Enhancement of Motor Cortical Gamma Oscillations and Sniffing Activity by Medial Forebrain Bundle Stimulation Precedes Locomotion

Visual Abstract


Introduction
The medial forebrain bundle (MFB) is a neural fiber tract in rats and humans that connects and passes through various brain regions of the reward system, including the ventral tegmental area (VTA), nucleus accumbens, lateral/medial hypothalamus, sublenticular regions, lateral/medial preoptic regions, diagonal band, and septal area Zahm, 2006;Coenen et al., 2012;Gálvez et al., 2015). A principal component of the MFB is a mesolimbic pathway, a collection of fibers that ascend from dopaminergic neurons in the VTA and terminate in the nucleus accumbens and medial prefrontal cortex (Fenoy et al., 2022). Psychologically, the MFB is considered to serve as the neural substrate for motivation and pleasure, and thus, stimulation of the MFB and surrounding regions has been behaviorally used as a neural and "virtual" reward (Olds and Milner, 1954;Margules and Olds, 1962;Beninger et al., 1977). MFB stimulation ignites hedonic feelings and elicits pleasant bodily sensations in animals, thus highly motivating them to perform a variety of operant and spatial tasks (Carlezon and Chartoff, 2007;Lee et al., 2010;Sun et al., 2012;Farakhor et al., 2019;Kong et al., 2019). Electrical stimulation of the reward system, including the MFB, has also allowed for (tele)control of the spatial navigation of rodents and birds (Talwar et al., 2002;Sun et al., 2012;Huai et al., 2016;Khajei et al., 2019).
Although anatomic and behavioral evidence has led to the assumption that central and peripheral activity is modified by MFB stimulation, little is known about how this activity is indeed affected by MFB stimulation. As both operant conditioning and spatial navigation are accompanied by motor execution, we hypothesized that neural activity in the primary motor cortex (M1) would be modulated by MFB stimulation, which is supported by a previous immunohistochemical study (Hosp et al., 2011). Moreover, because locomotion and breathing have been postulated to be closely correlated, a phenomenon called locomotor-respiratory coupling (Bramble and Carrier, 1983;Potts et al., 2005), we also examined respiratory activity as an index of peripheral function.
To this end, we set out to chronically implant recording electrodes into the M1, primary somatosensory cortex (S1), and olfactory bulb (OB) and insert a stimulation electrode into the MFB (i.e., MFB group) or a neighboring region (i.e., sham group) of rats. We then simultaneously recorded electrocorticograms (ECoGs) in the M1, S1, and OB of freely moving rats while the MFB was repeatedly and regularly stimulated.

Ethical approval
Animal experiments were performed with the approval of the Animal Experiment Ethics Committee at the University of Tokyo (approval number P29-7) and according to the University of Tokyo guidelines for the care and use of laboratory animals. These experimental protocols were conducted in accordance with the Fundamental Guidelines for the Proper Conduct of Animal Experiments and Related Activities in Academic Research Institutions (Ministry of Education, Culture, Sports, Science and Technology, Notice No. 71 of 2006), the Standards for Breeding and Housing of and Pain Alleviation for Experimental Animals (Ministry of the Environment, Notice No. 88 of 2006) and the Guidelines on the Method of Animal Disposal (Prime Minister's Office, Notice No. 40 of 1995). All efforts were made to minimize animal suffering.

Animals
A total of 20 male 8-to 10-week-old Wistar rats (Japan SLC) with a preoperative weight of 180-300 g were individually housed under conditions of controlled temperature and humidity (22 6 1°C; 55 6 5%) and maintained on a 12/12 h light/dark cycle (lights off from 7 A.M. to 7 P.M.) with ad libitum access to food and water. Rats were habituated to an experimenter via daily handling before experiments were conducted.

Electrodes
A recording interface assembly was prepared as previously described (Okada et al., 2016;Sasaki et al., 2017;Shikano et al., 2018;Yoshimoto et al., 2021b). In short, the assembly was composed of an electrical interface board (EIB; EIB-36-PTB, Neuralynx) and custom-made shell and core bodies created by three-dimensional (3-D) printers. The EIB had a sequence of metal holes for connections with wire electrodes. A particular individual hole was conductively connected with one end of the insulated wire (;5 cm) using attachment pins, whereas the opposite end was soldered to a corresponding individual electrode during surgery. Bipolar stimulating electrodes were made from pairs of stainless-steel insulated wires (TOG217-049c, Unique Medical). The distal end of the stimulation electrode was soldered to a two-pin connector protected by epoxy glue to prepare a stimulating electrode assembly (Shibata et al., 2022).

