The abilities to optogenetically manipulate and monitor neuronal activity with cell-type specificity are indispensable methods for neuroscientists to study neural circuits in behaving animals^1-4. In particular, heterogeneous populations of neurons and various neurotransmitters⁵ comprise neural circuits of different functions, exhibiting distinctive neuronal activities⁶ corresponding to specific behaviors^{4, 7}. For example, dopamine (DA) neurons are known to have two different firing patterns: the phasic pattern and the tonic pattern. These patterns induce distinct activities in their downstream neurons, which are involved in diverse physiological functions^{6, 8, 9}. Moreover, the dentate ‘gate’ hypothesis of seizures has been experimentally evaluated in vivo using a closed-loop system to record and optogenetically manipulate the activity of granule cells during seizures in mice¹⁰. Thus, simultaneous optogenetic manipulation and multi-color recording to monitor feedback effects resulting from optogenetic circuit interventions are now recognized as the method of dreams for the causal investigations of neural circuits and neurological diseases.

However, previous techniques or systems to optogenetically manipulate or monitor neuronal activity in behaving animals are mostly separated and work independently^11-23. Few studies have attempted to combine optogenetic manipulation and neuronal activity recording into a single system^24-28. The optrode, which integrates electrophysiological recording²⁹ with the optogenetic stimulations^{24, 30}, is the most widely used method to record and manipulate neurons, but it cannot directly monitor neurotransmitter activities. Enabled by the development of genetically encoded fluorescent indicators (GEFI)^31-35, optical imaging technologies including the miniature fluorescence microscope^{18, 27} and the miniaturized two-photon microscope^{17, 36} are capable of imaging and manipulating neurons at the same time. By implanting a gradient refractive index (GRIN) lens in the brain, these imaging techniques support direct imaging and optogenetic manipulation of specific types of neurons at single-cell resolution in deep brain regions of freely moving animals³⁷. However, an implanted GRIN lens is usually in a diameter of 500 μm to 1 mm, thus inducing substantial damages to the brain^{38, 39}. Furthermore, the current systems implemented with endoscopic imaging techniques are only capable of optogenetic manipulation and single-channel recording at a single wavelength.

As an alternative technique that induces less invasion, fiber photometry^{14, 15, 25, 26, 40-42} provides a unique feature for an easy and stable recording of population activities with cell-type specificity in freely moving animals. Various approaches have been proposed to incorporate optogenetics with fiber photometry, enabling wide applications and the discovery of many functional insights^{25, 26, 41}. However, since the excitation spectrum of Channelrhodopsin-2 (ChR2, 473 nm)^{1, 43} is close to the emission spectrum of the Ca²⁺ indicator Oregon green 488 BAPTA-1 (OGB-1) or GCaMP (both peak at ~510 nm), and considering that it is quite challenging to completely filter out optogenetic stimulation light with milliwatt, obvious artifacts from optogenetic stimulation are inevitable in a weak (pico watt) fluorescent signal during recording (Supplementary information Fig. S1(a))²⁵. Replacing ChR2 with a fast red-shifted channelrhodopsin, (e.g., bReaChES, 594 nm) can significantly reduce the artifacts observed with ChR2 (Supplementary information Fig. S1(b))²⁶. However, only one type of neuron or neurotransmitter can be monitored in this case, and stimulation artifact is still a problem when involving dual-color recording, as the excitation spectrum of bReaChES overlaps with the emission spectrum of the RFP-based GEFIs (Supplementary information Fig. S1(c)).

