1 

Title:

 Intercalated amygdala dysfunction drives extinction deficits in the 

Sapap3

 mouse 

model of obsessive-compulsive disorder 

 

Authors:

 Robyn St. Laurent

1,3

, Kelly M Kusche

1

, Anatol C Kreitzer

1,2

, Robert C 

Malenka

3†

 

 

Affiliations:

  

1

 Gladstone Institutes, San Francisco, CA, USA 

2

 Department of Physiology, Department of Neurology, Weill Institute for Neurosciences, 

University of California, San Francisco, CA, USA 

3

 Nancy Pritzker Laboratory, Department of Psychiatry and Behavioral Sciences, 

Stanford University, Stanford, CA, USA

 

 

†

Corresponding author: 

Robert C Malenka (electronic address: 

malenka@stanford.edu

) 

 

Short/running title:

 Intercalated Amygdala in Negative Reinforcement and OCD 

 

Keywords: 

amygdala, avoidance, reinforcement, extinction, obsessive-compulsive 

disorder, 

Sapap3

 

 

 

2 

ABSTRACT 

Background  

The avoidance of aversive stimuli due to negative reinforcement learning is critical for 

survival in real-world environments, which demand dynamic responding to both positive 

and negative stimuli that often conflict with each other. Individuals with obsessive-

compulsive disorder (OCD) commonly exhibit impaired negative reinforcement and 

extinction, perhaps involving deficits in amygdala functioning. An amygdala subregion of 

particular interest is the intercalated nuclei of the amygdala (ITC) which has been linked 

to negative reinforcement and extinction, with distinct clusters mediating separate 

aspects of behavior. This study focuses on the dorsal ITC cluster (ITC

d

) and its role in 

negative reinforcement during a complex behavior that models real-world dynamic 

decision making. 

Methods 

We investigated the impact of ITC

d

 function on negative reinforcement and extinction by 

applying fiber photometry measurement of GCamp6f signals and optogenetic 

manipulations during a platform-mediated avoidance task in a mouse model of OCD-like 

behavior: the 

Sapap3

-null mouse. 

Results 

We find impaired neural activity in the ITC

d

 of male and female 

Sapap3

-null mice to the 

encoding of negative stimuli during platform-mediated avoidance. 

Sapap3

-null mice also 

exhibit deficits in extinction of avoidant behavior, which is modulated by ITC

d

 neural 

activity. 

Conclusions 

 

3 

Sapap3

-null mice fail to extinguish avoidant behavior in platform-mediated avoidance, 

due to heightened ITC

d

 activity.  This deficit can be rescued by optogenetically inhibiting 

ITC

d

 during extinction. Together, our results provide insight into the neural mechanisms 

underpinning negative reinforcement deficits in the context of OCD, emphasizing the 

necessity of ITC

d

 in responding to negative stimuli in complex environments. 

 
 

 

 

4 

Introduction 

Appropriate selection of actions in complex environments requires behavioral flexibility, 

while persistent patterns of inappropriate action selection are often detrimental to well-

being.  Optimization of action selection is facilitated by reinforcement learning and is 

impaired in many neuropsychiatric disorders (1). Obsessive compulsive disorder (OCD) 

is one such brain disorder characterized by repetitive behaviors (i.e. compulsions), 

which likely develop as a form of negative reinforcement to reduce anxiety (2) and are 

maintained because of the inability to extinguish such learned behaviors (3-7). Thus, 

elucidating the circuit mechanisms mediating the extinction of negative reinforcement is 

one strategy for advancing our understanding of OCD pathophysiology.  

 

A key brain area involved in positive and negative reinforcement learning is the 

amygdala, which is critical for deciphering the significance of positive- and negative- 

valence stimuli (8-10). The amygdala is comprised of distinct subregions, the functions 

of which are an active area of research (11). Negative reinforcement learning and 

extinction in a simple environment have been linked specifically to circuits involving the 

intercalated nuclei of the amygdala (ITC) (12-17), a conserved region also present in 

humans (18). The ITC consists of anatomically segregated clusters of tightly packed 

GABAergic neurons (19-22), which surround the basolateral amygdala (BLA). They 

regulate amygdala output via inhibition of the BLA and central amygdala (16, 23-26) 

(CeA), and reciprocal inhibition between clusters (16, 27). The distinct ITC clusters also 

play unique roles in regulating behavior, including the dorsal and ventromedial clusters 

(ITC

d

 and ITC

vm

, respectively), which have been implicated in fear learning and 

extinction (16).  

 

5 

 

Here, we examine the hypotheses that, (1) negative reinforcement is regulated 

specifically by the ITC

d

 in

 

a dynamic reinforcement task and, (2) impaired ITC

d

 function 

results in negative reinforcement deficits. To better model the types of action selection 

for negative reinforcement required of organisms living in a dynamic environment, we 

used platform-mediated avoidance, a behavioral task that requires active avoidance at 

the cost of a positive reward (28, 29).  This task is particularly appropriate for probing 

the role of the ITC because extinction of platform-mediated avoidance behavior has 

been linked to the BLA and infralimbic prelimbic cortex, regions that innervate the ITC 

(30). Furthermore, to investigate the relevance of extinction of negative reinforcement 

learning to OCD, we examined extinction of platform-mediated avoidance in the 

Sapap3-

null mouse, which is commonly used to investigate behaviors related to OCD 

(31, 32).  

Sapap3-

null mice lack the postsynaptic scaffolding protein SAP90/PSD95-

associated protein 3 (

Sapap3

) (33) and display excessive grooming, increased anxiety, 

behavioral inflexibility, and impaired fear extinction (31, 34-38). Here, we show that ITC

d

 

activity encodes threats during platform-mediated avoidance and that dysfunctional 

neural signals in the ITC

d 

of 

Sapap3-null

 mice contribute to impaired extinction of 

platform-mediated avoidance.  

 

METHODS AND MATERIALS 

Animals:

 

Both male and female adult mice were used for all experiments and cared for 

in accordance with the guidelines set by the Institutional Animal Care and Use 

Committee at the Gladstone Institutes and Stanford University. FoxP2tm1.1(cre)Rpa/J 

mice (39) were originally obtained from Jackson Laboratories and the colony was 

 

6 

maintained and bred in-house. 

Sapap3

 mice were originally obtained from Jackson 

Laboratories and heterozygotes were bred to 

FoxP2

-Cre mice. We then bred pairs of 

FoxP2

-cre

+

 x 

Sapap3

+/-

 mice to generate the experimental mice used for behavior. 

Offspring from this double transgenic cross were genotyped using Transnetyx.  For all 

experiments, 

FoxP2

-cre

+

 x 

Sapap3

+/+

 or 

FoxP2

-cre

+

 x 

Sapap3

-/-

 mice were used. To 

reduce the potential impact of skin lesions on behavioral performance, we trimmed the 

hind paw nails of all mice at the time of surgery (40). For behavior experiments, mice 

were a minimum of postnatal day 90 at the start of the procedure. Mice were maintained 

on a 12-h light/dark cycle and provided food and water 

ad libitum

, except where noted 

for water restriction in platform-mediated avoidance. 

 

Stereotaxic surgery: Adult male and female mice were anesthetized with 2% isoflurane 

gas, placed in a stereotaxic frame, and an adeno-associated virus infused through a 

glass microinjection pipette into the region of interest. When required, implants were 

slowly lowered into the brain. Optical fibers were secured to the skull with C&B 

Metabond and light-cured dental adhesive cement (Geristore A&B paste, DenMat). For 

viral injections, 100 nl of virus was infused into the ITC

d

 a rate of 50 nL/min with a 

borosilicate pipette coupled to a pump-mounted 5 μL Hamilton syringe (AP -0.9; ML 

+3.18; DV -4.6 from bregma). Pipettes remained in place for 10 min following infusion. 

For photometry recordings, optical fibers (Doric Lenses) with 400 μm core and 0.66 NA 

were unilaterally implanted over the ITC

d

 (AP -0.9; ML +3.18; DV -4.5 from bregma). For 

optogenetic experiments, optical fibers were made and polished using 1.25 mm 

diameter multimode ceramic ferrules (ThorLabs), 200 μm core fiber optic cable with 

 

7 

0.39 numerical aperture (NA) (ThorLabs), and 5-minute epoxy (Devcon). Experiments 

from mice where the injection site or implant was inaccurately placed or spread outside 

of the targeted region were discarded. 

 

Water restriction: Prior to platform-mediated avoidance, mice were placed on a water 

restriction regimen and slowly restricted until they achieved 85% of their original body 

weight. Body weight was monitored daily, before and after training sessions, and 

additional fluids were provided to maintain 85% body weight throughout the platform-

mediated avoidance behavioral sessions.  

