Inhibitory effect of 980-nm laser on neural activity of the rat’s cochlear nucleus

Abstract. Near-infrared radiation (NIR) has been described as one of the highest-resolution tools for neuromodulation. However, the poor tissue penetration depth of NIR has limited its further application on some of the deeper layer neurons in vivo. A 980-nm short-wavelength NIR (SW-NIR) with high penetration depth was employed, and its inhibitory effect on neurons was investigated in vivo. In experiments, SW-NIR was implemented on the rat’s cochlear nucleus (CN), the auditory pathway was activated by pure-tones through the rat’s external auditory canal, and the neural responses were recorded in the inferior colliculus by a multichannel electrode array. Neural firing rate (FR) and the first spike latency (FSL) were analyzed to evaluate the optically induced neural inhibition. Meanwhile, a two-layered finite element, consisting of a fluid layer and a gray matter layer, was established to model the optically induced temperature changes in CN; different stimulation paradigms were used to compare the inhibitory efficiency of SW-NIR. Results showed that SW-NIR could reversibly inhibit acoustically induced CN neural activities: with the increase of laser radiant exposures energy, neural FR decreased significantly and FSL lengthened steadily. Significant inhibition occurred when the optical pulse stimulated prior to the acoustic stimulus. Results indicated that the inhibition relies on the establishment time of the temperature field. Moreover, our preliminary results suggest that short-wavelength infrared could regulate the activities of neurons beyond the neural tissues laser irradiated through neural networks and conduction in vivo. These findings may provide a method for accurate neuromodulation in vivo.

Inhibitory effect of 980-nm laser on neural activity of the rat's cochlear nucleus Bin Jiang, a Wensheng Hou, a,b,c Nan Xia, a,d, * Fei Peng, a Xing Wang, a,b Chunye Chen, a,b Yi Zhou, e Xiaolin Zheng, a,b,c and Xiaoying Wu a,b,c, *

Introduction
Neuromodulation is an efficient technique to study how neural activities operate brain functions, and it requires not only effective neural stimulation but also high precision. [1][2][3][4] Among all neuromodulation techniques, pulsed near-infrared (NIR) radiation (wavelength between 1850 and 2120 nm) is considered a strong candidate for precise control of neural activity, due to its high stimulating spatial resolution and contact-free advantages. [5][6][7][8][9][10] NIR has been widely used to modulate the functionality of the peripheral [5][6][7]11,12 and the central nervous system. [8][9][10]13,14 Most previous studies have focused on NIR-induced neural activation; [6][7][8][9]15 neural suppression is also indispensable in neuroscience, especially for brain disorder treatment. [16][17][18] Recently, evidence has gradually appeared showing that NIR could reversibly inhibit unwanted neural activity 16,19,20 or block neural signal propagation with high precision. 21,22 However, current NIR-based neural inhibition experiments have mostly been carried out on peripheral nerves, 21,23 neurons cultured in vitro, 19,20,24 and superficial cortex; 10 research investigating the inhibitory effect of NIR on neurons in deeper layers in vivo has been limited. The main reason for this, in addition to the complexity of the central nervous system and complex neural responses to NIR, 8,10 has been the poor tissue penetration depth of NIR. The maximum penetration depth of NIR used in previous works was about 1100 μm, 25 but target neurons in vivo such as hippocampal neurons are located much deeper, 2 so it has been a challenge for the conventional NIRs to modulate such deeper neurons in vivo.
Compared with NIR, short-wavelength NIR (SW-NIR, wavelength from 800 to 1100 nm) has a much higher penetration depth in tissues (at least 2.5 cm), 26 which is very beneficial for neuromodulation by laser in vivo. Recently, SW-NIR with wavelengths of 808 and 980 nm has been successfully applied to activate the auditory nerve and neurons in the primary motor cortex. [27][28][29][30][31] Different from the mechanism of optically induced neural activation, neural inhibition of NIR relies on the optically induced tissue baseline temperature increasing, 16,19,20,32 and evidence shows that neural activities can be effectively inhibited when tissue temperature is increased by 3°C to 8°C. 16,32 Besides its strong penetrating ability, SW-NIR also has a low tissue absorption coefficient, which may affect the thermal interaction between laser and target tissue. Therefore, it is unknown whether SW-NIR could effectively elevate the tissue temperature to reach the temperature threshold for neural inhibition. Fortunately, the inhibition effect of 980-nm laser pulse on neurons was recorded in our preliminary experiment, 33 but whether this inhibitory effect of SW-NIR on neurons is associated with the tissue baseline temperature increase and how SW-NIR regulates neurons in vivo are as yet unexplored. Furthermore, most of the existing studies have directly recorded the responses of neurons irradiated by lasers; few studies have been done to evaluate the infrared neuromodulation of neural networks and conduction in vivo.
To better understand the effect of SW-NIR on neurons in vivo, a 980-nm fiber-coupled laser was used to modulate CN neural activities of the Sprague-Dawley rat (SD rat). Optical stimulation parameters and different stimulation paradigms were investigated. Laser fiber was placed on top of the rat's CN, the auditory pathway was activated by pure-tones given from the external auditory canal, and neural responses were recorded from inferior colliculus (IC). Neural firing rate (FR) and the first spike latency (FSL) were analyzed to evaluate the optically induced neural inhibition. Results showed that 980-nm laser pulse could reversibly and repeatedly inhibit acoustically induced neural activities in CN: with the increase of laser pulse energy, the neural inhibition rate increased steadily and reached up to 70% when the laser radiant exposure was 39.06 J∕cm 2 (pulse power: 6 W, pulse width: 5 ms). Experiments under different stimulation paradigms indicated that the inhibition relies on the establishment time of the temperature field in tissue induced by laser pulse.

