Leveraging Fiber Photometry to Decipher Neural Circuits Underlying Anxiety in Mice

ABSTRACT

Anxiety disorders rank among the most prevalent mental health conditions worldwide, significantly affecting patients' lives. They are frequently comorbid with other psychiatric disorders, often exacerbating their severity. Current pharmacological treatments; selective serotonin reuptake inhibitors (SSRIs) and benzodiazepines, remain limited in efficacy and are associated with undesirable side effects, underscoring the urgent need for alternative therapeutic approaches. However, progress in developing new treatments has been hindered by an incomplete understanding of the neural mechanisms underlying these disorders. Bridging this knowledge gap requires advanced research tools capable of providing deeper insight into the neural circuits involved in anxiety. Fiber photometry (FP) has emerged as a powerful and cost‐effective technique for measuring neural activity in freely moving animal models. By enabling real‐time monitoring of calcium dynamics in specific neural populations within defined brain regions, this method offers invaluable insights into both normal physiological processes and pathological states. In this review, we first present an accessible introduction to FP, detailing its apparatus, procedures, and key advantages and limitations. We then conducted a comprehensive analysis of 39 studies indexed in PubMed that have employed FP to investigate neural circuits implicated in anxiety. Our review reveals the techniques' significant contributions across different research domains, including physiological (33%), pathological (53%), and dual‐purpose studies (13%). Beyond summarizing its utility, our goal is to make FP more accessible to researchers. By providing a foundational guide for its integration into future scientific projects, we aim to facilitate advances in anxiety research and contribute to the development of novel therapeutic strategies.

Abbreviations

5‐HT

serotonin

AAVs

adeno‐associated viruses

ACC

anterior cingulate cortex

ACh

acetylcholine

ACx

auditory cortex

ad

Alzheimer's disease

aIC

anterior insular cortex

AirPLS

adaptive iteratively reweighted penalized least squares

ARCAgRP

hypothalamic arcuate nucleus neurons releasing agouti‐related peptide

ASD

autism spectrum disorder

BDNF

brain‐derived neurotrophic factor

BLA

basolateral amygdala

BNST

bed nucleus of the stria terminalis

BNSTPKCδ

protein kinase Cδ‐expressing cells in the BNST

CART

cocaine‐ and amphetamine‐regulated transcript peptide

CCK

cholecystokinin‐expressing

CeAL

central amygdala (lateral division)

ChAT

choline acetyltransferase

Chx10 Gi

neurons expressing the chx10 gene

CMOS

complementary metal‐oxide‐semiconductor

CNS

central nervous system

COX‐2

cyclooxygenase‐2

CPP

conditioned place preference

CRF

corticotropin‐releasing factor

CRH

corticotropin‐releasing hormone

CRHPVN

corticotropin‐releasing hormone‐producing neurons in the paraventricular nucleus

CSDS

chronic social defeat stress

CVS

chronic variable stress

D1R

dopamine 1 receptor

DA

dopaminergic

dDG

dorsal dentate gyrus

dHPC

dorsal hippocampus

dlBNST

dorsolateral bed nucleus of the stria terminalis

dmPFC

dorsomedial prefrontal cortex

DMS

dorsomedial striatum

DRN

dorsal raphe nucleus

DSM‐5

diagnostic and statistical manual of mental disorders

ELA

early life adversity

eLPB

lateral parabrachial nucleus

ERK

extracellular‐signal‐regulated kinase

FAM92A1

family with sequence similarity 92 member A1

FLIM

fluorescence lifetime imaging microscopy

fMRI

functional magnetic resonance imaging

GAD

generalized anxiety disorder

GCaMP

genetically encoded calcium indicator

GECIs

genetically encoded calcium indicators

GFP

green fluorescent protein

GluN2D

glutamate NMDA receptor subunit 2D

GRAB5‐HT

genetically encoded serotonin sensors

GRABACh

genetically encoded acetylcholine sensors

GRAB‐DA

genetically encoded dopamine sensors

HEK

human embryonic kidney

HFD

high‐fat diet

HPC

hippocampus

IC

insular cortex

ICA

independent component analysis

ICeA

intercalated central amygdala

IDO1

indoleamine 2,3‐dioxygenase 1

iGluSnFR

intensity‐based glutamate‐sensing fluorescent reporter

IP3R2

inositol 1,4,5‐trisphosphate receptor type 2

IPN

interpeduncular nucleus

KO

knockout

LA

lateral amygdala

LEDs

light‐emitting diodes

LH

lateral hypothalamus

LHb

lateral habenula

LID

L‐DOPA‐induced dyskinesia

LMX

lumiracoxib

MC

motor cortex

MCs

mossy cells

MDD

major depressive disorder

METH

methamphetamine

MG

medial geniculate

mGluR2

metabotropic glutamate receptor 2

MHb

medial habenula

mPFC

medial prefrontal cortex

MS

medial septum

NA

numerical aperture

NAc

nucleus accumbens

NIR

near‐infrared

NREM

non‐rapid eye movement

NTG

nitroglycerin

OFT

open field test

ORM

object recognition memory

ovBNST

oval subdivision of the bed nucleus of the stria terminalis

PBN

parabrachial nucleus

PCA

principal component analysis

PD

Parkinson's disease

pIC

posterior insular cortex

PKA

protein kinase A

PKCδ

protein kinase C delta

PL cortex

prelimbic cortex

PMT

photomultiplier tube

PTSD

posttraumatic stress disorder

PV

parvalbumin‐expressing

PVN

paraventricular nucleus

RCaMP

calcium‐binding messenger protein

REM

rapid eye movement

RSC

retrosplenial cortex

SDS

social defeat stress

SI‐res

social interaction resilient

SI‐sus

social interaction susceptible

SNI

spared nerve injury

SOM

somatostatin

SRS

subacute restraint stress model

SSRIs

serotonin reuptake inhibitors

SST

somatostatin peptide

US

unconditioned stimulus

vCA1

ventral cornu ammonis area 1

vDG

ventral dentate gyrus

Vgat

vesicular GABA transporter

vHPC

ventral hippocampus

vmPFC

ventromedial prefrontal cortex

vmPFC‐NAcSh

ventro‐median prefrontal cortex nucleus accumbens shell

VTA

ventral tegmental area

VTAVgat

ventral tegmental area vesicular GABA transporter

VTAVglut2

ventral tegmental area vesicular glutamate transporter 2

1. Introduction

Anxiety disorders rank among the most prevalent mental health conditions, affecting approximately 4% of the global population [1] and 21.6% in France [2], with women being 1.5 to 2 times more vulnerable than men [3]. According to the diagnostic and statistical manual of mental disorders (DSM‐5), these disorders are characterized by excessive and persistent fear and worry lasting for at least 6 months, often accompanied by symptoms such as restlessness, cognitive impairments, irritability, and sleep disturbances [4]. These conditions significantly impair daily functioning, reducing quality of life and overall well‐being. Moreover, anxiety disorders exhibit a high degree of comorbidity with other psychiatric conditions, particularly major depressive disorder (MDD), where they frequently serve as an aggravating factor [2], potentially exacerbating disease severity and treatment resistance [5]. The current first‐line pharmacological treatments for anxiety disorders primarily involve selective serotonin reuptake inhibitors (SSRIs), which typically require 3 to 6 weeks to achieve clinical efficacy [4]. Benzodiazepines are frequently prescribed for acute symptom relief, but their use is limited by significant adverse effects, including sedation, cognitive impairment, dependence, and withdrawal syndrome [6].

Technological advancements have significantly accelerated research in neuroscience, facilitating the development of innovative tools to investigate neural circuits in both physiological and pathological contexts [7]. Among these, high‐resolution imaging techniques such as two‐photon microscopy, electrophysiological approaches including patch‐clamp recordings and multielectrode arrays, and functional magnetic resonance imaging (fMRI) have provided valuable insights into brain function. Amid these innovations, fiber photometry has emerged as a particularly accessible, cost‐effective, and versatile technique for monitoring neural activity in freely moving animals.

First introduced in 2005 by researchers in Munich [8], Germany, fiber photometry has steadily gained recognition as a reliable method for investigating in vivo neural dynamics. By enabling real‐time calcium imaging, this technique allows researchers to track the activity of specific neural populations with high temporal resolution, offering insights into both normal and pathological brain function.

In this literature review, we first aim to establish a comprehensive yet accessible overview of fiber photometry, detailing its methodology, key advantages, and limitations. We then explore its contributions to anxiety research, focusing on studies conducted in mouse models, as mice remain the most widely used species for in vivo calcium imaging [9]. By demonstrating the potential of fiber photometry in unraveling the neural mechanisms of anxiety, we aim to encourage researchers to incorporate this technique into future studies, ultimately advancing our understanding of anxiety disorders and fostering the development of improved therapeutic strategies.

2. Fiber Photometry: A Cutting‐Edge Technique for Measuring Neural Activity

2.1. Principle of Fiber Photometry and Mechanism

Fiber photometry is an advanced calcium imaging technique designed to monitor neural activity in freely moving animals, particularly rodents. Rather than recording the activity of individual neurons, this technique measures bulk calcium activity, which captures the combined activity of neural populations within a targeted brain region [10]. This provides a global measure of neural dynamics, allowing researchers to correlate neural responses with behavioral states, external stimuli, or pharmacological interventions [10]. The technique relies on genetically encoded calcium indicators (GECIs), such as GCaMP, which fluoresce upon calcium binding due to a conformational change in their structure [11]. To achieve fluorescence‐based readouts, an optical fiber implanted in the brain delivers excitation light at a specific wavelength to activate the GECI. The resulting fluorescence emission, which reflects neural activity, is then collected and analyzed, enabling real‐time neural monitoring.

2.2. Experimental Procedure

2.2.1. Materials and Setup

2.2.1.1. Viral Vector for Biosensor Expression

Prior to calcium signal recording, a viral vector carrying the biosensor gene (e.g., GCaMP) must be injected into the brain region of interest to allow targeted expression in specific neural populations [12]. Among the various viral delivery systems, adeno‐associated viruses (AAVs) are most commonly used because of their high efficiency, safety, and well‐characterized properties [13]. The choice of an AAV vector depends on the specific properties needed to meet the experimental goals. Below is a description of several available AAV options.

• AAV1: high transduction frequency in the central nervous system (CNS)

• AAV2: the most well‐studied, broad tissue tropism

• AAV5: particularly effective for targeting the CNS, vector of choice for astrocytes

• AAV8 and AAV9: high transgene expression and can cross vascular barriers, with AAV9 being particularly efficient at crossing the blood–brain barrier

• AAV2 retro: engineered for specific neural tropism, allowing retrograde transport, useful for neural circuit mapping

It is important to carefully optimize the viral titer before starting the experiment. Insufficient expression may lead to weak or undetectable signals, whereas excessive expression can elevate background noise or, in severe cases, induce neurotoxicity and cell death [14].

Concerning calcium indicators, GCaMP6 is among the most validated ones, offering a balance between sensitivity and temporal resolution, with GCaMP6s optimized for higher signal amplitude and GCaMP6f for faster kinetics (Table 1).

TABLE 1.

Calcium indicators.