Surgery
General anesthesia in the rats was induced and maintained with 2-3% and 1-2% isoflurane gas, respectively, with careful inspection of the animal's condition during the whole surgical procedure. Veterinary ointment was applied to the rat's eyes to prevent drying. The skin was sterilized with povidone iodine and 70% ethanol whenever we made an incision.
After anesthesia, electrodes for electromyograms (EMGs) were implanted as previously described (Yoshimoto et al., 2021a). Briefly, a rat was mounted onto a stereotaxic apparatus (SR-6R-HT, Narishige). One wire electrode (AS633, Cooner Wire) was implanted into the trapezius to record EMGs. The scalp was then removed with a surgical knife. A circular craniotomy with a diameter of ;0.9 mm was performed using a high-speed dental drill. Epidural stainlesssteel screws (1.4 mm in diameter, 3 mm in length) were used to record electrocorticograms (ECoGs) from S1 and M1, whereas a smaller screw electrode (1.0 mm in diameter, 4 mm in length) was used to record ECoGs from the OB. The three screw electrodes were stereotaxically implanted into the S1 (2.1 mm posterior and 2.8 mm lateral to bregma), M1 (3.2 mm anterior and 3.0 mm lateral to bregma), and OB (10.0 mm anterior and 1.0 mm lateral to bregma; Yamashiro et al., 2020). In addition, another two stainless-steel screws were implanted into the bone above the cerebellum (9.6 mm posterior and 1.0 mm bilateral to bregma) to serve as ground and reference electrodes. Each of the open edges of the electrodes was soldered to the corresponding open edge of the insulated wires of the recording interface assembly. The bipolar stimulation electrodes (described in the previous section) were stereotaxically implanted unilaterally into the MFB (2.0 mm posterior and 2.0 mm lateral to bregma, and 7.8 mm below the cortical surface; "MFB group" or "MFB-novel group") or other regions (2.0 mm posterior and 2.0 mm lateral to bregma, and 5.0 mm below the cortical surface; "sham group"). The electrodes were secured to the skull using dental cement. Immediately after implantation, the rat was scanned, and 3-D images were reconstructed by an X-ray microcomputed tomography system (CosmoScan GXII, Rigaku). The parameters for X-ray tomography were as follows: tube voltage, 90 kV; tube current, 88 mA; absorbed dose, 106 mGy; field of view (FOV), 45 mm; voxel size, 90 mm (isotropic); and scan time, 2 min. The electrode placement was roughly located using the reconstructed images (Fig. 1B). Rats that did not have the stimulation electrode implanted in the target region were not tested in the following experiments.
Following surgery, each rat was allowed to recover from anesthesia and was individually housed with free access to water and food. For the first 2 d after surgery, the condition of the animals was carefully checked every 3 h except during the night (i.e., 8 P.M. to 8 A.M.). The animals were rehabituated to the experimenter by handling.
While our experimental protocols mandate the humane killing of animals if they exhibit any signs of pain, prominent lethargy, or discomfort, we did not observe such symptoms in any of the 20 rats used in this study.

Apparatus
An operant chamber (OP-3501, O'hara) with two nosepoke holes (20 mm in diameter) in a soundproof box was used for behavioral tests (described below). The box measured 40 cm in width, 30 cm in depth, and 40 cm in height. Nose pokes were detected by a photoelectric sensor in a hole and recorded using Arduino; note that only one "active" hole was connected to the sensor, whereas the other was not. During electrophysiological experiments, the nose-poke holes were closed (described below).

Behavioral test
After full recovery from surgery, the animals were habituated to the apparatus for at least 2 d. Following familiarization with the apparatus, rats in the MFB and sham groups performed a nose-poke test for 3 d. In contrast, rats in the MFB-novel group never performed the nosepoke test or underwent familiarization with the apparatus before electrophysiological recordings (also see the next section). The stimulating electrode assembly was attached to a two-core cable. The cable was further connected to an isolator (A365, World Precision Instruments; WPI) and a stimulator (A310, WPI).
Rectangular symmetrical biphasic electric currents were generated by the stimulator. Parameters for the electric currents were as follows: amplitude (for each of the positive and negative phases), 180-300 mA; phase duration (for each phase), 1.0 ms; interphase interval, 0 s; interpulse interval (time between onsets of a positive phase and the next), 100 ms (i.e., pulse frequency, 10 Hz); and burst duration, 500 ms. These stimulations were delivered to the MFB of the rats every 5 s for 3 min.