Materials and Methods
Twelve healthy adult female SD rats (weight 200-300 g, age 6-8 weeks, not in the estrous cycle) were used in this study because of their better hearing sensitivity than male SD rats. 34 All rats were purchased from the Experimental Animal Center of the Daping Hospital. The procedures were performed in accordance with protocols of the Care and Use of Laboratory Animals approved by the Army Medical University.

Modeling of Laser-Induced Temperature Changes
A two-layered finite-element method of CN was established in COMSOL Multiphysics (version 5.3, Stockholm, Sweden), including a cerebrospinal fluid (CSF) layer (thickness: 0.1 mm) and a gray matter (GM) layer (thickness: 1 mm). Laser pulse power varied from 0.1 to 9 W (pulse width: 5 ms), and pulse width varied from 0.1 to 30 ms (pulse power: 6 W). Laser fiber diameter and numerical aperture were set to 200 μm and 0.22, respectively. The distance of optical fiber tip to the CN was 250 μm and the irradiation radius of the laser fiber was calculated by Eq. (1). (A schematic diagram is shown in Fig. S1 in the Supplementary Material.) The thermal parameters used for simulation are summarized in Table 1.
E Q -T A R G E T ; t e m p : i n t r a l i n k -; e 0 0 1 ; 6 3 ; 2 2 6 R ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi where α is the spreading angle (sin α ¼ 0.22), R a is the fiber core radius (100 μm), and Z is the distance from fiber to tissue (250 μm).
Laser-induced temporal and spatial temperature changes in CN were calculated using the Pennes bioheat equation, as follows: 38 E Q -T A R G E T ; t e m p : i n t r a l i n k -; e 0 0 2 ; 6 3 ; 1 0 9 ρC where ρ is the tissue density, C is the tissue-specific heat, k is the tissue thermal conductivity, T is the tissue temperature, ρ b is the blood density, ω b is the blood perfusion, C b is the blood specific heat capacity, T b is the blood temperature, t is the calculating time, Q met is the metabolic heat generation, and Q laser is the optical heat source. The optical heat source Q laser was calculated by Eq. (3). 39 The scattering (μ s ), absorption (μ a ) coefficients of GM and CSF for 980-nm laser and 1850-nm laser were obtained from the literature [40][41][42] (see Table 2). Anisotropy (g) was assumed to be equal to 0.9: E Q -T A R G E T ; t e m p : i n t r a l i n k -; e 0 0 3 ; 3 2 6 ; 3 3 4  where μ α is the absorption coefficient, μ s is the scattering coefficient, μ t ¼ μ s þ μ α is the tissue attenuation coefficient, and φ 0 is the incident light flux density.