 

Platform-mediated avoidance: Training and testing is conducted in the same arena 

throughout the experiment: a clear, acrylic custom arena (26 x 30 x 26 cm) containing 3 

custom-made infrared nose-pokes to deliver sucrose solution, a clear acrylic platform 

(12.5 x 12.5 cm), a shock-grid floor (MedAssociates, Inc.), and a tone generator 

(SparkFun WAV trigger). Acrylic arena and components are housed in a sound-

attenuating chamber and illuminated with overhead LED lights. Mouse location was 

recorded via overhead USB camera. Task parameters and behavioral responses are 

triggered and recorded using custom scripts written in Statescript (Spike Gadgets) and 

MATLAB (Math Works) and an mbed microcontroller system connected to a computer 

via micro-USB. Behavior was analyzed using custom MATLAB scripts and DeepLabCut 

(41). This procedure is adapted slightly from previous reports using the task in rats (28, 

30, 42-48). Water-restricted mice are trained for 7 sessions on a variable-interval 30s 

schedule where the center port pokes after the assigned ITI deliver a 10% sucrose 

 

8 

reward to the reward port. A small light on the center port is illuminated after a correct 

center poke and terminates after reward retrieval. The third nose poke is inactive and 

triggers no outcome. On sessions 8-17, a series of 9 pairs of 30 s warning tones 

terminating in a 2 s (0.2 mA) footshock occur pseudo-randomly in blocks of 3 every 3 

minutes. On extinction days (sessions 18-19), the warning tone occurs pseudo-

randomly every 2 minutes for a total of 15 times in the absence of footshock. On recall 

test day (session 20), the parameters are the same as extinction days, but there are no 

manipulations. The sucrose schedule remains constant throughout the experiment and 

the clear acrylic platform is always present.  

 

Bulk fiber photometry imaging: Fiber photometry was performed as described previously 

(49). Unilateral injections of AAV1-Syn-Flex-GCaMP6f-WPRE-SV40 and implantation of 

a fiber optic ferrule were counterbalanced across hemispheres in separate mice. Bulk 

fluorescence imaging of calcium transients in freely moving mice was performed on a 

custom-built fiber photometry set-up. All photometry experiments were aligned to 

behavior using TTL pulse synchronization. Synapse software controlling an RZ5P lock-

in amplifier was used to collect fiber photometry data. GCaMP6f was excited by 

frequency-modulated 472 and 405 nm LEDs (Doric Lenses) to stimulate Ca

+2

-

dependent and isosbestic emissions, respectively. Optical signals were band-pass-

filtered with a fluorescence mini cube (Doric Lenses) and signals were digitized at 6 

kHz. Signal processing was performed with custom MATLAB scripts. Photometry 

signals were collected for each session of platform-mediated avoidance and analyzed 

offline by correcting the 472 nm signal for artifacts with the 405 nm isosbestic signal. 

 

9 

Ca

+2

 signals were then z-scored for comparison across mice. Using custom-written 

MATLAB scripts, Ca

+2

 signals were aligned to relevant events.  

 

Optogenetic stimulation 

Bilateral injections of AAV5-EF1α-DIO-mCherry, AAV5- EF1α-DIO-eNpHR3.0-mCherry, 

or AAV5-EF1α-DIO-ChR2-mCherry (titer ~3.3-6e

12 

particles/ml) were infused into the 

ITC

d

. For optogenetic experiments, optical fibers were connected to a 532 nm or 473 

nm laser diode (Shanghai Laser & Optics Century Co.) through a FC/PC adapter 

connected to a fiber optic rotary joint commutator (Doric Lenses). Statescript programs 

delivered continuous illumination at requisite timepoints. Fiber light output was 

calibrated to 

∼

1 mW at the fiber tip using a digital power meter console (ThorLabs). 

 

Histology: 

Mice were deeply anesthetized with ketamine (75 mg/kg) and dexmedetomidine (0.25 

mg/kg) intraperitoneally and then transcardially perfused with 10 mL of phosphate-

buffered saline (PBS) followed by 10 mL of 4% PFA. Whole brains were dissected out 

and post-fixed overnight, then transferred to 30% sucrose in PBS for 48 hours until 

sunk. Brains were cut to 50 μm thickness on a Vibratome Leica® 1000 plus Sectioning 

System and stored in PBS prior to mounting. For virus and optic fiber implant 

verification, slices were mounted on slides with DAPI Fluoromount-G. All brain injections 

and implants for behavior experiments were verified post-hoc. Slides were imaged at 4x 

on a Keyence slidescanner fluorescent microscope. Mistargeting of either the viral 

 

10 

injection or implant locations were used as criteria for exclusion from behavioral 

analysis. 

 

Statistical analyses: 

All statistical analyses were performed in Prism (GraphPad). Where applicable, 

appropriate post-hoc tests were performed and p-values of p < .05 were considered 

significant. Statistical significance was determined using t-tests or repeated measures 

ANOVA (Table 1).

 

 

RESULTS

 

Sapap3

 KO mice exhibit deficits in platform-mediated avoidance

 

Historically

, in vivo

 experiments involving manipulations of the ITC have been difficult 

because of the small and interspersed nature of the ITC clusters. Recently, because 

ITC neurons preferentially express the transcription factor, forkhead box protein P2, 

which is encoded by 

FoxP2,

 several groups have used 

FoxP2

-IRES-cre mice (39) to 

specifically target isolated ITC clusters (16, 25, 50, 51). Because our planned 

experiments required transgene expression specifically in ITC

d

 neurons, all behavioral 

assays were performed in 

FoxP2

-cre

+

 x 

Sapap3

+/

+

 mice (“

Sapap3

 WT”) or 

FoxP2

-cre

+

 x 

Sapap3

-/-

 mice (“

Sapap3

 KO”) (Fig 1A). These mice were subjected to platform-

mediated avoidance training, which consisted of 3 phases (Fig 1B). Phase 1 (days 1-7) 

trained mice on a 30 second variable interval schedule where nose pokes into a center 

“initiation” port trigger sucrose reward from a lateral “reward” port. Reward delivery is 

indicated by an audible click of the solenoid used to dispense the liquid while an LED 

 

11 

illuminated the initiation port until the reward was retrieved. During phase 2 (days 8-17), 

mice were presented with 9 pairs of 30 second warning tones terminating in a mild 0.2 

mA footshock in blocks of 3 every 3 minutes. A safe platform was present in the corner 

opposite to the ports and thus the mice had to choose between avoiding the shock 

versus poking for sucrose, which remained available throughout the procedure. Phase 3 

(days 18 and 19) extinguished avoidance behavior by providing 15 warning tones 

without shock at 2 min intervals.    

 

Sapap3

 KO mice exhibited several differences in the platform-mediated 

avoidance task with a striking difference in how much time they spent on the platform 

during phase 2 and 3 (Fig 1C-D). While both 

Sapap3

 WT and KO mice gradually spent 

more total time on the platform during the phase 2 avoidance sessions ([Session: F

1,10 

= 

10.49, p = .009]), 

Sapap3

 KO mice spent significantly more time on the platform than 

WT mice (Fig 1E, [Genotype F

1,10 

= 5.326, p = .044]). Similarly, during the warning tone, 

both genotypes increased avoidance (i.e. platform time) between the first and last 

avoidance session with KO mice having greater tone avoidance than WT mice (Fig 1F, 

[Genotype F

1,10 

= 9.284, p = .012]).  A complimentary metric for analyzing avoidance 

behavior is how many shocks the mice avoided following the warning tone. Avoided 

shocks increased over training sessions and was higher in 

Sapap3

 KO mice (Fig 1G, 

[Session F

1,10 

= 10.51, p = .0088, Genotype F

1,10 

= 7.50, p = .021]). Analyzing overall 

performance across all ten avoidance training sessions revealed that 

Sapap3

 KO mice 

spent more time on the platform during the tone (Fig 1H, [t

118 

= 4.22, p < .0001]) and 

avoided more shocks than 

Sapap3

 WT mice (Fig 1I, [t

118 

= 5.03, p < .0001]).  

 

12 

Importantly, during the phase 3 extinction sessions, 

Sapap3

 WT mice 

extinguished their avoidant behavior while the 

Sapap3

 KO mice did not. Compared to 

the tenth and final avoidance training session, only 

Sapap3

 WT mice reduced 

avoidance by spending less time on the platform during the session (Fig 1J, [Session 

F

2,20

 = 5.085, p = .016]), the tones (Fig 1K), [Session F

2,20

 = 4.357, p = .027]), and the 

shock omission period by the second extinction session (Fig 1L, [Session F

2,20

 = 4.059, 

p = .033]). Thus, 

Sapap3

 KO mice spent more time on the platform during the tones 

during extinction sessions (Fig 1M, [t

22 

= 4.336, p = .0003]). Similarly, 

Sapap3

 KO mice 

avoided more shock omission periods than 

Sapap3

 WT mice during extinction (Fig 1N, 

[t

22 

= 3.635, p = .0015]). 

 

Impaired ITC

d

 activity in

 Sapap3

 KO mice during platform-mediated avoidance 

To measure neuronal population activity in the ITC

d 

during platform-mediated 

avoidance, we injected AAV-DIO-GCamp6f into the ITC

d

 of 

Sapap3

 WT or 

Sapap3

 KO 

mice (Fig 2A-B). Consistent with previous work (16, 25, 50, 51), the expression of this 

Cre-dependent GCamp6f was restricted specifically to the ITC

d 

(Fig 2C). On avoidance 

training days 1 and 10, both genotypes exhibited consistent GCamp6f responses to foot 

shock when located on the shock grid and diminished or absent shock responses when 

located on the platform (Fig 2D-E). The magnitude of these responses decreased over 

the course of training with smaller responses on avoidance training day 10 compared to 

day 1. However, ITC

d

 GCamp6f responses were smaller in the 

Sapap3

 KO vs. WT mice 

on both avoidance training days 1 and 10 (Fig 2F, [Session x Genotype 

F

1,106 

= 4.636, 

p

 

= .035]).   

 

13 

Arguably the most intriguing behavioral difference between 

Sapap3

 WT and KO 

mice was the failure of KO mice to extinguish their avoidance during extinction sessions 

(Fig 1J-N).  To determine if differences in ITC

d

 activity contributed to this behavioral 

deficit, we examined ITC

d

 GCamp6f signals comprehensively using a variety of metrics.  

