**4. Discussion**

In this study, we investigated how nutritional states modulate the neural activity and neural representation of odors in the M/Ts of awake, head-fixed mice. Our data indicate that, in a fasted state, both the spontaneous firing rate of M/Ts and the number of cells showing an excitatory response to odors are increased at the single-cell level, suggesting that the excitability of neural activity is enhanced. Unexpectedly, the ability of M/Ts to discriminate different odors is slightly decreased in the fasted state, indicating that the effect of different nutritional states on olfaction is complex and that higher brain centers beyond the OB are likely involved.

Many previous studies have investigated behavioral odor detection and discrimination under different nutritional states in humans and rodents [2,5]. Fasting results in an increased perception of some food-related odors in humans [33] and olfactory sensitivity to a neutral odor increases in fasted rats [7]. The neural mechanisms underlying these behavioral observations are not clear, even though many studies have focused on this issue and a lot of data have been collected [16,17,21,34–37]. Since the OB is the first olfactory center and plays critical roles in odor information processing, most of the studies of different nutritional states examined neural activity in the OB [2,16–18,34,37]. In the 1970s, Pager found that the electrical activity of mitral cells was dependent on nutritional status: fasting selectively increased mitral cell multiunit responses to food odor [16]. This finding was partially supported by another study in which mitral cell single-unit responses to odor increased with fasting, regardless of the odorant [17]. Our results are consistent with the latter since we found that both the number of M/Ts showing an excitatory response and the amplitude of the response were greater in a fasted state, and that these increments were independent of odor type. In addition, we found that the baseline firing rate of single M/Ts was significantly increased in a fasted state. Therefore, the excitability of M/T neural activity is enhanced in the OB of fasted mice.

While the spikes from single unit in the OB have the capacity to encode odor identity in awake behaving rodents [13,22], LFP recorded form the OB reflect temporally coordinated neuronal ensembles and provides a reference for spike timing-based codes [38]. In the OB, the spikes from M/Ts are highly correlated with high-gamma oscillation and their correlation carries important information on odor identity [22,39]. Thus, it is not surprising that the increase in excitatory neural activity in the OB in a fasted state is further supported by our LFP recordings. Although the ongoing baseline LFP activity did not change from the satiated to the fasted state, the amplitude of the odor-evoked gamma oscillations was significantly decreased, indicating an increase in an excitatory component of the LFP or a decrease in an inhibitory component during odor stimulation in a fasted state. Since gamma oscillations in the OB are generated by the an interaction between M/Ts and granule cells [39,40], and since the firing rate of M/Ts is greater in the fasted state, the decrease in the odor response in the fasted state is likely due to the increased excitability of the M/Ts. Whether the firing of granule cells changes with fasting is an interesting question; cell-type-specific recording of the spikes from granule cells will be needed to address this, using, for example, the juxtacellular "loose-patch" recording method [41].

Although previous study performed in free-moving rats found that odor-evoked beta response was affected by fasting [34], we did not find significant changes after fasting during passive odor stimulation in present study. However, it is likely that fasting would change odor-evoked beta oscillation if the mice detect the odors actively, e.g., during the odor discrimination task, since beta oscillation is critically involved in the learning process [39,42]. Previous studies have demonstrated that beta oscillation is generated by the interaction between the OB and piriform cortex [43], and the centrifugal input from the piriform cortex to the OB should be critical for the beta oscillation. Strikingly, CB1 receptors are expressed in these centrifugal fibers and they play important roles in food intake [1]. Thus, odor-evoked beta oscillation is likely modulated by fasting during active rather than passive odor sampling.

Another important finding in our study is that the neural discrimination of odors by M/Ts is slightly decreased in the fasted state. Interestingly, for individual odor pairs, a significant decrease in odor discrimination was found only for food odor and 2,4,5-trimethylthiazole (Figure 3). Since behaviorally, odor discrimination is better during a fasted state [7,33], this finding was unexpected. One possible explanation is that the mice in our experiment were head-fixed when the odor was delivered rather than free-moving. The experiment was performed with head fixation since odor sampling is more stable with head fixation and more dynamic when mice are free moving [12]. Although the behavioral output of the animals is generally similar under the two conditions [44], the neural activity may be very di fferent because of the di fferent status. Another possible explanation for the unexpected decrease in neural discrimination of odors is that the mice in our study may have been over-fasted, since we observed that the sni ffing volume decreased slightly under a fasted state. Thus, the decrease in odor discrimination by single M/Ts in a fasted state is likely due to changes in the animals' general state, such as the metabolic condition. Furthermore, we used the firing rate to calculate odor representation in the present study; however, temporal information can also be used for odor representation in the OB [13,14,45]. Finally, the OB receives dense modulation from higher brain areas, including both feedback and centrifugal inputs, and all of these projections dramatically modulate cell activity, odor information processing, and odor representation in the OB [12,46–48]. The change in odor discrimination at the single cell level observed in the OB is the result of complex network interactions between the OB and other higher brain centers. Thus, the neural representation in the OB may not necessarily represent the final behavioral output.

In the OB, activity in the circuits and in the di fferent neuronal subtypes is modulated dramatically by sni ffing [32]. The changes in M/T neural activity and odor discrimination in the fasted state observed in the present study may be due to changes in the sni ffing pattern. However, we found only a slight decrease in sni ffing volume with fasting; both sni ffing frequency and MID remained unchanged. This indicates that the modulation of neural activity in the OB by nutritional state is independent of sni ffing. Interestingly, a higher sni ffing frequency has been reported in fasted free-moving rats [49], but we did not observe a significant change in sni ffing frequency in head-fixed mice. This discrepancy is likely due to di fferences between the free-moving and head-fixed states. In free-moving state, animals are under active condition and they can move around and locate potential food source by fast sni ffing. However, in head-fixed state, animals are under passive condition, cannot not move, and thus have no enthusiasms to find food and need not sni ff fast. Thus, this discrepancy between the two studies is likely due to di fferences between the free-moving and head-fixed states. Future studies could test this possibility by monitoring sni ffing signals in the same animals under free-moving and head-fixed conditions.

In summary, we found that the excitability of neural activity is enhanced but neural discrimination of odors is slightly decreased in the OB of fasted mice under awake, head-fixed conditions. Although a more detailed investigation into the underlying neural mechanisms is warranted, our results represent a first step toward understanding the neural circuit mechanisms by which olfaction is modulated by nutritional status.

**Author Contributions:** Data curation, J.W., P.L., F.C., L.G., and Y.L.; Formal analysis, J.W., P.L., L.G., Y.L., and A.L.; Funding acquisition, F.C., and A.L.; Writing—original draft, J.W., P.L., F.C., L.G., Y.L., and A.L.; Writing—review & editing, J.W., P.L., F.C., L.G., Y.L., and A.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was funded by the National Natural Science Foundation of China (NSFC, 31571082 and 31872771 to A.L), the Priority Academic Program Development of Jiangsu Higher Education Institutions (16KJA180007 to A.L), and the National Science Foundation of Jiangsu Higher Education Institutions of China (18KJD310006 to F.C).

**Acknowledgments:** We thank Han Xu and Xinsong Guo for technical assistance and Qing Liu for critical comments.

**Conflicts of Interest:** The authors declare no conflict of interest.
