In operant learning, animals modify their action repertoires to earn desired rewards. Previous work on the neural substrates of such learning has focused on the striatum and the midbrain dopaminergic projections that target the striatum (Wise, 2004; Yin et al., 2005; Kravitz et al., 2013; Rossi et al., 2013; Yttri and Dudman, 2016). Midbrain dopamine neurons have been implicated in reinforcement learning (Schultz et al., 1997; Tsai et al., 2009; Rossi et al., 2013). Such learning is often thought to be distinct from declarative or episodic learning, which requires the hippocampus and medial temporal lobe structures (Mishkin et al., 1984; Morris et al., 1986; Milner et al., 1998; Eldridge et al., 2000). On the other hand, work in both humans and rodents has also implicated the hippocampus in reward processing and motivated behavior, though the underlying mechanisms remain unclear (Adcock et al., 2006; Gauthier and Tank, 2018).

The hippocampus is also a target of dopaminergic projections. Dopamine receptors are expressed in the hippocampus, and in mice D1-class receptor expression is common in the dentate gyrus (DG) region (Gangarossa et al., 2012; Kempadoo et al., 2016). However, these dopaminergic projections come from locus coeruleus (LC) (Kempadoo et al., 2016; Takeuchi et al., 2016; Chowdhury et al., 2022), rather than the major dopamine cell groups in the ventral tegmental area (VTA) and substantia nigra pars compacta (SNc), which supply dopamine to the basal ganglia (Björklund and Dunnett, 2007; Ikemoto, 2007). The functional role of the dopaminergic LC-DG projection remains obscure.

In this study, we examined the contribution of dopaminergic signaling in the DG to operant learning and behavior. We found that mice could learn to perform a new action (pressing a lever) for optogenetic activation of D1 + neurons in the DG. In addition, using both optogenetics and in vivo pharmacological manipulations, we found that activation of LC dopaminergic neurons that project to the DG can also support self-stimulation. This effect depended on the activation of D1-like receptors. Finally, using in vivo calcium imaging in appetitive operant conditioning with food rewards, we found that D1 + DG neurons were more related to the goal-directed actions than simply non-contingent reward presentation.

To understand the role of D1 + neurons in the hippocampus, we tested whether selective stimulation of these neurons can reinforce operant behavior using a self-stimulation paradigm. We injected either a Cre-dependent channelrhodopsin (AAV5-DIO-ChR2) or a fluorescent control (DIO-eYFP) into D1-Cre mice (D1::ChR2^DG or D1::eYFP^DG), producing selective expression of the excitatory opsin in D1 + neurons in the dentate gyrus (Figure 1A–B). Mice received photo-stimulation (500 ms, 20 Hz, 15 ms pulse width) following lever pressing on a fixed ratio schedule of reinforcement (Figure 1C). All D1::ChR2^DG mice learned to press a lever for stimulation, whereas control mice did not (Figure 1D). These results suggest that D1::ChR2^DG stimulation is sufficient to reinforce lever pressing. Interestingly, this form of self-stimulation is remarkably resistant to extinction, persisting after 8 days without any photostimulation.

Figure 1

Download asset Open asset

Figure 1—source data 1

The press rate (presses/min) of D1::ChR2 and D1::eYFP mice across FR1, FR3, and FR5 sessions.

https://cdn.elifesciences.org/articles/83600/elife-83600-fig1-data1-v2.csv

Download elife-83600-fig1-data1-v2.csv

Figure 1—source data 2

The press rate (presses/min) of D1::ChR2 and D1::eYFP mice across extinction sessions.

https://cdn.elifesciences.org/articles/83600/elife-83600-fig1-data2-v2.zip

Download elife-83600-fig1-data2-v2.zip

Next, using retrograde tracing methods, we were able to map projections to the DG (Figure 2A–D & Figure 2—figure supplement 1). We confirmed significant LC projections to the DG, but we did not find significant VTA or SNc projections (Figure 2E and H & Table 1). Retrograde labeling of DG-projecting LC neurons is colocalized with tyrosine hydroxylase (TH), a marker for catecholamine neurons (e.g. dopamine, norepinephrine; Figure 2F–H). In contrast, there was no labeling in the VTA (Figure 2H, Figure 2—figure supplement 1). This finding suggests that the DG receives TH + projections from the LC rather than VTA.

Figure 2 with 1 supplement see all

Download asset Open asset

Figure 2—source data 1

Percent of colocalized (TH+ and tdTomato+) neurons in the DG and VTA.

https://cdn.elifesciences.org/articles/83600/elife-83600-fig2-data1-v2.csv

Download elife-83600-fig2-data1-v2.csv

We then tested whether the LC-DG projection is responsible for the self-stimulation effect observed. In order to manipulate the LC-DG pathway selectively, we injected AAV-Retro2-Cre into the DG and a Cre-dependent ChR2 (AAV5-DIO-ChR2) into the LC (Figure 3A–B). We found that ChR2^DG-LC (n=8) mice also showed self-stimulation that is comparable to the stimulation of D1::ChR2^DG neurons (Figure 3C).

Figure 3

Download asset Open asset

Figure 3—source data 1

The press rate (presses/min) of LC-DG::ChR2 LC-DG::eYFP mice during extinction days.

https://cdn.elifesciences.org/articles/83600/elife-83600-fig3-data1-v2.csv

Download elife-83600-fig3-data1-v2.csv

Figure 3—source data 2

The press rate (presses/min) of LC-DG::ChR2 LC-DG::eYFP mice across FR1, FR3, and FR5 sessions.

https://cdn.elifesciences.org/articles/83600/elife-83600-fig3-data2-v2.csv

Download elife-83600-fig3-data2-v2.csv

Figure 3—source data 3

Total presses during vehicle and propranolol I.P. administration (3mg/kg and 6mg/kg).

https://cdn.elifesciences.org/articles/83600/elife-83600-fig3-data3-v2.zip

Download elife-83600-fig3-data3-v2.zip

Figure 3—source data 4

Total presses during vehicle and SCH-23390 I.P. administration (0.1mg/kg and 0.2mg/kg).

https://cdn.elifesciences.org/articles/83600/elife-83600-fig3-data4-v2.zip

Download elife-83600-fig3-data4-v2.zip

The LC-DG projection releases both norepinephrine and dopamine (Kempadoo et al., 2016; Takeuchi et al., 2016). It is unclear which transmitter is responsible for the self-stimulation effect, though our observation on D1 + DG neurons (Figure 1) suggests that dopamine might be responsible. Consequently, to determine which of these transmitters is responsible for the observed effects, we used pharmacological manipulations in combination with pathway-specific optogenetic manipulations using the same self-stimulation paradigm (Figure 3E and Figure 4). Mice (n=8) trained on self-stimulation were tested after either receiving systemic injections of a β-adrenoceptor antagonist (propranolol), or a D1-antagonist (SCH 23390). β-adrenoceptor blockade did not produce any significant effects (Figure 3D and F). In contrast, D1-antagonist significantly impaired self-stimulation (Figure 3E).