Calcium indicators	Types	Characteristics
GCaMP6	s	Higher signal amplitude	Balance between sensitivity and temporal resolution [15]
	f	Faster kinetics, detecting rapid or bursty neural activity
	m	Intermediate amplitude and kinetics
GCaMP7	f	Improved kinetics and sensitivity	Improved brightness, photostability, and detection of smaller calcium transients: detecting subtle neural activity in vivo [16]
	s	Highest sensitivity to small calcium changes
	b	Increased brightness and photostability, long recordings, and deeper brain imaging
RCaMPs	jRCaMP1a	Good sensitivity, low phototoxicity, excellent stability for chronic recordings, long‐term imaging, and deep tissue experiments	Dual‐color imaging, deeper tissue penetration due to longer excitation wavelengths, and reduced phototoxicity → valuable in multiplexed recordings or in experiments involving dense tissue or chronic imaging setups [17, 18]
	jRCaMP1b	Improved kinetics compared with jRCaMP1a, but slightly lower brightness, tracking fast calcium transients
	jRGECO1a	Strong fluorescent response to calcium, detection of single action potential; more sensitive to pH changes; suitable for multiplexed imaging or dense tissue

Abbreviations: GCaMP: genetically encoded calcium indicator; RCaMPs: red protein indicators for calcium.

GCaMP7 variants build on these properties with improved brightness, photostability, and detection of smaller calcium transients, making them particularly useful for detecting subtle neuronal activity in vivo, with GCaMP7s offering the highest sensitivity to small calcium changes [15, 16].

In addition, red‐shifted calcium indicators (RCaMPs) offer advantages in dual‐color imaging, deeper tissue penetration due to longer excitation wavelengths, and reduced phototoxicity. For instance, RCaMP1a has an excellent stability profile for chronic recordings, whereas jRCaMP1b has improved kinetics compared with jRCaMP1a and jRGECO1a, with a strong fluorescence response to calcium. These properties make red indicators especially valuable in multiplexed recordings or in experiments involving dense tissue or chronic imaging setups [17, 18].

Other genetically encoded biosensors for neurotransmitters that extend beyond calcium are available, including indicators for neurotransmitters like glutamate, dopamine, acetylcholine, and serotonin (Table 2). These tools are highly relevant for fiber photometry applications targeting specific neuromodulatory systems and behavioral paradigms. In fact, for glutamate, iGluSnFR provides not only a fast response but also a high sensitivity and specificity [19]. Moreover, several genetically encoded sensors are available for monitoring neuromodulators. For dopamine, both dLight and GRAB‐DA sensors are commonly used. Among the dLight variants, dLight1.3b offers the highest dynamic range, whereas dLight1.1 is the most sensitive. GRAB‐DA sensors also come in different forms and are characterized by faster binding kinetics. For acetylcholine, the GRABACh sensor provides high sensitivity along with good photostability. Finally, for serotonin, GRAB5‐HT demonstrates both high sensitivity and strong specificity, making it a reliable tool for in vivo studies [22].

TABLE 2.

Indicators for neurotransmitters.

Neurotransmitter	Biosensors	Characteristics
Glutamate [19]	iGluSnFR	High sensitivity (4 μM), fast response (ms), high specificity, bright and stable (works well with both one‐photon and two‐photon microscopy), can be used in neurons, astrocytes, brain slices, and even live animals (mice, zebrafish, worms)
Dopamine [20]	dLight	Based on the D1 receptor, 4 types (dLight 1.3b with highest dynamic range), detects 40 nM to 17 μM DA (dLight 1.1 most sensitive), slower on‐rates, faster off‐rates
	GRAB‐DA	Derive from the D2 receptor, 3 types, faster binding, slower unbinding, variable selectivity
Acetylcholine [21]	GRABACh	High sensitivity, rapid on/off signals, good photostability, validated in vitro (HEK293T cells, cultured neurons), on brain slices and in vivo (olfactory system of living Drosophila, and visual cortex of freely behaving mice), easy to express via transfection, viral vectors, or transgenics
Serotonin [22]	GRAB5‐HT	High sensitivity, large signal change, fast kinetics, high specificity, no signaling interference

Abbreviations: D1 receptor: dopamine 1 receptor; D2 receptor: dopamine 2 receptor; HEK293T cells: human embryonic kidney 293 cells; iGluSnFR: intensity‐based glutamate‐sensing fluorescent reporter; GRAB‐DA: genetically encoded dopamine sensors; GRABACh: genetically encoded acetylcholine sensors; GRAB5‐HT: genetically encoded serotonin sensors; HEK: human embryonic kidney.

2.2.1.2. Apparatus and Equipment

Fiber photometry relies on specialized optical and electronic components to excite fluorescent indicators, capture emitted light, and record neural activity in real time [12]. The key components of a fiber photometry system include a light source, optical combiners and splitters, optical fibers, photodetectors, control systems, and data analysis software [8] (Table 3). By optimizing excitation parameters, fiber properties, photodetector sensitivity, and data processing methods, researchers can maximize signal quality and experimental accuracy. The integration of fiber photometry with other neuroimaging and behavioral techniques further enhances its potential for advancing neuroscience and psychiatric research.

TABLE 3.

Fiber photometry components, functions, and examples.

Apparatus	Function	Examples/specifications
Light source	Excitation to stimulate fluorescent biosensors	LEDs Lasers [8] Typical wavelengths: −470 nm: calcium‐dependent −410 nm: noise reduction control −560 nm: dual‐color recordings (RCaMP, dLight) [9]
Optical combiners and splitters	Direct excitation light Separate the received fluorescence from the light sent to the brain	Dichromic mirrors Beam splitters
Patch cords, optical fiber, and ferrule	Deliver excitation light from a light source to the target brain region, collect emitted fluorescence from indicators to direct it to a photodetector	Core diameter Numerical aperture Length Well‐polished fiber end
Photodetectors	Convert photons emitted by fluorescent indicators into a measurable electrical signal	Photomultiplier tubes (PMTs) Photodiodes [8] Cameras (CMOS) [12] Spectrometers [8]
Control and acquisition system	Manages the entire experiment: triggering light sources, controlling optical filters, recording the photodetector signal, and integrating with external signals like behavioral signals
Data analysis software	Process and interpret raw signals from fiber photometry, create visualizations of the processed signals	MATLAB (MathWorks) Bonsai (Open source)

Abbreviations: CMOS: complementary metal‐oxide‐semiconductor; LED: light‐emitting diode; PMT: photomultiplier tube; RCaMP: calcium‐binding messenger protein.

2.2.1.2.1. Light Source for Fluorescence Excitation

A light source delivers excitation light to stimulate the fluorescent biosensors expressed in neurons [12]. Upon excitation, these biosensors emit fluorescence, which is then detected to measure neural activity or other biological processes [8]. The choice of light source depends on the experimental requirements. Light‐emitting diodes (LEDs) are widely used because of their stability, cost‐effectiveness, and compact size. However, lasers provide higher power and precise wavelength control, making them ideal for specialized applications requiring strong excitation, even though it increases photobleaching, which may compromise signal stability over time [8].

2.2.1.2.2. Optical Combiners and Splitters for Signal Separation

Optical combiners and splitters are essential for directing excitation light to the implanted optical fiber while separating the emitted fluorescence [12]. Dichroic mirrors are designed to reflect excitation light while allowing fluorescence emission to pass through, ensuring efficient separation of excitation and emission signals. On the other hand, beam splitters separate fluorescence signals of different wavelengths before their detection, making them particularly useful for dual‐color imaging or experiments requiring simultaneous multichannel acquisition [8]. Additionally, beam splitters can selectively direct light to multiple optical fibers [12].

2.2.1.2.3. Optical Fibers and Patch Cords for Light Transmission

Optical fibers and patch cords transmit excitation light to the target brain region and collect fluorescence emissions. Their design significantly influences signal quality and experimental flexibility. Large core diameters (200–400 μm) improve light collection efficiency but may increase background noise. Additionally, numerical aperture (NA) determines the light‐gathering capability of the fiber, influencing signal transmission and resolution. Finally, fiber length and flexibility also play an important role. Shorter fibers reduce signal loss and motion artifacts but limit movement flexibility. In contrast, longer fibers enhance flexibility but introduce signal attenuation. It is crucial to have well‐polished fiber ends to minimize signal loss and secure connections between the patch cord ferrule and the implanted cannula to prevent signal fluctuations [12].

2.2.1.2.4. Photodetectors for Fluorescence Signal Detection

Photodetectors convert emitted photons into electrical signals, allowing precise measurement of fluorescence intensity, which correlates with neural activity [8]. Various photodetector types are available; each is suited to different applications. Photomultiplier tubes (PMTs) are highly sensitive and ideal for detecting low‐intensity fluorescence signals, often used in fluorescence lifetime imaging microscopy (FLIM). Photodiodes are compact and cost‐effective, suitable for standard fluorescence intensity measurements [8]. Complementary metal‐oxide‐semiconductor (CMOS) cameras enable multifiber imaging, though they may introduce higher background noise compared with PMTs [12]. Spectrometers capture the full emission spectrum and allow the analysis of multiple fluorophores [8].

2.2.1.2.5. Control and Data Acquisition System

A control and acquisition system manages key experimental parameters, including triggering light sources and adjusting excitation parameters, managing optical filters to optimize signal detection, recording photodetector signals for real‐time neural activity analysis, and synchronizing external behavioral signals with neural recordings [12]. This system ensures precise timing control and allows integration with other experimental tools such as optogenetics or electrophysiology.

2.2.1.2.6. Data Processing and Analysis Software

Data analysis software plays a dual role in raw signals preprocessing, including normalization, background noise filtration, and motion artifact correction [12]. Quantitative analysis and interpretation link fluorescence signals with behavioral events, pharmacological effects, or external stimuli [8]. One of the most popular software tools is Bonsai, an open‐source tool used for real‐time data processing and custom experimental setups [12]. It connects fiber photometry systems, cameras, and other sensors and allows personalized workflows by connecting nodes that represent various functions. Additionally, MATLAB (MathWorks, USA) provides an accessible programming environment for signal processing, visualization, and statistical analysis in fiber photometry research.

2.2.2. Methods

2.2.2.1. Viral Vector Injection and Fiber Implantation

Under isoflurane anesthesia (3% induction and 1% maintenance), the selected viral vector is stereotaxically injected into the target brain region to ensure precise transgene expression (ventral hippocampus (vHPC) as an example in Figure 1). Following the injection, a post‐surgical period of 2 to 5 weeks is necessary to allow for both recovery and optimal biosensor expression [8]. To establish a stable optical interface for fluorescence excitation and detection, a multimode optical fiber (typically 200 to 400 μm in core diameter) is implanted at the desired depth and firmly secured using dental cement and/or anchoring screws [8, 12].

FIGURE 1.

Protocol to study and confirm pyramidal neural activity in the ventral hippocampus. Step 1: Local injection (ventral hippocampus) of the virus (AAV9‐CamKII.GCaMP6s.WPRE.SV40) through stereotaxic surgery under isoflurane anesthesia ⟶ recovery 2–5 weeks. Step 2: Fiber photometry recording (see Figure 2). Step 3: Confirmation of virus expression by immunohistochemistry (5× and 10×).