Animal Surgery and Preparation
Surgical and neural data collection methods have been previously described. 33 Animals were anesthetized by an intraperitoneal injection of urethane (1.2 g∕kg body weight in sterile saline). Depth of anesthesia was assessed by toe pinch every 30 to 60 min, maintaining with 0.16 g urethane/kg body weight to ensure the anesthesia state of animals. The animal was moved to the sound-attenuating room, where the ambient noise floor was 25-dB sound pressure level (SPL). The hairs on the auricle, head, and neck of the animal were removed. Tracheotomy was performed on the animal to assist animal respiration. The animal's head was secured on a stereotaxic instrument (ST-SND-B, Chengdu Instrument Factory, Chengdu, China) with ear bars through both sides of the external auditory canal. Skin and membranous tissue were removed to expose the skull. Three holes were made adjacent to the bregma, about 3 mm apart, using a motorized dental drill (STRONG90+102L, South Korea) with a drill bit size of 1-mm diameter. Three 1.4-mm stainless steel self-tapping screws (1.4 × 3.6 mm 2 ) were fastened on the skull through the holes. The tip of a custom-made head holder was placed on the top of the screws and fixed by dental acrylic. After dental acrylic solidified, the ear bars were removed to free the animal ears. The skull and occipital bone above the cerebellum were ground off using a dental drill to expose the edge of the cerebrum and the cerebellum. Left cerebellar hemisphere was sucked away by a homemade syringe straw to expose the left cochlear nucleus (CN). Then, the right cerebrum was partially removed to expose the right IC (see Fig. 1). The entire experiment was carried out on a heating pad connected with the thermostatic bath (HSS-1B, Chengdu Instrument Factory, China), and the animal's body temperature remained constantly at 38°C.

Acoustic Stimulation
Sound stimulation was given by the earphone (Beyerdynamic DT 770, Germany) placed in the sound-attenuating room. The earphone was placed at a distance of 1 cm from the ipsilateral ear canal with the exposed CN. The acoustic stimulation signals were generated by LabVIEW (LabVIEW2012, National Instruments, Austin, U.S.A). Acoustic clicks (square waveforms with 100-μs pulse width, 9-Hz repetition rate) and pure tone bursts (12-ms duration, 1-ms rise/fall) were used to detect the animal hearing threshold before and after surgery. The frequency for the pure tone bursts started from 0.5 kHz and was increased in eight steps per octave over seven octaves to 32 kHz. The repetition rate of pure tone bursts was 4 Hz. Sound levels at each frequency began at 80 dB SPL and were attenuated in steps of 5 to 30 dB SPL. A 50-ms white noise was used to search neurons in IC, and pure tone from 0.5 to 32 kHz was used to determine the neural characteristic frequency (CF) for each neuron recorded by the electrode.