Figure 4

Download asset Open asset

Figure 4—source data 1

Total presses during vehicle and propranolol administration through cranial cannula (21mM and 10.5mM).

https://cdn.elifesciences.org/articles/83600/elife-83600-fig4-data1-v2.zip

Download elife-83600-fig4-data1-v2.zip

Figure 4—source data 2

Total presses during vehicle and SCH-23390 administration through cranial cannula (3.6mM and 1.8mM).

https://cdn.elifesciences.org/articles/83600/elife-83600-fig4-data2-v2.zip

Download elife-83600-fig4-data2-v2.zip

Figure 4—source data 3

The press rate (presses/min) of LC-DG::ChR2 + DG cannula and LC-DG::eYFP + DG cannula mice across FR1 sessions.

https://cdn.elifesciences.org/articles/83600/elife-83600-fig4-data3-v2.csv

Download elife-83600-fig4-data3-v2.csv

Figure 4—source data 4

Total head movement (meters) during SCH-23390(3.6mM) and propranolol(21mM) administration through cranial cannula.

https://cdn.elifesciences.org/articles/83600/elife-83600-fig4-data4-v2.zip

Download elife-83600-fig4-data4-v2.zip

To activate DG-projecting LC neurons robustly, we targeted the cell bodies of the DG-LC projections. But the LC has broad projections to many brain areas (Schwarz and Luo, 2015). To verify that our self-stimulation effects were not due to the activation of LC collaterals in other regions, we performed local infusions of antagonists (Figure 4A–C, n=8). Infusions of a D1-antagonist into the DG significantly impaired self-stimulation, whereas propranolol showed no significant group differences in self-stimulation (Figure 4D–E). These results suggest that the reinforcing effects of LC-DG stimulation are due to the activation of D1 receptors by dopamine, rather than by norepinephrine.

Based on our self-stimulation results, we hypothesized that DG D1 + neurons may be preferentially activated during operant conditioning in general, including during actions that result in natural rewards rather than simply optogenetic stimulation of the LC-DG pathway. To test this, we performed in vivo calcium imaging of DG D1 + neurons while performing an operant lever pressing task with a food reward. We implanted a gradient index lens above the DG in D1-Cre mice (n=5) and injected them with a Cre-dependent calcium indicator (AAV9-syn-FLEX-jGCamp7f) (Figure 5A–B).

Figure 5

Download asset Open asset

Figure 5—source data 1

Lever press rate (presses/min) during FR1, FR3 and FR5 sessions, number of modulated neurons during the first press, percent of modulated neurons during each press across days.

https://cdn.elifesciences.org/articles/83600/elife-83600-fig5-data1-v2.xls

Download elife-83600-fig5-data1-v2.xls

We then recorded calcium transients from DG D1 + neurons during operant lever pressing for food rewards (Figure 5A–C) during lever pressing for food reward on fixed-ratio (FR) schedules (FR1, FR3, and FR5). We found distinct populations of DG D1 + neurons that were modulated by lever pressing. To see if the neural activity is action-contingent, we also used a control task in which pressing is not required. The reward was delivered non-contingently every 20 s, preceded by 1 s of white noise. On this task, there were far fewer significantly modulated DG D1 + neurons (n=6, 3% of the total population) compared to the operant task (Figure 5F, Table 2). To verify that the virus targets D1 + neurons in the DG, we quantified the percentage of neurons that are virally targeted that express D1 receptors. Using RNA scope, we found that GcAMP-7f was colocalized with D1 receptors (Figure 6).

Figure 6

Download asset Open asset

Figure 6—source data 1

Percent of GFP cells colocalized with negative control, positive control and Drd1a probes.

https://cdn.elifesciences.org/articles/83600/elife-83600-fig6-data1-v2.xls

Download elife-83600-fig6-data1-v2.xls

To determine if the activity of these neurons reflected the spatial locations of the lever pressing or the action of the lever pressing itself, we used a discrete trial design with two levers (Figure 7). On each trial, one of the two levers was randomly selected to extend into the operant box. Once pressed, the lever would retract. The reward would then be delivered 1 s later. This task allowed us to compare the neural activity modulated by lever pressing and reward, as well as determine the spatial tuning of the same neurons. We found that several populations of dentate D1 + neurons (n=40, 16.5% of the total population) that were significantly modulated at the time of lever pressing (Figure 7C). One small population was modulated by reward delivery (n=14, 4.9% of the total population). Importantly, another population with significantly more neurons was responsive to lever pressing at either lever location (Figure 7C–D; n=22, 9.79% of the total population). These neurons were not spatially selective, as they were responsive when the lever was presented at different locations. However, we did find a small population that responded to only a single lever (Figure 7C; left lever: n=15, right lever: n=12; 4.2% of total population).

Figure 7 with 2 supplements see all

Download asset Open asset

Figure 7—source data 1

Percent of modulated neurons during lever press, both levers, left lever only, right lever only, reward, and extension.

https://cdn.elifesciences.org/articles/83600/elife-83600-fig7-data1-v2.zip

Download elife-83600-fig7-data1-v2.zip

It is difficult to assess the spatial activity of neurons in operant tasks, as animals do not cover the arena equally but instead preferentially occupy specific task-relevant locations (Figure 7—figure supplement 1). To examine the stability of spatially related activity during operant conditioning, we used an FR5 task with two levers (Figure 7—figure supplement 2). We then used the same methods as described in Skaggs et al., 1993 to identify spatially modulated neurons, and split the sessions into periods when the left or right lever was active. This allowed us to recalculate the center of mass of our identified spatial firing fields with two different lever locations. As mice mostly stayed close to the wall where the two levers and food port were located, we limited our analysis to the x-dimension which explained most of the variance in the neural activity. We found that the place field centers were significantly different when the lever is available. While we did find that occupancy varied in this switch task, the occupancy across other tasks was consistent, suggesting that spatial modulation does not depend on the type of task (e.g. operant vs Pavlovian) (Figure 7—figure supplement 1). In contrast, task-related neural activity in the DG depends on whether the task is action-contingent.

Together our results provide the first evidence that DG D1 + neurons may play a role in operant reinforcement. Using a cell-type-specific approach, we showed that activation of D1 + neurons in the DG is sufficient for self-stimulation (Gangarossa et al., 2012). Furthermore, our retrograde tracing identified that the DG receives TH + input from the LC and not from the VTA or SNc, suggesting that the LC is a source of dopaminergic projections to the DG. We found that mice will press a lever for optogenetic stimulation of the LC-DG projection. Blockade of D1 receptors, but not noradrenergic beta receptors, attenuated the self-stimulation of hippocampal-projecting LC neurons.

These findings build upon previous research that suggests that the LC supplies the primary dopaminergic input to the dorsal hippocampus (Kempadoo et al., 2016; Takeuchi et al., 2016). While previous work focuses on the LC-CA1 pathway, we examine a population of D1 + neurons in the DG that plays a role in operant reinforcement.

Both the hippocampus and the LC are known to be effective sites for intracranial self-stimulation (Ursin et al., 1966; Crow et al., 1972; Ritter and Stein, 1973). However, as non-selective electrical stimulation was used in classic studies, the precise circuit mechanisms underlying these observations remain unclear. In the present study, we have investigated a pathway for operant reinforcement that originates from LC neurons and targets D1 + DG neurons in the hippocampus. Future work will have to address whether this self-stimulation effect requires D1 + neurons or if it is a property of hippocampal neurons in general.