To achieve simultaneous optogenetic manipulation and dual-color recording, more wavelength-range fluorescent proteins with spectral spacing (the peak emission wavelength of GEFIs to the effective excitation wavelength of opsin-sensors should exceed ~60 nm, Supplementary information Fig. S1(d)) need to be involved and the optical system is thus required to fully cover the visible spectrum. However, the currently available fiber photometry systems only cover the spectrum of 405~600 nm by using the classical epifluorescence imaging architecture (Supplementary information Fig. S2)²⁶ consists of an objective lens and two or three dichroic mirrors. In this case, more dichroic mirrors are needed to couple the multiple wavelength light beams into one optical fiber, which makes the system more complex to extend the spectrum. Therefore, to make a compatible and flexible system, an all-fiber-transmission optical system can deliver multiple-wavelength light for simultaneous optogenetics and neuronal activity recording, and collect multi-color fluorescence signals.

Here, we developed an all-fiber-transmission photometry system based on a multi-branch fiber bundle^44-46 for simultaneous optogenetic manipulation and multi-color recording of neuronal Ca²⁺ or neurotransmitter signals in a freely moving animal. Briefly, a non-wavelength selective multi-branch fiber bundle is combined with a PMT to replace the traditional imaging structure. The multi-function fiber photometry system can deliver all of the required light using optical fibers, making the system more robust for use in freely behaving experimental contexts. Notably, we adopted a narrow linewidth laser and used a lock-in amplification method^{47, 48} to successfully manage issues with optogenetic stimulation artifacts and channel crosstalk. We show that our system has excellent light transmission performance. Moreover, it can be readily reconfigured for simultaneous multi-color recording of neuronal Ca²⁺ signals and neurotransmitter dynamics and precise optogenetic manipulations in freely moving animals.

Using a custom-designed bifurcated fiber bundle (AlphaTech), we developed an all-fiber-transmission system for single-channel fiber photometry (Fig. 1). The overall appearance and section image of each branch of the bundle is shown in Fig. 1(a). The bifurcated fiber bundle consists of 3 branches in total: the single-fiber branch i, the collection branch ii, and the common branch iii. All these 83 fibers (branch iii) are hexagonally distributed in the common branch and divided into two branches with 1 (branch i) and 82 (branch ii) fibers in each of them (as Fig. 1(a) sub-image shows). To decrease optical transmission loss, the bundle length was limited to about 350 mm. The structure of common branch iii is shown in Fig. 1(b). The fiber we use in our design is glass optical fiber with 50 μm diameter and 0.54 numerical aperture (NA). To match the currently available 200 μm or 400 μm multi-mode fiber and efficiently collect the returned fluorescence, the diameter of the bundle is designed at approximately 500 μm. The single fiber from branch i for excitation light (blue) transmission is placed in the center of branch iii surrounded by the other 82 fibers (branch ii) for collecting fluorescence signals. Fig. 1(c) is a schematic diagram of the single-channel fiber photometry system we constructed using a bifurcated fiber bundle. Excitation light from a 450-nm LED through the single-fiber branch i is coupled into the multi-mode fiber (200 μm in diameter and 0.37 in NA) connecting to the fiber bundle and the animal. The emission fluorescence is collected by the common branch and delivered to the PMT (Hamamatsu, H10720-210) behind the collimator (DHC, GCL-010131) and the fluorescence filter (Semrock, FF01-525/39-25).