In vivo electrophysiology
Two days after the behavioral test, rats in the MFB and sham groups underwent electrophysiological recordings. Each rat in the MFB, sham, and MFB-novel groups was allowed to freely explore the operant chamber with its nose-poke holes shut for electrophysiological recordings; note that rats in the MFB-novel group had not been exposed to the operant chamber before the electrophysiological experiments were performed. The stimulating electrode assembly was attached to a two-core cable and connected to an isolator and a stimulator as described in the previous section.
The EIB of the recording interface assembly was connected to a digital headstage (CerePlex M, Blackrock Microsystems), and the digitized signals were amplified and transferred to a data acquisition system (CerePlex Direct, Blackrock Microsystems; Okada et al., 2016;Kuga et al., 2019). ECoG signals were digitized at a sampling rate of 2 kHz.
On the recording day, ECoGs were first recorded for 3 min without any electrical stimulation (i.e., "baseline session"). After the baseline session, ECoGs were recorded for 3 min again, while electrical stimulation was delivered every five seconds (i.e., "stim session"). The parameters of the electric currents were the same as those used in the behavioral test. For analysis, the stim session was split into "prestimulation," "poststimulation," and other periods (described below).

Histology
After the recordings, the rats were anesthetized with an overdose of isoflurane gas and transcardially perfused with 0.01 M PBS (pH 7.4) and 4% paraformaldehyde (PFA) in 0.01 M PBS, followed by decapitation. The brains were soaked overnight in 4% PFA for postfixation and coronally sectioned at a thickness of 100 mm using a vibratome (DTK-1000N, Dosaka EM). Serial slices were mounted on glass slides and processed for cresyl violet staining. To achieve cresyl violet staining, the slices were rinsed in water, ethanol, and xylene; counterstained with cresyl violet; and coverslipped with a mounting agent. The positions of all electrodes were confirmed by identifying dents on the neocortical superficial layer or tracks in the subcortical region in the histologic tissue. Data were excluded from the subsequent analysis if the electrode position was outside the target brain region. Cresyl violet-stained images were acquired using a phase-contrast microscope (BZ-X710, Keyence).

Data analysis
All data analyses were performed using custom-made MATLAB routines (MathWorks). The summarized values are reported as mean 6 SEM. The significance level was set at 0.05, and the null hypothesis was statistically rejected when p , 0.05, unless otherwise specified. For comparison of the power of rhythmic activity in a specific frequency range (see below), common logarithms of the power were taken based on the concept of decibels in the electrophysiological field (Ray et al., 2013;Nakazono et al., 2019;Dubey and Ray, 2020;Reddy et al., 2021); more specifically, the subtraction of logarithms of given raw values practically equals the division of the raw values. Before pairwise comparisons were performed, normality of the sample dataset (calculated by subtraction between corresponding two values) was evaluated by the Shapiro-Wilk test, which tests the null hypothesis that the dataset is drawn from a normally distributed population (Shapiro and Wilk, 1965). If the null hypothesis was rejected, nonparametric tests (i.e., the Wilcoxon signed-rank test) were used for the pairwise comparisons; otherwise, parametric tests (i.e., the paired t test) Figure 1. Rats learn to perform a nose-poke test with MFB stimulation. A, A diagram of the experimental setup for the rat nosepoke test. B, A 3D skeletal reconstruction of a rat in which an electrode had been implanted. C, Left, Representative Nissl-stained section image showing the track of the stimulation electrode into the MFB (gray). The tip of the electrode is indicated by a red circle in the simplified brain atlas. Right, The same as left, but for the electrode into a region (indigo) outside of the MFB (gray). D, Left, The number of nose-pokes increased daily in rats in the MFB group (red). Right, The same as left, but for rats in the sham group (indigo). E, The intervals of nose-pokes decreased daily in rats in the MFB group (red). The p and t values were obtained by paired t tests (n = 8 and 6 rats in the MFB and sham groups, respectively). MFB, medial forebrain bundle.
were performed. When multiple pairwise comparisons were required, the significance level was adjusted in accordance with the Bonferroni correction ( Fig. 1D,E). The effect size was evaluated by Cohen's d to find the most effective parameters as needed (Cohen, 1988;Kline, 2004). Sample sizes were not predetermined using statistical methods, but the sample sizes used here were similar to those reported in the field for similar electrophysiological experiments (Konno et al., 2021;Yoshimoto et al., 2021b).