Self-stimulation behavior supported by stimulation of the LC-DG projections or the D1 + DG neurons appears to be different from that supported by stimulation of dopamine neurons in the VTA or SNc. First, the rate of lever pressing is much lower with LC-DG self-stimulation. Yet once established, the lever pressing was surprisingly resistant to extinction, persisting for many days after the termination of stimulation. This is very different from the self-stimulation of VTA or SNc DA pathways and their target regions; such classic self-stimulation behavior is rapidly extinguished in the absence of stimulation (Gallistel, 1964; Olds, 1977). Thus, classic self-stimulation supported by SNc and VTA DA neurons, which project mainly to the frontal cortex and basal ganglia, has a much stronger immediate effect on performance but rarely produces persistent actions in the long-term without stimulation. The effect on performance can largely be explained by a repetition of the action commands recently generated, but without DA release in the BG this repetition effect decays. Self-stimulation of the LC-DG pathway, in contrast, is less robust but is supported by some long-lasting memory of the stimulation, in accordance with recent findings that the LC projections to the hippocampus play a major role in long-term contextual memory (Chowdhury et al., 2022).

Our calcium imaging results showed that D1 + neurons in the DG are preferentially active during goal-directed instrumental actions compared to passive reward delivery. These findings suggest that the DA projections to the hippocampus contribute to the reinforcement of specific instrumental actions. They are broadly in agreement with recent findings on entorhinal grid cells (Butler et al., 2019) and hippocampal CA1 cells (Gauthier and Tank, 2018), as well as LC terminals in CA1 that were preferentially active near a novel reward location (Kaufman et al., 2020). It remains to be determined how the activity of DG D1 + neurons changes during activation of LC neurons that project to the hippocampus, and if LC activation induces plasticity in the DG that is important for learning.

All experimental procedures were conducted in accordance with standard ethical guidelines and were approved by the Duke University Institutional Animal Care and Use Committee.

All data generated or analysed during this study are included in the manuscript and supporting files.

1. 1. Adcock RA
   2. Thangavel A
   3. Whitfield-Gabrieli S
   4. Knutson B
   5. Gabrieli JDE

   (2006) Reward-motivated learning: mesolimbic activation precedes memory formation

   Neuron 50:507–517.

   https://doi.org/10.1016/j.neuron.2006.03.036

   • PubMed
   • Google Scholar
2. 1. Björklund A
   2. Dunnett SB

   (2007) Dopamine neuron systems in the brain: an update

   Trends in Neurosciences 30:194–202.

   https://doi.org/10.1016/j.tins.2007.03.006

   • PubMed
   • Google Scholar
3. 1. Butler WN
   2. Hardcastle K
   3. Giocomo LM

   (2019) Remembered reward locations restructure entorhinal spatial maps

   Science 363:1447–1452.

   https://doi.org/10.1126/science.aav5297

   • PubMed
   • Google Scholar
4. 1. Chowdhury A
   2. Luchetti A
   3. Fernandes G
   4. Filho DA
   5. Kastellakis G
   6. Tzilivaki A
   7. Ramirez EM
   8. Tran MY
   9. Poirazi P
   10. Silva AJ

   (2022) A locus coeruleus-dorsal CA1 dopaminergic circuit modulates memory linking

   Neuron 110:3374–3388.

   https://doi.org/10.1016/j.neuron.2022.08.001

   • PubMed
   • Google Scholar
5. 1. Crow TJ
   2. Spear PJ
   3. Arbuthnott GW

   (1972) Intracranial self-stimulation with electrodes in the region of the locus coeruleus

   Brain Research 36:275–287.

   https://doi.org/10.1016/0006-8993(72)90735-4

   • Google Scholar
6. 1. da Silva JA
   2. Tecuapetla F
   3. Paixão V
   4. Costa RM

   (2018) Dopamine neuron activity before action initiation gates and invigorates future movements

   Nature 554:244–248.

   https://doi.org/10.1038/nature25457

   • PubMed
   • Google Scholar
7. 1. Eldridge LL
   2. Knowlton BJ
   3. Furmanski CS
   4. Bookheimer SY
   5. Engel SA

   (2000) Remembering episodes: a selective role for the hippocampus during retrieval

   Nature Neuroscience 3:1149–1152.

   https://doi.org/10.1038/80671

   • PubMed
   • Google Scholar
8. Software

   1. Flatiron Institute

   (2023) CalmAn, version 80e1681

   GitHub.

   https://github.com/flatironinstitute/CaImAn
9. 1. Gallistel CR

   (1964) Electrical self-stimulation and its theoretical implications

   Psychological Bulletin 61:23–34.

   https://doi.org/10.1037/h0047017

   • PubMed
   • Google Scholar
10. 1. Gangarossa G
    2. Longueville S
    3. De Bundel D
    4. Perroy J
    5. Hervé D
    6. Girault JA
    7. Valjent E

    (2012) Characterization of dopamine D1 and D2 receptor-expressing neurons in the mouse hippocampus

    Hippocampus 22:2199–2207.

    https://doi.org/10.1002/hipo.22044

    • PubMed
    • Google Scholar
11. 1. Gauthier JL
    2. Tank DW

    (2018) A dedicated population for reward coding in the hippocampus

    Neuron 99:179–193.

    https://doi.org/10.1016/j.neuron.2018.06.008

    • PubMed
    • Google Scholar
12. 1. Ikemoto S

    (2007) Dopamine reward circuitry: two projection systems from the ventral midbrain to the nucleus accumbens-olfactory tubercle complex

    Brain Research Reviews 56:27–78.

    https://doi.org/10.1016/j.brainresrev.2007.05.004

    • PubMed
    • Google Scholar
13. 1. Kaufman AM
    2. Geiller T
    3. Losonczy A

    (2020) A role for the locus coeruleus in hippocampal CA1 place cell reorganization during spatial reward learning

    Neuron 105:1018–1026.

    https://doi.org/10.1016/j.neuron.2019.12.029

    • PubMed
    • Google Scholar
14. 1. Kempadoo KA
    2. Mosharov EV
    3. Choi SJ
    4. Sulzer D
    5. Kandel ER

    (2016) Dopamine release from the locus coeruleus to the dorsal hippocampus promotes spatial learning and memory

    PNAS 113:14835–14840.

    https://doi.org/10.1073/pnas.1616515114

    • PubMed
    • Google Scholar
15. 1. Kravitz AV
    2. Owen SF
    3. Kreitzer AC

    (2013) Optogenetic identification of striatal projection neuron subtypes during in vivo recordings

    Brain Research 1511:21–32.

    https://doi.org/10.1016/j.brainres.2012.11.018

    • PubMed
    • Google Scholar
16. 1. Lopes G
    2. Bonacchi N
    3. Frazão J
    4. Neto JP
    5. Atallah BV
    6. Soares S
    7. Moreira L
    8. Matias S
    9. Itskov PM
    10. Correia PA
    11. Medina RE
    12. Calcaterra L
    13. Dreosti E
    14. Paton JJ
    15. Kampff AR

    (2015) Bonsai: an event-based framework for processing and controlling data streams

    Frontiers in Neuroinformatics 9:7.

    https://doi.org/10.3389/fninf.2015.00007

    • PubMed
    • Google Scholar
17. 1. Mathis A
    2. Mamidanna P
    3. Cury KM
    4. Abe T
    5. Murthy VN
    6. Mathis MW
    7. Bethge M

    (2018) DeepLabCut: markerless pose estimation of user-defined body parts with deep learning

    Nature Neuroscience 21:1281–1289.

    https://doi.org/10.1038/s41593-018-0209-y

    • PubMed
    • Google Scholar
18. 1. Milner B
    2. Squire LR
    3. Kandel ER

    (1998) Cognitive neuroscience and the study of memory

    Neuron 20:445–468.

    https://doi.org/10.1016/s0896-6273(00)80987-3

    • PubMed
    • Google Scholar
19. Book

    1. Mishkin M
    2. Malamut B
    3. Bachevalier J

    (1984)

    Memories and habits: two neural systems

    In: Lynch G, McGaugh N, editors. .Neurobiology of Learning and Memory. Guilford Press. pp. 65–77.

