Transformation of valence signaling in a striatopallidal circuit

The ways in which sensory stimuli acquire motivational valence through association with other stimuli is one of the simplest forms of learning. Though we have identified many brain nuclei that play various roles in reward processing, a significant gap remains in understanding how valence encoding transforms through the layers of sensory processing. To address this gap, we carried out a comparative investigation of the olfactory tubercle (OT), and the ventral pallidum (VP) - 2 connected nuclei of the basal ganglia which have both been implicated in reward processing. First, using anterograde and retrograde tracing, we show that both D1 and D2 neurons of the OT project primarily to the VP and minimally elsewhere. Using 2-photon calcium imaging, we then investigated how the identity of the odor and reward contingency of the odor are differently encoded by neurons in either structure during a classical conditioning paradigm. We find that VP neurons robustly encode reward contingency, but not identity, in low-dimensional space. In contrast, OT neurons primarily encode odor identity in high-dimensional space. Though D1 OT neurons showed larger response vectors to rewarded odors than other odors, we propose this is better interpreted as identity encoding with enhanced contrast rather than as valence encoding. Finally, using a novel conditioning paradigm that decouples reward contingency and licking vigor, we show that both features are encoded by non-overlapping VP neurons. These results provide a novel framework for the striatopallidal circuit in which a high-dimensional encoding of stimulus identity is collapsed onto a low-dimensional encoding of motivational valence.

The ways in which sensory stimuli acquire motivational valence through association with other 2 stimuli is one of the simplest forms of learning. Though we have identified many brain nuclei that 3 play various roles in reward processing, a significant gap remains in understanding how value 4 encoding transforms through the layers of sensory processing. To address this gap, we carried 5 out a comparative investigation of the olfactory tubercle (OT), and the ventral pallidum (VP) -2 6 connected nuclei of the basal ganglia which have both been implicated in reward processing. 7 First, using anterograde and retrograde tracing, we show that both D1 and D2 neurons of the 8 OT project primarily to the VP and minimally elsewhere. Using 2-photon calcium imaging, we 9 then investigated how the identity of the odor and reward contingency of the odor are differently 10 encoded by neurons in either structure during a classical conditioning paradigm. We find that 11 VP neurons robustly encode value, but not identity, in low-dimensional space. In contrast, OT 12 neurons primarily encode odor identity in high-dimensional space. Though D1 OT neurons 13 showed larger response vectors to rewarded odors than other odors, we propose this is better 14 interpreted as identity encoding with enhanced contrast rather than as value encoding. Finally, 15 using a novel conditioning paradigm that decouples reward contingency and licking vigor, we 16 show that both features are encoded by non-overlapping VP neurons. These results provide a 17 novel framework for the striatopallidal circuit in which a high-dimensional encoding of stimulus 18 identity is collapsed onto a low-dimensional encoding of motivational valence. 19 Introduction 20 Animals exhibit an impressive ability to change how sensory inputs map onto behavioral 21 outputs. Understanding how animals learn to output different behaviors through experience is 22 one of the fundamental problems in neuroscience. Over the last half century, the field has 23 developed compelling frameworks to tackle this problem at both the algorithmic level  Wagner models, Q-learning models) (Rescorla, 1972;Sutton, 1988) and the mechanistic level 25 (Hebbian learning, STDP, neuromodulation) ( Dan and Poo, 2004). By comparison, we lack 26 frameworks through which to understand how the brain might implement learning algorithms 27 through the updating of synaptic weights. One strategy has been to identify neural correlates of 28 latent features assumed to be required for these algorithms (e.g. dopamine as a neural 29 substrate for reward-prediction-error) (Hollerman and Schultz, 1998;Schultz et al., 1997). These 30 results, however, can often be difficult to interpret because reward related signals are found 31 globally throughout the brain (Allen et al., 2019), and are likely multiplexed with signals about 32 motor output and/or stimulus identity. We propose that a more powerful approach is one that 33 compares 1) how the encoding of reward cues changes from one brain nucleus to its 34 downstream target and 2) how much of the encoding can be explained by value vs. other 35 features such as identity or motor output. In this present work, we implement this comparative 36 approach to the investigation of how encoding of olfactory reward cues is transformed between 37 cortices. In stark contrast, we found that hardly any OT neurons were labeled. The rare OT 118 neurons that did have CTB labeling were exclusively localized to the dorsalmost portion of layer 119 III, closely bordering the VP. Taken together, we conclude that both D1 and D2 SPN's of the OT 120 project primarily to the lateral portion of the VP and negligibly to other brain areas, including the 121 VTA. 122 Once we had identified that OT has extremely constrained outputs to the lateral VP, we 123 set out to comparatively characterize the encoding of reward cue in this striatopallidal circuit. 