Acoustical and Optical Co-Stimulation Protocol
The optical stimulation was generated by a 980-nm laser (Associated Opto-Electronics Corp., Chongqing, China) coupled with a 200-μm fiber, and the fiber was held by a three-dimensional (3-D) micromanipulator (WST-3, Chengdu Instrument Factory, Sichuan, China). The distal tip of the optical fiber was placed on the CN surface. The laser parameters included laser pulse width and laser pulse power: laser pulse width ranged from 1 to 5 ms (in 2-ms step, pulse power: 6 W), and laser pulse power ranged from 1 to 6 W (in 1-W step, pulse width: 5 ms). The corresponding radiant exposures were ranging from 6.51 to 39.06 J∕cm 2 at a 5-ms pulse width. The stimulation frequency was 4 Hz. The output power from the optical fiber was measured before experiment using a laser power meter (PowerMax-Pro, Santa Clara, U.S.A).
The acoustic stimulus was pure-tone. The frequency of puretone was matched to a certain frequency, which was consistent with CF of IC neurons. The sound level of pure-tone was fixed at 80 dB SPL, and the repetition rate was fixed 4 Hz.
Stimulation paradigms of acoustic and optical stimulus included four types: laser pulse was delivered 5 ms prior to acoustic stimulus; acoustic stimulus and laser pulse were given simultaneously; laser pulse and acoustic stimulus end at the same time; and laser pulse was given immediately right after the acoustic stimulus [see Fig. 9(a), row 1].

Auditory Brainstem Response Measurement
Three stainless steel needles (25-mm length and 0.35-mm diameter, Suzhou Hua Tuo Medical Instruments Co., Ltd., Suzhou, China) were used to record auditory brainstem response (ABR). The stainless steel needle electrodes were insulated with polydimethylsiloxane, which expects that the two ends of the electrodes were exposed. The record electrode was placed on the top of the skull. The reference electrode was placed at the left mastoid and the ground electrode was placed on the tip of the nose. The electrodes were connected to the Cerebus physiological signal acquisition system (Cerebus 6.01, Blackrock Microsystems, Salt Lake City, U.S.A). The sampling rate was set to 30 kHz. The recorded signal was bandpass filtered between 300 and 3000 Hz. ABRs were averaged at least 1350 times.  accuracy of 1 μm. The ground electrode was placed on the temporal bone under the skin.

Neural Response Recording
Multiunit and single-unit responses (spikes) from each channel were recorded by Cerebus multichannel physiological signal acquisition system. For spike recording, a bandpass filter with the range of 300 to 3000 Hz was used and the sampling rate was set at 30 kHz. The synchronization signals from the optical and acoustical stimuli were also collected for offline data analysis.

Histological Examination
Histological analyses were performed after laser stimulation experiments to evaluate stimulating safety. CN was removed right after animal death and kept in 4% paraformaldehyde for 48 h. Then, the tissues were dehydrated in increasing concentrations of ethanol and were embedded in paraffin before cutting. CN tissues were cut with cross-sections (4-μm thick) on a paraffin slicing machine (RM2016, Shanghai Leica Instruments Factory, Shanghai, China). CN slices were stained with Mayer's hemotoxylin and Putt's eosin (H&E, Servicebio, China), and the CN cell status was observed under the microscope.

Data Analysis and Statistics
All data were analyzed offline in MATLAB 2012b (MathWorks, Inc., Natick, Massachusetts) and OriginPro 8 (OriginLab Corporation, Northampton, Massachusetts). Peristimulus time histograms (PSTHs), average firing rate, FSL, and FR change were used to analyze the neural response. The data of 50 ms after the start of each optical stimulus were used to calculate the PSTHs (25 bins∕50 ms). FSL was defined as the time difference between the onset of a stimulus and the occurrence time of the first spike. The FR changes of the group of acoustical and optical co-stimulation were calculated by Eq. (4): where FR ðAÞ and FR ðAþOÞ represent the average FR before and after the onset of laser pulse, respectively. The FR changes of the group of acoustical stimulation were calculated by Eq.
where FR ðAÞ and FR ðA1Þ represent the average FR before the onset of laser pulse and the laser pulse is turned off, respectively.