2.2.2.2. Calcium Signal Recording

The implanted fiber is then connected to a patch cord that links the detection system. During behavioral experiments, excitation light (usually 470 nm for GCaMP biosensors) is delivered to stimulate the fluorophore, and the resulting fluorescence emission is continuously recorded in real time as the animal engages in task‐related activities (Figure 2). To enhance signal fidelity, a secondary isosbestic wavelength, which is calcium‐independent, is often used to control motion artifacts and eliminate background noise. This approach ensures accurate synchronization between fluorescence signals and behavioral events, allowing for precise neurobehavioral correlations [10].

FIGURE 2.

Fiber photometry apparatus, recording, and signal processing in the ventral hippocampus. (A) Fiber photometry apparatus and experimental setup. A laser beam is directed through a dichroic mirror into an objective lens and further down an optical patch cord, and to the brain region of interest. Within the brain, the laser excites calcium‐bound GCaMP, inducing a conformational change that triggers fluorescence emission by the green fluorescent protein (GFP). The emitted fluorescence travels back through the optical patch cord, passes through the objective lens again, is separated from the laser light by the dichroic mirror, and then is directed towards a photodetector. The resulting signal is amplified and processed by a data acquisition system. (B) Example of a signal recorded in the vHPC of a Balb/c mouse during basic home cage exploration. We used FP3002 apparatus (Neurophotometrics, MBF Bioscience, USA) for signal recordings as well as Bonsai (open‐source, V 2.9.0) for system control. The raw signal includes a calcium‐dependent fluorescence trace at 470 nm and an isosbestic control trace at 415 nm. Signal processing involves calculating the mean fluorescence (F) of both the calcium‐dependent and isosbestic signals. The data are then normalized as ΔF/F (%), where ΔF/F represents the percentage change in fluorescence relative to the baseline fluorescence (F ₀). This provides a clearer representation of calcium dynamics in the recorded region. Signal processing was done using MATLAB (MathWorks, USA, R2024b). (C) Zoom on calcium transients (red arrow).

2.2.2.3. Signal Processing and Data Analysis

The collected fluorescence data undergo preprocessing to extract neural activity signals while minimizing artifactual distortions (Figure 2). The processing of the signal involves a series of steps to extract clean signals that reflect underlying neural activity with minimal artifacts and noise.

2.2.2.3.1. Baseline Correction

The data processing pipeline typically begins with baseline correction, often using algorithms such as adaptive iteratively reweighted penalized least squares (airPLS). airPLS is a fully automated computational method designed to remove baseline drift from fluorescence signals without the need for prior knowledge of peak positions or manual input. This makes it particularly well suited for large datasets and high‐throughput analysis [23]. This step is essential for the removal of nonneural fluctuations in signal that could otherwise obscure transients and compromise analysis, such as those caused by photobleaching or fiber motion artifacts. The correction is performed by iteratively fitting a smooth baseline to the signal through the minimization of a penalized least squares cost function. The observed fluorescence signal is first decomposed into a baseline estimate and a residual component. The algorithm iteratively downweights points that deviate significantly from the estimated baseline, assuming they correspond to true signal peaks, while assigning greater weight to points more likely to represent the baseline [24]. After each iteration, the points that rise well above the current baseline estimate are downweighted, whereas points that align well with the baseline are kept at high weight. Through repeated cycles of fitting and reweighting, the separation of slow drifts and fast transient signals becomes more prominent, thereby enabling a more accurate quantification of neural events. Although other methods such as asymmetric least squares smoothing and polynomial baseline fitting are commonly used, airPLS offers more robustness and adaptability when dealing with noisy signals presenting large transients, a frequent occurrence with fiber photometry recordings [24].

2.2.2.3.2. Isosbestic Subtraction

For dual‐wavelength fiber photometry recordings, signal preprocessing involves isosbestic subtraction to correct rapid motion artifacts and calcium‐independent fluorescence. This correction relies on the use of two excitation wavelengths: one that is calcium‐dependent and another at the isosbestic point (commonly 405 nm), where calcium indicators emit calcium‐independent fluorescence [25]. This isosbestic channel thereby serves as an internal reference, capturing motion‐related changes, fiber bending and movement, as well as alterations in tissue autofluorescence. To isolate the true calcium‐dependent signal, a linear regression is performed where the calcium‐sensitive fluorescence trace is regressed onto the isosbestic reference [25]. The residuals are then taken as the correct trace. This step is particularly important for in vivo recordings involving free‐moving animals because of behavioral motions and environmental perturbations.

2.2.2.3.3. ΔF/F Normalization

Following motion artifact correction via isosbestic signal subtraction, the next critical step is ΔF/F normalization, which converts the corrected fluorescence signal into a relative measure of change over time. The baseline fluorescence (F ₀) is typically estimated using a running percentile method applied over a defined sliding window, ensuring a robust and adaptive baseline reference [10]. Although isosbestic subtraction removes noncalcium‐related fluctuations, ΔF/F further standardizes the signal by expressing it relatively to a defined baseline level of fluorescence.

$$\mathrm{\Delta }F/F=\frac{F-F_0}{F_0}$$

ΔF/F = (F − F ₀)/F ₀ is then computed to normalize the signal. This normalization compensates for the variability in baseline intensity across animals or sessions and enhances transient calcium events' visibility. This allows for comparisons across time, animals, and conditions [8, 12].

2.2.2.3.4. Z‐Scoring (Optional, for Cross‐Section Comparison)

Following the normalization of the signal, z‐scoring is often applied to further standardize the signal for cross‐session or cross‐subject comparisons. The difference between the two lies in the type of normalization they perform and the scale on which signal changes are expressed. ΔF/F normalization reflects a relative change in fluorescence compared with the baseline, whereas z‐scoring standardizes the signal across time, transforming the data into a unitless scale by referencing the distribution of the signal; specifically, its means μ and standard deviation σ as follows [26]:

$$z=\frac{\mathrm{\Delta }F/F-\mu }{\sigma }$$

Although optional, z‐scoring is especially useful when aggregating data across animals or sessions, as it accounts for the variability in signal magnitude.

2.2.2.3.5. Noise Correction

Noise correction strategies are often recommended, especially in noisy environments or for multichannel setups, to enhance the signal‐to‐noise ratio and improve the reliability of peak detection and event‐triggered analyses, which facilitate comparisons across trials. In this section, we present several approaches commonly used in fiber photometry to reduce high‐frequency noise or systematic fluctuation.

• Low‐pass filtering

Low‐pass filters can be used to remove high‐frequency noises, such as mechanical vibrations and electrical noise, while preserving biologically relevant transients. Butterworth filters are particularly favored analog filters because of their maximally flat frequency response in the passband, which prevents ripple introduction and preserves the original signal shape. These filters can be implemented in various data analysis environments using scripting functions.

• Moving average smoothing

Moving average smoothing is a signal processing technique used to reduce short‐term fluctuation by averaging data points over a rolling window. This simple method helps smooth noisy signals and facilitates pattern detection. Moving average smoothing is compatible with other filters and can be applied before or after depending on the analytical objective [10].

• Savitzky–Golay (SG) filter

The SG filter is a digital smoothing filter based on local polynomial regression. It acts by fitting a low‐degree polynomial to a moving window of data points using the least squares and then taking the value at the center of the window as the smoothed value [27]. The process is repeated as the window slides across the entire dataset. Two main parameters are to be considered: the window size and the polynomial degree. The precise adjustment of these parameters is essential to prevent distortion or exaggeration of certain signal characteristics, particularly near edges. A narrow window is fit to capture more accurate details across the signal but may fail to smooth the trace effectively, whereas a broad one may smooth out important signal features [28]. Concerning the polynomial, high degrees carry the risk of overfitting the noise, whereas low polynomial degrees may underfit, missing relevant patterns in the data [28].

Despite their flat passband and their ability to preserve peak shapes, SG filters present a weak stopband attenuation, resulting in poor suppression of high‐frequency noise, especially above the cutoff frequency. This limitation can be improved through numerous methods such as the use of fitting weights (Hann or Gaussian) to reduce high‐frequency noise more effectively [27].

• Principal component analysis (PCA) and independent component analysis (ICA)

PCA is a statistical method used to reduce the dimensionality of large datasets by identifying patterns (principal components) that capture the most variance. This is done by projecting data onto a lower dimensional space defined by the directions that capture the most variance, assuming relevant signals that tend to lie in high variance, whereas noise is spread across low‐variance components [29]. ICA is a blind source separation method that goes further by separating statistically independent sources from mixed signals, assuming these sources are non‐Gaussian and independent. This is particularly interesting when the observed data include mixed underlying sources [12]. PCA and ICA allow for isolating the true neural signals from systemic noise or movement artifacts.

To summarize, PCA is recommended when a reduction of dimensionality and suppression of low‐variance noise is desirable, whereas ICA is more appropriate when the goal is to eliminate independent sources from the signal.

2.2.2.3.6. Event Alignment

Event alignment aims to synchronize fiber photometry data with specific behavioral cues or experimental interventions during testing. Precise alignment is essential to assess neural responses before, during, and after these events, to identify consistent time‐locked or behavior‐dependent responses across trials. This alignment can be done through different methods. Timestamp‐based alignment can be used for key‐locked events, whereas event‐triggered analysis requires either the alignment of signals separately to each key event or time‐wrapping signals across entire trials to create a unified timeline of all events [10].

2.2.2.3.7. Open‐Source Pipelines for Data Processing

All these previously mentioned steps are routinely implemented into open‐source toolboxes (PyPhotometry, pMAT, GuPPy, or FibrePhotometryAnalysis) and in custom MATLAB/Python pipelines. We describe the main alternatives below, as well as their advantages and drawbacks.

• PyPhotometry

This Python‐based system includes both hardware and software for fiber photometry data acquisition and presents the advantages of real‐time data visualization and cost‐effectiveness. However, the hardware is required and is not interchangeable with other systems [30].

• pMAT

This MATLAB‐based toolbox is user‐friendly, negates the need to code, and is available as a standalone app. However, pMAT is less flexible and may be difficult to adapt for in‐depth analysis. Therefore, it is more suitable for simple protocols [31].

• GuPPy

GuPPy is a Python toolbox and is suitable for both beginners (GUI access) and advanced (Jupyter notebook support) users. This toolbox is adapted to a variety of acquisition systems and allows clean signal visualizations and complex analysis [32].

• FiberPhotometry analysis

This Script‐based MATLAB toolbox focuses on transient detection and detailed signal characterization. It is a powerful tool, especially for elucidating peak kinetics and event‐related signal breakdowns. However, it is more suitable for experienced users and requires adaptations to fit experimental objectives.

• Custom pipelines in Python or MATLAB

Custom‐made pipelines in Python offer total flexibility and can incorporate machine learning and behavioral alignment and integrate advanced tracking tools like DeepLabCut. Python offers large sets of libraries and functionalities but requires advanced coding knowledge and careful validation.

MATLAB‐based customizations are easier to implement but offer less flexibility and require a paid license.

2.3. Advantages and Limitations of Fiber Photometry

Fiber photometry presents numerous advantages for in vivo monitoring of neural activity compared to conventional techniques such as fluorescence imaging with head‐mounted mini‐scopes or head‐fixed microscopes [10]. In this section, we compare fiber photometry with two‐photon microscopy, an advanced imaging technique that relies on the simultaneous absorption of two low‐energy photons, leading to the emission of a single higher energy photon [33]. Two‐photon microscopy is widely used in neuroscience research, particularly for studying calcium dynamics in neurons [34], making it an appropriate benchmark for comparison (Table 4) [10].