The rats' behavior was monitored using a web camera operating at 30 fps throughout the experiment. The frame rate of the video was then downsampled to 6 fps. The downsampled data were used to manually mark the rats' moment-to-moment positions and to track the paths using ImageJ software (National Institutes of Health). The paths traveled by rats were quantified based on the x and y coordinates of the rats' heads ( Fig. 2B-G).
The "prestimulation" and "poststimulation" periods were defined as 2 s before and after each stimulation, respectively. To better clarify the precise neural oscillatory activity during both periods, the ECoG signals were further convoluted with a complex Morlet wavelet family (bandwidth parameter, 1.5; center frequency, 2; Figs. 6, 7).

Code accessibility
Custom-made MATLAB codes for computational analyses are available (Extended Data). To run the codes, Windows was used as the operating system throughout this study. Thus, Windows is recommended to run the MATLAB codes; however, the codes would also run well on either macOS or Linux systems. Both nose-poke holes were shut. C, The distance traveled by rats in the MFB group during the baseline (gray) and stim (red) sessions. D, The same as B, but for rats in the sham group during the baseline (left, gray) and stim (right, indigo) sessions. E, The same as C, but for rats in the sham group. F, The same as B, but for rats in the MFB-novel group during the baseline (left, gray) and stim (right, green) sessions. G, The same as C, but for rats in the MFB-novel group. The p and t values were obtained by paired t tests (n = 8, 6, and 6 rats in the MFB, sham, and MFB-novel groups, respectively). MFB, medial forebrain bundle.
After the rats were fully habituated to an open field ( Fig.  2A), we allowed them to freely explore the field and simultaneously recorded ECoGs in the S1, M1, and OB for rats in the MFB, sham, and MFB-novel groups during the baseline and stim sessions (Fig. 3A,B); note that rats in the MFB-novel group had not experienced the apparatus or performed the nose-poke test. During the stim session, electrical stimulation was delivered to rats in both groups at regular intervals (5 s). We tracked and quantified the animal trajectories during both sessions ( Fig. 2B-E). In the sham group, a rat was likely to prefer a certain location in an open field during both sessions (Fig. 2D), suggesting the existence of the rat's home base (Eilam and Golani, 1989). In contrast, in rats in the MFB group, such a home base disappeared during the stim session compared with the baseline session (Fig. 2B). The total distance traveled by rats in the MFB group was significantly longer during the stim session than during the baseline session [0.95 6 0.15 m/min (baseline) vs 4.89 6 0.52 m/min (stim), p = 2.21 Â 10 À4 , t (7) = 6.95, n = 8 rats, paired t test; Fig. 2C], whereas the total distance traveled by rats in the sham group was not significantly different between the baseline and stim sessions [1.58 6 0.42 m/min (baseline) vs 1.21 6 0.22 m/min (stim), p = 0.37, t (5) = 0.97, n = 6 rats, paired t test; Fig. 2E].
To rule out the possibility that the preceding nose-poke performance had an impact on the subsequent home-base behavior and locomotion, we allowed rats in the MFB-novel group to freely explore the open field without performing any nose-poke pretest. In rats in the MFB-novel group, a home base was evident during the baseline session (Fig. 2F). Indeed, the home base was still present during the stim session, but compared with the baseline session, the rats visited places other than the original home base more frequently (Fig. 2F) and walked a longer distance [1.12 6 0.32 m/min (baseline) vs 5.00 6 0.57 m/min (stim), p = 2.71 Â 10 À3 , t (5) = 5.50, n = 6 rats, paired t test; Fig. 2G]. These results suggest that MFB stimulation enhances locomotor activity regardless of the preceding nose-poke behavior.
Altogether, these results suggest that repeated MFB stimulation acutely affected the high-frequency sniffing component in the OB ECoGs and the gamma oscillations in the M1. . MFB stimulation does not have any effect on S1 ECoG signals. A, Power of S1 ECoG signals bandpass-filtered within 0.3-4 Hz (delta, left), 4-8 Hz (theta, middle), and 30-90 Hz (gamma, right) during the baseline (black) and stim (red) sessions for rats in the MFB group. B, The same as A, but for rats in the sham group. The p and t values were obtained by paired t tests (n = 8 and 6 rats in the MFB and sham groups, respectively). MFB, medial forebrain bundle; ECoG, electrocorticogram; S1, primary somatosensory cortex.