Simulation Result of Temperature Changes Under Different Laser Parameters
Simulation diagram is shown in Fig. 2(a). The width of the model is 3 mm, the thickness is 1.1 mm, and the irradiation radius of the laser is 0.1564 mm. The tissue temperature spatial distributions of two lasers at the maximum temperature level are shown in Figs. 2(b) and 2(c). Compared with 1850 nm, 980 nm does have deep penetration in tissue, but it needs nearly 15 times more radiant energy than 1850 nm laser to achieve the same tissue baseline temperature increase. The tissue temperature changes under different radiant energy of 980-nm laser are shown in Fig. 3. Tissue temperature increased monotonously with the increasing of the laser pulse power (pulse width: 5 ms) and pulse width (pulse power: 6 W) [see Figs. 3(a) and 3(b)]. A typical temperature rise curve over time is shown in Fig. 3(c); the laser radiant exposure was 39.06 J∕cm 2 (pulse power: 6 W, pulse width: 5 ms). The curve shows that the rise time of tissue temperature was longer than the pulse width, and the time difference (δt) was about 3.4 ms. The time difference became larger when the laser pulse width was getting longer [ Fig. 3(d)]. To achieve stable neural inhibition, the appropriate temperature increment induced by laser should be around 8°C. 16,32 Corresponding to our simulation result, the appropriate pulse power is about 7 W, and pulse width is about 5 ms. Considering our laser output power limitations (maximum pulse power: 6 W), pulse power in the following experiments ranged from 1 to 6 W and pulse width was from 1 to 5 ms. The corresponding radiant exposures ranged from 6.51 to 39.06 J∕cm 2 at pulse width of 5 ms.

In Vivo Study on Neuromodulation by SW-NIR
Twelve SD rats were used in this study. ABR was performed to evaluate the hearing sensitivity of rats before and after surgery. Animals with hearing threshold under 30 dB SPL and elevated not more than 10 dB SPL after surgery were used in the following experiment. Data from 10 animals were used in the following analysis. For each animal, neural activity induced by pure tones was recorded in the IC, and CF of each neuron was calculated offline. The CF was defined as the frequency at which the neuron is most sensitive. Neurons with a clear CF were used in the following experiments and others were rejected. Figure 4 shows the neural responses recorded one-time by a 16-channel electrode array. Each panel represents a neural response to pure tones at different pure-tone frequencies and different sound levels at one channel. The CF of neuron and electrode channel number was marked on each panel. The CFs of the recorded neurons ranged from 8 to 29.34 kHz. A total of 60 neurons were recorded in the IC of 10 SD rats. The CFs of these neurons are shown in Table 3. We mainly focused on high-frequency neurons in IC (CF more than 16 kHz) since the neurons with high CF were mostly located in dorsal cochlear nucleus (DCN), which corresponded to the site of laser irradiation. . Figure 6(a) shows the inhibitory effect of SW-NIR on continuous interval neural activity induced by acoustical stimuli. The laser radiant exposure was 39.06 J∕cm 2 (pulse width: 5 ms) and the stimulation frequency was 4 Hz. Before the laser was turned on, acoustic stimuli could effectively activate auditory pathway. When the laser was turned on, the acoustically induced neural activities in CN were inhibited, leading to the obviously decreased FR of IC neurons. After laser was turned off, the neuron FR quickly recovered to the initial level. The same results could be found in other neurons with different CFs [see Fig. 6(b)], which indicates that the inhibition is reversible.
The effect of different laser parameters on neurons was investigated, and results are shown in Fig. 7. First, the pulse width of stimulation was set to 5 ms, and laser pulse power from distal end of the optical fiber was adjusted from 1 to 6 W. The radiant exposures were ranging from 6.51 to 39.06 J∕cm 2 . Figure 7(a) shows an example of laser induced neural inhibition. The inhibitory effect enhanced with the increase of laser radiant exposures (pulse width: 5 ms). The averaged result is shown in Fig. 7(b). The inhibition rate increases from 20% to 71.5% when laser radiant exposures varied from 6.51 to 39.06 J∕cm 2 . Then, the pulse power was set to 6 W, and the pulse width was tested at 1, 3, and 5 ms, respectively. The inhibition rate increased from 54.2% to 71.5% when laser pulse width varied from 1 to 5 ms [see Fig. 7(c)].