TABLE 4.

Comparison between fiber photometry and two‐photon microscopy [10].

Parameters	Fiber photometry	Two‐photon microscopy
Signal type	Bulk fluorescence signal	Individual cellular or subcellular calcium transients
Spatial resolution	Low (multiple cells/processes)	High (subcellular)
Temporal resolution	Good to high (10–100 ms)	High (10 ms)
Cell specificity	High	High
Invasiveness	Less invasive (implantation of optic fibers into the brain)	Cranial window for in vivo imaging
Field of view	Larger (simultaneous recording of neural activity from different brain regions and pathways)	Limited by light scattering and absorption by tissue, single cells, or small networks
Imaging depth	Surface (400 μm)	Deep (up to 1 mm, red‐shifted sensors)
Usefulness in behavioral studies	Compatible with natural behaviors, in vivo recording, real‐time monitoring (bendable and lightweight fibers that allow free movement)	Requires immobilization or head fixation, which restricts behavior
Complexity of data analysis	Low size and complexity of the raw data, involves preprocessing of raw signals and normalization	Complex, requires image analysis, single‐cell extraction, and motion correction
Equipment	Relatively simple and accessible, can be custom‐built	More complex (requires complex optical setups)
Cost	$5.000–$25.000	$125.000–$300.000
Applications	Population‐level neural dynamics, neurotransmitter release tracking	High‐resolution single neuron imaging, dendritic and synaptic activity

2.3.1. Signal Resolution and Data Acquisition

A key distinction between these techniques lies in the nature of the recorded signal. Fiber photometry captures bulk fluorescence signals that represent the collective activity of all neurons within the targeted brain region at a given time, without distinguishing between individual cells. In contrast, two‐photon microscopy enables imaging of individual neurons and even subcellular calcium transients, offering superior spatial and temporal resolution. However, fiber photometry still achieves high spatial and temporal fidelity, with robust cell‐type specificity, making it a versatile and reliable tool for a range of applications [10].

2.3.2. Suitability for Behavioral Studies

One of fiber photometry's major advantages is its ability to record calcium transients in freely moving, awake animals [10]. Its lightweight and flexible optical fibers enable unrestricted movement, a crucial feature for studying naturalistic behaviors. In contrast, two‐photon microscopy requires head immobilization and significantly constrains behavioral paradigms. Although wireless miniaturized systems are under development to address this limitation [35], their high cost and technical complexity further increase the financial and logistical burden of an already expensive and sophisticated system. Two‐photon microscopy systems typically cost between $125,000 and $300,000, whereas fiber photometry provides a more cost‐effective alternative with equipment costs ranging from $5000 to $25,000. Additionally, the surgical procedures required for fiber photometry (optical fiber implantation) are less invasive than those for two‐photon microscopy (cranial window implantation) [10]. Moreover, the data generated through fiber photometry are significantly smaller in size and less computationally intensive, leading to faster data processing and analysis.

2.3.3. Limitations and Considerations

Despite its advantages, fiber photometry has certain limitations. Since it captures bulk fluorescence, it does not allow single‐cell resolution, which may be limiting for studies requiring precise neural mapping. Additionally, fluorescence signals can be influenced by environmental conditions and animal movement and require rigorous correction methods to ensure accuracy. Autofluorescence artifacts can also impact data quality, and photobleaching effects should be considered when designing experiments. Ultimately, the choice of technique should be guided by the specific objectives of the study. For high‐resolution, single‐cell or subcellular imaging, two‐photon microscopy remains superior. However, for cost‐effective, minimally invasive, and behaviorally relevant recordings, fiber photometry emerges as the technique of choice.

2.4. Application in Neurosciences

Fiber photometry has emerged as a powerful technique for investigating the link between neural dynamics and specific behaviors and physiological processes [10]. The following sections summarize key studies that have utilized fiber photometry, highlighting their major findings in physiology and pathology (Table 5).

TABLE 5.

Applications of fiber photometry in physiology.

	Topic	References	Main outcome after the use of fiber photometry
Physiology	Sleep	Yu et al. [36]	Both VTAVglut2 and VTAVgat neurons are selectively active during wakefulness and REM sleep.
	Nutrition	Mazzone et al. [37]	High‐fat‐diet exposure rapidly reshapes ARCAgRP population dynamics in mice independently of sensory food recognition.
	Metabolism	Chen et al. [38]	A central astrocyte‐adipocyte axis regulates metabolism by modulating sympathetic nervous system activity.
	Social interaction	Fetcho et al. [39]	vmPFC‐NAcSh circuit is sensitive to hierarchical status.
	Mobility	Cregg et al. [40]	Chx10 Gi left population activity is correlated with leftward movements and inversely correlated with rightward movements, whereas straight movements were linked to a decreased activity.
	Cognition	Dautan et al. [41]	The cortico‐cortical circuit of inhibitory somatostatin neuron projections from the mPFC to the RSC regulates emotion recognition.
	Development	Wang et al. [42]	FAM92A1 depletion disrupts the activity of hippocampal excitatory neurons in FAM92A1 knockout mice.
Physiopathology	Autism	Brumback et al. [43]	D2R+ neurons are selectively and consistently activated during social exploration, with a deficit in mouse models of autism.
	Pain	Sun et al. [44]	Glutamatergic lateral parabrachial nucleus neurons are responsible for pain hypersensitivity.
	Parkinson's disease	Tabah et al. [45]	D1R signaling pathways contribute to the development of dyskinesia.
	Alzheimer's disease	Sun et al. [46]	Extra‐telencephalic neurons in the mPFC of 5xFAD mice showed reduced neural activity.
	Anxiety and depression	Asim et al. [47]	The activity of glutamatergic neurons in the BLA increased when presented with an aversive stimulus, with distinct responses being observed for different subtypes of GABAergic neurons in the BLA.

Abbreviations: ARCAgRP: hypothalamic arcuate nucleus neurons releasing agouti‐related peptide; BLA: basolateral amygdala; Chx10 Gi: neurons expressing the chx10 gene; D1R: dopamine 1 receptor; D2R: dopamine 2 receptor; gad92A1: family with sequence similarity 92 member A1; mPFC: medial prefrontal cortex; REM: rapid eye movement. RSC: retro splenial cortex; vmPFC‐NAcSh: ventro‐median prefrontal cortex nucleus accumbens shell; VTAVgat: ventral tegmental area vesicular GABA transporter; VTAVglut2: ventral tegmental area vesicular glutamate transporter 2.

2.4.1. Applications in Physiology

2.4.1.1. Sleep

One of the key contributions of this technique lies in the study of sleep regulation. For instance, fiber photometry experiments showed that ventral tegmental area vesicular glutamate transporter 2 (VTAVglut2) and ventral tegmental area vesicular GABA transporter (VTAVgat) neurons display distinct patterns during transitions between wakefulness, non‐rapid eye movement (NREM), and rapid eye movement (REM) sleep, suggesting their involvement in sleep architecture and regulation [36].

2.4.1.2. Nutrition

In the field of feeding and metabolism, fiber photometry has been instrumental in uncovering the dynamic behavior of hypothalamic arcuate nucleus neurons releasing agouti‐related peptide (ARCAgRP), which rapidly adapts their activity in response to high‐fat diet (HFD) exposure, independently of sensory food recognition. This provides valuable insights into the neural mechanisms underlying dietary choices and obesity‐related behaviors [37].

2.4.1.3. Metabolism

Furthermore, the technique has demonstrated that hypothalamic astrocytes regulate white adipose tissue lipolysis through sympathetic nervous system modulation. By increasing norepinephrine levels in adipose tissue, astrocytes enhance sympathetic outflow and promote lipid breakdown, which represents a potential therapeutic target for metabolic disorders [38].

2.4.1.4. Social Interaction

Fiber photometry has also been pivotal in social behavior and hierarchical interactions, particularly in demonstrating the role of the ventro‐median prefrontal cortex nucleus accumbens shell (vmPFC‐NAcSH) circuit in social dominance behavior. This circuit is particularly active in subordinate mice engaging in social challenges and is essential for adaptive social approach behaviors, especially in stress‐resilient individuals [39].

2.4.1.5. Mobility

In the domain of motor control, the technique has allowed researchers to map neural activity patterns involved in locomotion. It has been shown that Chx10 Gi neurons coordinate left and right movements, forming a critical link between the basal ganglia and the spinal cord to ensure smooth motor transitions. This discovery enhances our understanding of movement disorders and motor coordination mechanisms [40].

2.4.1.6. Cognition

Another key application is in emotion recognition, where fiber photometry has provided insight into the role of the medial prefrontal cortex–retrospinal cortex (mPFC‐RSC) circuit, modulated by somatostatin (SOM) inhibitory neurons. This circuit displays distinct activation patterns before and during social encounters, and dysfunction in this pathway is associated with impaired emotional recognition in psychiatric vulnerability models. Targeted stimulation of these neurons has been shown to restore normal emotional processing, highlighting their potential as therapeutic targets [41].

2.4.1.7. Neurodevelopment and Memory

In the context of neurodevelopment and memory, fiber photometry revealed that “family with sequence similarity 92 member A1” (FAM92A1) depletion leads to structural abnormalities in hippocampal excitatory neurons, resulting in impaired synaptic activity and memory deficits. These findings contribute to the understanding of neurodevelopmental disorders and memory dysfunctions [42].

Overall, fiber photometry has proven to be a powerful tool for exploring a wide range of physiological functions, offering high temporal resolution and the ability to record neural activity in freely behaving animals.

2.4.2. Applications in Neurological and Psychiatric Disorder

2.4.2.1. Autism Spectrum Disorder (ASD)

Fiber photometry has played a crucial role in unraveling the neural mechanisms underlying various neurological and psychiatric disorders. Historically, fiber photometry has identified in ASD mouse models, dopamine D2 receptor neurons in the mPFC as key players in social behavior alterations. Normally activated during social interactions, these neurons show significantly reduced activation in ASD mouse models, pointing to dysfunctional neural encoding of social cues [43].

2.4.2.2. Pain Processing

In the field of pain processing, fiber photometry has provided insight into neuropathic pain, revealing that glutamatergic neurons in the lateral parabrachial nucleus (PBN) exhibit heightened activity in response to noxious stimuli. This suggests that these neurons play an important role in amplifying pain perception and may represent potential targets for therapeutic intervention [44].

2.4.2.3. Parkinson's Disease (PD)

In PD, fiber photometry has been used to investigate dopaminergic signaling disruptions. Using FRET‐based biosensors, researchers found that dopamine 1 receptor (D1R) signaling in striatal neurons becomes hypersensitive to dopamine depletion, excessively activating the protein kinase A (PKA) and extracellular‐signal‐regulated kinase (ERK1/2) pathways. Chronic L‐DOPA treatment, while desensitizing these pathways, paradoxically exacerbates dyskinesia, providing new insights into L‐DOPA‐induced dyskinesia (LID) pathophysiology [45].

