well as the number of nose pokes during the ITIs (two-way repeated
measures ANOVA, no main effect of group: F
1,52
=1.6, P= 0.22; no
group × trial interaction: F
4,52
=0.68, P= 0.6). These results suggest
that TS-projecting DA neurons are not required for associating cues
with rewards and hence are not critical for associative reward learning.
Taken together, these results reveal a highly selective role for TS-
projecting DA neurons in associative learning. We demonstrate that
a
e
AAV5-CA G-
dLight1.1
Optical fiber
b c
d
10 CS 5 CS-US 10 CS
Hab FC Fear Recall
DAY 1 DAY 2 DAY 3
TS
TS
% Freezing
80
60
40
20
0
DAY1 DAY2 DAY3
Hab FC Fear
Recall
FC Fear Recall
f g
-0.01
0.01
0.03
-0.01 0.01 0.03
Average dF/F
(Hab)
Average dF/F
(Fear Recall)
dLight TH DAPI
CEA
BLA
TS
GP
ASt
Hab
CS
CS
2
4
6
8
10
CS
dF/F
Time (s)
0.2
0.1
0
0 10 20
1
2
3
4
5
CS
2
4
6
8
10
CS
0
0.1
0.15
0.05
-0.05
Time (s)
CS
0 10 20
US CS
Time (s)
0 10 20
h
50% Reward
Probability
Reward Task
i j
0.1
dF/F
0.05
0
0
Time (s)
-2 2 4
NE
-0.05
0.05
0.15
-0.05 0.05 0.15
Average dF/F
(Unpredicted US)
Average dF/F
(Unpred. Reward)
k
Unpredicted
Reward
Average dF/F
0
-0.005
0.015
-0.02
0
0.02
0.04
Average dF/F
CS1 CS5
FC
**
0
0.1
0.2
US1 US5
FC
**
-0.01
0.01
0.03
Average dF/F
Hab Fear
Recall
**
US
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 6
Content courtesy of Springer Nature, terms of use apply. Rights reserved

they are requird selectively during the acquisition of fear learning, but
not once the CS-US association is learned. Importantly, we found that
these DA neurons are specifically important for learning discrete
CS−US, but not context−US or CS−reward, associations.
Temporally-precise activation of DA terminals in TS enhances
associative fear learning
If DA activity in TS at the time of the US acts as a teaching signal and
causes learning about the CS, then boosting this signal should enhance
associative fear learning. To test this, we optogenetically excited DA
terminals in TS precisely at the time of the US during FC. DAT-Cre mice
were bilaterally injected with a Cre-dependent AAV expressing either
channelrhodopsin-2 (ChR2) fused with EYFP (ChR2-EYFP) or EYFP only
(EYFP control) in the SN, and implanted bilaterally with optical fibers
above TS (Fig. 5a–c, Supplementary Fig. 8). We again observed a high
level of overlap between Cre-dependent ChR2-EYFP expression and
immunohistochemical staining against TH (Fig. 5c) suggesting DA
neuron-specific expression of ChR2.
In order to examine enhancement of fear learning, mice were
trained in a weak fear conditioning protocol (Fig. 5d) using a low US
intensity (0.35 mA). The experimental group consisted of ChR2-EYFP
expressing mice which received blue light stimulation specifically at
the time of the US (US Paired-ChR2, n= 12; Fig. 5e). The control group
expressing EYFP received the identical light delivery (US Paired-EYFP,
n= 8). Comparison of freezing levels to the CS in the two groups
revealed a significant difference between the ChR2 and the control
mice during both FC and fear recall (Fig. 5i, j). The ChR2 group
exhibited higher freezing levels to the CS compared to control mice
suggesting enhanced fear conditioning. However, it is also possible
that excitation of DA terminals in TS is aversive per se and pairing of
the CS with this aversive outcome results in learning of the association
between the CS and DA terminal excitation rather than enhancing the
CS-footshock US association. To address this, we performed excitation
of DA terminals in TS in the absence of the US (No-US Control; Fig. 5f)
and found that this did not induce fear learning (Fig. 5i, j), suggesting
that excitation of DA terminals in TS did not act as an aversive stimulus
or a threat per se. Together, these results demonstrate that excitation
of DA terminals in TS at the time of the US enhances associative fear
learning.
We next asked whether DA activity in TS during the CS was critical
for driving associative fear learning. To address this, we optogeneti-
cally excited DA terminals in TS during the CS presentations of FC
(Fig. 5g). The ChR2 (n= 7) and EYFP (n = 6) expressing mice underwent
the same FC protocol but this time received light stimulation specifi-
cally during the CS presentations (CS-paired). We found a significant
difference between the two groups during both FC and fear recall test
(Fig. 5k, l). Notably, light stimulation during the inter-trial intervals (ITI
Control, Fig. 5h) did not have a significant effect on freezing levels and
the ChR2-expressing mice froze more to the CS compared to EYFP and
ITI controls (Fig. 5k, l), indicating that excitation of DA terminals in TS
during the CS enhances associative fear learning.
However, it is alsopossible that excitation of DA neuron terminals
in TS per se could affect movement, increase anxiety levels or induce
aversive responses, potent ially leading to a general increase in freezing
rather than specifically enhancing the associative learning process. To
address this, we performed real-time place preference, open field and
elevated plus maze tests (Fig. 5m–o) to examine the effect of DA
terminal excitation on avoidance, movement and anxiety-like beha-
viors. Exciting DA terminals in TS (ChR2-EYFP mice n=8, EYFP mice
n= 8) did not cause real-timeplace avoidance(Fig. 5m) suggesting that
DA terminal excitation per se was not aversive. Furthermore, we also
did not find a significant difference between the two groups in their
anxiety-like behaviors when DA terminals in TS were illuminated in the
open field (Fig. 5n, two-way repeated measures ANOVA, no main effect
of group: F
1,28
=0.18,P= 0.67) and the elevated plus maze (Fig. 5o, two-
way repeated measures ANOVA, no main effect of group, F
1,28
=1.88,
P= 0.19). Finally, excitation of DA terminals in TS also did not have an
effect on the animal’s velocity in the open field (Fig. 5n, two-way
repeated measures ANOVA, no main effect of group, F
1,28
=1.03,
P= 0.32). Together, these results suggest that the observed effect on
fear learning cannot simply be due to aversion, increased anxiety or
changes in movement that was caused by the excitation of DA
terminals in TS.
Temporally-precise inhibition of DA terminals in TS impairs
associative fear learning
If DA input to TS at the time of the US acts as a teaching signal and
drives the learning about the CS, then inhibiting this signal is expected
to impair associative fear learning. To address this, we optogenetically
inhibited DA terminals in TS precisely at the time of the US during FC.
DAT-Cre mice were bilaterally injected with a Cre-dependent AAV
expressing either archaerhodopsin (eArch) fused with EYFP (eArch-
EYFP) or EYFP only (EYFP control) in the SN, and implanted bilaterally
with optical fibers above TS (Fig. 6a–c, Supplementary Fig. 9). We again
observed a high level of overlap between Cre-dependent eArch-EYFP
expression and immunohistochemical staining against TH (Fig. 6c)
suggesting DA neuron-specific expression of eArch.
The experimental group consisted of eArch-EYFP expressing mice
which received yellow light delivery specifically at the time of the US
(n=7;Fig.6d). The control group expressing EYFP received the iden-
tical light delivery (n= 6). Comparison of freezing levels to the CS in
the two groups revealed a significant difference between the eArch and
the control mice during both FC and fear recall (Fig. 6e, f). The eArch
group exhibited lower freezing levels to the CS compared to control
mice suggesting impaired fear conditioning. These results indicate
that activation of DA terminals in TS at the time of the US is required
for associative fear learning.
To test whether inhibition of DA neuron terminals in TS per se
could affect movement, decrease anxiety levels or have a rewarding
effect, potentially leading to a general decrease in freezing rather than
specifically impairing the associative learning process, we performed
real-time place preference, open field and elevated plus maze tests
Fig. 3 | DA release dynamics in TS during associative fear learning and
reward task. a Schematic of the surgical procedure. bExample histological image
showing expression of dLight (green), tyrosine hydroxylase (TH, red) and DAPI
(blue) staining. White vertical track indicates the optical fiber placement. ASt
amygdalostriatal transition area, BLA basolateral amygdala, CEA central nucleus of
the amygdala, GP globus pallidus. Scale bar: 0.5 mm. cSchematic of the behavioral
protocol. dFreezing to the CS (n= 8 mice) during Hab, FC and Fear Recall. eTop:
Average dLight activity around each CS during Hab, FC and Fear Recall across all
recording sites (n= 10). The heat map shows response amplitudes (dF/F). Bottom:
average c hange in fluorescence around the time ofCS presentation(gray area). The
red area during FC represents the US presentation. fLeft: comparison of average
change in fluorescence during CS1 and CS5 of FC (n=10,**P= 0.0 08, two-sided
signed-rank test). Right: comparison of average change in fluorescence during US1
and US5 of FC (n=10, **P=0.009, two-sided signed-rank test). gLeft: average
change in fluorescence in the 5 s after CS onset during Hab and Fear Recall
(**P= 0.002, two-sided signed-rank test). Right: scatter plot showing CS responses
of each recording site (n=10) during Hab and Fear Recall. Data points below the
unity line represent larger CS responses during Fear Recall. hSchematic of the
reward task. Animals received reward 50% of the time after entering the noseport.
Schematic reprinted from ref. 28, copyright (2023), with permission from Elsevier.
iAverage change in fluorescence during rewarded noseport entries (NE, n=10).
jAveragechange in fluorescence in the 3 s after noseport entryduring rewardedNE
(n= 10). kScatter plot showing responses from each recording site (n= 10) during
unpredicted US (first US of FC) and unpredicted reward. Data points below the
unity line represent larger responses forUnpredicted US. Shaded regions and error
bars represent mean ± s.e.m. Source data are provided as a Source Data file.
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 7
Content courtesy of Springer Nature, terms of use apply. Rights reserved

(Fig. 6g–i). Inhibiting DA terminals in TS (eArch-EYFP mice n=7,EYFP
mice n= 6) did not cause real-time place preference (Fig. 6g) sug-
gesting that DA terminal inhibition per se did not have a rewarding
effect. Furthermore, we also did not find a significant difference
between the two groups in their anxiety-like behaviors when DA
terminals in TS were inhibited in the open field (Fig. 6h; two-way
repeated measures ANOVA, no main effect of group: F
1,22
=1.57,
P= 0.23; no group × trial interaction: F
2,22
=1.57, p= 0.22) and the
elevated plus maze (Fig. 6i; two-way repeated measures ANOVA, no
main effect of group: F
1,22
=0.19, P= 0.67). Finally, inhibition of DA
terminals in TS also did not have an effect on the animal’s velocity in
the open field (Fig. 6h; two-way repeated measures ANOVA, no main
effect of group: F
1,22
=0.10, P= 0.75; no group × trial interaction:
F
2,22
=1.51,p= 0.24). Together, theseresults indicate that the observed
a b
6-OHDA
3 weeks
TS SNDA
TS SNDA
c d
f
6-OHDA
Lesions
3 weeks
Auditory FC
(5 CS-US)
24 hr
Fear Recall
(5 CS)
0
20
40
60
80
DAY 1 DAY 2
Fear
Conditioning
Fear
Recall
% Freezing
e
0
20
40
60
80
100
DAY 1
% Freezing
***
Auditory FC
(5 CS-US)
48 hr
6-OHDA
Lesions
3 weeks
Fear Recall
(5 CS)
g h
0
20
40
60
80
DAY 1 DAY 24
Fear
Conditioning
Fear
Recall
% Freezing
0
20
40
60
80
DAY 1
% Freezing
l n
6-OHDA
Lesions
4 weeks
Reward
Learning
50 CS-R
(5 Days) CS R
(Days 1-5)
0
20
40
60
80
100
12345
% Rewarded
Days
m
0
20
40
60
80
100
% Rewarded
DAY 1 DAY 5
0
20
40
60
80
100
DAY 2
***
0
20
40
60
80
DAY 24
Control
6-OHDA
Control
6-OHDA
Control
6-OHDA
i
6-OHDA
Lesions
3 weeks
Context FC
(5 US)
24 hr
Context Test
(10 min)
j
Context
FC
Context
Testing
Context
Hab
1s
US
US
DAY 1 DAY 2 DAY 3 k DAY 3
0
10
20
30
40
50
DAY 1
% Freezing
o
TH DAPI TH DAPI
TS
TH DAPI
Control
6-OHDA
BLA
CEA
ASt
BLA
CEA
ASt
TS
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 8
Content courtesy of Springer Nature, terms of use apply. Rights reserved