GCamp6f signals were not significantly different between 

Sapap3

 WT and KO mice 

during the extinction warning tones (Fig 2G-H, [Session x Genotype F

1,178

 = 0.017, p = 

.897]). As expected, the absence of shock following the tone resulted in reduced ITC

d

 

activity (Fig 2I), similar to what was observed on avoided trials during training. Although 

the signal was already greatly diminished in both genotypes during the shock omission 

period, the GCamp6f signal modestly decreased across extinction sessions in 

Sapap3

 

KO mice, with no change over sessions in

 Sapap3

 WT mice, (Fig 2J, [Session x 

Genotype F

1,178

 = 5.834, p = .017]).  However, the most notable difference was an 

increase in the amount of GCamp6f transients in 

Sapap3

 KO mice compared to 

Sapap3

 

WT mice (Fig 2K-L, [Genotype F

1,20

 = 6.963, p = .025]). Despite the increased rate of 

neural activity, the mean amplitude of GCamp6f calcium peaks were not significantly 

different between genotypes (Fig 2M, [(Amplitude) Genotype F

1,20

 = 0.731, p = .413]). 

 

Inhibition of ITC

d

 accelerates extinction 

The blunted ITC

d

 GCamp6f signals in response to shock and increased spontaneous 

ITC

d

 GCamp6f signals during extinction of platform-mediated avoidance in 

Sapap3

 KO 

mice suggest that the ITC

d

 may play a causal role in this extinction process.  To address 

this hypothesis, we performed optogenetic manipulations during the warning tones on 

extinction days in 

Sapap3

 WT and KO mice by injecting AAV-DIO-mCherry, AAV-DIO-

 

14 

eNpHR3.0-mCherry, or AAV-DIO-ChR2-mCherry bilaterally into the ITC

d

 and implanting 

optic fibers immediately above this structure (Fig 3A-B). The training protocol in these 

mice was identical to that performed in the GCamp6f-expressing mice except that we 

perform laser stimulation during warning tones on extinction days and added a recall 

test 24 h following the final extinction session to assess persistent effects of optogenetic 

manipulations (Fig 3C). Consistent with the results from the previous cohort of mice, 

Sapap3

 KO mice exhibited increased avoidance behavior (Fig 3D-H).  Specifically, 

Sapap3

 KO mice spent more time on the platform during warning tones on the first and 

final avoidance training sessions compared to 

Sapap3

 WT mice (Fig 3F, [Genotype F

1,45 

= 27.14, p < .0001) and avoided more shocks than 

Sapap3

 WT mice (Fig 3H, 

[Genotype F

1,45

 = 21.15, p < .0001]).  

 

Inhibiting ITC

d

 activity by stimulation of eNpHR3.0 during the warning tones in 

extinction sessions 1 and 2 enhanced extinction in both 

Sapap3

 WT and KO mice (Fig 

3I-K). 

Sapap3

 WT mice expressing eNphR3.0 spent less time on the platform in the 

extinction sessions compared to WT mice expressing mCherry that received the same 

light stimulation (Fig 3K, [Opto Group F

1,14

 = 8.113, p = .013]). The same pattern 

occurred in 

Sapap3

 KO mice with the eNpHR3.0 group spending less time on the 

platform during tones in the second extinction session ([Opto Group F

1,11

 = 9.554, p = 

.01). The enhanced extinction of avoidance to the warning tone persisted into the recall 

test performed the following day in the absence of optogenetic manipulation for both 

genotypes (Fig 3L, [WT: t

14 

= 2.292, p = .038], [KO: t

11 

= 3.568, p = .004]). 

 

Activation of ITC

d

 enhances extinction deficit in 

Sapap3

 KO mice

 

 

15 

Given that inhibition of ITC

d

 activity during the warning tones enhanced extinction, we 

next addressed whether increasing ITC

d

 activity during warning tones (Fig 4A) does the 

opposite, as evidenced by increased time spent on the platform during extinction 

sessions. Optogenetic activation of ChR2 in ITC

d

 in 

Sapap3

 KO mice dramatically 

increased time spent on the platform during the warning tone in both extinction sessions 

compared to mCherry expressing mice (Fig 4B-C, [Opto Group F

1,12

 = 29.49, p = 

.0002]). Surprisingly, optogenetic stimulation of ITC

d

 neurons in 

Sapap3

 WT mice had 

no effect on time spent on the platform during the warning tones compared to mCherry 

controls ([Opto Group F

1,18

 = 0.121, p = .732]). This manipulation in 

Sapap3

 WT mice 

also did not influence performance in the recall test following extinction (Fig 4D-E, [WT: 

t

18 

= 0.151, p = .882]). In contrast, in 

Sapap3

 KO mice, the failure to extinguish 

avoidance to the warning tone persisted into the recall test in the absence of 

optogenetic stimulation ([KO: t

12 

= 2.285, p = .041]).  

 

Discussion

: 

Summary 

Here, we compare the function of the ITC

d

 in 

Sapap3 

WT and KO mice using a 

platform-mediated avoidance task involving competing motivations for positive and 

negative stimuli. 

Sapap3

 KO mice spent more time on the platform during avoidance 

training and, most importantly, unlike WT mice, failed to extinguish this avoidance 

behavior.  These aberrant behaviors in the 

Sapap3

 KO mice were associated with a 

reduction in the ITC

d

 neural response to a footshock as well as an increase in 

spontaneous Ca

+2

 transients during extinction indicating heightened ITC

d

 activity.  

 

16 

Consistent with a causal role for increased ITC

d 

activity in the extinction deficit in 

Sapap3

 KO mice, optogenetic inhibition of ITC

d

 neurons rescued extinction of the 

avoidant behavior. The same manipulation also facilitated extinction in 

Sapap3

 WT 

mice. In contrast, optogenetic activation of the ITC

d

 resulted in complete impairment of 

extinction in 

Sapap3

 KO mice with no effect on WT mice. Importantly, these 

interventions provided lasting efficacy, with the modification of extinction persisting into 

a recall test 24 h after the optogenetic manipulations.  

 

Role of the ITC in negative reinforcement learning  

Reinforcement learning allows for optimization of behavior selection and is mediated in 

part by brain circuits involving the striatum and amygdala (52). The ITC has been 

suggested to play a particularly important role in negative reinforcement and extinction 

(12-17), which are processes impaired in OCD subjects (4-7, 53).  Here, our goal was to 

understand the potential role of a single ITC cluster, the ITC

d

, in encoding negative 

reinforcement and extinction during a complex approach-avoidance task. Much of our 

understanding of the role of the ITC in negative reinforcement has been limited by the 

inability to target individual ITC clusters. Most prior studies used classic fear 

conditioning which measures a passive response to a threat, with the exception of a 

correlative study showing increased neural activation after extinction of lever-pressing to 

avoid a shock (14). Using fear conditioning, previous reports showed that partial lesions 

of the ITC led to impaired fear extinction (12) and distinct ITC clusters exhibited neural 

activation after fear learning or extinction (13). ITC clusters might regulate the switch 

between fear and extinction behaviors via changes in synaptic strength, for instance, 

 

17 

BLA inputs to the ITC are altered with fear extinction (54-56). A recent report using 

FoxP2

-cre mice found that ITC

d

 neurons encode fear memory and conversely, ITC

vm

 

neurons encode extinction (16). Furthermore, inactivation of ITC

d

 reduced freezing and 

inactivation of ITC

vm

 increased freezing during fear memory retrieval (16). Our data are 

in line with these findings, suggesting that ITC

d

 encoding of responses to negative 

stimuli in simple environments extends to complex operant tasks, such as the platform-

mediated avoidance task used here. 

 

Effects of Sapap3 deletion  

Sapap3

 KO mice have been extensively investigated as a rodent model of OCD 

because they express repetitive and compulsive behaviors, reinforcement deficits, 

increased anxiety, as well as improvement in symptom severity with serotonin reuptake 

inhibitor treatment (31, 34, 36-38, 57-60).  Several of the 

Sapap3

 KO behavioral 

abnormalities appear to be due to deficits in cortico-striatal synaptic transmission (31, 

61, 62), consistent with the generally accepted hypothesis that OCD involves abnormal 

functioning of cortico-striatal-thalamic-cortical networks (63). In addition to its high 

expression in the striatum, 

Sapap3

 is expressed in many regions of the amygdala in 

rodents (31, 33, 64). Although further work will be necessary to delineate how 

Sapap3

 

deletion causes deficits in ITC

d

 activity, given its prominent role in maintaining excitatory 

synaptic function, it seems likely that a reduction in synaptic transmission at 

glutamatergic synapses onto ITC

d

 neurons is one likely contributing mechanism. 

Another possibility is changes in synaptic properties or intrinsic excitability of ITC

d

 

 

18 

neurons due to homeostatic plasticity because of the loss of 

Sapap3

 throughout 

development.  

 

We obtained two surprising results when manipulating ITC

d

 during extinction.  