124 Past analysis of value encoding is confounded by not accounting for the difficult-to-avoid 125 overlaps among identity, salience, and reward contingency. To address this, we carefully 126 designed a 6-odor conditioning paradigm where these factors could be decoupled (Fig2A). 127 During each trial, the animal is exposed to 1 of 6 odors for 2 seconds. At the end of odor 128 delivery, the animal either receives: 2 µl of a 10% sucrose solution (S), 50 ms of airpuff at 70 psi 129 (P) or nothing (X). 3 of the odors are ketones (hexanone, heptanone, octanone) and the rest are 130 terpenes (terpinene, pinene, limonene), but the pairing contingencies are chosen such that each 131 contingency group (S, P, or X) includes 1 ketone and 1 terpene. In a value-encoding population, 132 but not in an identity-encoding population, we should see that odor pairs of different reward-133 contingency (e.g. SK, a sucrose-paired ketone vs. SP, an airpuff-paired ketone) are more 134 different than odor pairs of same reward-contingency (e.g. SK, a sucrose-paired ketone vs. ST, a 135 sucrose-paired terpene). Additionally, because both sucrose-pairing and airpuff-pairing should 136 make the associated odor more salient, we can disambiguate between increased 137 discriminability due to salience vs. valence by comparing neural activity in response to sucrose-138 cues or airpuff-cues. 139 To record the activity of the OT and VP neurons across multiple days of pairing, we 140 injected C57BL/6 mice with AAV9-hSyn-jGCaMP7s-WPRE (lateral VP) and Drd1-Cre or 141 we saw the strongest behavioral evidence that animals learned odor-sucrose associations by 159 day 6, we focused our analysis on how reward cues are encoded on the last day of imaging. 160 The animals also showed trends of behavioral changes in response to airpuff-cues, though they 161 were not significant: during airpuff-cues, animals walked less and closed their eyes more than 162 during other odors (FigS6D-G). These behavioral changes for aversive cues were less robust 163 than that for reward association, however, animals show clear responses to the US indicating 164 that they perceive the aversive stimulus. 165 OT and VP neurons showed heterogeneous responses to 6 odors across all 6 days of 166 imaging (FigS3, FigS4, FigS5). To unbiasedly describe the difference between regions, we 167 performed hierarchical clustering on the pooled trial-averaged responses to the 6 odors on the 168 6th day of imaging (Fig3A). We observed both inhibitory (clusters I, II) and excitatory (clusters 169 III-VI) responses to odors as well as broad (clusters II, VI) and narrow (clusters IV, V) odor-170 tuning (Fig3B). Cluster I and cluster III most closely fit our description of putative valence-171 encoding neurons, i.e. neurons that had similar responses to 2 sucrose-cues (SK vs. ST) but 172 different responses to a sucrose-cue and a puff-cue or control odor (SK vs. PK or XK). Although 173 all clusters included neurons from all subpopulations, cluster I and cluster III, which showed 174 larger responses to odors predicting sucrose, were enriched for VP neurons (Fig3C), leading us 175 to hypothesize that neurons in the VP were more likely to be valence encoding neurons than in 176 either OT subpopulation. 177 To assess this hypothesis, we quantified the number of neurons that had statistically 178 significant responses to each of the 6 odors on the last day of imaging. We found that more VP 179 neurons were either excited (29.8±4.1%, 36.6±4.0% for SK, ST) or inhibited (24.5±3.0%, 180 29.4±3.8% for SK, ST) to either sucrose-paired odor than to control or puff-paired odors (7.6-181 11.1% excited, 8.1-12.9% inhibited) (Fig3D,E). A 2-way ANOVA revealed that sucrose-182 contingency, but not puff-contingency, affects the percentage of responsiveness among VP 183 neurons on day 3 and day 6 (FigS8; F=18.6, p=2.04x10 -4 and F=32.2, p=5.78x10 -6 , 184 respectively), but not on day 1. By comparison, there was no significant effect of sucrose pairing 185 on the number of responsive neurons among OTD2 or OTD1 animals on day 6 (p>0.05). OTD2 186 neurons were slightly more responsive to puff-cues on day 3 (FigS8; F=3.03, p=3.99x10 -2 ) and 187 OTD1 neurons were more responsive to sucrose-cues on day 3 (FigS8; F=7.33, p=1.08x10 -2 ), 188 but these effects were small when compared to the effect of sucrose pairing observed among 189 VP animals. When compared across days, we found that the percentage of VP neurons that 190 respond to both S odors increases from 6.1±2.2% on day 1 to 34.1±5.1% by the 6th day of 191 imaging (Fig3E). By comparison, the percentage of OT neurons that respond to both S odors in 192 the same direction (i.e. excited by both S odors or inhibited by both S odors) did not increase 193 through training. Furthermore, whereas OTD1 and OTD2 neurons were more likely to respond to 194 a single odor than they were to respond to both S odors (12.6 vs 31.3% in OTD1, 11.8 vs 21.7% 195 in OTD2), VP neurons were more likely to respond to both S odors than to a single odor (34.1 vs 196 23.3%). 