2.4.2.4. Alzheimer's Disease (AD)

For ad, fiber photometry studies in 5xFAD mice have shown that extra‐telencephalic neurons' activity in the mPFC is disrupted, leading to deficits in object recognition memory (ORM). Activation of these neurons restored ORM performance, whereas their inhibition in wild‐type mice induced memory impairments. Additionally, decreased cholinergic input from the basal forebrain was identified as a major factor contributing to neural dysfunction in ad [ 46].

2.4.2.5. Anxiety and Depression

Additionally, fiber photometry has been gaining popularity for mood disorder exploration, particularly anxiety and depression. Studies in the basolateral amygdala (BLA) revealed that specific GABAergic neurons regulate emotional states: Parvalbumin‐expressing (PV) neurons control depressive behaviors, whereas cholecystokinin‐expressing (CCK) neurons modulate both anxiety and depression. High‐frequency activation of these neurons effectively reduced symptoms, offering promising therapeutic targets for mood regulation [47].

By enabling real‐time monitoring of neural activity in freely behaving animals, fiber photometry has revolutionized our understanding of both physiological functions and disease mechanisms. With its ability to dissect circuit‐specific dysfunctions in disorders such as ASD, ad, PD, pain syndromes, and mood disorders, this technique continues to drive discoveries in neuroscience. Having explored its diverse applications, we now turn our focus to its contributions in anxiety research, where it has proven particularly valuable in recent years.

3. Contribution of Fiber Photometry in the Study of Anxiety Disorders

Fiber photometry has become an essential tool for exploring the neural mechanisms underlying anxiety and other emotional behaviors. Thanks to its ability to record real‐time neural activity in living contexts, it has led to significant advances in understanding the brain circuits involved in anxiety regulation. Several recent studies have focused on specific brain regions, shedding light on the complexity of neural networks responsible for processing emotional responses, particularly in anxiogenic environments (Figure 3 and Tables 6, 7, and 8).

FIGURE 3.

Illustration summarizing the brain regions recorded across 39 studies. The illustration represents the zones recorded using fiber photometry across the selected studies, categorized into three types: Studies using naïve mice (13 studies, 54%, green) focused on 16 zones, with the BNST and PVN being prominent. Studies using anxiety‐inducing pathophysiological model mice (17 studies, 43.5%, purple) examined 17 zones, notably the NAc, PVN, ACC, and mPFC. Studies using both naïve and anxiety‐inducing pathophysiological model mice (5 studies, 13%, blue) investigated eight zones, with emphasis on the HPC and related areas. The BLA, PVN, IPN, NAc, and dmPFC stand out as the most frequently studied zones overall. *Studies from various hippocampal subregions.

TABLE 6.

Literature review using PubMed database with naïve mice to investigate neural circuits implicated in anxiety.

Recorded brain area		Reference	Aim of the study	Animal strain	Main results
Prefrontal cortex (PFCx)	vmPFC	Munguba et al. [48]	Investigate the anxiolytic roles of presynaptic mGluR2 in prefrontal‐ and insula–amygdala synapses	Adult male and female GM2‐Cre mice	↓ vmPFC terminals in the BLA signal during transitions to anxiogenic environments and social interaction
	PL cortex	Brockway et al. [49]	Investigate the role of SST peptide signaling in the PL cortex and how it influences exploratory behavior	Adult female and male VIP::Cre;SST::Flp mice on a C57BL/8J background	↑ SST neurons' activity upon transition to an anxiogenic environment
	dmPFC BLA	Loewke et al. [50]	Investigate the frontostriatal projections from the dmPFC and their role in regulating innate avoidance	Adult male and female C57BL/6J mice	↑ dmPFC pyramidal neurons' activity in anxiogenic environments
Insular cortex	aIC pIC	Nicolas et al. [51]	Investigate how the IC and amygdala interact to encode both emotional valence and anxiety‐related behaviors in mice	Adult male and female C57Bl6/J mice	↑ aIC glutamatergic neuron activity in anxiogenic spaces ↑ pIC neuron activity in response to negative stimuli
	pIC	Gehrlach et al. [52]	Investigate the role of the pIC in processing aversive emotional and bodily states	Adult male and female transgenic mice (tetO‐GCaMP6s × Camk2a‐tTA)	↓ pIC activity in anxiogenic environments
BNST		Williford et al. [53]	Investigate the function of BNSTPKCδ cells in the context of anxiety and threat‐related behaviors	Adult male and female PKCδ‐Cre × C57BL/6J	↑ BNSTPKCδ cells activity associated with reduced exploratory behavior and increased anxiety‐like behavior
		Jaramillo et al. [54]	Investigate the dynamic nature of in vivo BNST activity associated with anxiety‐like behavior in a stress‐inducing context	Adult male and female C57BL/6J mice, CalcaCRE mice	↑ PKCδ‐expressing BNST cells' activity synchronized to the consummatory food approach
PVN		Yuan et al. [55]	Investigate the response patterns of the PVN CRH neurons during adaptive stress response	Adult male CRH‐ires‐Cre knock‐in mice	↑ PVN CRH neurons activity during stressors and are inhibited by reward consumption
		Qui et al. [56]	Investigate the role of CRHPVN in mediating the anxiolytic effects of dexmedetomidine	Adult male and female C57BL/6J mice	↓ CRHPVN neural excitability after administration of dexmedetomidine, mediating anxiolytic effects
vDG dDG		Wang et al. [57]	Investigate the circuit mechanisms by which MCs control neural excitability and their effects on anxiety‐related behavior	Adult male and female Crlr‐Cre mice	↑ MCs' neural activity in an anxiogenic environment
Midbrain structures	NAc	Rizzo et al. [58]	Investigate the way ketamine affects dopamine release and neurotransmission in the NAc	Adult male C57BL/6J mice Naïve	↑ duration of dopamine transients ↓ frequency of glutamate transients in NAc after acute esketamine injection
	IPN	Molas et al. [59]	Investigate the role of VTA‐IPN dopaminergic circuit in response to anxiety and social novelty	Adult male and female C57BL/6J mice	↑ IPN dopaminergic neuron activity in anxiety‐inducing environments
	DRN	Balasubramanian et al. [60]	Investigate the interaction between CART peptide and the DRN 5‐HT systems and how it influences anxiety‐related behavior	Adult male Sert‐cre and C57BL/6J mice	↓ neural activity in the DRN in SERT‐cre mice, and 5HTDRN signal in C57BL/6J mice after CART microinfusion
Cerebellum lobule VII		Pei Wern Chin et al. [61]	Investigate whether 5‐HT signaling within the cerebellum participates in anxiety‐like behavior	Adult male C56BL/6 mice	↑ Serotonin levels in lobule VII correlated with lower anxiety‐like behavior

Note: Symbols: ↓: decreased; ↑: increased; ↔: unchanged.

Abbreviations: aIC: anterior insular cortex; BLA: basolateral amygdala; BNST: bed nucleus of the stria terminalis; BNSTPKCδ: protein kinase Cδ‐expressing cells in the BNST; CART: cocaine‐ and amphetamine‐regulated transcript peptide; CRH: corticotropin‐releasing hormone; CRHPVN: corticotropin‐releasing hormone‐producing neurons in the paraventricular nucleus; dDG: dorsal dentate gyrus; dmPFC: dorsomedial prefrontal cortex; DRN: dorsal raphe nucleus; IPN: interpeduncular nucleus; MCs: mossy cells; NAc: nucleus accumbens; pIC: posterior insular cortex; PKCδ: protein kinase C delta; PL cortex: prelimbic cortex; PVN: paraventricular nucleus; SST: somatostatin peptide; vDG: ventral dentate gyrus; vmPFC: ventromedial prefrontal cortex; VTA: ventral tegmental area.

TABLE 7.

Literature review using PubMed database with anxiety‐inducing pathophysiological model mice to study circuits implicated in anxiety.

Recorded brain area		References	Aim of the study	Animal strain	Main results
Prefrontal cortex	mPFC	Kuang et al. [62]	Investigate the role of P2X2 receptors in pyramidal neurons in regulating vulnerability to chronic stress	Adult male C57BL/6J mice, P2rx2‐cKO CSDS	↑ mPFC pyramidal neurons' activity after CSDS in conditional P2X2 knockout mice‐increased resilience to stress
	dmPFC	Kajs et al. [63]	Investigate the activity of the dmPFC and its projections to the striatum and the BLA during active avoidance learning in mice	Adult male and female C57BL6/J mice Cued two‐way active avoidance	↑ dmPFC activity during avoidance ↓ dmPFC during cued freezing ↑ dmPFC‐DMS projections activity associated with active avoidance behavior ↓ dmPFC‐BLA projections activity during active avoidance tasks
	ACC	Gao et al. [64]	Investigate the effect of dexmedetomidine on anxiety‐like behavior and its underlying mechanism involving glutamatergic neurons in the anterior cingulate cortex	Adult male C57BL/6J mice Chronic pain model	↓ glutamatergic neuron hyperactivity in the ACC after acute dexmedetomidine administration‐decreased anxiety‐like behavior
		Li et al. [65]	Investigate the role of the ACC in the therapeutic effects of duloxetine on chronic pain‐induced anxiety	Adult male C57BL/6J mice Inflammatory pain model	↑ serotonin concentrations ↓ glutamatergic neuron activity in the ACC after acute duloxetine administration
		Hu et al. [66]	Investigate the role of the enzyme IDO1 in the modulation of pain sensitivity and comorbid anxiety in patients with chronic migraines	Adult male C57BL/6J mice Chronic migraine model	↑ ACC neural activity
NAc		Muir et al. [67]	Investigate vHPC‐NAc involvement in depressive‐ and anxiety‐like behavior before and after chronic stress	Adult male and female C57BL/6J CVS	↑ vHPC‐NAc amplitude in both sexes ↑ Frequency in males during general stress and in females during NTG bouts
		Flores‐Garcia et al. [68]	Investigate the effects of music exposure on pain‐induced anxiety and the mechanism involved	Adult male and female CD‐1 mice Chronic pain model	↓ NAc DA activity after chronic pain‐prevented by music exposure
BLA		Morgan et al. [69]	Investigate the effects of pharmacological COX 2 inhibition on BLA neural activity and stress‐induced anxiety‐like behavior	Adult male and female ICR (CD‐1) mice Acute stress	↑ BLA neurons' activity in acute restraint stress. Reversed with LMX pretreatment
		Huang et al. [70]	Investigate the role of BLA neurons in regulating the susceptibility to stress‐related psychiatric disorders and explore the link between CSDS‐induced changes in BLA neural activity	Adult male C57BL/6J mice CSDS	↑ BLA activity observed in both SI‐sus and SI‐res mice after CSDS ↓ BLA activity in SI‐res mice post‐CSDS
Extended amygdala system	dlBNST	Salimando et al. [71]	Investigate the way GluN2D‐presenting NMDA receptors regulate emotional behaviors and neural activity in the BNST	Adult male GluN2D KO mice	↑ BNST CRF neuron activity in GluN2D KO mice during novel environment exploration
	CeAL	Demaestri et al. [72]	Investigate how early life adversity alters the function of CeAL CRF+ neurons and contribute to heightened threat activity and sex differences in anxiety susceptibility	Adult male and female C57BL/6N mice and CRF‐ires‐Cre mice ELA	↑ CeAL CRF+ neural activity during startle responses in ELA mice
MHb MS		Wu et al. [73]	Investigate the role of MS‐MHb pathway in modulating anxiety‐like behaviors in mice with PTSD	Adult male C56BL/6 mice PTSD model	↑ Neural activity of MS and MHb glutamatergic neuron in response to PTSD
LHb		Tan et al. [74]	Investigate the potential role of astrocytes in the LHb in modulating anxiety	Adult male C57BL/6J and Mlc1‐tTA::tetO‐YC nano50 mice Stress model	↓ Astrocytic neuron activity
PVN		Li et al. [75]	Investigate the role of PVN‐CeAL oxytocinergic projections in anxiety‐like behaviors induced by inflammatory pain in mice	Adult male C57BL/6 mice Inflammatory pain model	↓ PVN oxytocinergic neuron signals in CFA‐injected mice
		Yu et al. [76]	Investigate whether the acute administration of propofol could cause anxiety‐like behaviors associated with pain, notably through modulating the excitability of PVN CRH neurons	Adult male CRH‐ires‐Cre knock‐in mice Acute inflammatory pain or SNI chronic pain model	↑ PVNCRH neuron activity after acute and chronic pain ↓ PVNCRH neuron activity by acute propofol administration
LA MG ACx		Peng et al. [77]	Investigate how chronic noise exposure affects the brain, particularly the LA, and how this leads to anxiety‐like behaviors	Adult male C57BL/6J mice Chronic noise exposure	↑ Spontaneous and sound‐evoked activity in LA, MG, and ACx in chronic noise‐exposed mice
VTA		Salina‐Hernandez et al. [78]	Investigate the implication of DA neurons in the VTA in fear‐extinction learning	Adult male C57BL/6N and DAT‐cre mice Foot‐shock conditioning	↑ VTA DA neuron activity triggered by unexpected omission of the aversive US during fear extinction learning
		Chen et al. [79]	Investigate the role of GABAergic neurons in the VTA in regulating anxiety‐like behaviors and promoting the intake of palatable food	Adult male C57BL/6J mice Individual housing	↑ VTA GABAergic neuron activity during feeding behavior (palatable food)
		Klenowski et al. [80]	Investigate the activation of GABAergic neurons in the IPN by acute stressors and how it influences stress‐coping behaviors and reward‐seeking behaviors	Adult male SST‐Cre mice (IPN), GAD2‐Cre and DAT‐Cre mice (VTA) Acute stress–restraint	↑ IPN GABAergic neuron activity in GAD2‐Cre mice but not SST‐Cre mice under stress ↓ IPN GABAergic neuron activity during stress‐induced grooming and sucrose consumption
vCA1		Ryan et al. [81]	Investigate how prophylactic administration of (R,S)‐ketamine affects fear generalization and extinction and whether these effects are mediated by BDNF	Adult male wild‐type C57BL/6N mice and BDNF Val66Met (BDNFMet/Met) knock‐in mice Weak shock conditioning Strong shock + auditory cue	↑ vCA1 activity during fear generalization for salient cues in wild‐type mice after acute (R,S)‐ketamine administration