Using a custom-designed four-branch fiber bundle (AlphaTech), we developed a multi-function fiber photometry system (Fig. 2) that can realize multi-color recording while also supporting precise optogenetic stimulation. Figure 2(a) shows a photo of the bundle consisting of 1 common branch v and 4 splitting branches i–iv. As illustrated in Fig. 2(b), 85 glass optical fibers (50 μm in diameter and 0.54 in NA) are hexagonally distributed in the common branch v. The three brighter fibers in the center of the bundle correspond to the three single-fiber branches i–iii for coupling to three different excitation lights. Considering that the small diameter (50 μm) of the single optical fiber will dramatically reduce the coupling efficiency, a higher power excitation light source is needed to achieve the excitation light intensity of GEFIs (15~40 μW) or opsin (2~20 mW). A schematic diagram of the multi-function fiber photometry system we constructed using this fiber bundle is shown in Fig. 2(c). In this system, we used three light sources with different wavelengths (Source 1∶450-nm LED; Source 2∶561-nm laser, Changchun New Industries Optoelectronics Technology, MGL-FN-561-100mW; Source 3∶660-nm laser, Changchun New Industries Optoelectronics Technology, MRL-III-633-200mW) and coupled them into the multi-mode fiber connecting to the animal using the three single-fiber branches i–iii. 450-nm and 561-nm excitation lights are used to excite the two different GEFIs (450-nm for dopamine sensors and 561-nm for the jGRECO1a Ca²⁺ sensor in this paper) for monitoring neuronal activities, and the 660-nm laser is used for the optogenetic stimulation. The emission fluorescence is then collected by the common branch v and delivered to the PMT (Hamamatsu, H11461-03) behind the collimator (DHC, GCL-010131) and the fluorescence filters (Dual bandpass filter: Chroma, 59012M; 633-nm short-pass filter: Semrock, BSP01-633R-25). The output signals of the PMT are preprocessed by a custom-designed lock-in amplifier (LIA)⁴⁷, digitized using a DAQ card (NI, USB-6003), and recorded by a custom LabVIEW script (50 Hz per channel).

As the power of optogenetic stimulation light (2 mW in our experiment) is much higher than the power of excitation light (20~40 μW) for recording, artificial signals may arise during optogenetic stimulation. In our system, we placed a narrow line-width bandpass filter (Thorlabs, FBH660-10) and a short-pass filter (Semrock, BSP01-633R-25) in front of the 660-nm laser and PMT respectively to decrease the stimulation artifacts. Moreover, to separate the fluorescence signals and further inhibit the stimulation artifacts during the dual-color recording and simultaneous optogenetic manipulation, we adopted the lock-in amplification method in our system. As shown in Fig. 2(d), these two excitation lights for recording are modulated by sinusoidal signals with different frequencies (450 nm, 211 Hz; 561 nm, 531 Hz; the modulation frequencies must avoid harmonic interference, eg: powerline harmonics at multiples of 50 Hz)^{49, 50}, and the optogenetic light is often modulated using relatively low frequency (0–50 Hz) square signals. The emission fluorescence signals from the brain are detected by the PMT, and the output signals of the PMT are then de-modulated by the LIA, using the corresponding demodulation signals (211 Hz and 531 Hz sinusoidal signals); the fluorescence signals of different GEFIs can thus be recorded without channel crosstalk and stimulation artifacts.

The fluorescence signals are de-modulated from the output signals (f_PMT) of the PMT by the LIA, using the corresponding demodulation signals (f_ref-greenandf_ref-red). As the frequency of the optogenetic stimulation artifact (0~50 Hz) is quite lower than the modulation frequency of these two excitation lights (f₀: 211 Hz;f₁: 531 Hz), the artifact component can be treated as a constant noise (n_noise) during the period of stimulation. Thus, the input signal f_i (the output signalsf_PMT of the PMT) can then be represented as:

fi=fPMT=Asin(ω0t)+Bsin(ω1t)+nnoise,

where A and B are the signal amplitude, and w₀ and w₁are the angular frequency (2πf₀ and 2πf₁ respectively).

The corresponding modulation signal f_ref-greenand f_ref-red can be set to be:

fref−green=Csin(ω0t),

fref−red=Dsin(ω1t),

where C and D are the amplitude of the corresponding modulation signals.

After the multiplier, the outputs (f_c-green andf_c-red) are:

fc−green=fi×fref−green=(Asin(ω0t)+Bsin(ω1t)+nnoise)×Csin(ω0t)=A×C2−A×C2sin(2ω0t)+B×C2(cos((ω1−ω0)t)−cos((ω1+ω0)t))+nnoise×Csin(ω0t),

fc−red=fi×fref−red=(Asin(ω0t)+Bsin(ω1t)+nnoise)×Dsin(ω1t)=B×D2−B×D2sin(2ω1t)+A×D2(cos((ω0−ω1)t)−cos((ω0+ω1)t))+nnoise×Dsin(ω1t).