impairment in fearlearning was due to inhibition of DAterminals in TS,
rather than nonspecificeffects.
Neuronal activity in TS exhibits a PE-like pattern and potentia-
tion of CS responses during associative fear learning
The requirement of a DA PE signal in TS during FC suggests that the
activity of TS neurons is likely critical for acquisition of associative
fear learning. To address this, we first examined the neuronal activity
in TS during FC by measuring activity-dependent Ca+2 signals using
fiber photometry. To this end, an AAV expressing the Ca+2 indicator
GCaMP6f was injected and an optical fiber was implanted in the
TS (Fig. 7a, b, Supplementary Fig. 10). To test whether the observed
changes in fluorescence depend on neuronal activity we injected
an AAV expressing the control fluorophore EYFP in a separate group
of control mice. Transient fluctuations in fluorescence were absent
in mice expressing the control fluorophore (n= 2, Supplementary
Fig. 2c, d). The GCaMP-expressing mice (n= 9) underwent the
same FC protocol as in fiber photometry experiments described
above (Fig. 7c), and exhibited successful fear learning (Fig. 7d).
Interestingly, we observed that the activity in TS exhibited a PE-like
pattern during FC (Fig. 7e), similar to the results of our recordings
of DA terminal activity as well as DA release in TS. While the
responses to the US decreased during the course of FC, the CS
responses gradually increased (Fig. 7e). Indeed, there was a sig-
nificant increase in responses to the CS from first to last CSs (Fig. 7f).
Conversely, the responses to the US were larger to the first compared
to the last US (Fig. 7f). Importantly, we found a significant increase in
responses to the CS from Hab to fear recall (Fig. 7g). In almost all
recording sites, the CS responses during fear recall were larger in
amplitude compared to Hab (Fig. 7g). Together, these results
demonstrated a PE-like activity pattern in TS during associative fear
learning.
Causal Contribution of Neuronal Activity in TS to Associative
Fear Learning
We next asked whether activity in TS was necessary for associative
fear learning. To address this, we first performed chemogenetic
inhibition of TS neurons during FC. Mice received injections of an
AAV expressing the inhibitory DREADD receptor (hM4D(Gi)) in TS
(Fig. 8a, b) and underwent the same auditory FC protocol as in the
6-OHDA experiment (Fig. 8c). Thirty minutes before the FC session,
mice received systemic injections of the DREADD agonist clozapine
N-oxide (CNO) to inhibit activity of TS neurons during fear con-
ditioning whereas control mice received saline injections. We found
that CNO-injected mice (n= 8) froze significantly less to the CS at the
end of FC compared to saline-injected controls (n=8; Fig. 8d, e)
suggesting impaired fear learning. Furthermore, impaired fear
learning resulted in a weaker fear memory when tested the next day
during the fear recall test (Fig. 8d, f). These effects were dependent
on inhibitory DREADD receptor expression since CNO injection had
no effect in mCherry-expressing control mice (Supplementary
Fig. 11). These results demonstrated that the activity of TS neurons is
critical for associative fear learning.
A PE-like activity pattern in TS for the aversive US suggests that
this signal might drive fearlearning. If thatis the case, the activity of TS
neurons during the aversive US is expected to act as a teaching signal
for associative fear learning. We therefore next investigated whether
TS neuronal activity during the US is necessary for driving fear learn-
ing. To this end, we performed temporally-precise optogenetic inhi-
bition of TS neurons at the time of the US during FC (Fig. 8i). An AAV
expressing eArch-EYFP or EYFP only was bilaterally injected and optic
fibers were bilaterally implanted in TS (Fig. 8g, h, Supplementary
Fig. 12). We found a significant difference between the eArch (n=7)
and the control (n=8)miceduringFC(Fig.8j). eArch-expressing mice
showed significantly lower freezing to the CS at the end of FC (Fig. 8k)
and the beginning of fear recall test (Fig. 8l), suggesting impaired fear
learning. These results indicate that the activity of TS neurons at the
time of the US acts as a teaching signal and is necessary for driving
associative fear learning.
If activity of TS neurons during the US drives associative fear
learning, then boosting this activity is expected to enhance fear con-
ditioning. We tested this by performing temporally-precise optoge-
netic excitation of TS neurons at the time of the US during FC (Fig. 8o).
To this end, an AAV expressing ChR2-EYFP or EYFP only was bilaterally
injected and optic fibers were bilaterally implanted in TS (Fig. 8m, n,
Supplementary Fig. 13). In order to examine enhancement of fear
learning, mice were trained in a weak fear conditioning protocol
(Fig. 8o), as during optogenetic excitation of DA terminals described
above (Fig. 5). We found a significant difference between the ChR2
(n= 8) and the control (n=8) mice during FC (Fig. 8p). ChR2-
expressing mice showed significantly higher freezing to the CS at the
end of FC (Fig. 8r) and the beginning of fear recall test (Fig. 8s), indi-
cating that excitation of TS neurons at the time of the US enhances
associative fear learning.
To test whether manipulation of TS neuronal activity per se could
affect movement, decrease anxiety levels or could have an aversive
effect, potentially leading to a general change in freezing rather than
specifically affecting the associative learning process, we tested ani-
mals on the control tasks while inhibiting (Supplementary Fig. 14a, b)
or exciting (Supplementary Fig. 14c-e) the activity of TS neurons.
Chemogenetic inhibition of TS neurons (CNO group n= 6, saline group
n= 6) did not cause anxiety-like behaviors in the open field and the
elevated plus maze (Supplementary Fig. 14a, b). Inhibition of TS neu-
rons also did not have an effect on the animal’s velocity in the open
field (Supplementary Fig. 14a). Conv ersely, exciting TS neurons did not
cause real-time place avoidance (Supplementary Fig. 14c), anxiety-like
behaviors in the open field (Supplementary Fig. 14d) or the elevated
plus maze (Supplementary Fig. 14e) and did not affect velocity in the
open field (Supplementary Fig. 14d). Together, these results indicate
that the observed impairment and enhancement in fear learning was
due to inhibition and excitation of TS neurons, respectively, rather
than nonspecific effects.
Fig. 4 | TS-projecting DA neurons are required selectively for the acquisitionof
cued associative fear learning. a Schematic of the surgical procedure. bExample
sections showing tyrosine hydroxylase (TH, green) and DAPI (blue) staining in
saline (top, left) and 6-OHDA (bottom, left) groups. Scale bar: 0.5 mm. Close-up
images showing TH-staining in TS (middle) and the amygdala (right; ASt, amyg-
dalostriatal transition area; BLA, basolateral amygdala; CEA, central amygdala) in
saline (top) and 6-OHDA (bottom) groups. cSchematic of the protocol. dFreezing
to the CS for saline (n= 12) and 6-OHDA (n= 10) groups (two-way repeated mea-
sures ANOVA, FC: main effectof group: F
1,80
=7.84,P= 0.011; fear recall: main effect
of group: F
1,80
=18.60,P= 0.0003). eLeft: freezing to the last CS during FC (two-
sided t-test, t(20) = 3.85,P= 0.001). Right: freezing to the first CS of fearrecall (two-
sided t-test, t(20) = 4.68, P=0.0001).fSchematicof the protocol.gFr eezing to the
CS for saline (n=8) and 6-OHDA (n= 8) groups (two-way repeated measures
ANOVA, FC: no main effect of group: F
1,56
=1.37,P= 0.26; fear recall: no main effect
of group: F
1,56
=0.12,P=0.72).hFreezing to the last CS during FC (left, two-sided t-
test, t(14) = 0.06, P= 0.95) and to the first CS during fear recall (right, two-sided t-
test, t( 14)= 0 .07, P= 0.94). iSchematic of the protocol. jSchematic of the task.
Schematic reprinted from ref. 28, copyright (2023), with permission from Elsevier.
kFreezingto the context in control (n=7)and6-OHDA(n= 7) groups. (two-sided t-
tests: Context Hab, P= 0.14; Context Test, P=0.38). lSchematic of the protocol.
mSchematic of the task. Schematic reprinted from ref. 28, copyright (2023), with
permission from Elsevier. nPercent rewarded CSs for saline (n=8) and 6-OHDA
(n= 7) groups (two-way repeated measures ANOVA, no main effect of group:
F
1,52
=1.7,P=0.21).oPercent rewarded CSs during the first and last days (saline,
n= 8; 6-OHDA, n = 7). Error bars represent mean ± s.e.m. Source data are provided
as a Source Data file.
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved

DA Input is Critical for PE-like Activity and Potentiation of CS
Responses in TS during Associative Fear Learning
Our results demonstrated that the activity in TS exhibits a PE-like
pattern during fear learning. A key question is whether DA input is
important for this fear learning-related activity observed in TS. To
address this question, we performed 6-OHDA ablation of TS-
projecting DA neurons followed by Ca2+ recordings in TS using
fiber photometry. Mice received 6-OHDA or saline injections, and
one week later, were injected with an AAV expressing GCaMP6f as
well as implanted with an optical fiber in the TS (Fig. 9a, Supple-
mentary Fig. 15). The animals underwent the same FC protocol
(Fig. 9b) as in fiber photometry experiments described above.
Importantly, 6-OHDA injected animals (n= 8) showed reduction in
DAergic innervation of TS compared to control group (n=10;
-10
10
30
50
Δ Time spent (%)
(Laser ON – OFF side)
ChR2 EYFP
n
a
d
b
5 CS-US-laser 5 CS
FC Fear Recall
DAY 1 DAY 2
Laser
1s
CS
US
30 s
1-3s
c
ChR2-EYFP TH
Merged
e f
i
k
m o
ChR2-EYFP TH DAPI
TS
BLA
CEA
TS
BLA
CEA
ChR2-EYFP TH DAPI
VTA
SNc
SNL
Laser
CS
No US
30 s
3s
US-paired No-US Control h ITI Control
Laser
CS 30 s
1s
US
j
DAY 1 DAY 2
Fear
Conditioning
Fear
Recall
% Freezing
US-Paired EYFP
US-Paired ChR2
0
20
40
60
80
100
No-US-Control
g
30s
Laser
1s
CS
US
30 s
CS-paired
l
30s
0
20
40
60
80
100
% Freezing
DAY 1 DAY 2
Fear
Conditioning
Fear
Recall
CS-Paired EYFP
CS-Paired ChR2
ITI-Control
0
20
40
60
80
100
% Freezing
DAY 1
***
% Freezing
DAY 2
0
20
40
60
80
100
***
0
20
40
60
80
100
% Freezing
DAY 1
***
0
20
40
60
80
100
% Freezing
DAY 2
*
0
2
4
6
8
10
Velocity (cm/sec)
OFF ON OFF
Open Field Test Elevated Plus Maze
Real-time Place Preference
Test
Laser
OFF
side
Laser
ON
side
Laser
ON
side
Laser
OFF
side
0
10
20
% Time in Center
OFF ON OFF
EYFP
ChR2
0
15
30
% Time in Open Arms
OFF ON OFF
EYFP
ChR2
AAV5-
DIO-
ChR2
DAT-Cre
Optical
fibers
TS
SNDA
SNDA
10 min 10 min
3 min 3 min 3 min
Laser
OFF
Laser
OFF
Laser
ON
3 min 3 min 3 min
Laser
OFF
Laser
OFF
Laser
ON
center
periphery
closed
open
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 10
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Fig. 9d, h), and consistently, lesioned mice showed impaired fear
learning compared to controls (Fig. 9c).
The PE-like activity pattern during FC that the control group
exhibited (n=14 recording sites; Fig. 9e–g) was not observed in the
6-OHDA lesioned group (n=14 recordingsites;Fig. 9i–k). In lesioned
mice, we did not find an increase in the CS responses from first to last
CSs (Fig. 9j), which was seen in control animals (Fig. 9f). Interestingly,
the significant decrease in responses between the first and the last USs
of FC, that was observed in control mice (Fig. 9f), was also absent in
lesioned animals (Fig. 9j). Furthermore, while CS responses were sig-
nificantly increased between Hab and fear recall in control animals, no
increase was observed in lesioned mice (Fig. 9g, k). Whereas CS
responses were larger in magnitude during fear recall compared to
Hab in the majorityof control animals, this was not the case in lesioned
mice (Fig. 9g, k). Notably, there was no difference between the two
groups in their CS responses during Hab, whereas during FC and fear
recall, CS responses were significantly larger in the control group
(Fig. 9l). The responses to unpredicted US (first US of FC) were not
significantly different between the two groups (Fig. 9l), suggesting that
US input to TS in general was not affected. Importantly, the difference
in US responses from the first to the last US of FC was significantly
different between the groups (Fig. 9m), indicating that the decrease in
US responses as fear conditioning progressed was absent in 6-OHDA
mice, consistent with a deficit in learning about the CS-US association
in these mice. Overall, these results indicate that DA input is required
for fear learning-induced activity patterns observed in TS during
associative fear learning.
Discussion
Whether midbrain DA neurons projecting to brain structures outside
the canonical amygdala circuitry encode a PE signal that is critical to
drive associative fear learning is largely unknown. Recent studies
have implicated TS in fear learning11,12, yet the precise role DA neu-
rons projecting to TS play in associative fear learning is incompletely
understood. Here, we demonstrated that DA projections to TS
drive associative fear learning by encoding a PE signal that is
important for generating fear learning-induced activity patterns in
TS. We first showed that DA projections to TS exhibit a positive PE
signal during FC, and that this PE signal is transmitted by DA release
in TS. Projection-specific lesioning of TS-projecting DA neurons
selectively impaired the acquisition of associative fear learning, but
not fear retrieval and expression. Notably, these neurons were
required specifically for acquiring the association between the cue
CS and US, but not between the context and US or between a CS
and reward. Bidirectional temporally-precise optogenetic manipula-
tions of DA projections to TS demonstrated the necessity of DA PE
signaling in associative fear learning. Furthermore, activity in TS
exhibited a PE-like pattern during FC. The neuronal activity in TS
was required for fear learning, and temporally-precise enhancement
of this activity enhanced fear learning. Finally, we demonstrated that
DA input was critical for the fear learning-induced activity pat-
tern in TS.
Previous studies using lesioning, pharmacological manipulations,
recordings of DA neuronal activity, mea surements of DA release as well
as research on DA-deficient mice have consistently indicated a critical
role of dopamine in fear learning and memory as well as aversion30–33.
Specifically, while research on DA-deficient mice demonstrated the
requirement of DA in fear learning41, early pharmacological studies
highlighted the role of DAergic signaling particularly in the basolateral
amygdala (BLA) during acquisition and expression of fear
memories42–46. Consistent with these findings, in DA-deficient mice,
restoring DA productionspecificallyin projectionsto BLA and striatum
reversed deficits in fear memory formation47. Furthermore, disrupting
phasic firing in dopamine neurons has been shown to impair fear
learning48,49; and midbrain dopamine neurons exhibit phasic firing in
response to aversive USs as well as CSs that predict them16,25,29,35,49–56.
Interestingly, recent studies showed that DA terminals in BLA are
activated by both aversive and rewarding stimuli ,suggesting that these
DA neurons encode the motivational salience of stimuli57,58. However,
despite this extensive literature on the role of DA in fear learning and
memory, it still remained elusive whether ventral midbrain dopamine
neurons, located in the VTA and SN, encode prediction error signals
necessary for driving associative fear learning.
The midbrain DA system is composed of functionally distinct and
mostly non-overlapping subpopulations of DA neurons, each of which
projects mainly to a specificbrainregion
15,59–63.Notably,DAneurons
projecting to TS have recently emerged as a unique subpopulation
based on their distinct input-output circuitry15. These DA neurons have
been shown to be activated by novel as well as external aversive stimuli
(e.g. air puffs and loud tones but not bitter taste16,35). In line with these
previous studies, we found that TS-projecting DA neurons are acti-
vated particularly strongly by aversive stimuli that are noxious such as
foot shocks, but only weakly by rewards. Interestingly, TS-projecting
DA neurons have been shown to exhibit a PE-like activity pattern
during Pavlovian association of olfactory cues with mildly aversive
stimuli such as air puffs16,35; and these DA neurons are particularly
important for novelty-induced threat avoidance16,17. Although, recent
studies showed enhancement of CS responses in TS-projecting DA
neurons during fear conditioning12, whether these neurons encode a
PE signal for strong painful stimuli such as foot shocks; and whether
this signal is necessary for associating cues with the aversive US during
FC has remained elusive.
During associative learning, a positive PE acts as a teaching signal
and drives learning about the CS that precedes the US19,24,64–66.Wehere
demonstrate that TS-projecting DA neurons encode a positive PE for
the aversive US during FC. These neurons respond more strongly to
aversive USs when they are unpredicted compared to predicted ones.
Furthermore, we observed that during the course of FC the US
responses decreased whereas the CS responses were enhanced as the
CS-US association was learned and the CS came to predict the US.
These changes in the activity of TS-projecting DA neurons were spe-
cifically dependent on the temporally contingent presentation of the
CS and the US: in mice trained with unpaired conditioning, during
which the CS and the US were presented in a temporally unpaired
Fig. 5 | Temporally-precise activation of DA terminals in TS enhances associa-
tive fear learning. a Schematic of the surgery. bExample histological image
showing expression of ChR2-EYFP (green), tyrosine hydroxylase (TH, red), and
DAPI (blue) staining in TS. White vertical tracks indicate optical fiber placements.
Scale bar: 0.5 mm. cTop: example image of the midbrain showing expression of
ChR2-EYFP, TH, and DAPI staining. Scale bar: 0.5mm. Bottom: confocal images.
Scale bar:20 μm. dSchematicof the behavioral protocol. Schematic of optogenetic
excitation for US-Paired (e), No-US-Control (f), CS-Paired (g), and ITI-Control (h)
groups. iFreezing to the C S in ChR2 (n= 12), EYFP (n=8)andNo-US-Control(n=4)
groups(two-way repeated-measures ANOVA,FC: main effectof group: F
2,84
=12.97,
P= 0.0002; Fear recall: main effect of group: F
2,84
= 21.99, P< 0.0001). jLeft:
freezing to the last-CS of FC (one-way ANOVA, F
2,23
= 19.78, P<0.0001).Right:
freezing to the first-CS of fear recall (one-way ANOVA, F
2,23
=17.16,P<0.0001).
kFreezing to the CS in ChR2 (n= 7), EYFP (n= 6) and ITI-Control (n= 4) groups
(two-way repeated-measures ANOVA, FC: main effect of group: F
2,56
= 11.04,
P= 0.0013;fear recall: main effect of group: F
2,56
=5.77,P=0.014).lLeft:freezing to
the last-CS of FC (one-way ANOVA, F
2,16
=19.78,P< 0.0001). Right: freezing to the
first-CS of fear recall (one-way ANOVA, F
2,16
=3.86,P= 0.04). Top: schematic of the
task. Bottom: (m), time spent in laser ON minus laser OFFside for EYFP (n=8)and
ChR2 (n= 8) groups during the real-time place preference test (P= 0.56, two-sided
t-test). nVelocity(left) and timespent in the centerof the open field (right) for EYFP
(n= 8) and ChR2 (n=8)groups.otime spent in the open arms of the elevated plus
maze for EYFP (n= 8) and ChR2 (n= 8) groups. Error bars represent mean ± s.e.m.
Source data are provided as a Source Data file.
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 11
Content courtesy of Springer Nature, terms of use apply. Rights reserved

manner and thus CS did not predict the occurrence of the US, the CS
responses failed to potentiate. Interestingly, these DA neurons have
previously been shown to lack inhibitory responses to the omission of
air puff USs35. Consistent with this, we a lso did not observe inhibition in
response to random footshock US omissions, suggesting that TS-
projecting DA neurons exhibit only positive PEs and lack negative PE
signaling for aversive stimuli. Notably, inhibitory responses to aversive
US omissions were also absent in the dorsal tegmental DA neurons that
have been shown to encode PE signals to gate fear learning29,further
supporting the notion of a general lack of negative PE signaling in DA
neurons encoding PEs for aversive stimuli. Taken together, these
findings suggest that DA prediction error signaling mechanisms may
differ between rewarding versus aversive outcomes.
A PE signal is expected to initiate new learning; consistent with
this, we demonstrate that ablation of TS-projecting DA neurons
selectively impaired the acquisition of the CS-US association during
0
5
10
15
20
0
5
10
Velocity (cm/sec)
OFF ON OFF
a
AAV5-
DIO-
eArch
DAT-Cre
Optical
fibers
b c
g h
Open Field Test
i
Elevated Plus Maze
Real-time Place Preference
Test
eArch-EYFP TH DAPI
VTA
SNc
SNL
% Time in Open Arms
d
US-
Paired
5 CS-US−laser 5 CS
FC Fear Recall
DAY 1 DAY 2
1s
CS
US
30 s
3 s
30 s
e f
0
20
40
60
80
100
% Freezing
DAY 1 DAY 2
US-Paired EYFP
US-Paired eArch
Fear
Conditioning
Fear
Recall
0
20
40
60
80
100
% Freezing
EYFP eArch
DAY 1
**
0
20
40
60
80
100
% Freezing
EYFP eArch
DAY 2
*
% Time in Center
0
5
10
OFF ON OFF
EYFP
eArch
OFF ON OFF
EYFP
eArch
-50
-30
-10
10
30
Δ Time spent (%)
(Laser ON – OFF side)
eArch EYFP
eArch-EYFP TH DAPI
TS
BLA CEA
eArch-EYFP TH Merged
SNDA SNDA
TS
Laser
OFF
side
Laser
ON
side
Laser
ON
side
Laser
OFF
side
10 min 10 min
center
periphery
3 min 3 min 3 min
Laser
OFF
Laser
OFF
Laser
ON
3 min 3 min 3 min
Laser
OFF
Laser
OFF
Laser
ON
closed
open
TS
TS
BLA
CEA
Fig. 6 | Temporally-precise inhibitionof DA terminals in TS impairs associative
fear learning. a Schematic of the surgery. bExample histological image showing
expression of eArch-EYFP (green), tyrosine hydroxylase (TH, red), and DAPI (blue)
staining in TS. Whitevertical tracks indicate the optical fiber placementsin TS. Scale
bar: 0.5 mm. cTop:example histological image showing expression of eArch-EYFP,
TH and DAPI staining in the midbrain. Scale bar:0.5 mm. Bottom: confocal images.
Scale bar: 20 μm. dTop: schematic of the behavioral protocol. FC: fear con-
ditioning. Bottom: schematic of the optogenetic inhibition paired to the US (US-
Paired). eFreezing to the CS during FC and fear recall for US-Paired eArch (n=7)
and EYFP (n=6) groups (two-way repeated measures ANOVA, FC: main effect of
group: F
1,44
= 15.03, p= 0.0026; fear recall: main effect of group: F
1,44
= 17.02,
p= 0.0017). fLeft: freezing to the last CS of FC (two-sided t-test, t(11) = 3.3,
p= 0.007). Right: freezing to the first CS of fear recall (two-sided t-test, t(11)= 2.95,
p= 0.013). gTop: schematic of the real-time place preference test. Bottom: dif-
ferencebetween the percent of timemice spent in laserON minus laser OFFside for
EYFP (n=6)andeArch(n= 7) groups (two-sided t-test, t(11)= 1.42, P = 0.18). hTop:
schematic of the open field test. Bottom: velocity(left) and time spent in thecenter
of the open field (right) during laser ON and OFF epochs for EYFP(n=6)andeArch
(n= 7) groups. Velocity and time in the center of the open field were comparable
between the groups. iTop: schematicof the elevated plus maze test. Bottom: time
mice spent in the open arms of the elevated plus maze during laser ON and OFF
epochs for EYFP (n=6)andeArch(n= 7) groups. Error bars represent mean± s.e.m.
across animals. Source data are provided as a Source Data file.
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 12
Content courtesy of Springer Nature, terms of use apply. Rights reserved