Effect of SW-NIR on neural FSL
FSL has been found to provide more information than interval spikes, which is very important for neural temporal coding, especially in sensory systems. 43,44 Thus, we further explored laser-induced FSL changes in IC neuron. Figure 8(a) shows an example of FSL shifts before and after the laser with the radiant exposure of 39.06 J∕cm 2 (pulse width: 5 ms) was turned on. The FSL of the IC neurons was obviously lengthened after the laser was turned on. Figure 8

Effect of different stimulation paradigms on acoustically induced CN neural activity
Four different stimulation paradigms used in this study are shown in Fig. 9(a) (the first row). In each stimulation paradigm, pure tone with a fixed frequency (neuron CF) was used to stimulate from external auditory canal. The laser radiant exposure was 39.06 J∕cm 2 (pulse width: 5 ms). Figure 9(a) shows IC neurons in response to four stimulation paradigms. Neural inhibitory rates under different stimulation paradigms were summarized in Fig. 9(b). One-way ANOVA showed that the effect of stimulation paradigms on the neural FR change was significant. And the largest inhibition occurred when the laser pulse was prior and synchronous to the acoustical stimulus (Fig. 9). Compared with acoustic and optical synchronous stimulation, optical prior to acoustic stimulus could obtain a higher inhibitory rate (67.4% versus 87.5%) (ANOVA, p ¼ 0.046). Similar results were obtained in the neural FSL statistics [see Fig. 9(c)]. This may result from the establishment time of the tissue temperature field induced by laser pulse. From our simulation result, we could see that tissue temperature rising time was longer than laser pulse width. With the first stimulation paradigm, tissue temperature rose to a higher level when the neural signal transmitted to CN, so the inhibition effect of laser on CN neuron was better than other stimulation paradigms.

Histology of SW-NIR in CN
Histological analysis of DCN was carried out after the experiment to verify the safety of SW-NIR. Results are shown in Fig. 10. The negative control group was given fake laser stimulation (laser off), the experimental group was given radiant exposure of 39.06 J∕cm 2 (pulse width: 5 ms). The irradiation position on the DCN and slice cutting orientation was shown in Fig. 10(a). Based on anatomical structure, we can find the edges of the CN [blue dash line in Fig. 10(a)], and the two edges of CN served as reference lines [also see red triangles in Fig. 10(d)]. After we found the reference lines, we stimulated CN along the middle line of CN [red dash line in Fig. 10(a)]. Then, we cut the tissue as 4-μm thick slices parallel with the black line of the Fig. 10(a) during the histology staining. Thus, the center between two edges of every slice is the stimulation site [red dash line in Fig. 10

Discussion
The present study demonstrated that SW-NIR with a wavelength of 980 nm could reversibly inhibit the CN neural activities induced by acoustical stimuli, and block the transmission of neural signals to the higher level of the central nervous system consequently. Under the laser parameters used in this experiment, the higher the laser pulse power, the more significant the inhibition effect. And the largest inhibitory effect on CN neurons could be observed when laser pulse was prior to acoustical stimulus, which indicates that the inhibition relies on the establishment time of the temperature field.

Optical Stimulation on Rat's Auditory System
The idea of using an auditory system to study the inhibitory effect of SW-NIR on the central nervous system in vivo was from our previous study on auditory brainstem stimulating method based on SW-NIR. 33 In that study, we tried to use 980-nm laser pulse to activate CN neurons in vivo, and we occasionally found that 980-nm laser pulse could inhibit CN neural activity and cause associated IC neural FR reduction. Second, the auditory system has a specific tonotopic organization; 45 auditory nerves from different regions of the cochlea transmit tonotopically encoded sound frequency information through associated midbrain structures and finally to the primary auditory cortex. In other words, the response of IC neurons depends on the input of the dorsal and ventral CN. 46 In the auditory system, each auditory pathway has its own CF, therefore, the corresponding neurons with the same CF are relatively easily found. Moreover, IC tonotopic maps are very clear in the most mammals and were widely used to evaluate auditory responses decades ago. 27,47,48 In addition to the special tonotopic organization of the auditory pathway described above, the purpose of this experiment was to investigate how SW-NIR regulates neural activities and nerve conduction in vivo. Thus, we used neural responses in IC to evaluate the effect of SW-NIR on CN neural suppression and nerve conduction.