197 Similarly, we found that the magnitude of trial-averaged odor responses in the VP were 198 significantly higher for S odors than X or P odors on the last day of imaging (FigS9; F=59.5,  199 p=3.83x10 -14 ). By comparison, neither sucrose-pairing nor airpuff-pairing had any impact on the 200 magnitude of odor responses in OTD2 neurons on day 6. And though we did observe a 201 significant effect of sucrose-pairing on response magnitudes in OTD1 neurons (FigS9; F=6.34,  202 p=7.97x10 -3 ), both the effect size and significance were weaker than observed in VP. Given that 203 an ideal value-encoding neuron should respond similarly to 2 odors of equal reward-contingency 204 but disparate molecular structure, we looked at the correlation between each neuron's response 205 to the sucrose-paired ketone (SK) and to the sucrose-paired terpene (ST In parallel, we also performed decoding analysis using linear classifiers to assess how 299 reliably a given pair of odors could be decoded from population-level activity (Fig5C-D). To 300 quantify this, we extracted the average ΔFi,k/F values for each trial i [1,m] and each neuron k 301 [1,n]. The resulting matrix of size m x n was used to train a binary linear classifier with a logistic 302 learner. For each classifier, we looked at the average accuracy across 5-fold cross-validation 303 (CV accuracy). Classifiers were trained on neurons that were recorded together (i.e. neurons 304 from the same animal recorded on the same day) to capture biological variability. A total of 765 305 pairwise linear classifiers were trained (15 pairwise comparisons, 17 animals and 3 days). When 306 compared against 10,000 shuffles, 569 of these classifiers showed bootstrapped p-value less 307 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made strictly encodes value since the identity of 2 sucrose-cues can be decoded well. To address the 324 possibility that our results are due to the limitations of linear classification, we repeated the 325 analysis using support vector machines (SVM's) with a radial basis function kernel and found 326 we could draw the same conclusions (FigS12E). Similarly, to verify our results are not 327 epiphenomena of forcing the data into binary classification, we looked at population-level MNR 328 classifiers trained on day 6 data. Importantly, we observe high confusion between 2 sucrose 329 cues in MNR classifiers trained on VP data, but not those trained on OTD2 or OTD1 data 330 (FigS12F), corroborating through an alternate analysis method that VP population activity 331 encodes reward contingency whereas either OT subpopulations likely encode identity. 332 The fact that VP populations showed higher decoding for odor pairs of unequal sucrose-333 contingency provides strong evidence that VP encodes reward-contingency more than identity. 334 Results from OT decoder analyses, however, are less intelligible: all 15 odor pairs, regardless of 335 sucrose-contingency, could be decoded with above-chance success. Though this result is 336 consistent with OT populations encoding identity rather than valence, it does not rule out the 337 possibility that valence and identity are both encoded. In the context of cue-association, 2 cues 338 of different valence cannot have the same identity, meaning that good decoding of {SK vs. PK} 339 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made Analyses at the single-neuron and population levels showed that VP activity encodes 384 reward contingency, rather than the identity, of the olfactory stimulus. However, due to the task 385 design, the reward-contingency of a stimulus was highly correlated with the vigor of licking 386 Briefly, headfixed animals were presented with 1 of 6 odors in pseudorandomized order. 402 During the presentation of 3 of these odors, the lick spout was moved away from the mouse 403 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint with a linear stepper motor. These odors are denoted as N odors (N for No-lick spout). During 404 the presentation of the other 3 odors, the lick spout remained within licking distance of the 405 mouse's tongue. These odors are referred to as L odors (L for lick spout). 1 odor from each 406 group served as a control odor that had 0% reward-contingency (LX, NX). The other 2 odors in 407 each group were paired with sucrose at low (50%) or high (100%) probability (Llo, Lhi, Nlo, Nhi). 408 We reasoned that this contingency could allow us to make pairwise comparisons where one 409 odor has a higher value but lower anticipatory licking than the other (e.g. Nhi vs. Llo). To test this directly, we quantified single neuron decodability of odor pairs and examined 431 how correlated decoding along the reward-contingency axis is to decoding along vigor axis 432 (Fig6F-G). We reasoned that auROC values for {Lhi vs. Nhi} would be high for vigor encoding 433 neurons but not value encoding neurons given these 2 odors have the same reward-434 contingency but disparate licking behaviors. Similarly, we reasoned that auROC values for {Nhi 435 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint vs. NX} would be high for reward-contingency encoding neurons but not vigor encoding neurons 436 given there is a large difference in value but small difference in licking between these 2 odors. 437 First, we saw that while single neuron decodability along the reward-contingency axis (e.g. {Nhi 438 vs NX}) was higher than along the lick axis (e.g. (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint

477
Our anatomical investigations demonstrate that the primary output of the OT is to the 478 VP, with minimal connections to the VTA. Given its constrained connectivity, we propose that 479 the OT to VP circuit is an ideal model system for examining how the encoding of reward cues is 480 transformed across brain circuits. Utilizing comparative longitudinal imaging, we found that VP, 481 but not OTD2, robustly encodes the sucrose-contingency of odors. Although our analyses 482 revealed that sucrose-contingency influences odor-evoked responses in OTD1 neurons more so 483 than in OTD2 neurons, other evidence suggests value encoding is not the appropriate framework 484 for interpreting OTD1 activity. Specifically, information about sucrose-contingency in OTD1 485 resides in a high-dimensional space and generalizes poorly, whereas VP encodes reward-486 contingency robustly in a low-dimensional and generalizable manner. Thus, we suggest that the 487 changes in OTD1 activity are more likely to reflect increased contrast than value. Finally, using a 488 novel classical conditioning paradigm, we assigned motor-related signals and expected-value 489 signals to non-overlapping VP subpopulations. 