MFB stimulation induces sniffing and facilitates motor cortical gamma oscillations preceding locomotion
Since MFB stimulation facilitated gamma oscillations in the M1, induced locomotion and provoked highfrequency sniffing activity, we investigated the temporal relationship among these phenomena. For the MFB group, we bandpass-filtered the original ECoGs in the OB between 4 and 9 Hz (Fig. 6A,B) and estimated the timevarying high-frequency sniffing components during the poststimulation period based on spectral analysis (Fig.  6C). The sniffing component peaked immediately after the stimulation terminated and declined following that peak.
We then investigated how MFB stimulation modulated motor cortical neural activity because we found enhanced gamma power in the M1 ECoGs but not in the S1 ECoGs (Figs. 3-5). We bandpass-filtered the M1 ECoGs within three frequency ranges (i.e., delta, theta, and gamma) and convoluted the raw signals using a complex Morlet wavelet family (Fig. 6D), which demonstrated sustained enhancement of gamma power in the M1 after MFB stimulation (Fig. 6D,E). In the same manner as the sniffing components (Fig. 6C), we calculated the time-varying changes in the gamma component in the M1 ECoGs (Fig.  6F). We further quantified locomotor activity during the prestimulation and poststimulation periods (Fig. 6G,H) and revealed that the gamma power peaked before the locomotor activity reached the maximum level ( Fig. 6E-H).
We calculated a time lag based on the time when the sniffing, gamma power, and locomotor activity reached the maximum. For the MFB group, the time lag from sniffing activity to gamma power enhancement was significantly larger than 0 s (0.30 6 0.08 s, p = 1.98 Â 10 À3 , t (7) = 4.80, n = 8 rats, one-sample t test vs 0 s; Fig. 6I), whereas the time lag from sniffing to gamma was not significantly different from 0 s for the MFB-novel group (0.01 6 0.17 s, p = 0.97, t (5) = 0.04, n = 6 rats, one-sample t test vs 0 s; Fig. 6J). These results suggest that the facilitation of sniffing activity and the enhancement of gamma oscillatory activity in the M1 partially overlap during the period following MFB stimulation; the sniffing facilitation does not necessarily precede the motor cortical gamma enhancement. In contrast, the time lag from the gamma enhancement to the peak of locomotor activity was significantly larger than 0 s for the MFB group (0.49 6 0.06 s, p = 1.98 -Â 10 À3 , t (7) = 4.80, n = 8 rats, one-sample t test vs 0 s; Fig.  6I) and the MFB-novel group (0.67 6 0.16 s, p = 8.18 -Â 10 À3 , t (5) = 4.24, n = 6 rats, one-sample t test vs 0 s; Fig.  6J). Moreover, the time lag from the sniffing to the locomotion was also significantly larger than 0 s for the MFB group (0.85 6 0.11 s, p = 9.91 Â 10 À5 , t (7) = 7.90, n = 8 rats, one-sample t test vs 0 s; Fig. 6I) and the MFB-novel group Figure 5. MFB stimulation in a novel environment enhances sniffing components in the OB ECoGs and gamma power in the M1 ECoG signals. A, Power of OB ECoG signals bandpass-filtered within 1-4 Hz (low-frequency sniffing, left), 4-9 Hz (high-frequency sniffing, middle), and 30-90 Hz (gamma, right) during the baseline (black) and stim (green) sessions for rats in the MFB-novel group. Note that rats in the MFB-novel group did not perform the nose-poke test or undergo familiarization with the apparatus before electrophysiology. B, Power of M1 ECoG signals bandpass-filtered within 0.3-4 Hz (delta, left), 4-8 Hz (theta, middle), and 30-90 Hz (gamma, right) during the baseline (black) and stim (green) sessions for rats in the MFB-novel group. C, The same as B, but for the S1 ECoG signals. The p and t values were obtained by paired t tests or Wilcoxon signed-rank tests (n = 6 rats in the MFB-novel group). MFB, medial forebrain bundle; ECoG, electrocorticogram; OB, olfactory bulb; M1, primary motor cortex; S1, primary somatosensory cortex; LF, low-frequency; HF, high-frequency. Figure 6. MFB stimulation facilitates sniffing activity and gamma power in the M1, and subsequently induces locomotion. A, Representative raw (top; first) and bandpass-filtered [low-frequency sniffing (1-4 Hz), second; high-frequency sniffing (4-9 Hz), third; gamma (30-90 Hz), fourth] traces of ECoG signals in the OB during the prestimulation and poststimulation periods in the MFB group. The raw trace was convoluted with a Morlet wavelet family and transformed into pseudocolored matrices in the time-frequency domain (fifth). B, Average (black) of the OB ECoG traces bandpass-filtered