    • Google Scholar
20. 1. Morris RG
    2. Anderson E
    3. Lynch GS
    4. Baudry M

    (1986) Selective impairment of learning and blockade of long-term potentiation by an N-methyl-D-aspartate receptor antagonist, AP5

    Nature 319:774–776.

    https://doi.org/10.1038/319774a0

    • PubMed
    • Google Scholar
21. Book

    1. Olds J

    (1977)

    Drives and Reinforcements: Behavioral Studies of Hypothalamic Functions

    Raven Press.

    • Google Scholar
22. 1. Ritter S
    2. Stein L

    (1973) Self-Stimulation of noradrenergic cell group (A6) in locus coeruleus of rats

    Journal of Comparative and Physiological Psychology 85:443–452.

    https://doi.org/10.1037/h0035289

    • PubMed
    • Google Scholar
23. 1. Rolls ET
    2. Kelly PH
    3. Shaw SG

    (1974) Noradrenaline, dopamine, and brain-stimulation reward

    Pharmacology, Biochemistry, and Behavior 2:735–740.

    https://doi.org/10.1016/0091-3057(74)90103-8

    • PubMed
    • Google Scholar
24. 1. Rossi MA
    2. Sukharnikova T
    3. Hayrapetyan VY
    4. Yang L
    5. Yin HH
    6. Zhuang X

    (2013) Operant self-stimulation of dopamine neurons in the substantia nigra

    PLOS ONE 8:e65799.

    https://doi.org/10.1371/journal.pone.0065799

    • PubMed
    • Google Scholar
25. 1. Schultz W
    2. Dayan P
    3. Montague PR

    (1997) A neural substrate of prediction and reward

    Science 275:1593–1599.

    https://doi.org/10.1126/science.275.5306.1593

    • PubMed
    • Google Scholar
26. 1. Schwarz LA
    2. Luo L

    (2015) Organization of the locus coeruleus-norepinephrine system

    Current Biology 25:R1051–R1056.

    https://doi.org/10.1016/j.cub.2015.09.039

    • PubMed
    • Google Scholar
27. Conference

    1. Skaggs WE
    2. McNaughton BL
    3. Gothard KM

    (1993)

    An information-theoretic approach to deciphering the hippocampal code

    Advances in neural information processing systems. pp. 1030–1037.

    • Google Scholar
28. 1. Soudais C
    2. Laplace‐Builhe C
    3. Kissa K
    4. Kremer EJ

    (2001) Preferential transduction of neurons by canine adenovirus vectors and their efficient retrograde transport in vivo

    The FASEB Journal 15:1–23.

    https://doi.org/10.1096/fj.01-0321fje

    • Google Scholar
29. 1. Takeuchi T
    2. Duszkiewicz AJ
    3. Sonneborn A
    4. Spooner PA
    5. Yamasaki M
    6. Watanabe M
    7. Smith CC
    8. Fernández G
    9. Deisseroth K
    10. Greene RW
    11. Morris RGM

    (2016) Locus coeruleus and dopaminergic consolidation of everyday memory

    Nature 537:357–362.

    https://doi.org/10.1038/nature19325

    • PubMed
    • Google Scholar
30. 1. Tervo DGR
    2. Hwang B-Y
    3. Viswanathan S
    4. Gaj T
    5. Lavzin M
    6. Ritola KD
    7. Lindo S
    8. Michael S
    9. Kuleshova E
    10. Ojala D
    11. Huang C-C
    12. Gerfen CR
    13. Schiller J
    14. Dudman JT
    15. Hantman AW
    16. Looger LL
    17. Schaffer DV
    18. Karpova AY

    (2016) A designer AAV variant permits efficient retrograde access to projection neurons

    Neuron 92:372–382.

    https://doi.org/10.1016/j.neuron.2016.09.021

    • PubMed
    • Google Scholar
31. 1. Tsai H-C
    2. Zhang F
    3. Adamantidis A
    4. Stuber GD
    5. Bonci A
    6. de Lecea L
    7. Deisseroth K

    (2009) Phasic firing in dopaminergic neurons is sufficient for behavioral conditioning

    Science 324:1080–1084.

    https://doi.org/10.1126/science.1168878

    • PubMed
    • Google Scholar
32. 1. Ursin R
    2. Ursin H
    3. Olds J

    (1966) Self-Stimulation of hippocampus in rats

    Journal of Comparative and Physiological Psychology 61:353–359.

    https://doi.org/10.1037/h0023253

    • PubMed
    • Google Scholar
33. 1. Wise RA

    (2004) Dopamine, learning and motivation

    Nature Reviews. Neuroscience 5:483–494.

    https://doi.org/10.1038/nrn1406

    • PubMed
    • Google Scholar
34. 1. Yin HH
    2. Ostlund SB
    3. Knowlton BJ
    4. Balleine BW

    (2005) The role of the dorsomedial striatum in instrumental conditioning

    The European Journal of Neuroscience 22:513–523.

    https://doi.org/10.1111/j.1460-9568.2005.04218.x

    • PubMed
    • Google Scholar
35. 1. Yttri EA
    2. Dudman JT

    (2016) Opponent and bidirectional control of movement velocity in the basal ganglia

    Nature 533:402–406.

    https://doi.org/10.1038/nature17639

    • PubMed
    • Google Scholar
36. 1. Zhou P
    2. Resendez SL
    3. Rodriguez-Romaguera J
    4. Jimenez JC
    5. Neufeld SQ
    6. Giovannucci A
    7. Friedrich J
    8. Pnevmatikakis EA
    9. Stuber GD
    10. Hen R
    11. Kheirbek MA
    12. Sabatini BL
    13. Kass RE
    14. Paninski L

    (2018) Efficient and accurate extraction of in vivo calcium signals from microendoscopic video data

    eLife 7:e28728.

    https://doi.org/10.7554/eLife.28728

    • PubMed
    • Google Scholar
37. Software

    1. Zhou P

    (2020) CNMF_E, version ccca6f9

    GitHub.

    https://github.com/zhoupc/CNMF_E

Decision letter after peer review:

[Editors’ note: the authors submitted for reconsideration following the decision after peer review. What follows is the decision letter after the first round of review.]

Thank you for submitting the paper "Elucidating a locus coeruleus-hippocampal dopamine pathway for operant reinforcement" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and a Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Kei Igarashi (Reviewer #2).

Comments to the Authors:

We are sorry to say that, after consultation with the reviewers, we have decided that this work will not be considered further for publication by eLife.

As you will see below, all the reviewers thought that this work contains significant findings that are of interest to the neuroscience community. In particular, the finding of the reinforcing effect of locus coeruleus (LC) neurons projecting to the dentate gyrus (DG) is potentially of great interest. However, the reviewers found several substantive issues that limit the validity of the conclusions. Specifically, the reviewers are concerned about a lack of experiments to establish the release of dopamine (or still potentially norepinephrine) from LC neurons projecting to DG. Currently, this conclusion was made using intraperitoneal injections of dopamine receptor antagonist. It is important to perform local pharmacological experiments for both norepinephrine and dopamine antagonists within the DG in self-stimulating mice to fully address this. Furthermore, many experiments reported in this paper, including experiments using norepinephrine and dopamine antagonists (Figure 3F), are underpowered. The number of animals (samples) needs to be increased throughout this work, in particular, the data supporting the specificity of dopamine as opposed to norepinephrine (Figure 3F). Another important concern raised by the reviewers is that it remains unclear whether there is something unique about putative D1R-expressing neurons in the DG for controlling instrumental actions.