490 Some of our findings were unexpected. For example, we found no evidence that either 491 OTD1 or OTD2 have significant extrapallidal outputs. This is in direct contrast to a previous study 492 which reported that OTD1 neurons, and to a lesser extent, OTD2 neurons, project to the LH and 493 VTA (Zhang et al., 2017b). We suspect that at least some of the VTA labeling Zhang and 494 colleagues observed from anterograde viral tracing experiments could be due to backflow of the 495 tracer virus in nuclei immediately dorsal to the OT (e.g. AcbSh). As a critical control, we provide 496 evidence that retrograde tracing from VTA robustly labels AcbSh neurons but hardly any OT Though we found little difference in the output patterns of OTD1 and OTD2 neurons, we 527 observed differences in how these 2 subpopulations encode odor valence. Consistent with a 528 previous report (Martiros et al., 2022), we found that OTD1 activity, more than OTD2 activity, is 529 modulated by reward contingency. For example, OTD1 neurons, but not OTD2 neurons, were 530 more likely to respond to sucrose-paired odors than other odors. And the magnitude of 531 responses in OTD1 but not OTD2 neurons were significantly larger to sucrose-paired odors than 532 to other odors. We refrain, however, from concluding that the primary feature encoded in OTD1 533 neurons is value or reward contingency, for the following reasons. First, the above-mentioned 534 effects of sucrose-contingency on neural activity are much stronger for VP than for OTD1. 535 Additionally, whereas more than 50% of VP neurons could be categorized as reward-536 contingency encoders, this figure was less than 20% for OTD1. Lastly, population-level decoders 537 trained on odor pairs of different valence can generalize in the case of VP populations, but not 538 OTD1 populations. While we acknowledge that there is poor standardization when it comes to 539 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint defining value encoding, it is unlikely that discrepancies between our conclusions and those of 540 unique insights that cannot be obtained from recordings alone, we note that SPN's have 558 extensive inhibitory collaterals and exhibit high-dimensional activity. Given these peculiarities of 559 the striatum, we predict that bulk stimulation leads to activity patterns so far outside the 560 physiologically relevant range that it warrants conservative extrapolations regarding their 561 endogenous role. 562 An exciting conclusion from our work is that, within the context of our conditioning 563 paradigm, the dimensionality of neural activity was much lower in VP than in OT. Furthermore, 564 the dimensionality of the imaged subpopulations were anti-correlated with the robustness of 565 sucrose-contingency encoding: OTD2 displayed the highest dimensionality and lowest value 566 encoding whereas VP displayed the lowest dimensionality and highest value encoding. As 567 discussed elegantly by others (Chu et al., 2016;Shannon, 1948), there is generally a tradeoff 568 between the efficiency of a neural population (i.e. its total information capacity) and the 569 robustness of its encoding scheme (i.e. redundancy of encoding). Consistently, it is likely that 570 VP neurons display such robust encoding of valence, in large part, due to the loss of odor 571 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint identity information. By comparison, OT populations may be able to encode information about 572 the large olfactory identity space due to their high dimensionality. We speculate that the 573 extensive inhibitory collaterals among SPN's play a role in enforcing the high dimensionality of 574 OT activity. Though it is entirely unknown what anatomical or physiological strategies are used 575 to reduce VP dimensionality, we consider this an important piece of the puzzle in understanding 576 VP computations. 577 We saw little evidence of negative valence neurons in any of the 3 populations that 578 were imaged. This was surprising given previous reports of negative valence neurons in the VP 579 (Stephenson-Jones et al., 2020). We consider 2 potential explanations for this discrepancy. 580 First, it is possible that our conditioning paradigm was not sufficiently aversive for the animals. 581 Although our behavioral evidence for aversive association is significant, it is less robust than 582 sucrose association raising the possibility that the learning was insufficient. This could be due to 583 the fact that we targeted the airpuff to the animal's hindquarters rather than to the face. But we 584 note that in a previous report, airpuff delivery to the snout and to hindquarters elicited similar is difficult to explain how these 2 responses are related. We consider 3 possible explanations for 601 this paradox. First, in addition to large excitatory responses that were specific to the sucrose-602 cues, we also observed inhibitory responses that were specific to the sucrose-cues. It is 603 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. In this work, we present evidence that contradict previous anatomical and physiological 632 characterizations of the OT. We conclude that the anteromedial portion of the OT sends high-633 dimensional information about odor identity primarily to the VP and not the VTA. By directly 634 comparing OT and VP population-level activity in the same paradigm, we bridge together, for 635 