FC, but not retrieval and expression of fear memories. In line with our
results, ablation of these DA neurons has previously been shown to
impair learning about and avoiding aversive air puffs16. Interestingly,
we found that TS-projecting DA neurons were not required for
acquisition of contextual fear memories, highlighting the selective role
of these DA neurons in cued associative fear learning. Furthermore, we
demonstrate thatexciting DA projections to TS at the time of the CS or
the US caused enhancement of FC. Conversely, optogenetic inhibition
of these DA projections during the US impaired FC. However, in a
recent study optogenetic inhibition of these DA projections during
fear conditioning affected retrieval but not the acquisition of fear
memories12. The reasons for these varying results are not clear.
Nevertheless, our results indicate that the PE signal encoded by TS-
projecting DA neurons for aversive stimuli indeed acts as a teaching
signal to drive the acquisition of CS-US association during fear
learning.
In contrast to theirstrong responses to the aversive US, we show
that DA projections to TS are only weakly activated by rewards.
Consistent with this, we found that TS-projecting DA neurons were
not required for associating an auditory cue with reward; although
they have been implicated in more complex forms of reward learning
involving auditory discrimination67. It is important to note that
although the operant reward task that we used is not fully compar-
able with classical conditioning, our results nevertheless indicate
a b
AAV5-CamKII-
GCaMP6f
Optical fiber
TS
TS
c
d
10 CS 5 CS-US 10 CS
Hab FC Fear Recall
DAY 1 DAY 2 DAY 3
% Freezing
80
60
40
20
0
DAY1 DAY2 DAY3
Hab FC Fear
Recall
e
f
-0.01
0.02
0.05
-0.01 0.02 0.05
Average dF/F
(Hab)
Average dF/F
(Fear Recall)
GCaMP TH DAPI
TS
CEA
BLA
GP
ASt
g
FC Fear Recall
Hab
CS
CS
2
4
6
8
10
CS
dF/F
Time (s)
0.1
0
0 10 20
1
2
3
4
5
CS
2
4
6
8
10
CS
Time (s)
CS
0 10 20
US
CS
Time (s)
0 10 20
0
0.1
0.15
0.05
0.1
0
0.1
0
-0.01
0.04
0.09
Average dF/F
CS1 CS5
FC
** *
0
0.1
0.2
FC
US1 US5
-0.005
0.025
0.055
Average dF/F
Hab Fear
Recall
**
US
Fig. 7 | Activity in TS exhibits a PE-like pattern, and CS responses are poten-
tiated during associative fear learning. a Schematic of the surgical procedure.
bExample histological image showing expression of GCaMP (green) along with
immunostaining for tyrosine hydroxylase (TH, red) and DAPI (blue) staining in TS.
The white vertical track indicates the optical fiber placement in TS. ASt amygda-
lostriatal transition area, BLA basolateral amygdala, CEA central nucleus of the
amygdala; GP globus pallidus. Scale bar: 0.5 mm. cSchematic of the behavioral
protocol. Hab: tone habituation;FC: fear conditioning. dBehavioral freezing to the
CS (n= 9 mice) during Hab, FC, and Fear Recall. eTop: average GCaMP activity
aroundeach CS during Hab, FC and FearRecall across allrecording sites(n= 9). The
heat map shows response amplitudes (dF/F) around each CS presentation.Bottom:
average change in fluorescence around the time of CS presentation (gray area)
during Hab, FC, and Fear Recall. The red area during FC represents the US
presentation. fLeft: comparison of average change in fluorescence in the 5 s after
CS onset during CS1 and CS5 of FC. Note the significant increase in the Ca2+ signal
from CS1 to CS5(n=9,**P= 0.0039,two-sided signed-rank test).Right: comparison
of averag e change in fluorescence during the US (1 s) for US1 and US5 of FC. Note
the significant decrease in the Ca2+ signal from US1 to US5 (n=9, *P=0.02, two-
sided signed-rank test). gLeft: average change in fluorescence in the 5 s after CS
onset during Hab and Fear Recall. Note the significant increase in the Ca2+ signal
fromHabtoFearRecall(**P= 0 .0078, two-sided signed-rank te st). Right: scatter
plot showing CS responses of each recording site (n= 9) during Hab and Fear
Recall. Data points below the unity line represent larger CS responses during Fear
Recall. Shaded regions and error bars represent mean ± s.e.m. Source data are
provided as a Source Data file.
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 13
Content courtesy of Springer Nature, terms of use apply. Rights reserved

that TS-projecting DA neurons show weak reward responses and are
not required for forming CS-reward associations. Furthermore, our
results are also consistent with previous studies in head-fixed mice
which showed that TS-projecting DA neurons had weak reward
responses, did not exhibit RPE signaling during Pavlovian reward
learning35, and were not required for learning of reward value16.
Consistently, we also found that increasing the reward value did not
affect the magnitude of reward responses in these DA neurons.
Overall, these findings highlight the consistency of results across
both head-fixed and freely-behaving conditions. Importantly, our
results indicate an aversive bias in the responses of TS-projecting DA
neurons. This distinguishes the responses of these DA neurons from
other DA subpopulations that project to the striatum, for example,
the ones projecting to the ventral subregion of NAc which have also
a b
5 CS-US 5 CS
FC Fear Recall
DAY 1 DAY 2
CNO
Saline
AAV5-
hSyn-
hM4D(Gi)-
mCherry
TS
c
d
g h i
US-
Paired
1s
CS
US
30 s 30 s
DAY 2
DAY 1
Fear
Recall
0
20
40
60
80
Fear
Conditioning
% Freezing
Saline
CNO
AAV5-CaMKII-
eArch-EYFP
Optical fiber
TS
TS
5 CS-US−laser 5 CS
FC Fear Recall
DAY 1 DAY 2
1s
CS
US
30 s
DAY 2
DAY 1
% Freezing
EYFP
eArch
j
TH
hM4D(Gi)-mCherry DAPI
TS
TS
BLA
BLA
f e
l
k
eArch-EYFP TH DAPI
m n o
US-
Paired
AAV5-CaMKII-
ChR2-EYFP
Optical fiber
TS
TS
1s
CS
US
30 s 30 s
DAY 2
DAY 1
% Freezing
EYFP
ChR2
p
3s
s r
ChR2-EYFP TH DAPI
TS
GP
TS
GP
DAY 1
% Freezing
0
20
40
60
80
**
CNO
Saline
0
20
40
60
80
% Freezing
CNO
Saline
DAY 2
*
3 s
30 s
0
20
40
60
80
100
Fear
Recall
Fear
Conditioning
% Freezing
DAY 1
0
20
40
60
80
100
eArch
EYFP
*
% Freezing
DAY 2
0
20
40
60
80
100
eArch
EYFP
*
0
20
40
60
80
100
Fear
Recall
Fear
Conditioning
5 CS-US−laser 5 CS
FC Fear Recall
DAY 1 DAY 2
% Freezing
ChR2
DAY 1
**
0
20
40
60
80
100
EYFP
% Freezing
DAY 2
*
0
20
40
60
80
100
ChR2 EYFP
TS
GP
BLA
TS
GP
BLA
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 14
Content courtesy of Springer Nature, terms of use apply. Rights reserved

been shown to be activated by foot shocks28,68,69. However, those DA
neurons also exhibit strong, and even larger, responses to rewards
compared to aversive stimuli28,68, suggesting that they likely encode
salience rather than aversion. Our results raise the question of whe-
ther activation of TS-projecting DA neurons induces aversion
per se18,33. However, we found that exciting DA terminals in TS does
not cause place avoidance or anxiety. Likewise, activation of these
terminals at the end of the CS but in the absence of the US did not
induce fear learning, suggesting that this DA input is likely not
aversive by itself. Instead, our results suggest that DA input induces
plasticity in TS that biases associative learning to cues that are spe-
cifically paired with aversive outcomes. Furthermore, we found that
activation of DA input to TS during the CS enhances the CS-US
association during fear learning, supporting the hypothesis pro-
posed by Menegas et al.35 that DA input to TS may enhance “CS
associability”.
Traditionally, the amygdala, in particular LA, is established to be
the critical site for plasticity mediating the CS-US association during
fear learning5–8,10. LA receives inputs from both the thalamus and the
cortex, relaying information about the auditory CS70. Considerable
evidence indicates that potentiation of CS-evoked responses in LA
neurons underlies the acquisition of associative fear learning71–76.
Moreover, plastic changes in the activity of neurons located in the
central nucleus of the amygdala (CEA) have also been shown to
underlie acquisition of fear memories77–80. In line with these findings,
DA projections to the amygdala exhibit activation during fear
learning29,57,58; and in particular, non-canonical dorsal tegmental DA
neurons, that are located in the periaqueductal gray (PAG)/dorsal
raphe (DR)and project specifically to CEA, have been shown to encode
a PE signal to gate fear learning29. However, whether DA PE signaling
outside the amygdala circuitry contributes to the acquisition of CS-US
association during FC has remained largely unknown. Our findings
reveal that DA neurons projecting to a brain structure outside the
canonical amygdala circuitry encode a PE signal critical for driving
associative fear learning.
While the role of the amygdala is well-established, it is becoming
increasingly clear that acquisition of fear memories also involves
plasticity in brain structures beyond the traditional amygdala
circuitry81–84. Notably, despite earlier lesioning and anatomical studies
hinting at the involvement of the posterior striatal areas, including the
TS, in fear conditioning85–87, the contribution of TS in fear learning
remained largely elusive and has recently begun to be investigated11,12.
TS, also known as the auditory striatum, is a distinct subregion of the
dorsal striatum, characterized by the unique set of inputs it
receives13,14. Similar to LA, TS receives auditory inputs from both the
thalamus and the cortex85,87; and we here showed that auditory CS-
evoked responses are potentiated also in the TS asthe animals learned
the CS-US association during FC, consistent with previous reports11,12.
Previous studies have shown that activity of TS neurons is required for
fear retrieval11,12. Importantly, we further demonstrate that neuronal
activity in TS exhibits a PE-like activity pattern and is important for
acquisition of fear conditioning, indicating that a broader neural net-
work than the amygdala circuitry is indeed involved in acquiring the
CS-US association during fear learning.
One important question is whether and how TS is connected with
the canonical amygdala fear learning circuitry. Although, TS receives
strong input from the BLA13, whether these amygdala projections to TS
are critical for fear learning is largely unknown. Notably, BLA input to
the posterior striatum have been shown to facilitate plasticity in cor-
ticostriatal synapses88 suggesting that amygdala input might be critical
for the plasticity underlying fear acquisition in TS neurons. Future
studies investigating the role of amygdala input to TS during fear
acquisition will be important to address this question. Conversely,
although TS sends only sparse projections to the amygdala these
projections have recently been shown to be important for fear
retrieval12, suggesting a direct connection to the canonical amygdala
circuitry during fear conditioning. Further research on how TS input
affects neuronal activity in distinct components of the amygdala cir-
cuitry will be crucial for elucidating the role of TS in fear retrieval. In
addition, the role of TS projections to downstream structures outside
the amygdala circuitry during fear acquisition remains elusive.
An interesting finding in our study is the PE-like activity pattern
that we observed in TS. Similar to TS-projecting DA neurons, TS neu-
ronal activity exhibited larger responses to the US at the beginning of
FC when US presentation was unpredicted. As the CS-US association
was learned, the US responses decreased, and conversely, responses to
the CS gradually increased. Importantly, we found that PE coding
during FC was dependent on the DA input to TS. Activity of TS-
projecting SN neurons was shown to be important for the potentiation
of CS responses during fear conditioning in a recent study. However,
whether this was dependent on DA input to TS remained unclear as
inhibition of neuronal activity in SN lacked DA neuron specificity12.
Here, we demonstrated that ablation of specifically the TS-projecting
DA neurons prevented potentiation of CS responses in parallel with
impaired fear learning. In addition, we also found that the US
responses did not decrease during the course of FC in mice with
ablation of TS-projection DA neurons, consistent with the deficit in
acquiring the CS-US association in these mice. Notably, responses to
the first unpredicted US were not different between the lesioned and
control mice indicating that the somatosensory inputs to TS relaying
the US information were not affected by ablation of the DA input, but
rather the predictive coding of US responses in TS was impaired.
Together, these results reveal a PE-like activity in TS during fear
learning and that these fear learning-induced activity patterns in TS
were dependent on DA input during FC.
The principal neurons in TS are the medium spiny neurons
(MSNs), categorized into two types based on their expression of
dopamine D1 and D2 receptors, which constitute the direct and
indirect pathways, respectively14. Whether and how D1- and D2-MSNs
contribute to associative fear learning is an important question. A
recent study investigating the role of these MSNs in the ventralmost
portion of the TS, a region largely including the amygdalostriatal
Fig. 8 | Bidirectional manipulation of neuronal activity reveals the causal
contribution of TS in associative fear learning. a Schematic of the surgery.
bExample histological image showing expression of hM4D(Gi)-mCherry (red), TH
(green), and DAPI (blue) staining. Scale bar: 0.5 mm. cTop: schematic of the pro-
tocol. FC fear conditioning. Bottom: Schematic of CS and US presentations.
dFreezing to the CS for saline (n=8) andCNO(n= 8) groups (two-way repeated
measures ANOVA, FC: main effect of group, F
1,56
=5.08, P= 0.04; Fear recall: main
effect of group, F
1,56
=19.61,P=0.0006).eFreezing to the last CS during FC (two-
sided t-test, P=0.006).fFreezing to the first CS during fear recall (two-sided t-test,
P=0.03). gSchematic of the surgery. hExample histological image showing
expression of eArch-EYFP (green),TH (red), and DAPI (blue) staining.White vertical
tracks indicate optical fiber placements. Scale bar: 0.5 mm. iTop: schematic of the
protocol. Bottom: schematic of the US-Paired optogenetic inhibition. jFreezing to
the CS for EYFP (n=8) andeArch(n= 7) groups (two-way repeated measures
ANOVA, group × trial interaction, F
4,52
=4.40,P= 0.0039). kFreezing to the last CS
during FC (two-sided t-test, *P=0.04).lFreezing to the first CS d uring fear recall
(two-sided t-test, *P=0.01).mSchematic of the surgery. nExample histological
image showing expression of ChR2-EYFP(green), TH (red)and DAPI (blue) staining.
Whitevertical tracksindicate theoptical fiber placements.Scale bar: 0.5 mm. oTop:
schematic of the protocol. Bottom: schematic of the US-Paired optogenetic exci-
tation. pFreezing to the CS in EYFP (n= 8) and ChR2 (n=8)groups (two-way
repeatedmeasures ANOVA, maineffect of group, F
1,56
=5.8,P=0.03).rFreezing to
the last CS during FC (two-sided t-test,**P= 0.008). sFreezing to the firstCS during
fear recall (two-sided t-test, *P=0.03). Error bars represent mean ± s.e.m. Source
data are provided as a Source Data file.
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 15
Content courtesy of Springer Nature, terms of use apply. Rights reserved