Note: Symbols: ↓: decreased; ↑: increased; ↔: unchanged.

Abbreviations: ACC: anterior cingulate cortex; ACx: auditory cortex; BDNF: brain‐derived neurotrophic factor; BLA: basolateral amygdala; BNST: bed nucleus of the stria terminalis; CeAL: central amygdala (lateral division); COX‐2: cyclooxygenase‐2; CRF: corticotropin‐releasing factor; CRH: corticotrophin‐releasing hormone; CSDS: chronic social defeat stress; CVS: chronic variable stress; DA: dopaminergic; dmPFC: dorsomedial prefrontal cortex; DMS: dorsomedial striatum; IDO1: indoleamine 2,3‐dioxygenase 1; IPN: interpeduncular nucleus; LA: lateral amygdala; LHb: lateral habenula; LMX: lumiracoxib; MG: medial geniculate; MHb: medial habenula; mPFC: medial prefrontal cortex; MS: medial septum; NAc: nucleus accumbens; NTG: nitroglycerin; PTSD: posttraumatic stress disorder; PVN: paraventricular nucleus; vCA1: ventral cornu ammonis area 1; vHPC: ventral hippocampus; VTA: ventral tegmental area.

TABLE 8.

Literature review using PubMed database showed that 5 out of 39 (13%) studies used both naïve and anxiety‐inducing pathophysiological model mice (2020–2024) to investigate neural circuits implicated in anxiety.

Recorded brain area	References	Aim of the study	Animal strain/model	Main results
ICeA, ovBNST, eLPB	Chen et al. [82]	Investigate the role of specific projections from eLPB ChAT neurons in METH withdrawal anxiety and primed reinstatement of conditioned place preference CPP	Adult male C57BL/6 Naïve METH withdrawal model	↑ Ach release in the lCeA and the ovBNST in naïve mice upon eLPB activation ↓ PKCδ+ neurons' activity in both METH‐withdrawn and control mice upon inhibition of eLPBChAT‐lCeAPKCδ ↑ Upon activation of eLPBChAT‐lCeA terminal
BLAGAD	Asim et al. [47]	Investigate the role of BLA GABA neurons in modulating depressive and anxiety‐like phenotypes	Adult male C57BL/6J mice CCK‐ires‐Cre, PV‐ires‐Cre, C57 SST‐ires‐Cre, Vgat‐Cre: Ai14 Naïve SDS	↓ BLA‐GABA neuron activity after chronic stress
LH	Owens‐French et al. [83]	Investigate the role of LH galanin neurons in regulating anxiety‐like behaviors in mice	Adult male and female Galanin‐IRES‐Cre knock‐in mice Naïve Fear conditioning	↑ LH galanin neurons activity in anxiogenic situations, but not during tone
dDG	Qi et al. [84]	Investigate a new optical manipulation approach to stimulate hippocampal neurogenesis and understand the mechanisms underlying its anxiolytic effects	Adult male C57BL/6 mice Naïve Chronic restraint stress	↑ Hippocampal astrocyte activity after NIR treatment
VHPc dHPC	Li et al. [85]	Investigate the role and mechanism of vHPC astrocytes in the regulation of normal and pathological anxiety‐like behaviors	Adult male C57BL/6J mice IP3R2loxp/loxp mice Naïve SRS	↑ vHPC and dHPC astrocytic activity in anxiogenic environments for naïve mice ↔ Astrocytic activity for knockout mice despite anxiolytic effects

Note: Symbols: ↓: decreased; ↑: increased; ↔: unchanged.

Abbreviations: ACh: acetylcholine; BLA: basolateral amygdala; CCK: cholecystokinin‐expressing; ChAT: choline acetyltransferase; CPP: conditioned place preference; dDG: dorsal dentate gyrus; dHPC: dorsal hippocampus; eLPB: lateral parabrachial nucleus; ICeA: intercalated central amygdala; IP3R2: inositol 1,4,5‐trisphosphate receptor type 2; LH: lateral hypothalamus; METH: methamphetamine; NIR: near‐infrared; ovBNST: oval subdivision of the bed nucleus of the stria terminalis; PKCδ: protein kinase C delta; PV: parvalbumin; SDS: social defeat stress; SRS: subacute restraint stress model; SST: somatostatin peptide; Vgat: vesicular GABA transporter; vHPC: ventral hippocampus.

3.1. Naïve Mice (Table 6)

3.1.1. Prefrontal Cortex (PC)

• Prelimbic cortex (PL cortex) and ventromedial prefrontal cortex (vmPFC)

The role of neuropeptides in regulating anxiety behaviors was also highlighted in studies by Brockway et al. [49] and Munguba et al. [48]. The role of somatostatin peptide (SST) signaling in the PL reveals that SST neuron activity increased in response to anxiogenic environments [49]. This study emphasizes the importance of SST signaling in modulating anxiety responses and potentially influencing exploratory behaviors. Munguba et al. [48] focused on the presynaptic metabotropic glutamate receptor 2 (mGluR2) in the prefrontal cortex and its role in the insula–amygdala pathway. Their findings suggest that mGluR2 signaling plays an important anxiolytic role, especially in regulating synaptic activity in the amygdala during anxiety‐provoking situations.

• Medial prefrontal cortex and basolateral amygdala circuit (dmPFC‐BLA)

Loewke et al. [50] explored fronto–striatal projections between the dmPFC‐BLA, which play a key role in regulating innate avoidance behaviors. They observed an increase in the activity of pyramidal neurons in the dmPFC during exposure to anxiogenic environments, suggesting that this fronto–limbic pathway is critical for managing fear responses and anxiety in the face of potential threats.

3.1.2. Insular Cortex (IC)

• Anterior insular cortex (aIC) and pre‐insular cortex (pIC)

The interaction between the IC and the amygdala in encoding emotional valence and anxiety‐related behaviors was investigated by Nicolas et al. [51]. In their study, they observed that both the aIC and the pIC play crucial roles in anxiety processing. They found that activity in the aIC was particularly heightened in anxiogenic contexts, whereas the pIC showed increased activity in response to negative stimuli. These findings suggest that the IC, in conjunction with the amygdala, is integral in encoding both the emotional value and the anxiety‐related responses to environmental threats.

One of the first relevant studies was conducted by Gehrlach et al. [52], who investigated the role of the pIC in processing aversive emotional and bodily states. Their research revealed that pIC activity significantly decreased in anxiogenic environments. This result suggests that this brain region may play a modulatory role in processing negative emotions or stress, and that this reduced activity could indicate dysfunction in the processing of emotional information in anxious individuals.

3.1.3. Bed Nucleus of the Stria Terminalis (BNST)

The BNST is another key brain area involved in the regulation of anxiety [53]. Their study focused on PKCδ‐expressing neurons in the BNST, which they found to be activated in response to threat‐related behaviors and anxiety‐inducing situations. This activation was linked to reduced exploratory behavior, a hallmark of anxiety‐like responses. Their findings suggest that BNST PKCδ cells may be important in regulating anxiety in response to perceived threats, furthering our understanding of how specific cell types in the BNST contribute to anxiety processing.

The role of the BNST in anxiety regulation was also explored by Jaramillo et al. [54]. They showed that protein kinase C delta (PKCδ)‐expressing cells in the BNST were activated in synchrony with a consummatory food approach in a stress‐inducing context. This BNST activation may be linked to the regulation of reward‐seeking behaviors in anxiogenic situations, thus providing a connection between the activation of emotional circuits and reward‐motivated behaviors.

3.1.4. Paraventricular Nucleus (PVN)

Yuan et al. [55] focused on corticotropin‐releasing hormone (CRH) neurons in the PVN of the hypothalamus to understand their role in the stress response. The researchers observed an increase in the activity of these neurons during stressors, an activity that was inhibited during reward consumption. These results highlight the importance of modulating neural excitability in the PVN, suggesting that stress circuits are tightly regulated by the presence of aversive or rewarding stimuli, and that an imbalance of these signals may play a role in the development of anxiety disorders.