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint the first time, the fields of OT and VP. This provides valuable context which not only helps us 636 evaluate past conclusions about valence encoding in the OT but also consider the implications 637 of the stimulus-evoked activity in the OT. Lastly, we propose a novel framework for 638 understanding the contributions of striatopallidal projections in reward processing. In our 639 framework, OT neurons guide VP's integration of its excitatory afferents rather than being the 640 primary drivers of VP activity. We believe this framework effectively addresses the observed 641 paradox that both OTD1 and VP show increased excitatory activity to reward cues despite being 642 connected through inhibitory projections. Furthermore, we speculate this framework will be 643 useful for understanding striatopallidal projections at large. Behavior 713 Mice were water restricted to reach 85-90% of their initial body weight and given access 714 to water for 5 minutes a day in order to maintain desired weight. Prior to imaging, mice were 715 habituated to the head fixation device (Neurotar) and treadmill for 3-5 days, 15-30 minutes per 716 session. The treadmill parts were 3D printed using a LCD printer (X1-N, EPAX) from publicly 717 available designs (Jackson et al., 2018). During habituation, mice were provided 10% sucrose 718 from the water spout. Walking and licking behaviors were measured using a quadrature encoder 719 (HEDR-5420-es214, Broadcom) and a capacitance sensor (1129_1, Phidgets), respectively. A 720 video feed of the animal's face was also recorded using a camera (acA1300-30um, Basler) with 721 a 8-50mm zoom lens (C2308ZM50, Arducam) at 20 Hz with infrared illumination (VQ2121, 722 Lorex Technology). 723 Odor was delivered to the mouse using a custom-built olfactometer. Compressed 724 medical air was split into 2 gas-mass flow controllers (GFC17, Aalborg). One flow controller 725 directed a constant rate of 1.5 L/min to a hollowed out teflon cylinder. The other flow regulator 726 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint was connected to a 3-way solenoid valve (LHDB1223418H, The Lee Co.). Prior to odor delivery, 727 the 3-way valve directs clean air at 0.5 L/min to the teflon cylinder. During odor delivery, the 3-728 way valve directs air to an odor manifold, which consists of an array of 2-way solenoid valves 729 (LHDB1242115H, The Lee Co.), each connected to a different odor bottle. Depending on the 730 trial type, the appropriate 2-way valve opens, directing 0.5 L/min of air flow through the odor 731 bottle containing a kimwipe blotted with 50 ul of diluted odor. All odors were diluted in mineral oil 732 (M5310, Sigma-Aldrich) to 1.5 mmHg. The kinetics and consistency of odor delivery were 733 characterized for 30 trials of terpinene delivery using a miniature Photoionization Detector 734 (mPID) (Aurora Scientific, Inc). 735 During classical conditioning, animals were exposed to the following odors for 2 736 seconds: 3-hexanone, 3-heptanone, 3-octanone, ⍺-terpinene, ⍺-pinene, and (R)-(+)-limonene 737 (all odors were purchased from Sigma with the highest available purity). In days 1-3 of training, 738 each of the 6 odors and associated outcomes were provided 30 times with 12-18 seconds of 739 inter-trial interval. Hexanone and terpinene were not associated with any outcome, heptanone 740 and pinene were associated with 2 ul of 10% sucrose, and octanone and limonene were 741 associated with a 75 psi airpuff delivered to their hindquarters. Sucrose or airpuff was delivered 742 100-300 ms after the end of odor delivery. Trials were organized into 30 blocks, each of which 743 consisted of 1 trial of each of the 6 odors in randomized order. In days 4-6 of training, the 744 outcome contingencies were switched such that heptanone and limonene were not associated 745 with any outcome, octanone and terpinene were associated with 2 ul of 10% sucrose, and 746 hexanone and pinene were associated with 75 psi airpuff. 747 In the lick-no-lick paradigm, trials were also structured into 30 blocks, each of which 748 consisted of 1 trial of each of the 6 odors in randomized order. Hexanone and terpinene were 749 not associated with any outcome, heptanone and pinene were paired with 2 ul of 10% sucrose 750 at 50% chance, and octanone and limonene were paired with 2 ul of 10% sucrose at 100% 751 chance. 200 ms prior to the onset of 3 of the odors (terpinene, octanone, and limonene), the lick 752 spout was retracted 30 mm away from the animal's mouth using a linear stepper motor (BE073-753 1, Befenybay) and driver (A4988, BIQU). The lick spout would return to its original position 100 754 ms prior to the earliest possible time of sucrose delivery. 755 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. Olympus), using a Galvo-Galvo scanner at 5Hz. Stimulus delivery and behavioral 774 measurements were controlled through a custom software written in LabVIEW (National 775 Instruments) and operated through a DAQ (USB-6008, National Instruments). Each imaging 776 session lasted between 30-45 minutes and was synchronized with the stimulus delivery 777 software through a TTL pulse. The imaging depth was manually adjusted to closely match that 778 of the first imaging day such that we recorded from overlapping populations across days of 779 imaging. Animals were excluded from analysis if a) histology showed that either the GRIN lens 780 or the jGCaMP7s virus was mistargeted or b) the motion during imaging was too severe for 781 successful motion-correction. 2 animals were excluded due to mistargeting and 2 animals were 782 excluded due to excessive motion. 783 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023.