Temperature Distribution of SW-NIR and NIR used in Neuromodulation
Compared with NIR, the SW-NIR has a much lower absorption coefficient and stronger scattering coefficient, 41 which is less favorable for infrared neuromodulations. And if we only consider the effectiveness of neuromodulation by laser, NIRs do have better behaviors than SW-NIRs. However, when we use laser pulse to modulate neural activities or nerve conduction in vivo, the penetrability of laser pulse is indispensable, especially in deep brain stimulation. 2 From our simulation results, we could see that the 980-nm laser has a temperature distribution in tissue wider than 1850 nm, and it could easily reach the deeper layer neurons in tissue. Recently, it has been reported that nanoparticles could effectively enhance the stimulation efficiency and spatial resolution of laser stimulation. 49 From our result, although it needs nearly 15 times more radiant energy for 980-nm laser than 1850-nm laser to achieve the same tissue baseline temperature increase, nanoparticles could be added into target tissue to enhance the photothermal interactions between laser and tissue, thereby reducing the laser energy and improving the targeting ability of the 980-nm laser.
Simulating results also showed that the temperature falls slowly after laser stimulation (Fig. 3), and it takes about 1.2 s for the temperature to return to the baseline (see Fig. S2 in the Supplementary Material), which exceeds the stimulation time interval (250 ms) in animal experiments and might cause temperature accumulation in tissue. However, from the existing experimental results (see Fig. 6 and Fig. S3 in the Supplementary Material), there is no obvious difference of the neural firing rate along the stimulating period, which indicates that even optical pulses may cause temperature accumulation in tissue; it did not affect the stimulation effect in this study.

Inhibitory Effect of Optical Stimulation Parameters on CN Neurons
The detailed mechanism of infrared neural modulation is far from clear, and the photothermal effect between laser and tissue is generally accepted as the basic mechanism. 6,16,19,20,32,50,51 Different from optically induced neural activity, the neural inhibition induced by laser pulse is based on the tissue baseline temperature rise. 16,19,20,32 Temperature changes in tissue could be controlled by adjusting laser parameters, including laser wavelength, pulse power, pulse width, repetition rate, and so on, 23,32 or by changing tissue's ability to absorb laser radiation, for instance by adding certain nanoparticles into target tissue. 16,19 This work was mainly focused on the effect of SW-NIR (wavelength of 980 nm) parameters on neural inhibition. Our results showed that with the increasing of laser radiant exposures from 6.51 to 39.06 J∕cm 2 (pulse width: 5 ms), the increment temperature of the CN increased from 1°C to 6.5°C [ Fig. 3(a)] and the neural inhibition rate increased from 25% to 70% [ Fig. 7(b)], which is similar to the results from Yoo et al. 16 In their studies, a continuous 785-nm infrared laser was used to modulate cultured hippocampal neurons, and temperature induced by laser pulse was enhanced by adding gold nanorods into neurons. Their results showed that when the temperature rose by 2°C to 8°C, the inhibition rate varied from 40% to 89.6%. Similar results were also found in other infrared neuromodulation experiments carried out in vitro. 19,20  With the increase of our laser pulse width (pulse power: 6 W), the inhibition rate increased from 54.2% to 72.5%. Although the change was slight, the inhibition became more stable with larger pulse width [ Fig. 7(c)], which may also be related to tissue baseline temperature induced by SW-NIR. For infrared neuromodulation, laser parameters are always limited by a maximum temperature that neural tissue can withstand, which is also known as the tissue ablation threshold. 52,53 Beyond this threshold, tissue will be thermally ablated. According to previous studies, this threshold is about 56°C. 52 Our simulation result showed that the longer the pulse duration, the higher the tissue temperature induced. When the optical parameter was set to 6-W pulse power and 14-ms pulse width and the corresponding radiant exposure was 109.37 J∕cm 2 , the CN temperature rose about 55°C, which is very close to the ablation threshold. To avoid nerve tissue damage, we only focused on the shorter pulse width of SW-NIR.