transition area (ASt), found that D1- but not D2-MSNs exhibited
enhanced CS responses following fear conditioning11.Furthermore,
while optogenetic inhibition of D1-MSNs impaired fear retrieval, inhi-
bition of D2 neurons had no significant effect. However, the role of D1-
and D2-MSNs in associative fear learning across the larger, moredorsal
TS, where our study was focused, currently remains an open question.
If D1-, but not D2-, MSNs are similarly involved in fear learning in the
dorsalTS, it is possible that theaverage population activity observedin
our study was mainly driven by the activity of the D1-MSNs. It is also
possible that our optogenetic and chemogenetic manipulations of all
TS neurons actually influenced fear learning primarily through their
effect on D1-MSNs. Future research investigating the activity of these
neurons across the larger TS during fear learning will be necessary for
addressing this question.
a b c
6-OHDA
SNDA
TS
4 weeks
6-OHDA
Lesions
Hab FC Fear Recall
(10 CS) (5 CS-US) (10 CS)
DAY 1 DAY 2 DAY 3
d
% Freezing
e
f g
h i
-0.01
0
0.01
0.02
0.03
-0.01 0.01 0.03
Average dF/F
(Hab)
Average dF/F
(Fear Recall)
AAV5-
CamKII-
GCaMP6f
TS
Optical
Fiber
k
-0.01
0
0.01
0.02
-0.01 0 0.01 0.02
Average dF/F
(Hab)
Average dF/F
(Fear Recall)
GCaMP TH DAPI
TS
CEA
BLA
ASt
GP
TH DAPI
TS
CEA
BLA
ASt
GP
Control
TH DAPI
TS
CEA
BLA
ASt
GP
j
l
80
60
40
20
0
DAY1 DAY2 DAY3
Hab FC Fear
Recall
6-OHDA
Control
6-OHDA
GCaMP TH DAPI
TS
CEA
BLA
ASt
GP
CS
dF/F
Time (s)
Hab
0.1
0.05
0
0 10 20
CS
Hab
dF/F
Time (s)
0.1
0.05
0
0 10 20
CS US
Time (s)
FC
0 10
20
0.1
0.05
0
CS
Time (s)
Fear Recall
0 10
20
0.1
0.05
0
CS US
FC
Time (s)
0 10
20
0.1
0.05
0
CS
Fear Recall
Time (s)
0 10
20
0.1
0.05
0
-0.01
0
0.01
0.02
0.03
Average dF/F
Hab Fear
Recall
*
-0.05
0
0.05
0.1
Average dF/F
CS1 CS5
FC
**
0
0.1
0.2
FC
US1 US5
**
-0.01
0
0.01
0.02
Average dF/F
Hab Fear
Recall
-0.04
0
0.04
0.08
Average dF/F
CS1 CS5
FC
0
0.1
0.2
0.3
FC
US1 US5
-0.01
0
0.01
0.02
0.03
Control 6-OHDA
Hab.
Average dF/F
-0.05
0
0.05
0.1
Control 6-OHDA
FC (last CS)
**
Average dF/F
0
0.1
0.2
0.3
Control 6-OHDA
FC (first US)
Average dF/F
-0.01
0
0.01
0.02
0.03
Control 6-OHDA
Fear Recall
*
Average dF/F
m
-0.05
0
0.05
0.1
Control 6-OHDA
FC
(CS5 − CS1)
∆ Average dF/F
**
-0.1
-0.05
0
0.05
Control 6-OHDA
∆ Average dF/F
FC
(US5 − US1)
*
n
-0.015
0.005
0.025
Control 6-OHDA
∆ Average dF/F
CS
(FR − Hab)
**
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 16
Content courtesy of Springer Nature, terms of use apply. Rights reserved

PE signaling during FC has previously been demonstrated in LA
neurons, where the PE coding was shown to set the associative
memory strength89. In line with this notion, we here demonstrated
that optogenetic excitation and inhibition of TS neuronal activity at
the time of the US enhanced and impaired fear learning, respectively.
These results indicate that the PE coding by TS neurons drives
learning about the CS-US association. PE coding in LA neurons was
shown to be mediated by a feedback neuronal circuitry involving
projections from ventrolateral PAG (vlPAG) to LA89,90, and likely also
involves projections from the cerebellum to vlPAG91,92. Whether
feedback neural circuits are recruited to drive PE coding in TS neu-
rons, as well as in TS-projecting DA neurons, and the components of
these neural circuits will be important questions for future research.
Another important question that remains to be investigated is the
differential contributions of LA versus TS neuronal activity in driving
associative fear learning. Overall, our study reveals that a DA PE
signal in a non-canonical nigrostriatal circuitry is crucial for driving
associative fear learning.
Methods
Subjects
All procedures were conducted in accordance with the guidelines of
the German Animal Welfare Act and were approved by the local
authorities (Regierungspräsidium Darmstadt; protocol number 1256).
Adult male C57BL/6N (Charles River or Janvier Labs) and heterozygous
DAT-Cre mice93 (backcrossed with C57BL/6N) aged older than
3 months at the start of experiments were used. All experimental
groups were matched for age. For 6-OHDA lesioning, optogenetic and
chemogenetic experiments, littermate mice were allocated to experi-
mental and control groups.All mice were individually housed on a 12-h
light/dark cycle. All experiments were performed during the
light cycle.
Viral constructs
We obtained AAV5-CAG-Flex-GCaMP6m-WPRE-SV40, pENN-AAV5-
CaMKII-GCaMP6f-WPRE-SV40, AAV5-CAG-dLight1.1, AAV5-hSyn-
hM4D(Gi)-mCherry and AAV5-hSyn-mCherry from Addgene, and
AAV5-EF1a-DIO-hChR2(H134R)-EYFP, AAV5-EF1a-DIO-eArch3.0-EYFP,
AAV5-EF1a-DIO-EYFP, AAV5-CaMKIIa-hChR2(H134R)-EYFP, AAV5-CaM-
KII-ArchT-GFP, AAV5-CaMKIIa-EYFP and AAV5-CaMKII-GFP from the
University of North Carolina Vector Core.
Surgical procedures
Animals were anesthetized using isoflurane (1–2%) and placed in a
stereotaxic frame. At the onset of anesthesia, all animals received
intraperitoneal injections of atropine (0.05 mg/kg) and subcutaneous
injections of carprofen (4 mg/kg) and dexamethasone (2 mg/kg). Eye
gel was applied on the eyes to prevent dehydration of the cornea.
Lidocaine cream was applied on the scalp as a local anesthetic. The
animal’s temperature was maintained for the duration of the surgical
procedure using a heating blanket. Anesthesia levels were monitored
throughout the surgery, and the concentration of isoflurane adjusted
so that the breathing rate never fell below 1 Hz.
For GCaMP fiber photometry recordings of DA terminal activity in
the TS, DAT-cre mice were injected with 0.5–1μl of AAV5-CAG-Flex-
GCaMP6m-WPRE-SV40 (final titer ~1 × 1013 pp per ml) in the SN (3.2 mm
posterior to bregma, 1.2 mm lateral to the midline and 4.5 mm ventral
to bregma)94 at 50 nl/min using a 10 μl syringe with a 33-gauge needle
controlled by an injection pump. The needle was left in place for an
additional 10-15 min before slowly being withdrawn. Following infu-
sion of the virus, optical fibers (400 μmcorediameter,0.48NA,Doric
Lenses) were slowly inserted through the craniotomy above the TS (AP:
−1.3 mm, ML: 2.95 mm and DV: 2.75–3.0 mm)94. In a subset of animals,
recordings were performed bilaterally. The optical fiber was then
anchored to the skull using skull screws and dental cement (Paladur).
For GCaMP fiber photometry recordings of neuronal activity in
the TS, wild-type C57BL/6 N mice were injected with 150 nl of AAV5-
CaMKII-GCaMP6f-WPRE-SV40 (final titer ~1 × 1013 pp per ml) in the TS
(AP: −1.3 mm,ML: 2.95mm and DV: 3.25 mm) at 50 nl/min using a 10μl
syringe with a 33-gauge needle controlled by an injection pump. The
needle was left in place for an additional 10–15 min before slowly being
withdrawn. Following infusion of the virus, optical fibers (400 μmcore
diameter, 0.48 NA, Doric Lenses) were slowly inserted through the
same craniotomy to a depth of 2.75–3.0 mm below the bregma. The
optical fiber was then anchored to the skull using skull screws and
dental cement (Paladur).
For dLight fiber photometry experiments, wild-type C57BL/6N
mice were injected with 150–300 nl of AAV5-CAG-dLight1.1 (final titer
4.2 × 1012 pp per ml) in the TS (AP: −1.3 mm, ML: 2.95 mm and DV:
3.25 mm) as described above. Following infusion of the virus, an
optical fiber (400 μm core diameter, 0.48 NA, Doric Lenses) was slowly
inserted through the same craniotomy to a depth of 2.75–3.0 mm
below the bregma. In a subset of animals, recordings were performed
bilaterally. The optical fibers were then anchored to the skull using
skull screws and dental cement (Paladur).
6-OHDA lesioning of TS-projecting DA neurons was performed as
previously described16,95. Thirty min before the surgery, wild-type
C57BL/6N mice received intraperitoneal injections of a solution
(10 ml/kg) containing desipramine hydrochloride (2.5 mg/ml, Sigma-
Aldrich) and pargyline hydrochloride (0.5 mg/ml, Sigma-Aldrich) dis-
solved in0.9% saline (pH 7.4), in order to prevent uptake of 6-OHDA in
noradrenaline neurons. 6-OHDA (10 mg/ml; 6-OHDA hydrobromide,
Sigma-Aldrich) was dissolved in a vehicle solution immediately before
the surgeries. The vehicle contained saline (0.9%) and a small amount
of ascorbic acid (0.2%) to prevent oxidation of 6-OHDA. To further
minimize oxidation of 6-OHDA, the solution was kept on ice, wrapped
in aluminum foil and used within 4 h after preparation. During the
surgery, animals were injected bilaterally with 150–300 nl of the
6-OHDA solution or the vehicle in the TS using the coordinates
described above. The injections were performed at 50 nl/min speed
using a 10 μl syringe with a 33-gauge needle, which was controlled by
an injection pump. The needle was left in place for an additional 10-
Fig. 9 | DA input is critical for PE-like activity and potentiation of CS responses
in TS during associative fear learning. a Schematic of the surgeries. bSchematic
of the procedure. Hab tone habituation, FC fear conditioning. cFreezing to the CS
in control (n= 10) and 6-OHDA (n= 8) groups. Example image showing expression
of GCaMP (green; top), TH (red), and DAPI (blue) staining of a control (d)anda6-
OHDA-lesioned (h) animal. White vertical track indicates the optical fiber place-
ment. Scale bar: 0.5 mm. Average change in fluorescence around the CS pre-
sentation (gray area) in control (e,n= 14) and 6-OHDA (i,n= 14) groups. Left:
average change in fluorescence during CS1 and CS5 of FC (f,control:**P=0.0011,
and j6-OHDA: P= 0.5, two-sided signed-rank test). Right: average change in fluor-
escence during US1 and US5 of FC in the control (**P= 0.0012, two-sided signed-
rank test) and 6-OHDA-lesioned (P= 0.62, two-sided signed-rank test) mice. Left:
average change in fluorescence during the CS in the control (g,n=14, *P=0.012,
two-sided signed-rank test) and 6-OHDA lesioned (k,n= 14) mice. Right: Scatter
plots showing CS responses of each recording site in the control (g) and 6-OHDA
(k) mice during Hab and Fear Recall. lComparison of the average signal to the CS
(left: Hab,P= 0.76; right: Fear recall, P= 0.0054, two-sided rank-sum tests), and CS
(last-CS; P= 0.0012, two-sided rank-sum test) and unpredicted US (first-US,
P= 0.10, two-sided rank-sum test)during FC (middle) betweencontrol and 6-OHDA
groups. mComparison of change in the average signal between first and last CSs
(left) and USs (right) between control and 6-OHDA groups (ΔCS: **P=0.004;ΔUS:
*P= 0.04, two-sided rank-sum test). nComparison of change in the average signal
between Hab and Fear Recall(FR) between the two groups (**P= 0.0019, two-sided
rank-sum test). Shaded regions and error bars represent mean ± s.e.m. Source data
are provided as a Source Data file.
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 17
Content courtesy of Springer Nature, terms of use apply. Rights reserved