According to the Eqs. (4) and (5), when the signals pass through the low-pass filter, the high-frequency components are filtered out, and the fluorescence signals f_out-green and f_out-redcan be separated from the mixed fluorescence signals f_i:

fout−green=A×C2,

fout−red=B×D2.

All experiments were approved and were conducted in accordance with the guidelines of the committee on animal health and care of the National Institute of Biological Sciences (NIBS), Beijing.

DAT-Cre mice (JAX Strain 006660) were obtained from the Jackson Laboratory. Male wild-type C57BL6/N mice were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd (China). All mice were kept under standard conditions of 12/12 hours light-dark cycle (light on 8 PM), with ad libitum food and water, except for the sucrose delivery during the fiber photometry experiments, in which mice were deprived of water for 36 hours before the test.

The pAAV-hSyn-GRAB_DA2m plasmid (9.4 × 10¹² vg/mL) we used to detect the DA release in NAc was kindly provided by Dr. Yulong Li (Peking University). The AAV-hSyn-Flex-ChrimsonR-mCherry (10¹³ vg/mL) for VTA-NAc terminal stimulation was purchased from Shanghai Taitool Bioscience Co., China. pAAV9-hsyn-NES-jRGECO1a-WPRE (5.06 × 10¹³ vg/mL) was purchased from Shandong Vigene Biosciences. Inc., China. The AAV-hSyn-EGFP-RPL10 (4.3 × 10¹² vg/mL) plasmid in our control experiment was kindly provided by Dr. Fei Zhao (Chinese Institute for Brain Research, Beijing).

Mice were anesthetized by Avertin (i.p., 250 mg/kg) and fixed in a standard stereotaxic frame (RWD Instruments, China). After disinfection with medical alcohol and exposure of the skull, a small craniotomy was made and ~300 nL viruses were delivered using a calibrated pulled-glass pipette (Sutter Instrument); injection into the unilateral targeted brain areas (46 nL/min). All coordinates for the injection sites, as measured from the bregma, include: Nucleus Accumbens core (NAcC): 0.98 mm anterior, 1.3 mm lateral, 4.3 mm ventral; NAc lateral shell (NAcLat): 1.1 mm anterior, 1.3 mm lateral, 4.6 mm ventral; Ventral tegmental area (VTA): 3.4 mm posterior, 0.3 mm lateral, 4.1 mm ventral. After pausing for an additional 5 minutes, the pipette was removed slowly from the brain.

For the fiber photometry and optogenetic experiments, optical fiber (NA: 0.39, diameter: 200 μm; Thorlabs) was implanted 100 μm above the injection site. The ceramic ferrule outside the fiber was then fixed to the skull with super bond C&B dental silicate cement (Parkell Prod). The animals were maintained in their home cage for 2 weeks for recovery and virus expression prior to any testing.

After completion of the recording and testing, mice were anesthetized by an overdose of pentobarbital and perfused intracardially with 0.9% phosphate buffer saline and then 4% paraformaldehyde (PFA). The brains were removed carefully and postfixed in 4% PFA for 4 h at room temperature. After cryoprotection in 30% sucrose, coronal sections (50 μm thickness) were cut using a cryostat (Leica, CM1950). The slices were imaged using a slide scanner (Olympus, VS120) with 10x objectives to confirm virus expression and fiber placements. Images were analyzed using ImageJ (NIH).

For fiber photometry recording, the data were acquired by a data acquisition card (National Instrument, USB-6001) and was processed by a custom MATLAB (MathWork, Inc.) program for further analysis. We defined the value of fluorescence change (∆F/F) by calculating the following equation:

ΔF/F=(F−F0)/F0,

where F₀ is the mean value of the 2 s control time immediately before the behavioral event. The ∆F/F value was displayed as a heat map or as an average plot (with a shaded area representing the SEM).