Another significant contribution to the field was made by Qui et al. [56], who examined the role of CRH in the PVN in mediating the anxiolytic effects of dexmedetomidine. Their study, conducted on adult male and female C57BL/6J mice, demonstrated that administration of dexmedetomidine reduced the excitability of CRH neurons in the PVN, which in turn mediated the drug's anxiolytic effects. This work highlights the importance of CRH signaling in the PVN as a potential therapeutic target for anxiety relief, especially in the context of pharmacological interventions.

3.1.5. Ventral Dentate Gyrus (vDG) and Dorsal Dentate Gyrus (dDG)

Other studies have further expanded our understanding of more specific mechanisms, such as Wang et al. [57], who examined the activity of motor cortex (MC) neurons in anxiogenic environments. They observed an increase in neural excitability under these conditions, emphasizing that mossy cells (MCs) could influence anxious behaviors by modulating neural activity in brain regions critical for anxiety, such as the vDG and dDG.

3.1.6. Midbrain Structures

• Nucleus accumbens (NAc)

The effect of pharmacological agents has also been a subject of interest, and fiber photometry provided valuable insights into their influence on neurotransmission pathways associated with anxiety. In that context, Rizzo et al. [58] studied the impact of ketamine on dopamine release in the NAc. Their findings indicated that ketamine increased the duration of dopamine transients and decreased the frequency of glutamate transients in the NAc, suggesting that ketamine modulates dopamine signaling in a way that could alleviate anxiety symptoms. This research points to the potential of ketamine as a novel treatment for anxiety disorders and provides insights into how it may alter the brain's reward and emotional circuits.

• Interpeduncular nucleus (IPN)

One major area of study is the role of the dopaminergic circuit between the ventral tegmental area (VTA) and the IPN, which has been implicated in both anxiety and social behaviors. Molas et al. [59] investigated this VTA‐IPN circuit in adult male and female C57BL/6J mice. Their study found that dopaminergic neurons in the IPN were activated during exposure to anxiety‐inducing environments and social novelty. This suggests that the VTA‐IPN circuit may be involved in encoding emotional responses to both fear and social stimuli, indicating its potential as a target for future therapeutic interventions aimed at anxiety disorders.

• Dorsal raphe nucleus (DRN)

The study of serotonergic systems has also provided valuable insights into anxiety regulation. The interaction between cocaine‐ and amphetamine‐regulated transcript (CART) peptides and serotonin (5‐HT) systems in the DRN was also explored by Balasubramanian et al. [60]. Their research revealed that CART peptide signaling influenced 5‐HT release in the DRN, affecting anxiety‐like behaviors in mice. This work underscores the complexity of serotoninergic regulation in anxiety and highlights the potential for targeting 5‐HT systems in the development of anxiety treatments.

3.1.7. Cerebellum Lobule VII

Finally, Pei Wern Chin et al. [61] examined the role of serotonin signaling in the cerebellum, specifically in lobule VII, in mediating anxiety‐like behaviors. Their study showed that increased serotonin levels in this cerebellar region were associated with reduced anxiety‐like behaviors, suggesting that serotonergic modulation in the cerebellum could play a role in emotional regulation. This finding opens new avenues for exploring the cerebellum's involvement in anxiety and emotional processing, a less‐explored area in the field of anxiety research.

Together, these studies represent a broad spectrum of research contributing to deciphering the neural circuits involved in anxiety regulation. The diverse methodologies and experimental approaches used, including fiber photometry, continue to shed light on the complex interplay of neurotransmitters, neuropeptides, and brain regions that govern anxiety responses. As research in this field progresses, these insights could lead to more targeted and effective treatments for anxiety and related disorders.

3.2. Anxiety‐Inducing Pathophysiological Model Mice

To understand the mechanisms and neural circuits involved in anxiety disorders, particularly generalized anxiety disorder (GAD), it is essential to study anxiety models rather than relying on naïve animals. Research has shown that animal models of anxiety, such as those induced by chronic stress or exposure to anxiogenic stimuli, more accurately reflect the neurobiological changes associated with anxiety disorders. In contrast, naïve animals do not exhibit persistent and maladaptive responses typical of clinical anxiety, which makes them less suitable for studying the complex interactions between brain regions, neurotransmitter systems, and behavior that underpin anxiety pathophysiology (Table 7).

3.2.1. Prefrontal Cortex (Table 7, Part 1)

• mPFC

Investigating the molecular underpinnings of resilience to chronic stress, Kuang et al. [62] studied the role of P2X2 receptors in pyramidal neurons in the mPFC. They found that the knockout (KO) of the P2X2 receptor increased mPFC pyramidal neuron activity and enhanced resilience to chronic stress, suggesting that P2X2 receptor signaling might influence the ability to cope with stress.

• Dorsomedial prefrontal cortex (dmPFC)

Kajs et al. [63] explored the role of the dmPFC in active avoidance learning in mice. They found that dmPFC activity increased during active avoidance but decreased during cued freezing. Additionally, projections from the dmPFC to the dorsomedial striatum (DMS) were more active during avoidance behavior, whereas projections to the BLA showed reduced activity. These results suggest that the dmPFC coordinates both emotional responses and goal‐directed avoidance behavior through distinct neural pathways.

• Anterior cingulate cortex (ACC)

The role of the ACC in anxiety regulation has been highlighted in several studies. Gao et al. [64] investigated how dexmedetomidine, an α2‐adrenergic agonist, affects glutamatergic neurons in the ACC during a chronic pain model. They observed a decrease in glutamatergic neuron hyperactivity in the ACC following dexmedetomidine administration, which was accompanied by a reduction in anxiety‐like behaviors. This work suggests that ACC glutamatergic signaling plays a key role in anxiety, and its modulation could provide new therapeutic strategies for pain‐associated anxiety.

In addition, the role of serotonin in anxiety and pain was explored by Li et al. [65], who examined the therapeutic effects of duloxetine in an inflammatory pain model. Duloxetine increased serotonin concentrations in the ACC while reducing glutamatergic neuron activity. These changes were associated with a reduction in anxiety‐like behaviors, further supporting the role of serotoninergic systems in the modulation of anxiety in the context of chronic pain.

Hu et al. [66] focused on the enzyme indoleamine 2,3‐dioxygenase's (IDO1) role in chronic migraine‐induced anxiety, finding an increase in ACC neural activity linked to pain sensitivity and comorbid anxiety.

3.2.2. NAc (Table 7, Part 1)

The involvement of brain areas in both anxiety and depressive‐like behaviors has also been extensively studied. Muir et al. [67] examined the ventral hippocampus nucleus accumbens (vHPC‐NAc) circuit in C57BL/6J mice exposed to chronic variable stress (CVS). Their findings showed increased amplitude and frequency of vHPC‐NAc activity in both sexes, with significant sex differences observed during stress exposure. This suggests that the vHPC‐NAc circuit is involved in the expression of stress‐induced behaviors, potentially linking hippocampal and striatal circuits to emotional regulation.

Moreover, Flores‐Garcia et al. [68] showed that music exposure could counteract chronic pain‐induced anxiety by preventing reductions in NAc dopamine activity.

3.2.3. BLA (Table 7, Part 2)

Morgan et al. [69] explored the effects of pharmacological inhibition of cyclooxygenase‐2 (COX‐2) on neural activity in the BLA and its impact on stress‐induced anxiety‐like behaviors. In their experiments with ICR (CD‐1) mice subjected to acute restraint stress, they found that COX‐2 inhibition resulted in an increase in BLA neural activity, which was reversed by lumiracoxib (LMX) pretreatment. This work underscores the importance of COX‐2 signaling in the amygdala as a potential target for modulating anxiety, particularly in stress‐related contexts.

Similarly, Huang et al. [70] focused on the BLA and its role in the susceptibility to stress‐related psychiatric disorders. After exposure to chronic social stress (CSDS), BLA activity was elevated in susceptible mice (SI‐sus) but decreased in resilient mice (SI‐res). These results point to the BLA as a critical region in determining vulnerability or resilience to stress, with differential activity patterns potentially underlying individual differences in anxiety susceptibility.

3.2.4. Extended Amygdala System (Table 7, Part 2)

• BNST

Similarly, the role of glutamatergic signaling in the brain's response to stress was investigated by Salimando et al. [71], who examined how GluN2D‐containing NMDA receptors modulate neural activity in the BNST. Using GluN2D KO mice, they observed heightened activity of corticotropin‐releasing factor (CRF) neurons in the dorsolateral BNST during novel environment exploration. These findings suggest that GluN2D receptors are involved in the regulation of stress and anxiety‐like behaviors, as their absence leads to heightened CRF neural activity, which is often associated with anxiety responses.

• Central amygdala lateral subdivision (CeAL)

Demaestri et al. [72] examined the impact of early life adversity (ELA) on CRF+ neurons in the CeAL, showing increased neural activity during startle responses in ELA mice, which contributed to heightened threat sensitivity and sex differences in anxiety susceptibility.

3.2.5. Medical Habenula (MHb) and Lateral Habenula (LHb) (Table 7, Part 2)

The involvement of the mesolimbic system in post‐traumatic stress disorder (PTSD) has been explored by Wu et al. [73], who investigated the role of the medial septum to medial habenula circuit (MS‐MHb) pathway in modulating anxiety‐like behaviors in a PTSD model. Their findings revealed that the two regions showed increased neural activity in response to PTSD, suggesting that these regions play a critical role in the expression of anxiety‐related behaviors associated with trauma. In addition, Tan et al. [74] studied the role of astrocytes in the LHb in anxiety modulation, finding a reduction in astrocytic neuron activity during stress.

3.2.6. PVN (Table 7, Part 2)

Further studies have examined the role of oxytocinergic signaling in the regulation of anxiety‐like behaviors. Li et al. [75] focused on the projections from the PVN to the CeAL and found that oxytocinergic signaling was reduced in CFA‐injected mice, which were exhibiting anxiety‐like behaviors. This suggests that oxytocin might play a modulatory role in anxiety responses, particularly in the context of pain and inflammation.

Yu et al. [76] explored the effects of propofol on anxiety‐like behavior in mice with inflammatory pain, revealing that acute propofol administration decreased PVN CRH neural activity, thereby alleviating anxiety.

3.2.7. Lateral Amygdala (LA), Medial Geniculate Nucleus (MG), and Auditory Cortex (ACx) (Table 7, Part 3)

Environmental factors, such as chronic noise exposure, can also influence brain activity and anxiety responses. Peng et al. [77] demonstrated that chronic noise exposure resulted in increased spontaneous and sound‐evoked activity in the LA, MG, and ACx, which was associated with anxiety‐like behaviors. These findings suggest that environmental stressors can alter neural circuits involved in emotional processing, leading to heightened anxiety responses.

3.2.8. VTA (Table 7, Part 3)

A key study by Salina‐Hernandez et al. [78] investigated the role of the VTA in fear extinction learning. Using adult male C57BL/6N and DAT‐Cre mice, the authors showed that the unexpected omission of the aversive unconditioned stimulus (US) during extinction learning led to a marked increase in the activity of VTA dopamine (DA) neurons. These findings underscore the critical role of the VTA in encoding prediction error signals during fear extinction, suggesting that dopaminergic signaling in this region is essential for adaptive emotional learning and the suppression of fear‐related memories.