within 4-9 Hz during the prestimulation and poststimulation periods for rats in the MFB group, superimposed on the traces from each rat (gray). Note that the OB ECoG traces bandpass-filtered within 4-9 Hz indicate high-frequency sniffing activity. C, The raw trace of the OB ECoG signals during the poststimulation period was convoluted with a Morlet wavelet family and transformed into pseudocolored matrices in the time-frequency domain (bottom), yielding the power of the high-frequency sniffing activity based on the wavelet coefficients (middle). The mean (red) and mean 6 2 Â SD (pink) of the power during the prestimulation period are shown as thresholds; any suprathreshold values during the poststimulation period are significantly higher than values during the prestimulation period. Note that each of the power trace (middle) and pseudocolored spectrogram (bottom) is not those made from an averaged trace but an average of those made from individual traces. The time that gives the peak power is shown for all rats in the MFB group (top). D, Representative raw [top (first)] and bandpass-filtered [delta (0.3-4 Hz), second; theta (4-8 Hz), third; gamma (30-90 Hz), fourth] traces of ECoG signals in the M1 during the prestimulation and poststimulation periods for rats in the MFB group. The raw trace was convoluted with a Morlet wavelet family and transformed into pseudocolored matrices in the time-frequency domain (fifth). E, Average (black) of the M1 ECoG traces bandpass-filtered in a gamma (30-90 Hz) frequency band during the prestimulation and poststimulation periods for rats in the MFB group, superimposed on the traces from each rat (gray). F, The same as C, but for the gamma power in the M1 ECoG signals. G, The average distance traveled (black) by rats in the MFB group during the prestimulation and poststimulation periods, superimposed on the traces from each rat (gray). H, The expanded trace of the average distance traveled (shown in G) during the poststimulation period (bottom). The time that gives the highest locomotor activity is shown for all rats in the MFB group (top). I, Using the time with the largest values (in C, F, H), the time lag was calculated for all pairs (i.e., "sniffing to gamma," "gamma to locomotion," and "sniffing to locomotion") in the MFB group. J, The same as I, but for the MFB-novel group. The p and t values were obtained by paired t tests (n = 8 and 6 rats in the MFB and MFB-novel groups, respectively). MFB, medial forebrain bundle; ECoG, electrocorticogram; OB, olfactory bulb; M1, primary motor cortex; LF, low-frequency; HF, high-frequency.

Discussion
In this study, we found that electrical stimulation of the rat MFB increased exploratory behavior, sniffing activity, and extracellular gamma oscillatory power in the M1. Moreover, the time-series analysis confirmed that MFB stimulation enhanced sniffing activity and gamma power in the M1, and subsequently induced locomotion.
As a neural reward, MFB stimulation motivates animals so powerfully that their behaviors are dynamically modified (Talwar et al., 2002). Rats learned to exhibit instrumental (e.g., nose-poking and lever-pressing) behavior faithfully and quickly (Fig. 1D). Additionally, the locomotor activity of the rats in the MFB group was enhanced (Fig. 2B,C); this activity may be mediated by dopamine D1 receptors (Tran et al., 2005). Moreover, it is well known that rats alternately run and stop when they are placed in an environment, but they are likely to stop at one or two specific places, defined as their home base (Eilam and Golani, 1989); home bases can be modulated to some extent by salient stimuli and environmental geometry (Thompson et al., 2018). Here, the trajectory of a rat in the sham group during the stim session confirmed that the rat frequently crossed a specific location (Fig. 2D), which can be regarded as the rat's home base. Home bases were also observed in rats in the sham, MFB, and MFB-novel groups during the baseline session (Fig. 2B,D,F). However, rats in the MFB and MFB-novel groups during the stim session explored not only around their specific home bases but also around every corner and beside every wall in the open field, suggesting that acute MFB stimulation diminished home base behavior. We assumed that reward-seeking responses evoked by the preceding MFB stimulation resulted in the disappearance of the home base behavior (Margules and Olds, 1962;Wise, 2005), regardless of preexposure to the conditioning apparatus with nose-poke holes.