801
To determine how many neurons were responsive to a given odor, we compared ΔF/F at 802 each frame during the 2 second odor period against a pooled distribution of ΔF/F values from 803 the 2-seconds prior to odor onset using a Wilcoxon rank sum test. The resulting p-values were 804 evaluated with Holm-Bonferroni correction to ensure that familywise error rate (FWER) was 805 below 0.05. We then calculated the percentage of responsive neurons for each animal to show 806 the mean and the standard error as a function of time. We also counted the number of neurons 807 that were significantly responsive for at least 4 frames during the odor period to report the total 808 percentage of responsive neurons during odor. 809 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. We thank members of Root lab for discussions, M. Aoi for discussions on data analysis, and T. 865 Komiyama and for comments on the manuscript. This research was supported by grants from 866 the NIH (R00DC014516, R01DC018313), and C.M.R. was a Hellman Fellow. 867 868 869 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure 1: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; Two-way ANOVAs followed by post-hoc t-tests for each output region were performed to generate p-values, which were then corrected for FDR by Benjamini-Hocherg procedure. ***p<0.001, **p<0.01, *p<0.05.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ;

Figure 2:
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; Figure 2. Head-fixed 2-photon Ca 2+ imaging of OTD1, OTD2, or VP neurons during 6-odor conditioning paradigm. (A) State-diagram of odor conditioning paradigm. Each trial begins with 2 seconds of odor delivery. Odors are chosen in pseudorandomized order such that the same odor is not repeated more than twice in a row. At the end of odor delivery, there is a variable delay (100-300ms), after which the animal is given either a 10% sucrose solution (SK and ST), a 70 psi airpuff (PK and PT), or nothing (XK and XT). Trials are separated by a variable intertrial interval (ITI; 12-18s). Schematic representation of (B) lens implant surgery and (C) headfix 2-photon microscopy setup. An example of spatial (D) and temporal (E) components extracted by CNMF from Drd1-Cre animal on day 3 of imaging. (D) The spatial footprints of 20 example neurons are shown on top of a maximum-correlation pixel image that was used to seed the factorization. The number displayed over each neuron matches the row number of the temporal components in (E). (F) An example raster plot (top) and averaged-across-trials trace (bottom) of the licking behavior recorded concurrently as (D) and (E). The timing of odor delivery is shown as shaded rectangles. The timing of US delivery is shown as arrowheads. (G) The mean total licks during each of the odors is shown averaged across all animals (n=17) after application of a moving-average filter with a window size of 10 trials. Red line marks the sucrose and airpuff contingency switch between day 3 and day 4. (H) Bar graph showing the ratio of licks during either SK or ST to licks during any odor. The effects of imaging day and lens target site were analyzed using a 2-way ANOVA. The imaging day, but not lens target site, was a significant factor on licking accuracy (Fday=27.64 vs. Flens location=2.30).
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; Figure 3: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; Figure 3. VP neurons encode reward-contingency more robustly than OTD1 or OTD2 neurons. (A) Heatmap of odor-evoked activities in OTD1, OTD2, and VP neurons from day 6 of imaging. The fluorescence measurements from each neuron were averaged over trials, Zscored, then pooled for hierarchical clustering. Neurons are grouped by similarity, with the dendrogram shown on the right and a raster plot on the left indicating which region a given neuron is from. Horizontal white lines demarcate the boundaries between the 6 clusters. Odor delivered at 0-2 seconds marked by vertical red lines and US delivery is marked by arrowheads. From left to right, the columns represent neural responses to sucrose-paired ketone and terpene, control ketone and terpene, and airpuff-paired ketone and terpene (SK, ST, XK, XT, PK, PT). . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; Odor delivery period is shown with 2 red vertical lines and sucrose/airpuff timing is shown with downward arrowhead. (C) An example neuron's responses on day 1 across 30 trials to 6 different odors. Individual trial traces are shown in light gray whereas the averaged-across trials trace is shown in black. Odor delivery period is depicted as shaded rectangles and US delivery is marked by arrowheads. (D-F) Same as (A-C), respectively, but for day 3. (G) Percentage of all tracked neurons that were both sucrose-responsive on day 1 and odor-responsive in the same direction on day 3. (H) Scatter plot of averaged-over-trials responses to SK or ST on day 1 (x-axis) and day 3 (y-axis). Each point is a neuron that was successfully matched from day 1 and day 3. Neurons from OTD2, OTD1, and VP are plotted as pink circles, blue crosses, and yellow squares, respectively. Neurons that have increased response magnitudes on day 3 would fall between the 2 dotted lines. (I) Violin plot showing the distributions of day 3 responsive magnitude -day 1 response magnitude.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; Figure 5: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; Figure 6: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S2: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S3: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S3. Pooled averaged-over-trials neural activity of all neurons from OTD2 animals across days. Heatmap of odor-evoked activity in OTD2 neurons from day 1, day 3, and day 6 of imaging. The fluorescence measurements from each neuron were averaged over trials, Zscored, then pooled for hierarchical clustering. Neurons are grouped by similarity, with the dendrogram shown on the right. Horizontal white lines demarcate the boundaries between the 6 clusters. Odor delivered at 0-2 seconds marked by vertical red lines. From left to right, the columns represent neural responses to sucrose-paired ketone and terpene, control ketone and terpene, and airpuff-paired ketone and terpene (SK, ST, XK, XT, PK, PT). Data is pooled from 6 animals.