Overall, the reviewers thought that the above issues are essential. Although these issues could be addressed by additional experiments, it will likely take more than 2-3 months. Because of these reasons, we decided to reject this work, at least in the current form. If the authors can fully address the above issues and each of the reviewers individual concerns, we would be happy to consider a revised manuscript as a new submission.

Reviewer #1 (Recommendations for the authors):

This manuscript by Petter et al., examines the role of putative D1R expressing neurons in the dentate gyrus in regulating operant behaviors in mice. The authors show that mice will learn to press a lever to receive optogenetic stimulation of D1R-Cre targeted cells compared to eYFP control mice. Next, they show that the DG receives input from the LC as previously described. They then show that mice will self-stimulate for activation of LC neurons that project to the DG, and that optogenetic inhibition of D1-DG neurons reduces the amount of self-stimulation of LC-DG neurons. Finally, the authors use miniscope based calcium imaging to show that DG D1R expressing neurons show changes in their activity timelocked to lever pressing. Overall, the work builds upon previous studies to support a role of LC projections to the DC in instrumental learning and reinforcement. While many of the findings support these conclusions, there are a number of issues with the data and paper in its current form.

1) Targeting DG cells in the D1R-Cre mouse. The authors imply that somehow D1R expressing cells in the DG are critical for instrumental behavior. However, it is unclear whether there is robust co-localization with D1R expression and the virally targeted cells. One image is presented to suggest this, but the authors need to quantify the percent of cells that are virally targeted that express D1R vs. those that do not. This has been performed for striatal tissue, but I am not aware of studies that have done this for the DG.

2) The amount of nosepoking to receive optical stimulation in all of the self-stimulation experiments seems low, and it is somewhat difficult to really assess this as it is presented as a lever press rate. I would be helpful to present the data in the total number of nosepokes per session for all groups. Even if the level of self-stimulation is low, they do use the appropriate control group (eYFP expressing controls) show the behavior is different between the groups.

3) The tracing experiments, although simple and straightforward, suffer from low n's. 2 hemispheres from 2 mice are presented in figure 2 for example. It would be preferable if at least 3-4 biological replicates are included in each dataset. These should also include both male and female mice whenever possible.

4) The antagonist experiments presented in Figure 3 are consistent with their hypothesis that dopamine is mediating the self-stimulation effect. However, these drugs are given IP and there is no way to know whether they are actually acting in the DG vs. other areas that may also reduce reinforcement.

5) For the self-stimulation of LC neurons experiment (Figure 3) it is unclear why a non-specific viral strategy was employed. The authors could have easily targeted TH expressing neurons that project from the LC to the DG instead of only targeting neurons based on their projection target. At the very least, the authors should perform a careful quantification of the percentage of the LC cells that are targeted using the retroAAV that are TH⁺.

6) The presentation of the miniscope data could use some improvement. For example, how many cells for each animal were recorded? How many trials were analyzed per session? All of the image data looks processed, and it would be nice to see some raw data presented alongside the processed data. I was also struck by how short the recording sessions were (10 minutes). Why did the authors record for such a short period of time?

Reviewer #2 (Recommendations for the authors):

In this manuscript, Petter and colleagues identify the neuromodulatory pathway from the locus coeruleus to the dentate gyrus of the hippocampus. The manuscript has a novelty on focusing in this previously undescribed circuit. However, I have several major concerns that some of the claims are not justified by the results.

1. I have a concern on the propranolol experiment (Figure 3F). Although the authors state that they did not see significant effect with propranolol, there is a trend of decreases (as mentioned by the author). The sample number here looks like n=5, which would not be achieving a sufficient power. Thus, it may be the case that this circuit functions using both dopamine AND norepinephrine.

2. I have a difficulty in interpreting LC-stimulation + DG inhibition experiment (Figure 4), as this can be interpreted with many possibilities. Rather, LC axon inhibition would achieve simple conclusion about the involvement of this circuit.Reviewer #3 (Recommendations for the authors):

In this study, the authors show that stimulation of Drd1-expressing dentate gyrus (DG) neurons in the mammalian brain promotes operant reinforcement, thereby expanding the role of the hippocampus traditionally studied with respect to episodic/spatial memory to instrumental learning. This is a very interesting finding, supported by a substantial array of behavioural observations. Yet, I do have three sets of comments/questions regarding the selectivity of the main observations, and thus the strength of the related conclusions.

First, the authors conclude from their experiments that D1-expressing DG cells constitute a specific population of hippocampal neurons that supports operant reinforcement learning. But how cell-type selective is this behavioural contribution? Notably, can Drd1-non-expressing DG neurons also be involved in this behavioural effect? That is, can self-stimulation be achieved with other DG neuron types? Interestingly, many retro-Cre labelled LC neurons that project to DG are not expressing tyrosine hydroxylase (see Table 1). Likewise, is this behavioural effect selective to LC-to-DG inputs? That is, can this effect be obtained with any input targeting Drd1-expressing DG neurons (see Supplementary Figure 1)?

Second, while the data supporting the direct contribution of the LC-to-DG pathway in operant reinforcement is convincing (irrespective of whether or not such a contribution is selective), the dopaminergic identity of this pathway remains elusive. Throughout the manuscript the authors keep making this sort of strong statement: e.g., "surprisingly, these neurons receive dopaminergic projections from the locus coeruleus …" (e.g., in the Abstract). As far as I understand the results, neither the release of dopamine from, nor the dopaminergic identity of, the LC inputs to DG are demonstrated. This claim is central to the work but seems to rely on two indirect observations. First, DG-targeting LC neurons are immuno-positive for tyrosine hydroxylase. Indeed, because this marks catecholamine neurons and is thus expressed by both dopamine and norepinephrine neurons. Second, systemic injections of the β-adrenoceptor antagonist propranolol does not prevent self-stimulation while the D1-antagonist SCH 23390 does. But the direct blockade of the LC-DG pathway is not established in these pharmacological experiments. Should other pathways be recruited in this behaviour (e.g., involving VTA or SNC to other, non-hippocampal circuits), then the current interpretation of this pharmacological blockade would be misleading.

Finally, hippocampal activity is strongly coupled to animal's speed (and thus the corresponding network states that report active exploration versus immobility). The observation that D1+ DG neurons are modulated by lever pressing during operant conditioning is interesting. The inclusion of the passive reward delivery is a good control. But speed should be formally controlled for (analyses in Figure 5 and 6). This can be done in at least two ways: by using speed-matched time windows across the two reward delivery conditions (i.e., active lever press versus passive delivery), and by reporting the measure of DG activity (Δ F/F) as a function of speed at the active reward delivery.

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Elucidating a locus coeruleus-hippocampal dopamine pathway for operant reinforcement" for further consideration by eLife. Your revised article has been evaluated by Kate Wassum (Senior Editor) and a Reviewing Editor.

The manuscript has been improved but there are some remaining issues that need to be addressed, as outlined below:

As you can see, we very much appreciated your thorough response to the prior review. The manuscript is much improved. However, there are some remaining concerns that need to be addressed. These are summarized for your convenience below but are described in more detail in the reviewers' comments. In your revision please provide a point x point response to each reviewer's comment.