First, the optogenetic manipulations that were applied only during the two extinction 

sessions caused behavioral changes that were still present 24 hours later.  This 

suggests that suppression of ITC

d

 activity during extinction allows for generation of 

more robust extinction memory, perhaps via removal of inhibition onto ITC

vm

 extinction 

neurons through the reciprocal connection of ITC

d

à

ITC

vm

.  Second, optogenetic 

activation of ITC

d

 in 

Sapap3

 KO mice exacerbated their extinction deficits yet had no 

effects on extinction of platform-mediated avoidance in 

Sapap3

 WT mice.  Similar to 

human OCD subjects (37), this difference could be explained by 

Sapap3

 KO mice 

responding to negative stimuli more robustly. Additionally, deficits in ITC

vm

 or its inputs 

could lead to an imbalance between the reciprocally connected ITC

d

 and ITC

vm

 in 

Sapap3

 KO mice, allowing for our moderate activation with ChR2 to overpower the 

inhibition originating in the ITC

vm

. Conversely, in WT mice, endogenous ITC

vm

à

ITC

d

 

inhibition may be sufficient to override our optogenetic manipulations in WT but not KO 

mice. 

 

Conclusions 

Morphological and functional abnormalities in the amygdala of OCD subjects have been 

reported (65-73) but little is known about the potential role of amygdala dysfunction in 

contributing to negative reinforcement deficits in OCD. Using a mouse model for OCD, 

we present evidence that abnormalities in the functioning of a specific subregion of the 

 

19 

amygdala, the ITC

d

, importantly contributes to impairment in the extinction of an 

avoidant behavior.  Specifically, we identified diminished neural responses in the ITC

d

 to 

an aversive stimulus and increased neural activity during extinction. Most importantly, 

we demonstrate that inhibition of ITC

d

 accelerates the extinction process and 

persistently rescued the extinction deficit in 

Sapap3

 KO mice, suggesting that targeting 

ITC

d

 circuits in patients with impaired negative reinforcement deficits has therapeutic 

potential. 

 

 

 

 

20 

ACKNOWLEDGEMENTS

 

This work was supported by a grant from the Foundation for OCD Research (AK, 

RM) and the Berkelhammer Award for Excellence in Neuroscience (RS). The authors 

would like to thank Kendall Raymond for assisting in imaging and Aphroditi Mamilagas 

for assistance with MATLAB scripts to analyze calcium imaging peak frequency & 

amplitude. Some schematics included in figures were created with BioRender. 

CRediT author statements for this manuscript: Robyn St. Laurent participated in 

Conceptualization, Methodology, Software, Validation, Formal Analysis, Investigation, 

Resources, Data Curation, Writing – Original Draft, Writing – Review & Editing, 

Visualization, and Funding Acquisition. Kelly Kusche participated in Investigation and 

Writing – Review & Editing. Anatol Kreitzer participated in Conceptualization, Writing – 

Review & Editing, Supervision, Project Administration, and Funding Acquisition. Robert 

Malenka participated in Conceptualization, Writing – Review & Editing, Supervision, 

Project Administration, and Funding Acquisition. 

Data and code used for statistical analyses presented here are available upon 

request to the corresponding author. 

 

 

 

21 

DISCLOSURES 

Authors RS and KK have nothing to disclose. RM is on the scientific advisory boards of 

MapLight  Therapeutics,  MindMed  and  BrightMinds  Biosciences.  Currently,  RM  is  the 

Chief  Scientific  Officer  at  Bayshore  Global  Management.  AK  is  currently  the  Chief 

Discovery Officer at MapLight Therapeutics. 

 

 

 

22 

REFERENCES 

1. 

Maia TV, Frank MJ (2011): From reinforcement learning models to psychiatric and 

neurological disorders. 

Nat Neurosci

. 14:154-162. 

2. 

Fontenelle  LF,  Oostermeijer  S,  Harrison  BJ,  Pantelis  C,  Yücel  M  (2011): 

Obsessive-compulsive disorder, impulse control disorders and drug addiction: common 

features and potential treatments. 

Drugs

. 71:827-840. 

3. 

McGuire  JF,  Orr  SP,  Essoe  JK,  McCracken  JT,  Storch  EA,  Piacentini  J  (2016): 

Extinction  learning  in  childhood  anxiety  disorders,  obsessive  compulsive  disorder  and 

post-traumatic stress disorder: implications for treatment. 

Expert Rev Neurother

. 16:1155-

1174. 

4. 

McGuire JF, Orr SP, Wu MS, Lewin AB, Small BJ, Phares V, et al. (2016): Fear 

conditioning and extinction in youth with obsessive-compulsive disorder. 

Depress Anxiety

. 

33:229-237. 

5. 

Geller DA, McGuire JF, Orr SP, Pine DS, Britton JC, Small BJ, et al. (2017): Fear 

conditioning  and  extinction  in  pediatric  obsessive-compulsive  disorder. 

Ann  Clin 

Psychiatry

. 29:17-26. 

6. 

Geller DA, McGuire JF, Orr SP, Small BJ, Murphy TK, Trainor K, et al. (2019): Fear 

extinction learning as a predictor of response to cognitive behavioral therapy for pediatric 

obsessive compulsive disorder. 

J Anxiety Disord

. 64:1-8. 

7. 

Milad MR, Furtak SC, Greenberg JL, Keshaviah A, Im JJ, Falkenstein MJ, et al. 

(2013):  Deficits  in  conditioned  fear  extinction  in  obsessive-compulsive  disorder  and 

neurobiological changes in the fear circuit. 

JAMA Psychiatry

. 70:608-618; quiz 554. 

 

23 

8. 

Baxter  MG,  Murray  EA  (2002):  The  amygdala  and  reward. 

Nat  Rev  Neurosci

. 

3:563-573. 

9. 

Murray EA (2007): The amygdala, reward and emotion. 

Trends Cogn Sci

. 11:489-

497. 

10. 

Tye  KM,  Prakash  R,  Kim  SY,  Fenno  LE,  Grosenick  L,  Zarabi  H,  et  al.  (2011): 

Amygdala  circuitry  mediating  reversible  and  bidirectional  control  of  anxiety. 

Nature

. 

471:358-362. 

11. 

Janak  PH, Tye  KM  (2015):  From  circuits  to  behaviour  in  the  amygdala. 

Nature

. 

517:284-292. 

12. 

Likhtik  E,  Popa  D, Apergis-Schoute  J,  Fidacaro  GA,  Paré  D  (2008): Amygdala 

intercalated neurons are required for expression of fear extinction. 

Nature

. 454:642-645. 

13. 

Busti D, Geracitano R, Whittle N, Dalezios Y, Mańko M, Kaufmann W, et al. (2011): 

Different fear states engage distinct networks within the intercalated cell clusters of the 

amygdala. 

J Neurosci

. 31:5131-5144. 

14. 

Jiao X, Beck KD, Myers CE, Servatius RJ, Pang KC (2015): Altered activity of the 

medial  prefrontal  cortex  and  amygdala  during  acquisition  and  extinction  of  an  active 

avoidance task. 

Front Behav Neurosci

. 9:249. 

15. 

Knapska E, Maren S (2009): Reciprocal patterns of c-Fos expression in the medial 

prefrontal cortex and amygdala after extinction and renewal of conditioned fear. 

Learn 

Mem

. 16:486-493. 

16. 

Hagihara  KM,  Bukalo  O,  Zeller  M, Aksoy-Aksel A,  Karalis  N,  Limoges A,  et  al. 

(2021): Intercalated amygdala clusters orchestrate a switch in fear state. 

Nature

. 594:403-

407. 

 

24 

17. 

Rajbhandari AK, Octeau CJ, Gonzalez S, Pennington ZT, Mohamed F, Trott J, et 

al. (2021): A Basomedial Amygdala to Intercalated Cells Microcircuit Expressing PACAP 

and Its Receptor PAC1 Regulates Contextual Fear. 

J Neurosci

. 41:3446-3461. 

18. 

Braak H, Braak E (1983): Neuronal types in the basolateral amygdaloid nuclei of 

man. 

Brain Res Bull

. 11:349-365. 

19. 

Nitecka L, Ben-Ari Y (1987): Distribution of GABA-like immunoreactivity in the rat 

amygdaloid complex. 

J Comp Neurol

. 266:45-55. 

20. 

Nitecka  L,  Frotscher  M  (1989):  Organization  and  synaptic  interconnections  of 

GABAergic and cholinergic elements in the rat amygdaloid nuclei: single- and double-

immunolabeling studies. 

J Comp Neurol

. 279:470-488. 

21. 

McDonald AJ, Augustine JR (1993): Localization of GABA-like immunoreactivity in 

the monkey amygdala. 

Neuroscience

. 52:281-294. 

22. 

Paré D, Smith Y (1993): Distribution of GABA immunoreactivity in the amygdaloid 

complex of the cat. 

Neuroscience

. 57:1061-1076. 

23. 

Mańko M, Geracitano R, Capogna M (2011): Functional connectivity of the main 

intercalated nucleus of the mouse amygdala. 

J Physiol

. 589:1911-1925. 

24. 

Gregoriou GC, Kissiwaa SA, Patel SD, Bagley EE (2019): Dopamine and opioids 

inhibit synaptic outputs of the main island of the intercalated neurons of the amygdala. 

Eur J Neurosci

. 50:2065-2074. 

25. 

Asede  D,  Bosch  D,  Lüthi  A,  Ferraguti  F,  Ehrlich  I  (2015):  Sensory  inputs  to 

intercalated cells provide fear-learning modulated inhibition to the basolateral amygdala. 

Neuron

. 86:541-554. 

 

25 

26. 

Asede D, Doddapaneni D, Bolton MM (2022): Amygdala Intercalated Cells: Gate 

Keepers and Conveyors of Internal State to the Circuits of Emotion. 

J Neurosci

. 42:9098-

9109. 

27. 

Royer  S,  Martina  M,  Paré  D  (2000):  Polarized  synaptic  interactions  between 

intercalated neurons of the amygdala. 