For the statistical significance of the event-related fluorescence (ERF) change analysis³ (Statistically evaluate the degree of correlation between neuronal activity and event stimulation), we performed Wilcoxon signed-rank tests to test the significance of neuronal responses by comparing the ∆F/F value at each time point with the baseline values (the mean ∆F/Fvalue of the 2 s control time immediately before the behavioral event). The P-value of each time was superimposed on the average ERF curve with orange lines, indicating a statistically significant (P<0.05) increase.

In the all-fiber-transmission photometry system, a multi-channel fiber bundle was designed and manufactured to realize the all-fiber transmission of both the excitation light and the emission light. The spectral transmittance of the multi-channel bundle was about 60% per meter in the range of 400 to 900 nm. To decrease optical transmission loss, the bundle length was limited to about 350 mm, thus about 84% of the light was effectively transmitted.

When using the common branch to collect the fluorescence from the multi-mode fiber, part of the fluorescence is guided back to the light source through the single fiber branch, therefore cannot be detected by the PMT. To characterize the loss in fluorescence signal, we tested the collection efficiency of these two systems using a series of sodium fluorescein solutions with different concentrations and compared the results to those acquired using the traditional epi-fluorescence system used in our previous study³. We used the same laser, PMT, multi-mode fiber, emission filter, and sodium fluorescein solution in the testing with various systems.

Under the same excitation intensity (40 μW), we tested five concentrations (10~50 ng/mL) and calculated the intensity variation (ΔF) by subtracting the value obtained with pure water. As shown in Fig. 3, our system was linearly correlated with the traditional setup: the red lines represent the fitting curves for the single-channel fiber photometry system (Fig. 3(a), f(x) = 1.297x, R-square: 0.9938) and the multi-function fiber photometry system (Fig. 3(b), f(x) = 1.233x, R-square: 0.9982). The slopes of the fitting curve are 1.297 and 1.233, indicating that the fluorescence collection efficiency of our method was 20–30% better than the traditional setup. These results demonstrated that our all-fiber transmission fiber photometry system based on a multi-branch fiber bundle was capable of effectively exciting and collecting fluorescence signals.

We then evaluated the potential crosstalk between two channels. As shown in Fig. 3(c), we measured the fluorescence intensity of each channel while opening (“on”) and closing (“off”) the 450-nm and 561-nm excitation lights. We observed no significant change in both channels when the other excitation light was “on” or “off”, which demonstrated that no substantial channel crosstalk existed in our all-fiber-transmission photometry system.

We further evaluated the potential stimulation artifacts by measuring the fluorescence intensity of the two channels without the 450-nm and 561-nm excitation light. We observed that the fluorescence intensity of both channels exhibited no response to the optogenetic stimulation (Fig. 3(d)). This demonstrated that the stimulation artifacts can be inhibited to a basal noise level in our all-fiber-transmission photometry system.

We then tested the performance of our single-channel fiber photometry system based on a bifurcated fiber bundle by recording DA dynamics at a 100 Hz sampling rate in the nucleus accumbens core (NAcC) of a freely moving mouse. DA is an important neurotransmitter in the brain that exerts functions related to reward, motivation, and motor control^{8, 51}. The dysfunctions of the DA system are associated with neurological disorders such as Parkinson's disease and schizophrenia⁵². The NAc is known to be a major brain area that accepts DA released from the ventral tegmental area (VTA) DA neurons⁵³.

We used the single-channel fiber photometry system shown in Fig. 1(c) to monitor DA dynamics when delivering a sucrose solution (12.5 μL, 5% w/v) to a 36-hour water-deprived mouse through a cheek fistula (Fig. 4(a)). The genetically encoded GPCR-activation-based-DA sensor (GRAB_DA2m)³³ was expressed in NAcC neurons to monitor the DA release from DA neurons in the VTA in a wild-type mouse. An optical fiber was implanted 100 μm above the injection site for both the delivery of excitation lights and the collection of emitted light. Histology confirmed the locations of the fiber and GRAB_DA2m expression in the NAcC two weeks later (Fig. 4(b)).