15 min before slowly being withdrawn. The scalp incision was then
closed using sutures. In experiments with GCaMP recordings, mice
received virus injections and optical fiber implantation as described
above one week after 6-OHDA injections. In a subset of animals in the
control and 6-OHDA groups, recordings were performed bilaterally.
For optogenetic excitation or inhibition of DA terminals in the TS,
DAT-cre mice were injected bilaterally in the SN with 0.5–1μlofAAV5-
EF1a-DIO-hChR2(H134R)-EYFP (final titer 4.4 x 1012 pp per ml), AAV5-
EF1a-DIO-eArch3.0-EYFP (final titer 5.6 × 1012 pp per ml) or AAV5-EF1a-
DIO-EYFP (final titer 4.3 x 1012 pp per ml) per hemisphere using the
coordinates described above. Virus injection was performed as
described above and followed by implantation of optical fibers (200
μm core diameter, 0.22 NA, Thorlabs) bilaterally above the TS (AP:
−1.3 mm, ML: 2.95 mm) to a depth of 2.75–3.0 mm bel ow br egma. The
optical fibers were then anchored to the skull using skull screws and
dental cement (Paladur).
For optogenetic excitation or inhibition of neuronal activity in the
TS, wild-type C57BL/6N mice were injected bilaterally in the TS with
150 nl of AAV5-CaMKIIa-hChR2(H134R)-EYFP (final titer 4.1 x 1012 pp per
ml), AAV5-CaMKII-ArchT-GFP (final titer 5.2 ×1012 pp per ml) or AAV5-
CaMKIIa-EYFP (final titer 3.6 x 1012 pp per ml) per hemisphere using the
coordinates described above. Virus injection was performed as
described above. Optical fibers (200 μm core diameter, 0.22 NA,
Thorlabs) were then implanted bilaterally above the TS using the
coordinates described above. The optical fibers were then anchored to
the skull using skull screws and dental cement (Paladur).
For chemogenetic experiments, C57BL/6N mice were bilaterally
injected with 150 nl of AAV5-hSyn-hM4D(Gi)-mCherry (final titer 8.6 x
1012 particles per ml) or AAV5-hSyn-mCherry (final titer 2.3 × 1013 par-
ticles per ml) per each hemisphere in the TS using the coordinates
described above. Virus injection was performed as described above.
The scalp incision was then closed using sutures at the end of the
surgery.
Behavior
Fear conditioning. Fear conditioning and fear recall took place in two
different contexts (A and B). Context A consisted of a square chamber
with an electrical grid floor (Med Associates) used to deliver the
footshock US. Context B consisted of a white teflon cylindrical cham-
ber with bedding material on the floor. The chambers were located
inside a sound attenuating box and were cleaned with 1% acetic acid
before and after each session. Before fear conditioning experiments
started, all mice were habituated to handling and being connected to
the patchcord and habituated tocontexts A and B for 10-15 mineach in
a counterbalanced fashion. On day 1, mice received a tone habituation
session which started following a 2min baseline period in context A
and consisted of 10 presentations of the CS (4kHz tone, 75dB) with a
random intertrial interval (ITI) of 40–120 s. On day 2, mice underwent
fear conditioning consisting of five pairings of the CS with a US (1 s
footshock, ITI: 40-120 s). The CS was 10 s and 30 s long in photometry
and optogenetic/chemogenetic experiments, respectively. In 6-OHDA
experiments, both 10 s and 30 s long CSs were used. Comparable
results were obtained with both CS durations (p> 0.05) and hence the
data was pooled. Furthermore, in 6-OHDA experiments, a group of
mice underwent visual fear conditioning in which the CS was a 30 s
long light cue (yellow LED diode, 0.25 W). The US intensity was
0.4–0.5 mA for photometry and 0.5 mA for 6-OHDA, optogenetic
inhibition and chemogenetic experiments. For optogenetic excitation
experiments where mice received a weak training, the US intensity was
0.35 mA. The onset of the US coincided with the offset of the CS. On
day 3, mice received a fear reca ll session consisting of 10 presentations
of the CS alone in context B for photometry experiments. For 6-OHDA,
optogenetic and chemogenetic experiments, fear recall test consisted
of 5 presentations of the CS on day 3. For contextual fear conditioning,
mice received context habituation (context A) for 10min on day 1. On
day 2, following a 2 min baseline period mice received 5 presentations
of the US (0.5 mA) with a random intertrial interval (ITI) of 40–120 s in
context A. On day 3, mice were placed back to context A for 10 min for
their context testing. For the unpaired training protocol, mice received
five presentations of the CS and the US inan explicitlyunpaired fashion
(ITI: 40-120 s) on day 2. At the end of fiber photometry experiments,
one subset of mice was further tested for predicted versus unpredicted
US presentations (5 CS-US pairings and 5 USs presented in a random
order). Another subset of mice were further trained on a partial fear
conditioning protocol in which they received 5 CS-USpairings and 5 CS
presentations in a random order to examine US omission responses.
Throughout the experiments, the behavior of mice was recorded to
video and scored by experienced observers blind to the experimental
condition. Behavioral freezing, defined as the absence of all bodily
movements except breathing-related movement96,wasusedasthe
measure of fear.
Reward Tasks. Mice in the photometry experiments underwent a
reward task following the fear conditioning protocol. To motivate
animals, their access to water was restricted until they reached 85% of
their body weight. Animals were placed in an operant chamber con-
taining a liquid delivery port. Nosepokes into the delivery port that
followed the previous nosepoke byat least a variable inter-trial interval
of 3–5 s triggered the delivery of liquid reward (10% or 25% sucrose
solution). Animals were first trained on 90% probability of reward
delivery on day 1, in which they quickly learned the task. The next day,
photometry recordings were performed while rewards were delivered
randomly at a 50% probability, making reward delivery unpredictable.
To test whether prior fear conditioning affected reward responses, a
separate cohortof animals underwent the rewardtask both before and
after fear conditioning. Reward responses before fear conditioning
were comparable to ones after fear conditioning (n=7; p=0.37,
signed-rank test).
In 6-OHDAexperiment,mice underwenta reward learning task for
5 consecutive days. Animals were placed in an operant chamber con-
taining a liquid delivery port. Each session consisted of 50 pairings of a
CS (8 kHz tone, 75 dB, 5 s long) and a reward (10% sucrose solution)
with a random ITI of 20–30 s. If the animal accessed the reward port
during the CS presentation, a 10 μl reward was delivered.
Real-time place preference test.Attheendoffearconditioning
protocol, a subset of mice in the optogenetic experiments underwent
place preference, open field, and elevated plus maze tests. Real-time
place preference test was conducted in a custom-made chamber (50 ×
50 x 50 cm, wooden gray box) divided into two compartments. The
test consisted of two 10 min phases (Fig. 5m). During the first phase,
one side of the chamber was randomly assigned as the laser ON side.
Mice were individually connected to the patch cords and placed in the
laser OFF side of the chamber. Each time the mouse entered the laser
ON side laser light was delivered until the mouse crossed back to the
OFF side. In the second phase, the sides were switched and the pre-
viously laser OFF side became laser ON side in order to counterbalance
each side.
Open field test. The custom made open field chamber (50 × 50 x
50 cm, wooden gray box) was divided into a central area (center 25 ×
25 cm) and an outer area (periphery). The open field testconsisted of a
9 minute session with three alternating 3-minute epochs (OFF-ON-OFF
epochs) in which laser was delivered during the ON epoch (Fig. 5n).
Elevated Plus Maze (EPM). EPM consisted of two open arms (30 ×
5 cm), twoclosed arms (30 × 5 × 15 cm) and a central area (5 × 5 × 5 cm).
Themazewasplaced40cmabovethefloor. Mice were individually
connected to the patch cords and placed in the center. The test con-
sisted of a 9 minute session with three alternating 3-minute epochs
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 18
Content courtesy of Springer Nature, terms of use apply. Rights reserved