J Neurophysiol

. 83:3509-3518. 

28. 

Bravo-Rivera  C,  Roman-Ortiz  C,  Brignoni-Perez  E,  Sotres-Bayon  F,  Quirk  GJ 

(2014):  Neural  structures  mediating  expression  and  extinction  of  platform-mediated 

avoidance. 

J Neurosci

. 34:9736-9742. 

29. 

Diehl  MM,  Bravo-Rivera  C,  Quirk  GJ  (2019):  The  study  of  active  avoidance: A 

platform for discussion. 

Neurosci Biobehav Rev

. 107:229-237. 

30. 

Bravo-Rivera  C,  Roman-Ortiz  C,  Montesinos-Cartagena  M,  Quirk  GJ  (2015): 

Persistent  active  avoidance  correlates  with  activity  in  prelimbic  cortex  and  ventral 

striatum. 

Front Behav Neurosci

. 9:184. 

31. 

Welch JM, Lu J, Rodriguiz RM, Trotta NC, Peca J, Ding JD, et al. (2007): Cortico-

striatal  synaptic  defects  and  OCD-like  behaviours  in  Sapap3-mutant  mice. 

Nature

. 

448:894-900. 

32. 

Chamberlain BL, Ahmari SE (2021): Animal Models for OCD Research. 

Curr Top 

Behav Neurosci

. 49:55-96. 

33. 

Welch  JM,  Wang  D,  Feng  G  (2004):  Differential  mRNA  expression  and  protein 

localization of the SAP90/PSD-95-associated proteins (SAPAPs) in the nervous system 

of the mouse. 

J Comp Neurol

. 472:24-39. 

34. 

Hadjas LC, Lüscher C, Simmler LD (2019): Aberrant habit formation in the Sapap3-

knockout mouse model of obsessive-compulsive disorder. 

Sci Rep

. 9:12061. 

 

26 

35. 

Lamothe H, Schreiweis C, Mondragón-González LS, Rebbah S, Lavielle O, Mallet 

L, et al. (2023): The Sapap3. 

Transl Psychiatry

. 13:26. 

36. 

Manning  EE,  Dombrovski  AY,  Torregrossa  MM,  Ahmari  SE  (2019):  Impaired 

instrumental  reversal  learning  is  associated  with  increased  medial  prefrontal  cortex 

activity 

in 

Sapap3 

knockout 

mouse 

model 

of 

compulsive 

behavior. 

Neuropsychopharmacology

. 44:1494-1504. 

37. 

van den Boom BJG, Mooij AH, Misevičiūtė I, Denys D, Willuhn I (2019): Behavioral 

flexibility  in  a  mouse  model  for  obsessive-compulsive  disorder:  Impaired  Pavlovian 

reversal learning in SAPAP3 mutants. 

Genes Brain Behav

. 18:e12557. 

38. 

Kajs  BL,  van  Roessel  PJ,  Davis  GL,  Williams  LM,  Rodriguez  CI,  Gunaydin  LA 

(2022):  Valence  processing  alterations  in  SAPAP3  knockout  mice  and  human  OCD. 

J 

Psychiatr Res

. 151:657-666. 

39. 

Rousso DL, Qiao M, Kagan RD, Yamagata M, Palmiter RD, Sanes JR (2016): Two 

Pairs of ON and OFF Retinal Ganglion Cells Are Defined by Intersectional Patterns of 

Transcription Factor Expression. 

Cell Rep

. 15:1930-1944. 

40. 

Lamothe H, Schreiweis C, Mondragón-González LS, Rebbah S, Lavielle O, Mallet 

L, et al. (2023): The Sapap3 −/−  mouse reconsidered as a comorbid model expressing a 

spectrum of pathological repetitive behaviours. 

Transl Psychiatry

. 13:26. 

41. 

Mathis A, Mamidanna P, Cury KM, Abe T, Murthy VN, Mathis MW, et al. (2018): 

DeepLabCut: markerless pose estimation of user-defined body parts with deep learning. 

Nat Neurosci

. 21:1281-1289. 

 

27 

42. 

Bravo-Rivera H, Rubio Arzola P, Caban-Murillo A, Vélez-Avilés AN, Ayala-Rosario 

SN,  Quirk  GJ  (2021):  Characterizing  Different  Strategies  for  Resolving  Approach-

Avoidance Conflict. 

Front Neurosci

. 15:608922. 

43. 

Diehl MM, Bravo-Rivera C, Rodriguez-Romaguera J, Pagan-Rivera PA, Burgos-

Robles A, Roman-Ortiz C, et al. (2018): Active avoidance requires inhibitory signaling in 

the rodent prelimbic prefrontal cortex. 

Elife

. 7. 

44. 

Diehl MM, Iravedra-Garcia JM, Morán-Sierra J, Rojas-Bowe G, Gonzalez-Diaz FN, 

Valentín-Valentín  VP,  et  al.  (2020):  Divergent  projections  of  the  prelimbic  cortex 

bidirectionally regulate active avoidance. 

Elife

. 9. 

45. 

Martínez-Rivera FJ, Bravo-Rivera C, Velázquez-Díaz CD, Montesinos-Cartagena 

M, Quirk GJ (2019): Prefrontal circuits signaling active avoidance retrieval and extinction. 

Psychopharmacology (Berl)

. 236:399-406. 

46. 

Martínez-Rivera  FJ,  Pérez-Torres  J,  Velázquez-Díaz  CD,  Sánchez-Navarro  MJ, 

Huertas-Pérez CI, Diehl MM, et al. (2022): A Novel Insular/Orbital-Prelimbic Circuit That 

Prevents  Persistent  Avoidance  in  a  Rodent  Model  of  Compulsive  Behavior. 

Biol 

Psychiatry

. 

47. 

Rodriguez-Romaguera  J,  Greenberg  BD,  Rasmussen  SA,  Quirk  GJ  (2016): An 

Avoidance-Based  Rodent  Model  of  Exposure  With  Response  Prevention  Therapy  for 

Obsessive-Compulsive Disorder. 

Biol Psychiatry

. 80:534-540. 

48. 

Rosas-Vidal LE, Lozada-Miranda V, Cantres-Rosario Y, Vega-Medina A, Melendez 

L,  Quirk  GJ  (2018): Alteration  of  BDNF  in  the  medial  prefrontal  cortex  and  the  ventral 

hippocampus  impairs  extinction  of  avoidance. 

Neuropsychopharmacology

.  43:2636-

2644. 

 

28 

49. 

Heifets BD, Salgado JS, Taylor MD, Hoerbelt P, Cardozo Pinto DF, Steinberg EE, 

et al. (2019): Distinct neural mechanisms for the prosocial and rewarding properties of 

MDMA. 

Sci Transl Med

. 11. 

50. 

Stern DB, Wilke A, Root CM (2023): Anatomical Connectivity of the Intercalated 

Cells of the Amygdala. 

eNeuro

. 10. 

51. 

Carvour HM, Roemer CAEG, Underwood DP, Padilla E, Sandoval O, Robertson 

M,  et  al.  (2023):  Mu-opioid  receptor  knockout  on  Foxp2-expressing  neurons  leads  to 

reduced aversion-resistant reward seeking. 

bioRxiv

. 

52. 

Averbeck  BB,  Costa  VD  (2017):  Motivational  neural  circuits  underlying 

reinforcement learning. 

Nat Neurosci

. 20:505-512. 

53. 

Cooper SE, Dunsmoor JE (2021): Fear conditioning and extinction in obsessive-

compulsive disorder: A systematic review. 

Neurosci Biobehav Rev

. 129:75-94. 

54. 

Jüngling  K,  Seidenbecher  T,  Sosulina  L,  Lesting  J,  Sangha  S,  Clark  SD,  et  al. 

(2008):  Neuropeptide  S-mediated  control  of  fear  expression  and  extinction:  role  of 

intercalated GABAergic neurons in the amygdala. 

Neuron

. 59:298-310. 

55. 

Amano T,  Unal  CT,  Paré  D  (2010):  Synaptic  correlates  of  fear  extinction  in  the 

amygdala. 

Nat Neurosci

. 13:489-494. 

56. 

Huang  CC,  Chen  CC,  Liang  YC,  Hsu  KS  (2014):  Long-term  potentiation  at 

excitatory  synaptic  inputs  to  the  intercalated  cell  masses  of  the  amygdala. 

Int  J 

Neuropsychopharmacol

. 17:1233-1242. 

57. 

Hadjas LC, Schartner MM, Cand J, Creed MC, Pascoli V, Lüscher C, et al. (2020): 

Projection-specific  deficits  in  synaptic  transmission  in  adult  Sapap3-knockout  mice. 

Neuropsychopharmacology

. 45:2020-2029. 

 

29 

58. 

Manning EE, Geramita MA, Piantadosi SC, Pierson JL, Ahmari SE (2021): Distinct 

Patterns of Abnormal Lateral Orbitofrontal Cortex Activity During Compulsive Grooming 

and Reversal Learning Normalize After Fluoxetine. 

Biol Psychiatry

. 

59. 

Manning EE, Wang AY, Saikali LM, Winner AS, Ahmari SE (2021): Disruption of 

prepulse  inhibition  is  associated  with  compulsive  behavior  severity  and  nucleus 

accumbens dopamine receptor changes in Sapap3 knockout mice. 

Sci Rep

. 11:9442. 

60. 

Burguière E, Monteiro P, Feng G, Graybiel AM (2013): Optogenetic stimulation of 

lateral  orbitofronto-striatal  pathway  suppresses  compulsive  behaviors. 