We measured DA signals in a freely moving mouse upon a 1 s unexpected sucrose solution that was delivered at random inter-trial intervals of 15–25 s (Fig. 4(c)). As shown in Fig. 4(d, e) (Supplementary information Fig. S3), the DA fluorescence signals, excited by 40 μW of 450-nm LED light, significantly increased immediately following the delivery of the unexpected reward (∆F/F = 11.5 ± 0.9%, mean ± SEM; P<0.05, Wilcoxon’s signed-rank test;n = 20 trials with one mouse). It indicated that the dynamics of DA release from the VTA-NAc DA neurons were positively correlated with the reward, consistent with previous studies^{33, 53}. These results together demonstrated that our all-fiber transmission fiber photometry system can effectively detect temporal fluorescence dynamics and can be used for real-time monitoring of neurotransmitter dynamics in freely moving animals.

We then tested the performance of a multi-channel fiber photometry system based on a four-branch fiber bundle by simultaneously recording DA dynamics and neuronal Ca²⁺ signals in the NAc lateral shell (NAcLat). Multiple neuron types that release different neurotransmitters can show various activities in response to a specific function in the NAc or other nucleus; thus, the ability to simultaneously record neuronal activities labeled by multiple-color was essential to study the interactions amongst distinct neuron types in behaving animals⁵⁴. The DA sensor GRAB_DA2m (450-nm excitation) and the red Ca²⁺ indicator jGRECO1a (561-nm excitation)³² were co-expressed in the NAcLat of a wild-type mouse (Fig. 5(a)), and the excitation lights (20 µW) were delivered to the brain through the same optical fiber. Histology confirmed the locations of the fiber and the expression of both GRAB_DA2m and jGRECO1a in the NAcLat two weeks later (Fig. 5(b)).

We recorded the DA signals and neuronal activities in response to unexpected rewards in a freely moving mouse following the same protocol as described above (Fig. 5(c)). As shown in Fig. 5(d, e), we observed that DA fluorescence signals and neuronal Ca²⁺ signals both significantly increased upon the delivery of unexpected sucrose solutions (DA: ∆F/F = 11.2 ± 0.4%, jGRECO1a: ∆F/F = 8.5 ± 0.8%, mean ± SEM; P<0.05, Wilcoxon’s signed-rank test;n = 20 trials with one mouse). It indicated that NAcLat neurons and the VTA-NAc DA neurons were positively correlated with the rewards, consistent with previous studies^{33, 34, 53}. However, we noted that these activities had a clear difference in the time course³ for unexpected rewards (Supplementary information Fig. 4). Throughout the recording session, the DA signals were correlated with, yet lagged behind, the neuronal Ca²⁺ signals (DA: peak time = 1.91 ± 0.14 s, median latency to peak = 0.99 ± 0.14 s; jGRECO1a: peak time = 0.96 ± 0.02 s, median latency to peak = 0.34 ± 0.04 s, mean ± SEM). Further, the DA signals displayed an overall prolongment of activation (DA: decay duration = 3.81 ± 0.16 s; jGRECO1a: decay duration = 1.08 ± 0.04 s, mean ± SEM). Given that GRAB_DA2m and jGRECO1a exhibited similarly rapid rising times^{32, 33}, the result thus suggested that NAc neurons received reward signals from two separate sources: a yet-to-identified input triggered the earlier Ca²⁺ increase, and another input likely from the VTA triggered the slightly later DA release. Together, the results showed strong evidence of heterogeneous neuronal activities in local neuronal circuits. These results also demonstrated that our all-fiber transmission fiber photometry system based on the four-branch fiber bundle can simultaneously detect dynamic fluorescence signals from different GEFIs and was suitable for real-time monitoring of multiple, distinct neuronal activities in freely moving animals.