We also scrutinized how this behavioral modification was associated with neural activity in the OB and M1 (Figs. 3,5). Regarding OB activity, high-frequency sniffing is often observed when animals are motivated to explore an external environment (Wesson et al., 2008;Kuga et al., 2019) and may play a role in the acquisition of olfactory information to guide their ongoing behavior (Kepecs et al., 2006;Ranade et al., 2013;Kleinfeld et al., 2016). Consistent with a previous study on sniffing responses based on thermal changes in the rat nasal cavity (Waranch and Terman, 1975), we observed intense high-frequency sniffing activity immediately after MFB stimulation (Fig. 6A-C). Since there were no odor cues in our experimental setup as a matter of course, we considered that rats were driven to "virtually" incorporate sensory information into themselves to search for the origin of rewards; this approach contributed to reward-seeking behavior and a gradual increase in subsequent locomotion (Fig. 6G,H). Moreover, the sniffing activity of mice is increased in anticipation of future reward delivery (Wesson et al., 2008). Thus, enhanced sniffing activity preceding locomotion appears to signify reward-seeking and reward-anticipating behavior.
In addition to sniffing activity, we found that MFB stimulation enhanced the power of delta, theta, and gamma oscillations in M1 ECoGs of well-trained rats in the MFB group (Fig. 3); however, we should also hasten to add that the only gamma power was increased in completely naive (i.e., preexposure-free) rats in the MFB-novel group (Fig.  5). Regarding the mechanism underlying the MFB stimulation-induced enhancement of M1 gamma oscillations, we considered neural projections to the M1 via the MFB, although we cannot completely exclude the possibility that sniffing activity directly affected M1 ECoGs. The MFB is considered to connect several brain areas, including the VTA, lateral and medial hypothalamus, and ventral striatum (Gálvez et al., 2015). Among these brain areas, dopaminergic neurons in the VTA project to the M1 in rats (Lindvall et al., 1978;Luft and Schwarz, 2009;Hosp et al., 2011) and humans (Hosp et al., 2019). Intrinsic properties and synaptic transmission of M1 parvalbumin-positive interneurons are modulated by dopaminergic signals via dopamine D2 receptors (Cousineau et al., 2020;Duan et al., 2020); of note, VTA neurons innervate parvalbuminpositive interneurons in the M1 (Duan et al., 2020). Fast-spiking activity of parvalbumin-positive neurons is believed to produce gamma oscillations via synchronized inhibitory synaptic currents in cortical pyramidal cells (Buzsáki and Wang, 2012). Thus, we speculate that the MFB stimulation-induced enhancement of M1 gamma oscillations is mediated by dopaminergic signals sent from the VTA to the M1.
Despite the MFB stimulation-induced enhancement of M1 gamma oscillations, it is surprising, to some extent, that we did not observe either enhancement or impairment of oscillatory power in the S1 ECoGs in rats in the MFB or MFB-novel group because sniffing and whisking activities are tightly coupled with each other and both activities are involved in reward-seeking and reward-anticipating behaviors. Although we did not provide experimental proof, we speculate that the possible mechanism underlying the lack of an effect of MFB stimulation on the S1 ECoGs is related to the release of acetylcholine. Previous studies demonstrated that higher concentrations of acetylcholine are released in the S1 than the M1 in the nocturnal phase (Jiménez-Capdeville and Dykes, 1996) and that cholinergic neuronal activity is associated with desynchronized extracellular oscillations (Blake and Boccia, 2016). Therefore, compared with the M1, more desynchronization of neural activity in the S1 may have brought about more variability in the oscillatory change and overwhelmed sniffing/whisking-induced neural activity.