Chen et al. [79] shifted focus to the VTA again, but this time to investigate the role of GABAergic neurons in regulating anxiety‐like behaviors and food intake. Their study in C57BL/6J mice revealed that GABAergic neurons in the VTA were more active during feeding behavior, particularly when consuming palatable food. This suggests that GABAergic signaling in the VTA plays a role in modulating both emotional and motivational aspects of behavior, with potential implications for understanding the link between anxiety and eating behaviors.

Furthermore, Klenowski et al. [80] investigated how acute stressors activate GABAergic neurons in the IPN and their influence on stress‐coping and reward‐seeking behaviors. They found that GABAergic neuron activity in the IPN increased in GAD2‐Cre mice but not in SST‐Cre mice under stress, with a decrease in IPN activity during stress‐induced grooming and sucrose consumption.

3.2.9. Ventral Cornu Ammonis Area 1 (vCA1) (Table 7, Part 3)

The brain‐derived neurotrophic factor (BDNF) has been implicated in anxiety and fear responses [81]. They showed that acute (R,S)‐ketamine administration increased the vCA1 activity during fear generalization in wild‐type mice, but not in BDNF Val66Met knock‐in mice. This finding suggests that BDNF plays a crucial role in modulating fear and anxiety responses, and that ketamine's effects on fear generalization may be mediated through BDNF signaling.

3.3. Both Naïve and Anxiety‐Inducing Pathophysiological Model Mice

To better understand the neural circuits involved in anxiety, as previously described, many studies focus on anxiety‐inducing pathophysiological models, which allow the examination of how specific brain regions and pathways are activated under conditions of stress or anxiety. However, a smaller subset of research has begun to explore the use of both naïve and anxiety‐inducing pathophysiological model mice, offering insights into how pre‐existing neural states may interact with anxiety‐induced changes in brain activity (Table 8).

3.3.1. Intermediate Central Amygdala (ICeA), Oval Bed Nucleus of the Stria Terminalis (ovBNST), and External Lateral Parabrachial Nucleus (eLPB)

Moreover, Chen et al. [82] investigated the role of specific projections from eLPB ChAT neurons in a naïve and a METH withdrawal anxiety model in adult male C57BL/6 mice. This study revealed an increase not only of Ach release in the lCeA and in the ovBNST in naïve mice upon eLPB activation but also upon activation of eLPB ChAT‐lCeA terminal and a decrease of PKCδ+ neurons activity in both METH‐withdrawn and control mice upon inhibition of eLPB ChAT‐lCeAPKCδ.

3.3.2. BLA

Asim et al. [47] investigated the role of BLA GABA neurons in modulating depressive and anxiety‐like phenotypes in adult naïve C57BL/6, CCK‐ires‐Cre, PV‐ires‐Cre, and SST‐ires‐Cre, Vgat‐Cre: Ai14 male mice before and after undergoing SDS. They showed that different aversive stimuli had different effects on BLA‐GABA neuron activity, depending on the stimuli's nature and duration and that chronic aversive stimuli appear to inhibit the activity of BLA‐GABA neurons.

3.3.3. Lateral Hypothalamus (LH)

Owens‐French et al. [83] investigated the role of LH galanin neurons in regulating anxiety‐like behaviors in adult male and female Galanin‐IRES‐Cre knock mice, both in their naïve state and after undergoing fear conditioning. An increase in LH galanin neuron activity during anxiogenic situations, but not during tone even after conditioning, was observed. This suggests an involvement in the initial anxiety response rather than a response to a conditioned stimulus.

3.3.4. dDG

Qi et al. [84] investigated a new optical manipulation approach to stimulate hippocampal neurogenesis and understand the mechanisms underlying its anxiolytic effects in both naïve and chronic restraint stress models of adult male C57BL/6 mice. They showed that there was an increase in hippocampal astrocyte activity after NIR treatment.

3.3.5. vHPC and Dorsal Hippocampus (dHPC)

Finally, Li et al. [85] studied the role and mechanism of vHPC astrocytes in the regulation of normal and pathological anxiety‐like behaviors in naïve and transgenic IP3R2loxp/loxp mice. This study revealed an increase in vHPC and dHPC astrocytic activity in anxiogenic environments for naïve mice. However, the conditional KO of astrocytic Itpr2 in the vHPC produced no change in astrocytic activity despite clear anxiolytic effects.

To summarize, fiber photometry has significantly advanced our understanding of the neural circuits involved in anxiety and emotional behaviors (Figure 3). Through real‐time recordings of neural activity, studies have identified key brain regions and pathways, such as the pre‐IC, PVN, and amygdala, which play critical roles in anxiety regulation. These studies have also underscored the complexity of how emotional and stress‐related circuits interact, providing insights into the underlying mechanisms of anxiety. As research continues to explore various neuropeptides, neurotransmitters, and neural circuits, these findings pave the way for more targeted and effective treatments for anxiety disorders, including potential pharmacological interventions. By focusing on anxiety‐inducing pathophysiological models, researchers are gaining deeper insights into the brain's response to stress and emotional stimuli, which could lead to novel therapeutic strategies for managing anxiety‐related conditions.

4. Challenges and Limits of Fiber Photometry in Anxiety Circuit Investigations

Despite offering valuable insight into anxiety‐related neural circuits, the application of fiber photometry is challenged by the complexity of anxiety behaviors and the inherent limitations associated with the technique. Below, we outline key technical and interpretive obstacles and potential strategies to address them.

4.1. Behavioral Artifacts

Fiber photometry recordings rely on detecting fluctuations of fluorescence over time, which can be compromised by movement. Freezing, a core anxiety‐related behavior in rodents, may affect the mechanical stability of the optical fiber and potentially reduce signal noise. In contrast, avoidance may increase motion artifacts and elevate the amount of photonic noise recorded. These behavior dynamics can confound ΔF/F measurements and analyses.

The occurrence of behavioral artifacts can be mitigated by integrating fiber photometry with inert fluorophore controls (such as tdTomato), real‐time motion correction algorithms, and a precise behavioral tracking system to improve signal fidelity. Additionally, the preprocessing pipeline needs to be designed in a way that sufficiently accounts for freezing‐induced signal stabilization or, if such behaviors might amplify or suppress interpreted neural engagement.

4.2. Biological and Experimental Variability

Fiber photometry signals are influenced by both experimental conditions (stress vs. control) and by sex and hormonal status (intact vs. ovariectomized rats [86]). For example, Johnson et al. demonstrated that male and female mice exposed to 6 to 56 days of variable stress exhibited distinct timelines and magnitudes of anxiety‐like behaviors. This suggests that behavioral endophenotypes are shaped by sex and stress exposure, showing that signal variability may arise from these factors that are important to interpret photometry results [87].

Wallace et al. further showed that chronic stress induces sex‐specific cellular and activity changes in vmPFC‐dendritic remodeling, interneuron modulation, which impacts photometry signals [88].

Additionally, to the sex‐based variability in behavior and stress response, a higher gene AAV‐driven transgene expression was described for male mice. For instance, Davidoff et al. [89] showed that they had 5 to 13 times higher AAV‐driven transgenes than females, which were androgen‐regulated and persisted across different AAV serotypes.

We recommend verifying equal viral expression post‐mortem using histology or fluorescence. Furthermore, the normalization of photometry data by expression levels allows for the control of signal amplitude bias. Ultimately, sex should be included as a biological variable during data analysis and interpretation.

4.3. Circuit Ambiguity

Brain regions implicated in anxiety, such as the amygdala [90], often serve multiple functions, and the bulk nature of fiber photometry signals makes it difficult to disentangle anxiogenic activity from neural responses related to other processes. Additionally, the differentiation between anxiety, aversion, and novelty‐driven responses can be challenging due to the capture of overlapping neural activities without resolving the underlying motivational or emotional context.

Circuit ambiguity can be resolved by combining orthogonal behavioral tasks with projection‐specific labeling strategies, such as dual‐virus retrograde systems (e.g., AAV‐FLEX‐GCaMP in conjunction with retro¹¹V‐Cre). This approach enables selective monitoring of defined neural pathways, improving the interpretability of circuit‐level activity.

4.4. Cell‐Type Specificity and Causal Inference

Fiber photometry allows the measurement of the average fluorescent activity across a population of neurons within the illuminated brain region but does not allow distinguishing how different cell types are contributing to the signal without the use of targeted strategies [10]. This limited cell‐type specificity may represent a significant constraint in interpreting activity at the resolution of defined neural populations. Indeed, the summation of multiple, sometimes opposing, activities may obscure the recorded signal. Furthermore, causal inference can be difficult to determine as the observed calcium activity could be the cause of an anxious state, its consequence, or completely unrelated.

To address these limitations, we recommend the use of dual‐color calcium indicators alongside genetically targeted expression systems, such as Cre‐loxP or cell‐type–specific viral vectors, to restrict GCaMP expression to defined neuronal populations or circuits. Furthermore, combining imaging with neuromodulatory techniques, including optogenetics or DREADDs, would enable precise spatial and temporal control of neural activity, thereby facilitating more accurate mechanistic dissection of circuit function. For causal inference, the use of genetically modified models (e.g., KO mice) can provide deeper insights into the contribution of specific genes or pathways involved in anxiety‐related behaviors. Finally, the inclusion of appropriate control groups is essential to ensure the reliability and interpretability of experimental findings.

4.5. Spatio‐Temporal Resolution and State Transitions

The technique's bulk sampling volume may obscure spatial heterogeneity and limit the ability to resolve microcircuit‐specific dynamics [91]. Furthermore, the technique struggles to detect tonic or sustained signal changes in fluorescence due to common processing steps such as z‐score normalization, which may inadvertently remove meaningful baseline shifts [92]. Temporal resolution is also constrained by the slow kinetics of calcium indicators like GCamP, which can blur or miss rapid state transitions and phase neural events [93]. Additionally, although early phases of the photometry signal correlate with population spiking, prolonged calcium elevations often reflect dendritic activity rather than precise spike timing [94].

The use of tapered optical fibers that can collect light over a larger and more structured volume may help overcome spatial resolution limits [95]. Complementary electrophysiology could improve temporal resolution and capture fast dynamics.

In the light of its technical and interpretive constraints, we advise caution when relying solely on fiber photometry to investigate anxiety‐related circuits. Depending on the experimental goals, a multimodal framework may offer a path to circumvent the limitations and challenges previously described.

5. Conclusion

In conclusion, anxiety disorders are a prevalent and significant public health concern, yet current treatments remain insufficient. A comprehensive understanding of neural circuits is essential for elucidating the pathophysiology of these disorders and developing novel, more effective, and safer treatments. In this context, we demonstrated that fiber photometry is a powerful, versatile, and cost‐effective tool to explore anxiety‐related neural circuits and establish a link between neural activity and behavior in awake, freely moving subjects, both naïve animals and experimental models. This approach already provides valuable insights into the physiological, pathological, and pharmacological mechanisms underlying anxiety and holds the potential to contribute even further.

Author Contributions

All authors contributed to the design of the review paper. All the authors read and approved the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Abdennebi S., Touihri N., Corruble E., David D., and Mendez‐David I., “Leveraging Fiber Photometry to Decipher Neural Circuits Underlying Anxiety in Mice,” Fundamental & Clinical Pharmacology 39, no. 5 (2025): e70043, 10.1111/fcp.70043.

Data Availability Statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.