We then tested the performance of our multi-function fiber photometry system for simultaneous dual-color recording and precise optogenetic manipulation of neuronal activity in the NAc of a freely moving mouse. Many studies have demonstrated the power of combining simultaneous recording of cell-type-specific neurons and precise optogenetic manipulation of specific neuron types during animal behavior for studies of functional connectivity, activity dynamics, causation between neuronal activities and behaviors, and the physiological functions of neural circuits^{4, 25, 54, 55}.

VTA dopaminergic neurons are known to project to the NAc and to regulate a variety of physiological processes, including reward, motivation, and motor control^{48, 56}. We first verified that our system does not elicit stimulation artifacts when simultaneously performing optogenetic manipulation and real-time dual-color recording in vivo. A schematic diagram of the experiment is shown in Fig. 6(a): we specifically labeled the VTA DA neurons in a DAT-Cre mouse using a DIO-ChrimsonR-mCherry AAV vector and expressed GFP in all NAcLat neurons. Histology conducted two weeks later confirmed the location of the fiber, the expression of GFP in the somata of NAcLat neurons, and the expression of mCherry in the VTA-NAc dopaminergic terminals (Fig. 6(b)). No significant fluorescence changes were observed in the GFP or mCherry channels during optogenetic stimulation (2 mW, 20 ms per pulse, and 20 Hz for 1 s) (Fig. 6(c), P>0.05, Wilcoxon’s signed-rank test;n= 10 trials with one mouse), supporting that no obvious stimulation artifacts occurred with our system in this experiment.

Finally, we recorded neuronal Ca²⁺ signals and DA dynamics simultaneously while activating the VTA-NAC dopaminergic terminals in a freely moving mouse. We specifically labeled the VTA DA neurons in a DAT-Cre mouse using a DIO-ChrimsonR-mCherry virus, and co-expressed the DA sensor GRAB_DA2m and the red Ca²⁺ indicator jGRECO1a in NAcLat neurons. We recorded the DA signals and jRGECO1a fluorescence signals in a freely moving mouse (50 Hz per channel). As we know, DA neurons exhibit two different firing patterns to induce distinct activities in their downstream neurons and involve diverse physiological functions^{6, 8}. Then, two different patterns of light pulse stimulations were delivered to the VTA-NAC dopaminergic terminals to induce tonic or phasic firing patterns⁴⁸. We observed that optogenetic phasic stimulation (20-ms per pulse and 20 Hz for 1 s) evoked a significant, time-locked rising trend in the DA signals (DA: ∆F/F = 6.3 ± 0.7%, mean ± SEM; P<0.05, Wilcoxon’s signed-rank test;n = 10 trials with one mouse), consistent with a previous study³³. However, we noted that neuronal Ca²⁺ signals displayed only a slight rising trend (Fig. 6(e), Supplementary. Fig. 5). Upon applying the tonic stimulation (10-ms per pulse and 4 Hz for 5 s), there were no significant changes in either the DA signals or the neuronal Ca²⁺ signals (Fig. 6(f),P>0.05, Wilcoxon’s signed-rank test;n= 10 trials with one mouse).

These results support previous findings that dopamine is acutely released in the NAcLat when evoked by a burst stimulation of the VTA-NAC dopaminergic terminals^{48, 57}, and suggest that Ca²⁺ signals from population activity are not sensitive to DA release from the VTA neurons in the NAcLat. Notably, to our understanding, such results could not be obtained using conventional electrophysiological recordings: it cannot directly monitor the level of neurotransmitter release⁵⁸. These results support that our system can perform simultaneous dual-color recording of various neuronal populations while also enabling precise optogenetic manipulation; these combined capacities seem likely to make our system useful for diverse neuroscience functional investigations of various neuron populations and circuits.