(OFF-ON-OFF epochs) in which laser was delivered during the ON
epoch (Fig. 5o).
GCaMP and dLight recordings using fiber photometry
Animals were injected with viral vectors and implanted with optical
fibers in the TS, as described above. After a waiting period of 3-4 weeks
to allow for surgical recovery and virus expression, mice were con-
nected to 400 μm patch cords (Doric Lenses). Fluorescence was
measured by delivering 465 nm excitation light through the patch cord
and separating the emission light at 525 nm with a beamsplitter
(Fluorescence MiniCube FMC3, Doric Lenses). The emission light was
collected using a Femtowatt Silicon Photoreceiver (Model # 2151,
Newport). The voltage output of the photoreceiver was then digitized
at 2 kHz (Digital Lynx SX, Neuralynx). After animals were habituated to
handling and being connected to the patch cord, they underwent the
behavioral protocol as described above. Photometry recordings were
performed throughout the protocol.
Analysis of fiber photometry data
The voltage output of the photoreceiver, representing fluctuations in
fluorescence, was downsampled to 10 Hz. The change in fluorescence
evoked by the CS (dF/F) was then calculated by subtracting from
each trace the baseline fluorescence (average during the 5 s before
CS onset) and dividing it by the baseline fluorescence. dF/F traces
were then averaged separately for each animal for tone habituation
(Hab: average of 10 CSs), fear conditioning (Fear Cond., average of 5
CSs) and fear recall (average of 10 CSs). To examine responses to
the CS, we further averaged dF/F values in the 5 s following CS onset
in each session. To quantify responses to the US, we averaged the
dF/F values at the time of the US (1 s) during Fear Cond. for each
animal. To examine the change in responses to the CS and the US
during Fear Cond., we compared responses to the first and last
CSs and USs for each animal. To examine responses to reward,
average dF/F was calculated for rewarded trials using the baseline
fluorescence 3 s before noseport entry. Reward responses were
quantified by averaging the dF/F in the 3s following noise port entry
for each animal.
Chemogenetic experiments
Three to four weeks after viral injections, mice underwent the che-
mogenetic experiment. Thirty min before the fear conditioning ses-
sion, mice expressing the inhibitory DREADD hM4D(Gi)-mCherry or
the control fluorophore mCherry received systemic injections of the
DREADD agonist clozapine N-oxide (CNO, Sigma-Aldrich; 1–1.5 mg/kg
dissolved in saline) or saline. Fear conditioning was performed as
described above. The next day, animals underwent the fear recall test
drug free.
Optogenetic experiments
For bilateral optogenetic manipulations during behavior, the implan-
ted optical fibers (200 μm core diameter, 0.22 NA, Thorlabs) were
connected to 200 μm patch cords (Thorlabs) with zirconia sleeves and
the patch cords were connected to a light splitting rotary joint (FRJ
1x2i, Doric Lenses) that was connected to a laser with a 200 μmpatch
cord (Thorlabs). For mice expressing the light-activated excitatory
opsin ChR2 (ChR2-EYFP) and their EYFP controls, blue light pulses
were delivered from a 473 nm laser (LuxX473, Omicron). Laser power
at the tip of the optic fiber was 5–10mW. DA terminals in TS were
excited at the time of the CS (CS-Paired), the US (US-Paired), during the
ITIs (ITI Control) or in the absence of the US (No-US Control), and TS
neuronal activity was excited at the time of the US (US-Paired) during
fear conditioning. For CS-Paired animals, 5-ms light pulses were
delivered at 20 Hz during the CS presentation. For US-Paired animals,
5-ms light pulses were delivered at 20 Hz for 1 or 3 seconds. For ITI
Control group, 5-ms light pulses were delivered at 20 Hz for 30 s
during the ITIs. For No-US Control group, 5-ms light pulses were
delivered at 20 Hz for 3s at the end of the CS. The laser was turned on
1 s before the US onset to 1 s after the US offset.
For mice expressing the light-activated inhibitory opsin eArch
(eArch-EYFP) or ArchT (ArchT-GFP) and their EYFP or GFP controls,
yellow light pulses were delivered from a DPSS 594 nm laser (Omi-
cron). Laser power at the tip of the optic fiber was 5–10 mW. DA
terminals in TS or TS neurons were inhibited atthe timeof the US from
1 s before to 2 s after CS offset. The laser was then turned off gradually
using a 2s ramp to avoid rebound excitation.
Histology
At the end of the experiments, mice were deeply anesthetized with
sodium pentobarbital and were transcardially perfused with 4% par-
aformaldehyde and 15% picric acid inphosphate-buffered saline (PBS).
Brains were removed, post-fixed overnight and coronal brain slices
(60 µm) were sectioned using a vibratome (VT1000S, Leica). Standard
immunohistochemical procedures were performed on free-floating
brain slices. Briefly, sections were rinsed with PBS and then incubated
in a blocking solution (10% horse serum, 0.5% Triton X-100 and 0.2%
BSA in PBS) for 1 h at room temperature. Slices were then incubated in
a carrier solution (1% horse serum, 0.5% Triton X-100 and 0.2% BSA in
PBS) containing the primary antibody overnight at room temperature.
The next day, the sections were washed in PBS and then incubated in
the same carrier solution containing the secondary antibody overnight
at room temperature. The following primary antibodies were used:
polyclonal rabbit anti-tyrosine hydroxylase (TH, catalog # 657012,
1:1000, Calbiochem), monoclonal mouse anti-TH (catalog # MAB318,
1:1000, Millipore), polyclonal guinea pig anti-TH (catalog # 213004,
1:1000, Synaptic Systems), polyclonal rabbit anti-GFP (catalog #
A11122, 1:1000, Life Technology), polyclonal chicken anti-GFP (catalog
# AB13970, Abcam), and mCherry monoclonal anti-rat (catalog #
M11217, 1:1000) Invitrogen). The following secondary antibodies were
used: Alexa Fluor 568 goat anti-rabbit (catalog # A11011, 1:1000,
Thermo Fisher Scientific, Invitrogen), Alexa Fluor 568 goat anti-mouse
(catalog # A11004, 1:1000, Thermo Fisher Scientific, Invitrogen), Alexa
Fluor 568 goat anti-guinea pig (catalog # A11075, 1:1000, Invitrogen),
Alexa Fluor 405 goat anti-rabbit (catalog # A31556, 1:750, Invitrogen),
Alexa Fluor 488 goat anti-rabbit (catalog # A11008, 1:1000, Thermo
Fisher Scientific, Invitrogen), Alexa Fluor 488 goat anti-chicken (cata-
log # AB150173, 1:1000, Abcam), Alexa Fluor 568 goat anti-rat (catalog
# A11077, 1:1000, Invitrogen). For DAPI staining, sections were incu-
bated for 10 min in 0.1 M PBS containing 0.02% DAPI (catalog # D1306,
Molecular Probes, Invitrogen). Finally, all sections were washed with
PBS, mounted on slides embedded with a mounting medium for
fluorescence (VECTASHIELD®, Vector Laboratories), and coverslipped.
Statistics
Data were statistically analyzed using GraphPad Prism (GraphPad
Software) and MATLAB (Mathworks). All statistical tests were two-
tailed and had an αlevel of 0.05.All error bars shows.e.m. All ANOVAs
were followed by Bonferroni post hoc tests if significant main or
interaction effects were detected. No statistical methods were used to
predetermine sample size, but our sample sizes were similar to those
generally used in the fear conditioning field. Animals were randomly
assigned to experimental groups before the start of each experiment
after ensuring that all experimental groups were matched for age. For
6-OHDA lesioning, chemogenetic and optogenetic experiments,
experimental and control groups were matched from littermate mice.
All results were obtained using groups of mice that were run in several
cohorts.
Reporting summary
Further information on research design is available in the Nature
Portfolio Reporting Summary linked to this article.
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 19
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Data availability
All data in the figures and supplementary data are provided in the
Source Data file. Raw data areavailable from the corresponding author
upon request. Source data are provided with this paper.
Code availability
This paper does not report original code. The custom codes used for
data analysis in this study are available from the corresponding author
upon request.
References
1. LeDoux,J.E.Comingtotermswithfear.Proc. Natl Acad. Sci. USA
111,2871–2878 (2014).
2. Craske, M. G. et al. Anxiety disorders. Nat. Rev. Dis. Prim. 3,17024
(2017).
3. Duits, P. et al. Updated meta-analysis of classical fear conditioning
in the anxiety disorders. Depress. Anxiety 32,239–253 (2015).
4. Shin, L. M., Rauch, S. L. & Pitman, R. K. Amygdala, medial prefrontal
cortex, and hippocampal function in PTSD. Ann. N. Y Acad. Sci.
1071,67–79 (2006).
5. Duvarci,S.&Pare,D.Amygdalamicrocircuitscontrollinglearned
fear. Neuron 82,966–980 (2014).
6. Johansen, J. P., Cain, C. K., Ostroff, L. E. & LeDoux, J. E. Molecular
mechanisms of fear learning and memory. Cell 147,509–524 (2011).
7. LeDoux, J. E. Emotion circuits in the brain. Annu Rev. Neurosci. 23,
155–184 (2000).
8. Maren,S. & Quirk, G. J. Neuronal signaling of fear memory. Nat. Rev.
Neurosci. 5,844–852 (2004).
9. Sigurdsson, T., Doyère, V., Cain, C. K. & LeDoux, J. E. Long-term
potentiation in the amygdala: a cellular mechanism of fear learning
and memory. Neuropharmacology 52,215–227 (2007).
10. Tovote, P., Fadok, J. P. & Lüthi, A. Neuronal circuits for fear and
anxiety. Nat. Rev. Neurosci. 16,317–331 (2015).
11. Kintscher, M., Kochubey, O. & S chneggenburger, R. A striatal circuit
balances learned fear in presence and absence of sensory cues.
Elife 12, e75703 (2023).
12. Chen, A. P. F. et al. Nigrostriatal dopamine modulates the striatal-
amygdala pathway in auditory fear conditioning. Nat. Commun. 14,
7231 (2023).
13. Hunnicutt, B. J. et al. A comprehensive excitatory input map of the
striatum reveals novel functional organization. Elife 5, e19103 (2016).
14. Valjent, E. & Gangarossa, G . The tail of the striatum: from anatomy to
connectivity and function. Trends Neurosci. 44,203–214 (2021).
15. Menegas, W. et al. Dopamine neurons projecting to the posterior
striatum form an anatomically distinct subclass. eLife 4,
e10032 (2015).
16. Menegas,W.,Akiti,K.,Amo,R.,Uchida,N.&Watabe-Uchida,M.
Dopamine neurons projecting to the posterior striatum reinforce
avoidance of threatening stimuli. Nat. Neuro. 21,1421–1430 (2018).
17. Akiti, K. et al. Striatal dopamine explains novelty-induced behavioral
dynamics and individual variability in threat prediction. Neuron 110,
3789–3804 (2022).
18. Zafiri, D. & Duvarci, S. Dopaminergic circuits underlying associative
aversive learning. Front. Behav. Neurosci. 16, 1041929 (2022).
19. Rescorla, R. A., & Wagner, A. R. A theory of Pavlovian conditioning:
variations in the effectiveness of reinforcement and nonreinforce-
ment. in Classical Conditioning II: Current Research and Theory (eds.
Black, A. H. & Prokasy, W. F.) 64–99 (Appleton Century Crofts, New
York) (1972).
20. Bayer,H.M.&Glimcher,P.W.Midbraindopamineneuronsencode
a quantitative reward prediction error signal. Neuron 47,129–141
(2005).
21. Eshel, N. et al. Arithmetic and local circuitry underlying dopamine
prediction errors. Nature 525,243–246 (2015).
22. Eshel, N., Tian, J., Bukwich, M. & Uchida, N. Dopamine neurons share
common response function for reward prediction error. Nat. Neuro.
19,479–486 (2016).
23. Schultz, W., Dayan, P. & Montague, P. R. A neural substrate of pre-
diction and reward. Science 275,1593–1599 (1997).
24. Steinberg, E. E. et al. A causal link between prediction errors,
dopamine neurons and learning. Nat. Neurosci. 16, 966–973
(2013).
25. Salinas-Hernández, X. I. et al. Dopamine neurons drive fear extinc-
tion learning by signaling the omission of expected aversive out-
comes. Elife 7, e38818 (2018).
26. Luo, R. et al. A dopaminergic switch for fear to safety transitions.
Nat. Commun. 9,2483(2018).
27. Cai, L. X. et al. Distinct signals in the medial and lateral VTA dopa-
mine neurons modulate fear extinction at different times. Elife 9,
e54936 (2020).
28. Salinas-Hernández,X.I.,Zafiri, D., Sigurdsson, T. & Duvarci, S.
Functional architecture of dopamine neurons driving fear extinction
learning. Neuron 111, 3854–3870 (2023).
29. Groessl, F. et al. Dorsal tegmental dopamine neurons gate asso-
ciative learning of fear. Nat. Neuro. 21,952–962 (2018).
30. Duvarci, S. Dopaminergic circuits controlling threat and safety
learning. Trends Neurosci. 47,1014–1027 (2024).
31. Likhtik, E. & Johansen, J. P. Neuromodulation in circuits of aversive
emotional learning. Nat. Neurosci. 22,1586–1597 (2019).
32. Pezze,M.A.&Feldon,J.Mesolimbicdopaminergicpathwaysinfear
conditioning. Prog. Neurobiol. 74,301–320 (2004).
33. Verharen, J. P. H., Zhu, Y. & Lammel, S. Aversion hot spots in the
dopamine system. Curr. Opin. Neurobiol. 64,46–52 (2020).
34. Lammel, S. et al. Diversity of transgenic mouse models for selective
targeting of midbrain dopamine neurons. Neuron 85,429–438
(2015).
35. Menegas, W., Babayan, B. M., Uchida, N. & Watabe-Uchida, M.
Opposite initialization to novel cues in dopamine signaling in ven-
tral and posterior striatum in mice. ELife 6, e21886 (2017).
36. Roesch, M. R., Calu, D. J. & Schoenbaum, G. Dopamine neurons
encode the better option in rats deciding between differently
delayed or sized rewards. Nat. Neurosci. 10,1615–1624 (2007).
37. Adrover,M.F.,Shin,J.H.&Alvarez,V.A.Glutamateanddopamine
transmission from midbrain dopamine neurons share similar
release properties but are differentially affected by cocaine. J.
Neurosci. 34,3183–3192 (2014).
38. Tritsch, N. X., Ding, J. B. & Sabatini, B. L. Dopaminergic neurons
inhibit striatal output through non-canonical release of GABA.
Nature 490,262–266 (2012).
39. Patriarchi, T. et al. Ultrafast neuronal imaging of dopamine
dynamics with designed genetically encoded sensors. Science
360, eaat4422 (2018).
40. Maren, S., Phan, K. L. & Liberzon, I. The contextual brain: implica-
tions for fear conditioning, extinction and psychopathology. Nat.
Rev. Neurosci. 14,417–428 (2013).
41. Fadok, J. P., Dickerson, T. M. & Palmiter, R. D. Dopamine is necessary
for cue-dependent fear conditioning. J. Neurosci. 36, 11089–11097
(2009).
42. Greba, Q., Gifkins, A. & Kokkinidis, L. Inhibition of amygdaloid
dopamine D2 receptors impairs emotional learning measured with
fear-potentiated startle. Brain Res. 899,218–226 (2001).
43. Guarraci, F. A., Frohardt, R. J. & Kapp, B. S. Amygdaloid D1 dopamine
receptor involvement in Pavlovian fear conditioning. Brain Res. 827,
28–40 (1999).
44. Guarraci,F.A.,Frohardt,R.J.,Falls,W.A.&Kapp,B.S.Theeffectsof
intra-amygdaloid infusions of a D2 dopamine receptor antagonist
on Pavlovian fear conditioning. Behav. Neurosci. 114,647–651
(2000).
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 20
Content courtesy of Springer Nature, terms of use apply. Rights reserved