Compared with the completely awake animals used here, a previous study measured single-cell unit activity in the thalamus and brainstem and electroencephalograms (EEGs) in the frontal and occipital cortices of anesthetized Figure 7. Effects of MFB stimulation on sniffing, M1 gamma power, and locomotion are not different between the first and last halves of the recording sessions. A, For the first half of the whole recording sessions, the raw trace of the OB ECoG signals of rats in the MFB group during the poststimulation period was convoluted with a Morlet wavelet family and transformed into pseudocolored matrices in the time-frequency domain (bottom), yielding the power of the high-frequency sniffing activity based on the wavelet coefficients (middle). The mean (red) and mean 6 2 Â SD (pink) of the power during the prestimulation period are shown as thresholds; any suprathreshold values during the poststimulation period are significantly higher than values during the prestimulation period. Note that neither the power trace (middle) nor the pseudocolored spectrogram (bottom) is made from an averaged trace but an average of those made from individual traces. The time that gives the peak power is shown for all rats in the MFB group (top). B, The same as A, but for the last half of the whole recording sessions. C, The same as A, but for the gamma power in the M1 ECoG signals. D, The same as C, but for the last half of the whole recording sessions. E, The average distance traveled by rats in the MFB group during the poststimulation period of the first half of the whole recording sessions (bottom). The time that gives the highest locomotor activity is shown for all rats in the MFB group (top). F, The same as E, but for the last half of the whole recording sessions. G, Time lags from high-frequency sniffing to gamma enhancement in the M1 are not significantly different between the first and last halves of the recording sessions, but both lags are significantly larger than 0 s. Note that the positive value of the lag indicates that high-frequency sniffing precedes gamma enhancement in the M1. H, The same as G, but for the time lags from gamma enhancement to locomotor activity. I, The same as G, but for the time lags from high-frequency sniffing to locomotor activity. The p and t values were obtained by paired t tests (n = 8 rats in the MFB group). HF, high-frequency; M1, primary motor cortex. rats simultaneously with MFB stimulation (Rolls, 1971). This study indicated that, based on desynchronization of the cortical EEGs, the anesthetized rats were forced to be somewhat awake when receiving MFB stimulation at least at the firing and oscillatory activity levels (Rolls, 1971). The observation of attenuated EEG signals appears to contradict our current findings that the multiple oscillatory (i.e., delta, theta, and gamma) powers in the M1 ECoGs of the MFB group were enhanced by MFB stimulation (Fig.  3E). However, we assume that this contradiction originates from the fact that MFB stimulation excited only a subpopulation of neurons in anesthetized rats (Rolls, 1971). This previous study divided the MFB stimulationresponsive firing units into three types: antidromically driven (i.e., directly excited) brainstem units, monosynaptically driven brainstem units, and multisynaptically driven units in the brainstem and thalamus (Rolls, 1971). Importantly, MFB stimulation was considered to antidromically excite brainstem neurons and further excite neurons downstream of the "antidromically excited" neurons via synapses. The antidromically driven units did not exhibit firing rates that correlated with the real awake state, whereas some of the monosynaptically driven units and all multisynaptically units had firing rates that resembled firing rates under arousal (Rolls, 1971). In this sense, the completely waking state in this study and the MFB stimulation-induced pseudoarousal under anesthesia by urethane and equithesin are totally different (Rolls, 1971;Batzri-Izraeli et al., 1992). Hence, MFB stimulationinduced enhancement of a wide range of the power of the M1 ECoGs (of rats in the MFB group) is an awake statespecific phenomenon.
Although we demonstrated that the extracellular gamma oscillations in the M1 were facilitated by MFB stimulation, how MFB stimulation affects M1 neural activity at the synaptic level and contributes to behavioral functions remains to be fully elucidated. In this light, the previous histologic evidence provides insights that could address the question. Expression of c-Fos protein, an immediate early gene (i.e., c-fos) product, is induced in the M1 by dopamine release on electrical stimulation in the VTA (Hosp et al., 2011). Dopamine is also involved in long-term synaptic plasticity in the M1 (Rioult-Pedotti et al., 2015). These studies suggested that synaptic plasticity in the M1 may be induced when rewarding dopaminergic signals are sent from the VTA to the M1. Motor learning is accompanied by synaptic plasticity in the M1 (Rioult-Pedotti et al., 1998, 2000; thus, dopaminergic signals should contribute to motor learning (Molina-Luna et al., 2009).
In addition to the possible synaptic plasticity in the M1 induced by MFB stimulation, we assume that motor cortical gamma oscillations potentially impact learning. Indeed, we have not empirically demonstrated whether or how motor cortical gamma oscillations induced by MFB stimulation serve to promote motor learning; however, a previous study using rats showed that gamma oscillations in the M1 were dominant during a leverpressing task associated with rewards (Igarashi et al., 2013). A recent study suggested that gamma oscillations in the rat M1 regulate motor learning (Otsuka and Kawaguchi, 2021). Taken together, it is plausible that dopaminergic signals elicited by MFB stimulation facilitate motor learning via gamma oscillations and synaptic plasticity in the M1; this relation could be further elucidated by behavioral electrophysiology with an operant task.