We developed a flexible all-fiber transmission fiber photometry system that supports simultaneous dual-color recording of neuronal or neurotransmitter activities combined with precise optogenetic manipulation in freely moving animals. In this system, we designed and manufactured a non-wavelength selective multi-branch fiber bundle to realize the all-fiber transmission of both the excitation light and the emission light, which simplified the system and increased its robustness in freely moving animal experiments. Adopting a laser of narrow line-width for the stimulation and a lock-in amplification method, our system can extract fluorescence signals of two different colors (GFP-based and RFP-based GEFIs) while inhibiting the potential artifacts caused by the optogenetic manipulation (channelrhodopsin with far red-shifted excitation spectrum, 660 nm). Importantly, we showed that our system’s collection efficiency was better than a conventional epi-fluorescence system, and we successfully recorded DA dynamics in the NAc of a freely moving mouse. Finally, we successfully recorded both neuronal Ca²⁺ and neurotransmitter dynamic signals in the NAc while also applying precise optogenetic manipulation of the dopaminergic terminals in the same site in a freely moving mouse.

Optogenetic manipulation and monitoring of neuronal activity in a freely moving animal are essential methods for rodent-based neuroscience investigations of brain connectivity and function^{1, 2}. Several previous studies^24-27 have attempted to combine these two methods to achieve real-time feedback after applying optogenetic stimulation to directly test whether inferred models of dynamics, connectivity, and causation are accurate in vivo⁴. However, these methods—including the optrode²⁴, the miniature fluorescent miniscope²⁷, the miniaturized two-photon microscope¹⁷, and fiber photometry^{25, 26, 41}—only support the monitoring of one type of neuronal activity when applying optogenetic manipulation at present. A major reason for this is that the action spectrums of the commonly used opsin-sensors⁴³ like ChR2 (470 nm), eArch3.0 (566 nm), and ChrimsonR (590 nm) are relatively limited, which either have spectral overlap to the excitation spectrum or are too close to the emission spectrum of GEFIs^{31, 32} like GCaMP6 (510 nm) and jRGECO1a (588 nm). Therefore, when using both the GFP-based and RFP-based GEFIs to label the neurons simultaneously, it is very challenging to completely filter out high-power stimulation light (2~20 mW). And the residual optogenetic stimulation light detected by the detector will lead to obvious artifacts in the acquired signals^{4, 25, 26, 41}. Fluorescent proteins with more spectral spacing need to be involved and the optical system is required to fully cover the visible spectrum.

Our design adopted the strategy of all-fiber-transmission by using a small multi-branch fiber bundle to replace the dichroic mirrors and an objective lens, which realized—in a single fiber bundle—three excitation or stimulation lights and two emission lights. In this way, we dramatically simplified the multichannel system and made it more robust and flexible for use in freely behaving experimental contexts. We adopted a laser of narrow line-width of 660-nm to activate the red light–drivable channelrhodopsin (ChrimsonR), therefore inhibited the stimulation-induced artifacts using a lock-in amplification method, which also allowed us to separate the fluorescence signals of the dual-color channels. Our system can be easily modified into different variants for other fluorescent probes and opsin-based sensors by changing the light sources and filter positions before the PMT without adjusting the optical path.

With rapid developments of fluorescent indicators^{31, 32}, such as neurotransmitter probes^{33, 34, 54}, light-activated conversion technologies^{59, 60}, a multi-function system is required for neuroscience research^{4, 10}. Our all-fiber-transmission photometry system supports simultaneous dual-color recording of neuronal or neurotransmitter activities simultaneously with precise optogenetic manipulation in freely moving animals. Thus, it is especially well-suited for the study of interactions amongst different neuron types or neurotransmitters in behaving animals for multiple processes, such as reward³⁴ or sleep homeostasis⁵⁴. More importantly, our system has the potential to be modified into a closed-loop system that can guide optogenetic manipulation informed by real-time monitoring feedback effects^{4, 10, 25}. The all-fiber-transmission photometry system should be a valuable tool for neuroscientists seeking to comprehensively profile neural circuit functions^{3, 53, 54} and to study neurological diseases¹⁰.