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S4: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S4. Pooled averaged-over-trials neural activity of all neurons from OTD1 animals across days. Heatmap of odor-evoked activity in OTD1 neurons from day 1, day 3, and day 6 of imaging.. The fluorescence measurements from each neuron were averaged over trials, Zscored, then pooled for hierarchical clustering. Neurons are grouped by similarity, with the dendrogram shown on the right. Horizontal white lines demarcate the boundaries between the 6 clusters. Odor delivered at 0-2 seconds marked by vertical red lines. From left to right, the columns represent neural responses to sucrose-paired ketone and terpene, control ketone and terpene, and airpuff-paired ketone and terpene (SK, ST, XK, XT, PK, PT). Data is pooled from 6 animals.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S5: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; Figure S5. Pooled averaged-over-trials neural activity of all neurons from VP animals across days. Heatmap of odor-evoked activity in VP neurons from day 1, day 3, and day 6 of imaging. The fluorescence measurements from each neuron were averaged over trials, Zscored, then pooled for hierarchical clustering. Neurons are grouped by similarity, with the dendrogram shown on the right. Horizontal white lines demarcate the boundaries between the 6 clusters. Odor delivered at 0-2 seconds marked by vertical red lines. From left to right, the columns represent neural responses to sucrose-paired ketone and terpene, control ketone and terpene, and airpuff-paired ketone and terpene (SK, ST, XK, XT, PK, PT). Data is pooled from 5 animals.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S6: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S6. Extended behavioral analysis from imaging period. (A) mPID voltage reading in response to 30 trials of odor (terpinene) delivery. The time period during which the odor valve was turned on is shown by the yellow rectangle. Individual recordings are shown in gray and the average is shown in black. (B) On-kinetics of odor delivery. On delay refers to the interval between the valve turning on and the mPID voltage increasing by more than 10% of baseline. (C) Off-kinetics of odor delivery. Off delay refers to the interval between the odor valve turning off and the mPID voltage decreasing by more than 10% of its maximum. (D) Representative velocity of the head-fixed mouse in response to the 6 different odors measured by a digital encoder. The lines represent the average across 30 trials and the shaded areas represent the SEM. The black arrowhead marks when the US is delivered. (E) The difference in walking velocity in response to odor (left) and US delivery (right). Differences are calculated between the last second before odor delivery and the last second before the odor exposure (left) or the first half second after US delivery (right) grouped by US pairing. Circles represent the average across animals and the error bars show SEM. (F) Representative changes in range-normalized eye-size in response to the 6 different odors. (G) The difference in eye-size in response to odor (left) and US delivery (right). Differences are calculated between the last second before odor delivery and the last second before the odor exposure (left) or the first half second after US delivery (right) grouped by US pairing. Circles represent the average across animals and the error bars show SEM.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S7: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; Figure S7. Traces of example neurons and their corresponding metrics. (A) Example traces from an OTD1 neuron recorded on day 3. Each column shows this neuron's response to a given odor across 30 trials (gray). The average across all trials is shown in black. For green and red arrowheads mark the time at which sucrose and airpuff were delivered, respectively. auROC values from single-neuron binary classifiers for discriminating {SK vs. XK}, {SK vs. PK}, and {SK vs. ST} are displayed on the right. Additionally, the results of statistical analysis to determine if this neuron reliably responded to each odor (in. = significant inhibitory response, exc. = significant excitatory response, ns= no significant difference between baseline and odor period). (B) Example traces from an OTD2 neuron recorded on day 1. (C) Example traces from a VP neuron recorded on day 3. (D) Example traces from a VP neuron recorded on day 6.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S8: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; Figure S8. Percentage of neurons responsive to each odor across days. Bar graphs showing percentage of neurons from OTD2 on day 1 that were significantly excited or inhibited by each odor and each US (S for sucrose, P for airpuff). The average is shown by the bar while the + and x markers show individual biological replicates. For each region and imaging day pair, 3 one-way ANOVA's were run to test for the effect of sucrose, airpuff, or ketone vs. terpene on percentage responsiveness. Only the significant F-statistics are reported with their corresponding p-value.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S9: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S9. Distribution of response magnitudes to each odor across days. Violin plots showing the averaged-over-trials response magnitudes to each odor during the last second of odor exposure. For each region and imaging day pair, 3 one-way ANOVA's were performed to test for the effect of sucrose, airpuff, or ketone vs. terpene on response magnitude. Only the significant F-statistics are reported with their corresponding p-value.