45. Lamont, E. W. & Kokkinidis, L. Infusion of dopamine D1 receptor
antagonist SCH 23390 into the amygdala blocks fear expression in
a potentiated startle paradigm. Brain Res 795,128–136 (1998).
46. Nader, K. & LeDoux, J. E. The dopaminergic modulation of fear:
quinpirole impairs the recall of emotional memories in rats. Behav.
Neurosci. 113,152–165 (1999).
47. Fadok, J. P., Darvas, M., Dickerson, T. M. & Palmiter, R. D. Long-term
memory for pavlovian fear conditioning requires dopamine in the
nucleus accumbens and basolateral amygdala. PLoS One 5,
e12751 (2010).
48. Zweifel, L. S. et al. Disruption of NMDAR-dependent burst firing by
dopamine neurons provides selective assessment of phasic
dopamine-dependent behavior. Proc.NatlAcad.Sci.USA106,
7281–7288 (2009).
49. Zweifel, L. S. et al. Activation of dopamine neurons is critical for
aversive conditioning and prevention of generalized anxiety. Nat.
Neurosci. 14,620–626 (2011).
50. Brischoux, F., Chakraborty, S., Brierley, D. I. & Ungless, M. A. Phasic
excitation of dopamine neurons in ventral VTA by noxious stimuli.
Proc.NatlAcad.Sci.USA106, 4894–4899 (2009).
51. Cohen,J.Y.,Haesler,S.,Vong,L.,Lowell,B.B.&Uchida,N.Neuron-
type-specific signals for reward and punishment in the ventral
tegmental area. Nature 482,85–88 (2012).
52. Gore,B.B.,Soden,M.E.&Zweifel,L.S.Visualizationofplasticityin
fear-evoked calcium signals in midbrain dopamine neurons. Learn
Mem. 21,575–579 (2014).
53. Guarraci, F. A. & Kapp, B. S. An electrophysiological characteriza-
tion of ventral tegmental area dopaminergic neurons during dif-
ferential pavlovian fear conditioning in the awake rabbit. Behav.
Brain Res 99,169–179 (1999).
54. Jo, Y. S., Heymann, G. & Zweifel, L. S. Dopamine neurons reflect the
uncertainty in fear generalization. Neuron 100,916–925 (2018).
55. Matsumoto, M. & Hikosaka, O. Two types of dopamine neurons
distinctly convey positive and negative motivational signals. Nature
459,837–841 (2009).
56. Mileykovskiy, B. & Morales, M. Duration of inhibition of ventral teg-
mental area dopamine neurons encodes a level of conditioned fear.
J. Neurosci. 31,7471–7476 (2011).
57. Lutas, A. et al. State-specific gating of salient cues by midbrain
dopaminergic input to basal amygdala. Nat. Neurosci. 22,
1820–1833 (2019).
58. Tang, W., Kochubey, O., Kint scher, M. & Schneggenburger, R. A VTA
to basal amygdala dopamine projection contributes to signal sali-
ent somatosensory events during fear learning. J. Neurosci. 40,
3969–3980 (2020).
59. Lammel, S. et al. Unique properties of mesoprefrontal neurons
within a dual mesocorticolimbic dopamine system. Neuron 57,
760–773 (2008).
60. Roeper, J. Dissecting the diversity of midbrain dopamine neurons.
Trends Neurosci. 36, 336–342 (2013).
61. Beier,K.T.etal.Circuitarchitecture of VTA dopamine neurons
revealed by input-output maping. Cell 162,622–634 (2015).
62. Lerner, T. N. et al. Intact-brain analyses reveal distinct infor-
mation carried by SNc dopamine subcircuits. Cell 162,
635–647 (2015).
63. Farassat, N. et al. In vivo functional diversity of midbrain dopamine
neurons within identified axonal projections. Elife 8, e48408 (2019).
64. Glimcher, P. W. Understanding dopamine and reinforcement
learning: the dopamine reward prediction error hypothesis. Proc.
Natl Acad. Sci. USA 108, 15647–15654 (2011).
65. Montague, P. R., Dayan, P. & Sejnowski, T. J. A framework for
mesencephalic dopamine systems based on predictive Hebbian
learning. J. Neurosci. 16,1936–1947 (1996).
66. Schultz, W. & Dickinson, A. Neuronal coding of prediction errors.
Annu.Rev.Neurosci.23,473–500 (2000).
67. Chen, A. P. F. et al. Nigrostriatal dopamine pathway regulates
auditory discrimination behavior. Nat. Commun. 13, 5942 (2022).
68. de Jong, J. W. et al. A neural circuit mechanism for encoding
aversive stimuli in the mesolimbic dopamine system. Neuron 101,
133–151 (2019).
69. Yuan, L., Dou, Y. N. & Sun, Y. G. Topography of reward and aversion
encoding in the mesolimbic dopaminergic system. J. Neurosci. 39,
6472–6481 (2019).
70. Romanski,L.M.,Clugnet,M.C.,Bordi,F.&LeDoux,J.E.Somato-
sensory and auditory convergence in the lateral nucleus of the
amygdala. Behav. Neurosci. 107,444–450 (1993).
71. Quirk, G. J., Repa, C. & LeDoux, J. E. Fear conditioning enhances
short-latency auditory responses of lateral amygdala neurons:
parallel recordings in the freely behaving rat. Neuron 15,
1029–1039 (1995).
72. Rogan,M.T.,Stӓubli, U. V. & LeDoux, J. E. Fear conditioning induces
associative long-term potentiation in the amygdala. Nature 390,
604–607 (1997).
73. Collins, D. R. & Paré, D. Differential fear conditioning induces reci-
procal changes in the sensory responses of lateral amygdala neu-
rons to the CS(+) and CS(-). Learn. Mem. 7,97–103 (2000).
74. Repa, J. C. et al. Two different lateral amygdala cell populations
contribute to the initiation and storage of memory. Nat. Neurosci. 4,
724–731 (2001).
75. Rosenkranz, J. A. & Grace, A. A. Dopamine-mediated modulation of
odour-evoked amygdala potentials during pavlovian conditioning.
Nature 417,282–287 (2002).
76. Goosens, K. A., Hobin, J. A. & Maren, S. Auditory-evoked spike firing
in the lateral amygdala and Pavlovian fear conditioning: mnemonic
code or fear bias? Neuron 40,1013–1022 (2003).
77. Wilensky, A. E., Schafe, G. E., Kristensen, M. P. & LeDoux, J. E.
Rethinking the fear circuit: the central nucleus of the amygdala is
required for the acquisition, consolidation, and expression of Pav-
lovian fear conditioning. J. Neurosci. 26,12387–12396 (2006).
78. Ciocchi, S. et al. Encoding of conditioned fear in central amygdala
inhibitory circuits. Nature 468,277–282 (2010).
79. Haubensak, W. et al. Genetic dissection of an amygdala micro-
circuit that gates conditioned fear. Nature 468,270–276 (2010).
80. Duvarci,S.,Popa,D.&Paré,D.Centralamygdalaactivityduringfear
conditioning. J. Neurosci. 31,289–294 (2011).
81. Letzkus, J. J. et al. A disinhibitory microcircuit for associative fear
learning in the auditory cortex. Nature 480, 331–335 (2011).
82. Weinberger, N. M. The medial geniculate, not the amygdala, as the
root of auditory fear conditioning. Hear. Res. 274,61–74 (2011).
83. Herry, C. & Johansen, J. P. Encoding of fear learning and memory
in distributed neuronal circuits. Nat. Neurosci. 17,1644–1654
(2014).
84. Dalmay, T. et al. A critical role for neocortical processing of threat
memory. Neuron 104, 1180–1194 (2019).
85. LeDoux,J.E.,Sakaguchi,A.,Iwata,J.&Reis,D.J.Auditoryemotional
memories: establishment by projections from the medial genicu-
late nucleus to the posterior neostriatum and/or dorsal amygdala.
Ann. N. Y Acad. Sci. 444, 463–464 (1985).
86. LeDoux,J.E.,Sakaguchi,A.,Iwata,J.&Reis,D.J.Interruptionof
projections from the medial geniculate body to an archi-neostriatal
field disrupts the classical conditioning of emotional responses to
acoustic stimuli. Neuroscience 17,615–627 (1986).
87. LeDoux, J. E., Farb, C. R. & Romanski, L. M. Overlapping projections
to the amygdala and striatum from auditory processing areas of the
thalamus and cortex. Neurosci. Lett. 134,139–144 (1991).
88. Popescu, A. T., Saghyan, A. A. & Paré, D. NMDA-dependent facil-
itation of corticostriatal plasticity by the amygdala. Proc.NatlAcad.
Sci. USA 104,341–346 (2007).
89. Ozawa, T. et al. A feedback neural circuit for calibrating aversive
memory strength. Nat. Neurosci. 20,90–97 (2017).
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 21
Content courtesy of Springer Nature, terms of use apply. Rights reserved

90. Johansen, J. P., Tarpley, J. W., LeDoux, J. E. & Blair, H. T. Neural
substrates forexpectation-modulated fear learning in theamygdala
and periaqueductal gray. Nat. Neurosci. 13,979–986 (2010).
91. Frontera, J. L. et al. Bidirectional control of fear memories by cere-
bellar neurons projecting to the ventrolateral periaqueductal grey.
Nat. Commun. 11, 5207 (2020).
92. Urrutia Desmaison, J. D. et al. Cerebellar control of fear learning via
the cerebellar nuclei-Multiple pathways, multiple mechanisms?
Front Syst. Neurosci. 17, 1176668 (2023).
93. Zhuang, X., Masson, J., Gingrich, J. A., Rayport, S. & Hen, R. Tar-
geted gene expression in dopamine and serotonin neurons of the
mouse brain. J. Neurosci. Methods 143,27–32 (2005).
94. Franklin, K. B. J. & Paxinos, G. The Mouse Brain in Stereotaxic
Coordinates. (Elsevier, New York, 2007).
95. Thiele,S.L.,Warre,R.&Nash,J.E.Developmentofaunilaterally-
lesioned 6-OHDA mouse model of Parkinson’sdisease.J. Vis. Exp.
60, 3234 (2012).
96. Blanchard, D. C. & Blanchard, R. J. Innate and conditioned reactions
to threat in rats with amygdaloid lesions. J. Comp. Physiol. Psychol.
81,281–290 (1972).
Acknowledgements
We would like to thank Jochen Roeper for his support; Beatrice Fischer,
Jasmine Sonntag, Sebastian Betz, Günther Amrhein a nd Thomas Wulf for
technical assistance; and TorfiSigurdsson for helpful discussions. This
work was supported by the Deutsche Forschungsgemeinschaft (DFG
Grant DU 1433/5-1 to S.D.).
Author contributions
D.Z., X.I.S-H., and S.D. designed the experiments. D.Z., X.I.S-H., E.S.D.B.,
L.R., and S.D. performed the experiments. D.Z., X.I.S-H., and S.D. ana-
lyzed the data. X.I.S-H. helped with the supervision of E.S.D.B., L.R., and
D.Z. D.Z., X.I.S-H., and S.D. wrote the paper with input from all authors.
S.D. supervised the study and obtained funding.
Funding
Open Access funding enabled and organized by Projekt DEAL.
Competing interests
The authors declare no competing interests.
Additional information
Supplementary information The online version contains
supplementary material available at
https://doi.org/10.1038/s41467-025-58382-5.
Correspondence and requests for materials should be addressed to
Sevil Duvarci.
Peer review information Nature Communications thanks Avishek Adhi-
kari, Mitsuko Watabe-Uchida and the other anonymous reviewer(s) for
their contribution to the peer review of this work. A peer review file is
available.
Reprints and permissions information is available at
http://www.nature.com/reprints
Publisher’s note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as
long as you give appropriate credit to the original author(s) and the
source, provide a link to the Creative Commons licence, and indicate if
changes were made. The images or other third party material in this
article are included in the article's Creative Commons licence, unless
indicated otherwise in a credit line to the material. If material is not
included in the article's Creative Commons licence and your intended
use is not permitted by statutory regulation or exceeds the permitted
use, you will need to obtain permission directly from the copyright
holder. To view a copy of this licence, visit http://creativecommons.org/
licenses/by/4.0/.
© The Author(s) 2025
Article https://doi.org/10.1038/s41467-025-58382-5
Nature Communications | (2025) 16:3066 22
Content courtesy of Springer Nature, terms of use apply. Rights reserved

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com