– There are some concerns, especially regarding Figure 3, whether drug effects on behavior are specific to reinforcement and not confounded by general effects on motor behavior. We feel this could be ideally addressed experimentally, though there may be other avenues to alleviate this concern.

– There are concerns that the conclusions that the reinforcement effects are mediated by dopamine and not norepinephrine hinge on the strong conclusions made about null results from experiments that could be underpowered (e.g., N=5) and also have high variability. As important conclusions rely on these findings, we think this concern should be addressed with additional experiments to increase power.

– Reviewer 3 makes a number of additional important points that we all agree should be addressed.

Reviewer #1 (Recommendations for the authors):

Most of my previous concerns are now resolved.

Reviewer #2 (Recommendations for the authors):

All of my previous concerns have been addressed.

Reviewer #3 (Recommendations for the authors):

1. There are concerns with some of the reinforcement data that clouds interpretation. The biggest concern I have is in Figure 3 where it is not possible to tell whether responding during the drug treatment challenges is truly being reinforced by LC-DG stimulation in these tests. These tests were run after 8d of extinction of a previous operant response, and there is no difference in responding between EXT day 8 and test day 1. In fact, they respond less during the testing. In figure 4, basal response rates are very low at the vehicle (10 responses/30min). Again, these rates are lower than extinction levels previously established. Are the authors sure that the drug effects on behavior are reinforcement-specific and not just general effects on motor behavior?

2. The only evidence that supports their claim that this is mediated by dopamine and not norepinephrine is the microinjection study, which is quite underpowered and also involves the question of reinforcement (above). This hinges on the strong conclusions made about null results from an underpowered set of experiments. For the authors are trying to rule out the contributions of NE, which is hard to do given n=5 where the mean effect size is a reduction of ~5-6 responses vs. ~2-3 responses on a task with fairly high variance to begin with. In particular, the 21nM propranolol group may be lower than the vehicle if the outlier with the increased responding is removed. This is important as the conclusion of the manuscript is that an LC-DG circuit mediates reinforcement via a dopamine input

3. It is not clear why in Figure 3 they didn't use a TH-specific viral strategy here. Since the manipulation in not cell-type specific, they really should have quantified these neurons. How many of the ChR2+ neurons are TH⁺?

4. The imaging studies were conducted on D1+ neurons, but not demonstrably related to LC inputs. With such small cell numbers activated here, and such a small TH⁺ population of LC neurons projecting to the DG – this is potentially a very small group of cells and the authors have not linked these responses to this specific input.

5. The discussion overstates a large number of the conclusions.

a. The authors point out in the text that it is interesting that this behavior doesn't really extinguish… yet this observation is left entirely unexplored. Perhaps, this circuitry is important for the acquisition of [reward-related] memories… but perhaps it is not critical to the maintenance of conditioned behavior. Some inhibition (or DA depletion) studies may give some insight into this possibility

b. Authors write: "These findings suggest that the LC supplies the primary dopaminergic input to the dorsal hippocampus (1), especially to a population of D1+ neurons in the DG (2), and that this dopaminergic pathway plays a critical role in operant reinforcement (3)."

i. It is not clearly demonstrated here that the LC provides dopaminergic input to the DG; however, it has been shown more clearly in previous studies. This language could be softened to suggest that they have data to support previous work.

ii. There is no direct evidence that LC projections modulate the D1+ population, so this may need to be softened as well.

iii. There is no evidence that this pathway is necessary, or even involved in natural reinforcement. This would require some circuit inhibition studies or pharmacological manipulations during sucrose reinforcement. It's also not clear what the authors believe the function of D1+ DG neurons to be.

c. Authors note "These findings suggest that the hippocampus contributes to the reinforcement of specific instrumental actions.". While this is possible, there is no clear evidence for this function – only evidence the calcium events were organized around lever presses. What about lever presses prior to learning? What about lever presses after extinction?

6. Is figure 1C showing a response isolated from an FR5 (as stated in the legend)? Were there no other responses within the 15-second timeframe of each response? How were these signals isolated and how was it ensured that the same data traces weren't being averaged several times but slightly offset from each other because of the temporal relationship between responses?

7. The response and the reinforcer delivery are only separated by 1 second… so it is challenging to distinguish whether there is a calcium response to reward (+1s) or not since fluorescence is already elevated at the time of reward consumption.

8. Similarly, the lever protracted 2sec after reward delivery – are the animals still consuming the sucrose pellet at this time?

9. Ultimately, it's unclear how they were able to categorize the neurons given the condensed task parameters and high variability exact spike time for the neurons shown in Figure 7B.

[Editors’ note: the authors resubmitted a revised version of the paper for consideration. What follows is the authors’ response to the first round of review.]

As you will see below, all the reviewers thought that this work contains significant findings that are of interest to the neuroscience community. In particular, the finding of the reinforcing effect of locus coeruleus (LC) neurons projecting to the dentate gyrus (DG) is potentially of great interest. However, the reviewers found several substantive issues that limit the validity of the conclusions. Specifically, the reviewers are concerned about a lack of experiments to establish the release of dopamine (or still potentially norepinephrine) from LC neurons projecting to DG. Currently, this conclusion was made using intraperitoneal injections of dopamine receptor antagonist. It is important to perform local pharmacological experiments for both norepinephrine and dopamine antagonists within the DG in self-stimulating mice to fully address this.

We have performed local pharmacological experiments using norepinephrine and dopamine antagonists within the DG (Figures 3-4). The results support our original conclusions.

Furthermore, many experiments reported in this paper, including experiments using norepinephrine and dopamine antagonists (Figure 3F), are underpowered. The number of animals (samples) needs to be increased throughout this work, in particular, the data supporting the specificity of dopamine as opposed to norepinephrine (Figure 3F).

We have also increased the number of animals for pharmacological experiments as well as tracing experiments.

Another important concern raised by the reviewers is that it remains unclear whether there is something unique about putative D1R-expressing neurons in the DG for controlling instrumental actions.

We have included discussion of this issue (see discussion) – “Future work will have to address whether this self-stimulation effect requires D1+ neurons or if it is a property of hippocampal neurons in general.”

Reviewer #1 (Recommendations for the authors):

This manuscript by Petter et al., examines the role of putative D1R expressing neurons in the dentate gyrus in regulating operant behaviors in mice. The authors show that mice will learn to press a lever to receive optogenetic stimulation of D1R-Cre targeted cells compared to eYFP control mice. Next, they show that the DG receives input from the LC as previously described. They then show that mice will self-stimulate for activation of LC neurons that project to the DG, and that optogenetic inhibition of D1-DG neurons reduces the amount of self-stimulation of LC-DG neurons. Finally, the authors use miniscope based calcium imaging to show that DG D1R expressing neurons show changes in their activity timelocked to lever pressing. Overall, the work builds upon previous studies to support a role of LC projections to the DC in instrumental learning and reinforcement. While many of the findings support these conclusions, there are a number of issues with the data and paper in its current form.

1) Targeting DG cells in the D1R-Cre mouse. The authors imply that somehow D1R expressing cells in the DG are critical for instrumental behavior. However, it is unclear whether there is robust co-localization with D1R expression and the virally targeted cells. One image is presented to suggest this, but the authors need to quantify the percent of cells that are virally targeted that express D1R vs. those that do not. This has been performed for striatal tissue, but I am not aware of studies that have done this for the DG.

We used RNAscope to quantify the number of virally target cells that colocalized with D1R expression. We found robust colocalization (Figure 6).