Effect of SW-NIR on Neural FSL
Studies showed that temporal codes based on FSL contain more information, which is important for the study of neural coding mechanisms. 47,54 In the auditory system, it has been suggested that FSL may represent auditory features, such as signal amplitude, frequency, and threshold, more accuracy than the neural FR. 43,44,55,56 And FSL was more sensitive to the most stimuli. 43,44 Here, we explored laser-induced FSL changes in IC neuron. We found that SW-NIR does affect neural FSL in our experiment. The neural FSL was longer when the laser pulse energy was larger [Figs. 8(c) and 8(d)]. Studies demonstrated that the change of neural FSL may relate to any factors that affects nerve conduction. 57 In this study, nerve conduction was modulated by temperature deposited in CN. With the increasing of laser pulse energy, more radiation energy was absorbed by CN neurons, followed by the tissue baseline temperature increasing, which inhibited the neural activity and nerve conduction, and finally caused the prolonged neural FSL. Meanwhile, we found that the neural FSL varied greatly from neurons with different CFs [ Fig. 8

Inhibitory Effect of Different Stimulation Paradigms on CN Neurons
Four stimulation paradigms were investigated in this study, and results showed that optical stimulus prior and synchronous to acoustical stimulus could effectively inhibit the acoustically induced neural activities of CN neurons. Optical stimulus delivered 5 ms prior to acoustic stimulus had the most significant inhibitory effect on CN neurons [Figs. 9(b) and 9(c)]. This is mainly related to the establishment time of tissue temperature field. Optically induced tissue temperature increase is not a kind of instantaneous action but a process. 32 The simulation result showed that the tissue temperature reached the highest level, 3.4 ms, after the end of the optical stimulus with the radiant exposure of 39.06 J∕cm 2 (pulse width: 5 ms) [ Fig. 3(c)]. The larger the pulse width, the longer the time difference [ Fig. 3(d)]. Duke et al. also recorded similar results from sciatic nerve of rat's peripheral pathway. 32 Under the first stimulation paradigm, tissue temperature reaches the highest level when CN neurons were about to be activated by acoustical stimuli, thus the inhibitory effect was significant. In other stimulation paradigms, neural tissue temperature was relatively low so that the inhibitory effect on CN neurons was relatively poor. This result further demonstrates that the level of tissue baseline temperature is the main factor for infrared neural inhibition and the neural inhibition relies on the establishment time of temperature field.

Conclusion
In this paper, we first introduced SW-NIR with the wavelength of 980 nm into neuromodulation in the central nervous system in vivo. Both stimulation parameters and stimulation paradigms of SW-NIR were investigated on the rat's central auditory system. The study showed that 980-nm laser pulses could reversibly and repeatedly inhibit acoustically induced neural activities in CN, and that the inhibition relies on the establishment time of the tissue temperature field. Moreover, our preliminary results suggest that short-wavelength infrared could regulate the activities of neurons beyond the neural tissues laser irradiated through neural networks and conduction in vivo. Based on the above experimental results, a type of neuromodulation tool with a transcutaneous stimulation mode is quite promising. This may be of great importance for improving the treatment of neurological disorders, such as epilepsy, tinnitus, neuropathic pain, and Parkinson's disease. Further work will focus on the study of precise transcutaneous stimulation methods in vivo using SW-NIR; temperature accumulation induced by laser pulse will be tested and the stimulus parameters will be readjusted in the follow-up experiments.

Disclosures
The authors have no additional relevant financial interests or potential conflicts of interest.