Science

. 

340:1243-1246. 

61. 

Wan Y, Ade KK, Caffall Z, Ilcim Ozlu M, Eroglu C, Feng G, et al. (2014): Circuit-

selective striatal synaptic dysfunction in the Sapap3 knockout mouse model of obsessive-

compulsive disorder. 

Biol Psychiatry

. 75:623-630. 

62. 

Corbit VL, Manning EE, Gittis AH, Ahmari SE (2019): Strengthened Inputs from 

Secondary  Motor  Cortex  to  Striatum  in  a  Mouse  Model  of  Compulsive  Behavior. 

J 

Neurosci

. 39:2965-2975. 

63. 

Ting JT, Feng G (2011): Neurobiology of obsessive-compulsive disorder: insights 

into neural circuitry dysfunction through mouse genetics. 

Curr Opin Neurobiol

. 21:842-

848. 

64. 

Kindler  S,  Rehbein  M,  Classen  B,  Richter  D,  Böckers  TM  (2004):  Distinct 

spatiotemporal  expression  of  SAPAP  transcripts  in  the  developing  rat  brain:  a  novel 

dendritically localized mRNA. 

Brain Res Mol Brain Res

. 126:14-21. 

 

30 

65. 

Szeszko PR, Robinson D, Alvir JM, Bilder RM, Lencz T, Ashtari M, et al. (1999): 

Orbital frontal and amygdala volume reductions in obsessive-compulsive disorder. 

Arch 

Gen Psychiatry

. 56:913-919. 

66. 

Boedhoe  PS,  Schmaal  L, Abe  Y, Ameis  SH, Arnold  PD,  Batistuzzo  MC,  et  al. 

(2017): Distinct Subcortical Volume Alterations in Pediatric and Adult OCD: A Worldwide 

Meta- and Mega-Analysis. 

Am J Psychiatry

. 174:60-69. 

67. 

Atmaca  M,  Yildirim  H,  Ozdemir  H,  Ozler  S,  Kara  B,  Ozler  Z,  et  al.  (2008): 

Hippocampus and amygdalar volumes in patients with refractory obsessive-compulsive 

disorder. 

Prog Neuropsychopharmacol Biol Psychiatry

. 32:1283-1286. 

68. 

Zhang L, Hu X, Lu L, Li B, Bu X, Li H, et al. (2020): Anatomic alterations across 

amygdala  subnuclei  in  medication-free  patients  with  obsessive-compulsive  disorder. 

J 

Psychiatry Neurosci

. 45:334-343. 

69. 

Kwon JS, Shin YW, Kim CW, Kim YI, Youn T, Han MH, et al. (2003): Similarity and 

disparity  of  obsessive-compulsive  disorder  and  schizophrenia  in  MR  volumetric 

abnormalities  of  the  hippocampus-amygdala  complex. 

J  Neurol  Neurosurg  Psychiatry

. 

74:962-964. 

70. 

Rus OG, Reess TJ, Wagner G, Zaudig M, Zimmer C, Koch K (2017): Structural 

alterations in patients with obsessive-compulsive disorder: a surface-based analysis of 

cortical volume, surface area and thickness. 

J Psychiatry Neurosci

. 42:395-403. 

71. 

Rus OG, Reess TJ, Wagner G, Zimmer C, Zaudig M, Koch K (2017): Functional 

and  structural  connectivity  of  the  amygdala  in  obsessive-compulsive  disorder. 

Neuroimage Clin

. 13:246-255. 

 

31 

72. 

Göttlich M, Krämer UM, Kordon A, Hohagen F, Zurowski B (2015): Resting-state 

connectivity  of  the  amygdala  predicts  response  to  cognitive  behavioral  therapy  in 

obsessive compulsive disorder. 

Biol Psychol

. 111:100-109. 

73. 

Simon D, Adler N, Kaufmann C, Kathmann N (2014): Amygdala hyperactivation 

during  symptom  provocation  in  obsessive-compulsive  disorder  and  its  modulation  by 

distraction. 

Neuroimage Clin

. 4:549-557. 

 

 

 

 

 

 

 

 

 

32 

TABLE AND FIGURE LEGENDS 

Table 1.

 Statistical results. 

 

Figure  1.

 

Bulk  fiber  photometry  recording  of  ITC

d

  during  platform  mediated 

avoidance in 

Sapap3

 KO mice.

  

A.

 

FoxP2

-cre and 

Sapap3

 mice used for experiments. 

B. 

Behavior  apparatus  and  experimental  timeline  for  platform-mediated  avoidance. 

C.

 

Heatmaps of location throughout the session for an example WT or KO mouse. 

D-E.

 Time 

on  platform  during  entire  session. 

F.

  Time  on  platform  during  the  tone  for  avoidance 

training sessions. 

G.

 Percentage of shocks avoided. 

H.

 Time on platform during tones for 

all  10  avoidance  training  sessions.   

I.

  Number  of  shocks  avoided  for  all  10  avoidance 

training  sessions. 

J. 

  Change  in  time  on  platform  for  the  entire  session  from  final 

avoidance training session through extinction. 

K.

 Change in time on platform during tones 

from  final  avoidance  training  session  through  extinction. 

L.

  Change  in  percentage  of 

shock  [omission]  periods  where  the  mouse  was  on  the  platform  from  final  avoidance 

training  session  through  extinction. 

M.

  Time  on  the  platform  during  tones  for  both 

extinction sessions. 

N.

 Number of shock omission periods where the mouse was on the 

platform for both extinction sessions. 

n = 6 WT and 6 KO mice. Bars indicate mean ± SEM, * p < .05. 

 

Figure 2.

 

Disrupted ITC

d

 activity during platform mediated avoidance in 

Sapap3

 KO 

mice

. 

A.

 

FoxP2

-cre and 

Sapap3

 mice used for experiments. 

B.

 Intracranial surgery for 

fiber photometry recordings. 

C.

 Example GCamp6f expression and fiber placement. 

D.

 

Calcium signal aligned to 2 s shock on tone-shock avoidance training day 1 (first session). 

 

33 

Heatmap of all trials aligned to shock (left) and Gcamp6f signal depending on the location 

of the mouse on or off the platform during the shock (right). n = 6 WT, 6 KO mice. 

E.

 

Calcium  signal  aligned  to  2  s  shock  on  tone-shock  avoidance  training  day  10  (final 

session). 

F.

 Area  under  the  curve  during  the  2  s  shock  for  early  and  late  tone-shock 

avoidance  training.  Dots  represent  a  single  shock. 

G.

  Calcium  signal  aligned  to  30  s 

warning tone during extinction day 2 for all mice, all tones. 

H. 

 Area under the curve during 

the  30  s  warning  tone  for  extinction  sessions. 

I.

  Calcium  signal  aligned  to  2  s  shock 

omission period and separated by location of the mouse on or off the platform during the 

omission  period. 

J.

  Area  under  the  curve  during  the  2  s  shock  omission  period  for 

extinction sessions. 

K.

 Example calcium signals from extinction day 2 with peak detection 

(left)  and  inset  of  gray  bar  expanded  (right). 

L.

  Peak  frequency  and 

M.

  amplitude  of 

calcium signal on extinction days.  

n = 6 WT and 6 KO mice. 

Abbreviations:  basolateral  amygdala  (BLA);  lateral  amygdala  (LA);  central  amygdala 

(CeA); stria terminalis (st); external capsule (ec); extinction (Ext.) 

 

Figure  3

. 

Inhibition  of  ITC

d

  during  platform  mediated  avoidance  enhances 

extinction of avoidance in 

Sapap3

 WT and KO mice.

 

A.

 Intracranial surgery for bilateral 

viral  injections  and  optic  fiber  implantation. 

B.

  Example  viral  expression  and  fiber 

placement. 

C.

  Experimental  timeline  for  platform-mediated  avoidance  experiments. 

D.

 

Heatmaps of location throughout the session for an example WT or KO mouse. 

E-F

. Time 

on platform during the tone for avoidance training sessions.  

G-H.

  Number of shocks 

avoided  for  avoidance  training  sessions. 

I.

  Stimulation  parameters  for  inhibition 

34 

experiments  during  extinction  sessions. 

J.

  Heatmaps  of  location  for  an  example  KO 

control vs. KO eNpHR3.0 mouse on extinction day 2. 

K.

 Time on platform during the tone 

for extinction sessions. 

L.

 Time on platform during the tone in a recall test 24 h following 

the final extinction session. 

n = 7 WT eNpHR3.0, 9 WT mCherry, 6 KO eNpHR3.0, 7 KO mCherry mice 

Bars indicate mean ± SEM, * p < .05. 

Figure 4

. 

Activation of ITC

d

 during platform mediated avoidance impairs extinction 

of avoidance in 

Sapap3

 KO mice.

 

A.

 Stimulation parameters for silencing experiments 

during  extinction  sessions. 

B.

  Heatmaps  of  location  throughout  the  session  for  an 

example  KO  mouse. 

C

.  Time  on  platform  during  the  tone  for  extinction  sessions.   

D

. 

Heatmaps of location for an example WT ChR2 vs. KO ChR2 mouse on recall test day. 

E

. Time  on  platform  during  the  tone  in  a  recall  test  24  h  following  the  final  extinction

session. 

n = 11 WT ChR2, 9 WT mCherry, 7 KO ChR2, 7 KO mCherry mice 

Bars indicate mean ± SEM, * p < .05. 