2) The amount of nosepoking to receive optical stimulation in all of the self-stimulation experiments seems low, and it is somewhat difficult to really assess this as it is presented as a lever press rate. I would be helpful to present the data in the total number of nosepokes per session for all groups. Even if the level of self-stimulation is low, they do use the appropriate control group (eYFP expressing controls) show the behavior is different between the groups.

Lever pressing was used. We now include total number of presses for all data.

3) The tracing experiments, although simple and straightforward, suffer from low n's. 2 hemispheres from 2 mice are presented in figure 2 for example. It would be preferable if at least 3-4 biological replicates are included in each dataset. These should also include both male and female mice whenever possible.

We added 2 mice with retro-Cre injected into the DG of Ai-14 mice (n = 4).

4) The antagonist experiments presented in Figure 3 are consistent with their hypothesis that dopamine is mediating the self-stimulation effect. However, these drugs are given IP and there is no way to know whether they are actually acting in the DG vs. other areas that may also reduce reinforcement.

We have performed the same experiments using local infusion of these drugs (Figure 4). Similar results were found.

5) For the self-stimulation of LC neurons experiment (Figure 3) it is unclear why a non-specific viral strategy was employed. The authors could have easily targeted TH expressing neurons that project from the LC to the DG instead of only targeting neurons based on their projection target. At the very least, the authors should perform a careful quantification of the percentage of the LC cells that are targeted using the retroAAV that are TH⁺.

We added quantification of the percentage of LC cells that are TH⁺ (Figure 2) and found that ~40% of the retro-cre labeled cells are also TH⁺.

6) The presentation of the miniscope data could use some improvement. For example, how many cells for each animal were recorded? How many trials were analyzed per session? All of the image data looks processed, and it would be nice to see some raw data presented alongside the processed data. I was also struck by how short the recording sessions were (10 minutes). Why did the authors record for such a short period of time?

We updated the Figure to reflect how many neurons were recorded per animal and included a table (Table 2) to show the number of neurons recorded from each animal for each task. We added more imaging data from other tasks that show the number of presses (“trials”) per session for each animal (Figure 6D, Figure 7A). Sessions were short because we wanted to avoid photobleaching.

Reviewer #2 (Recommendations for the authors):

In this manuscript, Petter and colleagues identify the neuromodulatory pathway from the locus coeruleus to the dentate gyrus of the hippocampus. The manuscript has a novelty on focusing in this previously undescribed circuit. However, I have several major concerns that some of the claims are not justified by the results.

1. I have a concern on the propranolol experiment (Figure 3F). Although the authors state that they did not see significant effect with propranolol, there is a trend of decreases (as mentioned by the author). The sample number here looks like n=5, which would not be achieving a sufficient power. Thus, it may be the case that this circuit functions using both dopamine AND norepinephrine.

We added animals to bring the sample number to N=8 for the systemic injections (Figure 3). We also performed local infusions with both NE and DA antagonists (cannulae in DG) to further confirm our effects.

2. I have a difficulty in interpreting LC-stimulation + DG inhibition experiment (Figure 4), as this can be interpreted with many possibilities. Rather, LC axon inhibition would achieve simple conclusion about the involvement of this circuit.

We removed this data, as it could be difficult to interpret.

Reviewer #3 (Recommendations for the authors):

In this study, the authors show that stimulation of Drd1-expressing dentate gyrus (DG) neurons in the mammalian brain promotes operant reinforcement, thereby expanding the role of the hippocampus traditionally studied with respect to episodic/spatial memory to instrumental learning. This is a very interesting finding, supported by a substantial array of behavioural observations. Yet, I do have three sets of comments/questions regarding the selectivity of the main observations, and thus the strength of the related conclusions.

First, the authors conclude from their experiments that D1-expressing DG cells constitute a specific population of hippocampal neurons that supports operant reinforcement learning. But how cell-type selective is this behavioural contribution? Notably, can Drd1-non-expressing DG neurons also be involved in this behavioural effect? That is, can self-stimulation be achieved with other DG neuron types? Interestingly, many retro-Cre labelled LC neurons that project to DG are not expressing tyrosine hydroxylase (see Table 1). Likewise, is this behavioural effect selective to LC-to-DG inputs? That is, can this effect be obtained with any input targeting Drd1-expressing DG neurons (see Supplementary Figure 1)?

We focused on the role of D1+ neurons and the DA projections from LC. We did not examine the role of other cell types in the DG. Future work will have to investigate how specific the effects are to D1 neurons. We found ~41% of the retrogradely labeled neurons colocalized with TH.

Second, while the data supporting the direct contribution of the LC-to-DG pathway in operant reinforcement is convincing (irrespective of whether or not such a contribution is selective), the dopaminergic identity of this pathway remains elusive. Throughout the manuscript the authors keep making this sort of strong statement: e.g., "surprisingly, these neurons receive dopaminergic projections from the locus coeruleus …" (e.g., in the Abstract). As far as I understand the results, neither the release of dopamine from, nor the dopaminergic identity of, the LC inputs to DG are demonstrated. This claim is central to the work but seems to rely on two indirect observations. First, DG-targeting LC neurons are immuno-positive for tyrosine hydroxylase. Indeed, because this marks catecholamine neurons and is thus expressed by both dopamine and norepinephrine neurons. Second, systemic injections of the β-adrenoceptor antagonist propranolol does not prevent self-stimulation while the D1-antagonist SCH 23390 does. But the direct blockade of the LC-DG pathway is not established in these pharmacological experiments. Should other pathways be recruited in this behaviour (e.g., involving VTA or SNC to other, non-hippocampal circuits), then the current interpretation of this pharmacological blockade would be misleading.

As mentioned above, we also included experiments using local infusion of antagonists to address this problem.

Finally, hippocampal activity is strongly coupled to animal's speed (and thus the corresponding network states that report active exploration versus immobility). The observation that D1+ DG neurons are modulated by lever pressing during operant conditioning is interesting. The inclusion of the passive reward delivery is a good control. But speed should be formally controlled for (analyses in Figure 5 and 6). This can be done in at least two ways: by using speed-matched time windows across the two reward delivery conditions (i.e., active lever press versus passive delivery), and by reporting the measure of DG activity (Δ F/F) as a function of speed at the active reward delivery.

In order to address these concerns, we report the measure of DG activity (df/f) as a function of speed during the reward delivery period. Specifically, we used a 4 second window (+/- 2 seconds) centered on reward delivery. For each trial and neuron pair, we take the average df/f in this 4 second window and plot it as a function of speed. In order to compare neurons on the same scale df/f was normalized by the maximum df/f for each neuron.

The data is plotted for all three tasks in Author response image 1. There is no significant speed modulation for the non-contingent task (R²=0.0591, p value: 0.2924) or the delay task (R² = 0.0311 , p value: 0.4102). The FR5 task shows a weak negative relationship between speed and df/f (R² = -0.094 , p value: 0.0031)

Author response image 1

Download asset Open asset

[Editors’ note: what follows is the authors’ response to the second round of review.]

The manuscript has been improved but there are some remaining issues that need to be addressed, as outlined below:

As you can see, we very much appreciated your thorough response to the prior review. The manuscript is much improved. However, there are some remaining concerns that need to be addressed. These are summarized for your convenience below but are described in more detail in the reviewers' comments. In your revision please provide a point x point response to each reviewer's comment.

– There are some concerns, especially regarding Figure 3, whether drug effects on behavior are specific to reinforcement and not confounded by general effects on motor behavior. We feel this could be ideally addressed experimentally, though there may be other avenues to alleviate this concern.