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S10: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S11: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S11. Multinomial analysis of single neuron odor encoding (A) Confusion matrix of single-neuron MNR classifiers trained on neural activity during the last second of odor exposure on day 6 of imaging. Rows represent the true class while columns represent the predicted class. Each confusion matrix is averaged across 10 k-fold and across all neurons of a given region. ɸ represents data taken from 30 pre-odor bins randomly sampled from -1.5 to -0.5 seconds relative to odor delivery. (B) Violin plot of single-neuron MNR classifier accuracy, averaged across 10 k-fold, grouped by region. (C) Violin plot of the single-neuron MNR classifier accuracy trained on shuffled data. Each data point represents the average across 10 shuffles. (D) Violin plot of MNR S-cue confusion, i.e. confusion between SK and ST. This corresponds to 1) when the true class was SK but predicted class was ST and 2) when the true class was ST but the predicted class was SK. (E) Violin plot of MNR confusion among all ketones. This corresponds when the true class was a ketone and the predicted class was a different ketone (e.g. true class = XK and predicted class = PK). (F) Scatterplot of each neuron's ketone confusion on the x-axis and S-cue confusion on the y-axis on days 1, 3, and 6 of imaging. (G) Stacked bar \graph showing the distribution of neurons from each population that fall into each of the 4 quadrants across the 3 different imaging days.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S12: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S12. Analysis of population-level odor encoding (A) Scatterplot of CV-accuracy of linear classifiers trained on simultaneously-recorded neurons on the x-axis and their bootstrapped unadjusted p-values on the y-axis. Red horizontal line marks p = 0.001 and red red vertical line marks CV accuracy = 0.75. All classifiers with CV accuracy higher than 0.75 had p<0.001. In total, there are 765 binary classifiers (15 pairwise classifiers for each of the 51 recordings across 3 regions and days 1, 3, and 6 of imaging). Each classifier was compared against 10,000 shuffles. For auROC values that were greater than all 10,000 shuffles, a conservative p-value of 0.0001 was assigned. (E) Heatmap of CV accuracy from binary SVM's trained on day 6 of imaging with a radial basis function kernel. CV accuracy was averaged across biological replicates. (F) Confusion matrix of population-level MNR classifiers trained on neural activity during the last second of odor exposure on day 6 of imaging. Rows represent the true class while columns represent the predicted class. Each confusion matrix is averaged across biological replicates. ɸ represents data taken from 30 pre-odor bins randomly sampled from -1.5 to -0.5 seconds relative to odor delivery.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S13: . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint Figure S13. Camera-based detection of licking in head-fixed animals (A-C) Metrics of a representative neurons with activity that predicts licking. (A) Representative neuron's Z-scored ΔF/F (dark green) and Z-scored dF/dt (light green) aligned to the onset of a lickbout. Each line represents the same neuron's activity during an individual lickbout. (B) Stemplot showing an example of the lagged correlation between the onset of licking and the fluorescence of a neuron across 9 frames (1 frame = 0.2s). Time bins of various lags are shown on the x-axis (negative number denotes frames that precede onset of licking) and the resulting R 2 is plotted on the yaxis. Red vertical line marks the case where the 9 frames are centered on the onset of licking.
As an example, [-6:2] refers to fluorescence between 1.2 seconds prior to lickbout onset and 0.4 seconds after lickbout onset. (C) The output of a distributed lag model (DLM) that predicts the onset of a lickbout from ΔF/F of a single neuron. ΔF/F (dark green, top), DLM score (magenta, middle), and the licking recorded by a capacitive sensor (black, bottom) are shown in parallel. The DLM model was trained using 9 distributed frames ([-6:2]) of ΔF/F for each frame of lickbout onset. (D) Scatterplot of day 6 VP neurons' DLM lick classifier auROC on the y-axis plotted against their mean {SK vs. XK} auROC on the x-axis. The slope, intercept, R 2 , and p-value of the slope are shown on the top left corner. (E) 3 example snapshots of the camera feed during moving lick spout paradigm with overlay of DeepLabCut labeling. The coordinates of top lip, bottom lip, base of tongue (tbase), and tip of tongue (ttip) are displayed with a probability cutoff of 0.4. (F) Range-normalized metrics from DLC labeling. P(ttip) (red, top) is the probability score assigned to the labeling of the tongue tip. P(tbase) (yellow, middle) is the probability score assigned to the labeling of the tongue base. Mgap (purple, bottom) is the Euclidean distance between the top lip and the bottom lip. The 3 vertical magenta lines represent the timing of the 3 snapshots shown in (E). (G) The same range-normalized metrics as in (F) plotted against the ground truth licking data from capacitive sensor (black, bottom) and DLC-based licking classifier score (magenta, second from bottom). (H) Lineplot showing the difference in total licking to Lhi nd Nhi during the time bin (1.5-2.5 seconds after odor onset) used for most analyses plotted against imaging day for individual animals. The time of peak difference is circled in black.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 3, 2023. ; https://doi.org/10.1101/2023.08.01.551547 doi: bioRxiv preprint