Resource 

Type

Specific Reagent or 

Resource

Source or Reference

Identifiers

Additional 

Information

Reagent

DAPI Fluoromount-G®

Southern Biotech

Cat. No.: 0100-20

Bacterial or 
Viral Strain

AAV1-Syn-Flex-GCaMP6f-WPRE-SV40 

Stanford Neuroscience Gene Vector 
and Virus Core Facility or Addgene

RRID:SCR_023250; 
RRID:SCR_002037; Addgene #100833

Bacterial or 
Viral Strain

AAV5-EF1α-DIO-mCherry

Stanford Neuroscience Gene Vector 
and Virus Core Facility or Addgene

RRID:SCR_023250; 
RRID:SCR_002037; Addgene #114471

Bacterial or 
Viral Strain

AAV5- EF1α-DIO-eNphR3.0-mCherry

University of North Carolina 
Neuroscience Center and the BRAIN 
Initiative Viral Vector Core Facility

RRID:SCR_023280

Bacterial or 
Viral Strain

AAV5- EF1α-DIO-ChR2(H134R)-
mCherry 

University of North Carolina 
Neuroscience Center and the BRAIN 
Initiative Viral Vector Core Facility

RRID:SCR_023280

Organism/ 
Strain

FoxP2-cre mouse, male and female, 
FoxP2tm1.1(cre)Rpa/J

Jackson Laboratories

JAX #030541, RRID:IMSR_JAX:030541

Organism/ 
Strain

Sapap3

-null mouse, male and female, 

B6.129-Dlgap3tm1Gfng/J

Jackson Laboratories

JAX #008733, RRID:IMSR_JAX:008733

Software; 
Algorithm

Prism

GraphPad

RRID:SCR_002798

Software; 
Algorithm

Synapse

Tucker-Davis Technologies

RRID:SCR_006495

Software; 
Algorithm

Statescript

Spike Gadgets

RRID:SCR_021623

Software; 
Algorithm

MATLAB

MathWorks

RRID:SCR_001622

Software; 
Algorithm

DeepLabCut

Mathis Lab

RRID:SCR_021391

Other

mbed microcontroller

NXP

mbed NXP LPC1768

Other

Keyence BZ-X810 slidescanner 
microscope

Keyence

https://www.keyence.com/products/mic
roscope/fluorescence-microscope/bz-
x700/models/bz-x810/

Other

BioRender

BioRender

RRID:SCR_018361

KEY RESOURCES TABLE

Figure 

Description 

Test 

Result 

Post-hoc test 

Result 

1E 

Total session 

platform time 

(GCamp6f) 

Repeated 
measures 
Two-way 
ANOVA 

Interaction: F

1,10

 = 1.993, p = .188 

Session: F

1,10 

= 10.49, p = .009 

Genotype: F

1,10 

= 5.326, p = .044 

Šídák’s 
multiple 
comparisons 

WT vs. KO (Day 1) p = .612 
WT vs. KO (Day 10) p = .029 
Day 1 vs. 10 (WT) p = .400 
Day 1 vs. 10 (KO) p = .016 

1F 

Tone platform time 

(GCamp6f) 

Repeated 
measures 
Two-way 
ANOVA 

Interaction: F

1,10

 = 0.275, p = .612 

Session: F

1,10 

= 9.284, p = .012 

Genotype: F

1,10 

= 3.571, p = .088 

Šídák’s 
multiple 
comparisons 

Day 1 vs. 10 (WT) p = .199 
Day 1 vs. 10 (KO) p = .059 

1G 

Shocks avoided 

(GCamp6f) 

Repeated 
measures 
Two-way 
ANOVA 

Interaction: F

1,10

 = 2.462, p = .148 

Session: F

1,10 

= 10.51, p = .0088 

Genotype: F

1,10 

= 7.50, p = .021 

Šídák’s 
multiple 
comparisons 

WT vs. KO (Day 1) p = .50 
WT vs. KO (Day 10) p = .011 
Day 1 vs. 10 (WT) p = .458 
Day 1 vs. 10 (KO) p = .013 

1H 

Time on platform for 

all avoidance 

sessions 

(GCamp6f) 

Unpaired t-
test, 
two-tailed 

t

118 

= 4.22, p < .0001 

n/a 

1I 

Shocks avoided for 

all avoidance 

sessions 

(GCamp6f) 

Unpaired t-
test, 
two-tailed 

t

118 

= 5.03, p < .0001 

n/a 

1J 

Time on platform 

during extinction 

sessions 

(GCamp6f) 

Repeated 
measures 
Two-way 
ANOVA 

Interaction: F

2,20

 = 0.645, p = .535 

Session: F

2,20

 = 5.085, p = .016 

Genotype: F

1,10 

= 11.91 p = .0062 

Šídák’s 
multiple 
comparisons 

Day 10 vs. Ext. Day 1 (WT) p = .436 
Day 10 vs. Ext. Day 2 (WT) p = .032 
Day 10 vs. Ext. Day 1 (KO) p = .118 
Day 10 vs. Ext. Day 2 (KO) p = .152 

1K 

Time on platform 

during tone for 

extinction sessions 

(GCamp6f) 

Repeated 
measures 
Two-way 
ANOVA 

Interaction: F

2,20

 = 2.198, p = .137 

Session: F

2,20

 = 4.357, p = .027 

Genotype: F

1,10 

= 6.411, p = .0298 

Šídák’s 
multiple 
comparisons 

Day 10 vs. Ext. Day 1 (WT) p = .578 
Day 10 vs. Ext. Day 2 (WT) p = .007 
Day 10 vs. Ext. Day 1 (KO) p = .508 
Day 10 vs. Ext. Day 2 (KO) p = .658 

1L 

Shock omission 

periods spent on 

platform for 

extinction sessions 

(GCamp6f) 

Repeated 
measures 
Two-way 
ANOVA 

Interaction: F

2,20

 = 0.999, p = .386 

Session: F

2,20

 = 4.059, p = .033 

Genotype: F

1,10 

= 13.05, p = .005 

Šídák’s 
multiple 
comparisons 

Day 10 vs. Ext. Day 1 (WT) p = .840 
Day 10 vs. Ext. Day 2 (WT) p = .049 
Day 10 vs. Ext. Day 1 (KO) p = .199 
Day 10 vs. Ext. Day 2 (KO) p = .235 

1M 

Time on platform for 

all extinction 

sessions 

(GCamp6f) 

Unpaired t-
test, 
two-tailed 

t

22 

= 4.336, p = .0003 

n/a 

1N 

“Shocks” avoided 

for all extinction 

sessions 

(GCamp6f) 

Unpaired t-
test, 
two-tailed 

t

22 

= 3.635, p = .0015 

n/a 

2F 

Area under the 

curve (footshocks) 

Repeated 
measures 
Two-way 
ANOVA 

Interaction: F

1,106 

= 4.636, p = .035 

Session: F

1,106 

= 34.67, p < .0001 

Genotype: F

1,106 

= 27.58, p < .0001 

Šídák’s 
multiple 
comparisons 

WT vs. KO (Day 1) p < .0001 
WT vs. KO (Day 10) p = .033 
Day 1 vs. 10 (WT) p < .0001 
Day 1 vs. 10 (KO) p = .019 

2H 

Area under the 

curve (extinction 

tones) 

Repeated 
measures 
Two-way 
ANOVA 

Interaction: F

1,178

 = 0.017, p = .897 

Session: F

1,178 

= 0.111, p = .74 

Genotype: F

1,178 

= 0.155, p = .155 

n/a 

2J 

Area under the 

curve (shock 

omissions) 

Repeated 
measures 
Two-way 
ANOVA 

Interaction: F

1,178

 = 5.834, p = .017 

Session: F

1,178 

= 1.59, p = .209 

Genotype: F

1,178  

= 0.232, p = .631 

Šídák’s 
multiple 
comparisons 

WT vs. KO (Ext. Day 1) p = .373 
WT vs. KO (Ext. Day 2) p = .096 
Ext. Day 1 vs. 2 (WT) p = .658 
Ext. Day 1 vs. 2 (KO) p = .02 

Table 1.

 

 

2L 

Peak Frequency of 

GCamp6f during 

extinction sessions 

Repeated 
measures 
Two-way 
ANOVA 

Interaction: F

1,10

 = 1.144, p = .31 

Session: F

1,10 

= 1.227, p = .294 

Genotype: F

1,10 

= 6.963, p = .025 

Šídák’s 
multiple 
comparisons 

WT vs. KO (Ext. Day 1) p = .043 
WT vs. KO (Ext. Day 2) p = .025 

2M 

Peak Amplitude of 

GCamp6f during 

extinction sessions 

Repeated 
measures 
Two-way 
ANOVA 

Interaction: F

1,10

 = 3.700, p = .083 

Session: F

1,10 

= 0.293, p = .600 

Genotype: F

1,10 

= 0.731, p = .413 

n/a 

 

3F 

Time on platform 

during tone for early 

vs. late avoidance 

training (opto) 

Repeated 
measures 
Two-way 
ANOVA 

Interaction: F

1,45

 =0.139, p = .712 

Session: F

1,45 

= 24.94, p < .0001 

Genotype: F

1,45 

= 27.14, p < .0001 

Šídák’s 
multiple 
comparisons 

WT vs. KO (Day 1) p < .0001 
WT vs. KO (Day 10) p = .0003 
Day 1 vs. 10 (WT) p = .0003 
Day 1 vs. 10 (KO) p = .008 

3H 

Shocks avoided for 

early vs. late 

avoidance training 

(opto) 

Repeated 
measures 
Two-way 
ANOVA 

Interaction: F

1,45

 = 0.273, p = .604 

Session: F

1,45 

= 27.20, p < .0001 

Condition: F

1,45 

= 21.15, p < .0001 

 

Šídák’s 
multiple 
comparisons 

WT vs. KO (Day 1) p = .004 
WT vs. KO (Day 10) p = .0004 
Day 1 vs. 10 (WT) p = .0016 
Day 1 vs. 10 (KO) p = .0009 

3K 

(WT) 

Time on platform 

during tone for 

extinction sessions 

(opto) 

Repeated 
measures 
Two-way 
ANOVA (no 
interaction) 

Session: F

1,15

 = 0.474, p = .502 

Opto Group: F

1,14

 = 8.113, p = .013 

 

Šídák’s 
multiple 
comparisons 

mCherry vs. eNpHR3.0 p = .035 

3K 

(KO) 

Time on platform 

during tone for 

extinction sessions 

(opto) 

Repeated 
measures 
Two-way 
ANOVA  
(no 
interaction) 

Session: F

1,12

 = 1.17, p = .301 

Opto Group: F

1,11

 = 9.554, p = .010 

 

Šídák's 
multiple 
comparisons 
test 

mCherry vs. eNpHR3.0 p = .0046 

3L 

Time on platform 

during tone for 

recall test (opto) 

Unpaired t-
test, 
two-tailed 

(WT) t

14 

= 2.292, p = .038 

(KO) t

11 

= 3.568, p = .004 

n/a 

 

4C 

(WT) 

Time on platform 

during tone for 

extinction sessions 

(opto) 

Repeated 
measures 
Two-way 
ANOVA  
(no 
interaction) 

Session: F

1,19

 = 0.1856, p = .671 

Opto Group: F

1,18

 =,0.121 p = .732 

 

n/a 

 

4C 

(KO) 

Time on platform 

during tone for 

extinction sessions 

(opto) 

Repeated 
measures 
Two-way 
ANOVA  
(no 
interaction) 

Session: F

1,13

 = 0.532, p = .479 

Opto Group: F

1,12

 = 29.49, p = 

.0002 
 

Šídák's 
multiple 
comparisons 
test 

mCherry vs. ChR2 p < .0001 

4E 

Time on platform 

during tone for 

recall test (opto) 

Unpaired t-
test, 
two-tailed 

(WT) t

18 

= 0.151, p = .882 

(KO) t

12 

= 2.285, p = .041 

n/a 

 

A.

F.

E.

G.

H.

D.

platform

ports

platform

ports

platform

ports

platform

ports

platform

ports

platform

ports

platform

ports

platform

ports

Sucrose day 7

Avoidance day 1

Avoidance day 10

Extinction day 2

WT

KO

Platform location

0

200

400

600

800

1000

0

200

400

600

800

1000

I.

Session

Avoidance

Ext.

Sucrose

1

18

8

0

25

50

75

100

B.

T

ime on 

platform 

(%) 

J.

K.

L.

Entire Session

WT

KO

C.

M.

N.

WT

KO

1 2

Ext.

10

0

25

50

75

100

T

ime

on

platform

during tone (%)

WT

KO

X

X

X

X

X

X

X

X

X

X

X

1 2

Ext.

10

0

25

50

75

100

Shocks avoided /

omissions

avoided

(%)

0

25

50

75

100

T

ime

on

platform

during tone (%)

WT

KO

0

25

50

75

100

1

10

Avoidance

Training

WT

KO

Shocks avoided (%)

0

25

50

75

100

1

10

Avoidance

Training

0

25

50

75

100

T

ime

on

platform

during session (%)

1 2

Ext.

10

WT

KO

0

2

4

6

8

10

Shocks avoided (#)

WT

KO

Genotype

max

0

100

200

300

T

ime

on

platform 

during tones (s)

WT

KO

Genotype

max

0

25

50

75

100

T

ime

on

platform

during session (%)

1

10

Avoidance

Training

WT

KO

Shcok omissions avoided (#)

WT

KO

Genotype

)

X

X

X

X

X

X

X

X

X

X

X

X

X

X

0

5

10

15

max

T

ime

on

platform

during tones (s)

max

0

150

300

450

WT

KO

Genotype

ns

ns

ns

Figure 1.

0

8

E.

Shocked

GCamp6f (z)

-2

0

2

4

6

8

Avoided

GCamp6f (z)

-2

0

2

4

6

8

0

8

WT

KO

F.

G.

Avoidance Training Day 1

D.

Shocked

GCamp6f (z)

-2

0

2

4

6

8

Avoided

GCamp6f (z)

-2

0

2

4

6

8

WT

KO

-1

0

1

2

3

4

10

4

AUC

KO

WT

Day 1

Avoidance Training

Extinction

0

1

2

0

1

2

WT

KO

2 s shock

0

10

20

30

Time (s)

-0.5

0.5

GCamp6f (z)

0

10

20

Time (s)

-0.5

0

0.5

GCamp6f (z)

30

z

z

0

8

2 s

0

8

2 s

2 s

2 s

z

z

z

z

30 s tone

I.

0

4

0

200

400

600

800

1000

0

4

Time (s)

WT

KO

3 s

3 z

GCamp6f (z)

Extinction

Session

1 2

1 2

0

2

4

6

8

Peak Frequency (Hz)

K.

0

0.5

1

1.5

2

Peak

Amplitude

(z)

Extinction

Session

1 2

1 2

WT

KO

WT

KO

-4 -2 0

2

4

6

Time (s)

-4 -2 0

2

4

6

Time (s)

-4 -2 0

2

4

6

Time (s)

-4 -2 0

2

4

6

Time (s)

-4 -2

0

2

4

6

Time (s)

-4 -2

0

2

4

6

Time (s)

-4 -2

0

2

4

6

Time (s)

-4 -2

0

2

4

6

Time (s)

-1

0

1

On platform

-4 -2 0

2

4

6

Time (s)

-4 -2 0

2

4

6

Time (s)

Off platform

Omission

Omission

GCamp6f (z)

-1

0

1

GCamp6f (z)

-4 -2 0

2

4

6

Time (s)

-4 -2 0

2

4

6

Time (s)

WT

KO

Omission

Omission

L.

M.

TONE

WT

KO

Example recordings

Inset

Extinction Day 2

Extinction Day 2

ns

KO

WT

Day 10

30 s tone

H.

-5

0

5

10

4

AUC

KO

WT

KO

WT

1

2

Extinction Session

ns

WT

KO

Inset

WT

KO

WT

KO

WT

KO

WT

KO

Avoidance Training Day 10

-7

0

7

10

3

AUC

2 s omission

J.

X

KO

WT

KO

WT

1

2

Extinction Session

ns

ns

ns

A.

B.

C.

DAPI

GCamp6f

fiber

tip

BLA

CeA

BLA

LA

ec

st

ITC

d

ITC

vm

amygdala

ITC

d

Figure 2.

laser on during tones

Tone-Shock Avoidance Training

Stimulation

parameters

Extinction

A.

B.

C.

Recall Test

15x tone + laser

Extinction session: 

silencing

I.

J.

H.

K.

WT

mCherry
eNpHR3.0

0

25

50

75

T

ime

on

platform

during tone (%)

mCherry

eNpHR3.0

D.

KO

KO

0

25

50

75

mCherry

WT

E.

WT

1

10

Avoidance

training

0

25

50

75

100

T

ime

on

platform

during tone (%)

mCherry

eNpHR3.0

ChR2

KO

1

10

Avoidance

training

0

25

50

75

100

1

10

Avoidance

training

0

25

50

75

100

Shocks

avoided

(%)

mCherry

WT

eNpHR3.0

ChR2

Behavior Timeline: Platform-mediated avoidance with optogenetic stimulation

G.

1

10

Avoidance

training

0

25

50

75

100

KO

WT

KO

0

25

50

75

100

1

10

Avoidance

training

T

ime

on

platform

during tone (%)

L.

platform

por
ts

Avoidance day 10

Extinction day 2

200

400

600

800

1000

0

WT

KO

200

400

600

800

1000

0

KO - eNpHR3.0

platform

por
ts

platform

por
ts

KO - mCherry

platform

por
ts

mCherry
eNpHR3.0

0

25

50

75

100

T

ime

on

platform

during tone (%)

1

2

Extinction

session

0

25

50

75

100

1

2

Extinction

session

F.

0

25

50

75

100

Shocks

avoided

(%)

1

10

Avoidance

training

DAPI

mCherry

fiber tract

ITC

d

ITC

lateral

BLA

200

400

600

800

1000

0

WT

KO

eNpHR3.0

Figure 3.

15x tone + laser

Extinction session: 

activation

C.

D.

A.

B.

WT

KO

mCherry
ChR2

KO

0

25

50

75

100

mCherry

ChR2

WT

0

25

50

75

100

T

ime

on

platform

during tone (%)

mCherry

ChR2

ns

Extinction

Recall Test

platform

por
ts

platform

por
ts

0

200

400

600

800

1000

platform

por
ts

Ext. day 1

Ext. day 2

0

200

400

600

800

1000

platform

por
ts

Avoidance Day 10

E.

KO ChR2

KO ChR2

WT ChR2

0

25

50

75

100

T

ime

on

platform

during tone (%)

ns

ns

1

2

Extinction

session

mCherry
ChR2

0

25

50

75

100

1

2

Extinction

session

platform

por
ts

0

200

400

600

800

1000

Test Day

Figure 4.