We analyzed total distance traveled for each drug condition and did not find effects on motor behavior (Figure 4F).

– There are concerns that the conclusions that the reinforcement effects are mediated by dopamine and not norepinephrine hinge on the strong conclusions made about null results from experiments that could be underpowered (e.g., N=5) and also have high variability. As important conclusions rely on these findings, we think this concern should be addressed with additional experiments to increase power.

We have performed additional experiments to increase N.

Reviewer #3 (Recommendations for the authors):

1. There are concerns with some of the reinforcement data that clouds interpretation. The biggest concern I have is in Figure 3 where it is not possible to tell whether responding during the drug treatment challenges is truly being reinforced by LC-DG stimulation in these tests. These tests were run after 8d of extinction of a previous operant response, and there is no difference in responding between EXT day 8 and test day 1. In fact, they respond less during the testing. In figure 4, basal response rates are very low at the vehicle (10 responses/30min). Again, these rates are lower than extinction levels previously established. Are the authors sure that the drug effects on behavior are reinforcement-specific and not just general effects on motor behavior?

We believe that the drug effects on behavior are reinforcement specific. The animals used for LC-hippocampal pathway stimulation with intrahippocampal cannula were all naïve to start the experiments. These animals showed an increase in responding compared to controls (Figure 4C) and significant decreases in lever pressing in response to SCH23390 but not propranolol (Figure 4D,E). Further, we analyzed the movement of these animals using DeepLabCut and found that the head movement of these animals was unchanged by the drug infusions (Figure 4F). The change in lever pressing but not gross movement suggests these are likely reinforcement specific.

2. The only evidence that supports their claim that this is mediated by dopamine and not norepinephrine is the microinjection study, which is quite underpowered and also involves the question of reinforcement (above). This hinges on the strong conclusions made about null results from an underpowered set of experiments. For the authors are trying to rule out the contributions of NE, which is hard to do given n=5 where the mean effect size is a reduction of ~5-6 responses vs. ~2-3 responses on a task with fairly high variance to begin with. In particular, the 21nM propranolol group may be lower than the vehicle if the outlier with the increased responding is removed. This is important as the conclusion of the manuscript is that an LC-DG circuit mediates reinforcement via a dopamine input

We performed additional experiments and increased the N in each group to 8.

3. It is not clear why in Figure 3 they didn't use a TH-specific viral strategy here. Since the manipulation in not cell-type specific, they really should have quantified these neurons. How many of the ChR2+ neurons are TH⁺?

We did quantify the number of ChR2+ neurons that are TH⁺ in figure 2H.

4. The imaging studies were conducted on D1+ neurons, but not demonstrably related to LC inputs. With such small cell numbers activated here, and such a small TH⁺ population of LC neurons projecting to the DG – this is potentially a very small group of cells and the authors have not linked these responses to this specific input.

The imaging experiments merely attempted to show the properties of D1+ neurons in the DG during operant behavior. Whether these neurons receive TH⁺ LC projections cannot be determined, as doing pathway specific stimulation while imaging is not technically feasible. In the paper we did not make any specific claims about the relationship between TH⁺ LC projections and neurons we imaged. We do know from previous work that TH⁺ axons strongly innervate the hippocampus (Kempadoo et al., Kaufman et al.,2020, Takeuchi et al., 2016). While we have not linked the responses of the cells directly to LC activity, we find it probably that the LC modulates these cells based on the changes in behavior with D1-DG activation of LC-DG activation.

5. The discussion overstates a large number of the conclusions.

a. The authors point out in the text that it is interesting that this behavior doesn't really extinguish… yet this observation is left entirely unexplored. Perhaps, this circuitry is important for the acquisition of [reward-related] memories… but perhaps it is not critical to the maintenance of conditioned behavior. Some inhibition (or DA depletion) studies may give some insight into this possibility

Testing reward-related memories is beyond the scope of these studies. We agree it would be interesting for future studies to inhibit D1-DG neurons during acquisition and extinction of conditioned behavior.

b. Authors write: "These findings suggest that the LC supplies the primary dopaminergic input to the dorsal hippocampus (1), especially to a population of D1+ neurons in the DG (2), and that this dopaminergic pathway plays a critical role in operant reinforcement (3)."

i. It is not clearly demonstrated here that the LC provides dopaminergic input to the DG; however, it has been shown more clearly in previous studies. This language could be softened to suggest that they have data to support previous work.

We softened the language to reflect that this claim about DA inputs to the DG is shown by previous work, rather than our findings.

ii. There is no direct evidence that LC projections modulate the D1+ population, so this may need to be softened as well.

We have revised our discussion to point out the lack of direct evidence.

iii. There is no evidence that this pathway is necessary, or even involved in natural reinforcement. This would require some circuit inhibition studies or pharmacological manipulations during sucrose reinforcement. It's also not clear what the authors believe the function of D1+ DG neurons to be.

We have revised our discussion to point out the lack of direct evidence, but our calcium imaging data shows systematic changes in press-related activity during training.

c. Authors note "These findings suggest that the hippocampus contributes to the reinforcement of specific instrumental actions.". While this is possible, there is no clear evidence for this function – only evidence the calcium events were organized around lever presses. What about lever presses prior to learning? What about lever presses after extinction?

An additional piece of evidence supporting the role of these hippocampal neurons in operant conditioning is that there are more neurons responsive to lever pressing compared to noncontingent reward delivery. There are no lever presses prior to learning in such a task, since lever pressing is learned.

6. Is figure 1C showing a response isolated from an FR5 (as stated in the legend)? Were there no other responses within the 15-second timeframe of each response? How were these signals isolated and how was it ensured that the same data traces weren't being averaged several times but slightly offset from each other because of the temporal relationship between responses?

In figure 5C there was a typo and the example neuron was actually taken from an FR1 session.

We have updated the legend to reflect this. There are no other responses within the time frame.

7. The response and the reinforcer delivery are only separated by 1 second… so it is challenging to distinguish whether there is a calcium response to reward (+1s) or not since fluorescence is already elevated at the time of reward consumption.

This is partially addressed by the non-contingent reward task, where there is no operant conditioning component.

8. Similarly, the lever protracted 2sec after reward delivery – are the animals still consuming the sucrose pellet at this time?

It is possible that the animals are still consuming the sucrose pellet at this time. If they are still consuming the pellet they can wait to respond on the lever. The lever does not retract after a fixed amount of time, and therefore the animal can press it whenever it is ready.

9. Ultimately, it's unclear how they were able to categorize the neurons given the condensed task parameters and high variability exact spike time for the neurons shown in Figure 7B.

We used a non-contingent reward task so we could examine how many cells were responsive when no operant behavior was required for a reward. In this task we found significantly fewer modulated neurons compared to a task where lever pressing was required.

Author details

1. Elijah A Petter

   Department of Psychology and Neuroscience, Duke University, Durham, United States

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

2. Isabella P Fallon

   Department of Psychology and Neuroscience, Duke University, Durham, United States

   Contribution

   Formal analysis, Investigation, Methodology

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-1456-4152

3. Ryan N Hughes

   Department of Psychology and Neuroscience, Duke University, Durham, United States

   Contribution

   Investigation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-4999-0215

4. Glenn DR Watson

   Department of Psychology and Neuroscience, Duke University, Durham, United States

   Contribution

   Investigation

   Competing interests

   No competing interests declared

5. Warren H Meck

   Department of Psychology and Neuroscience, Duke University, Durham, United States

   Contribution

   Investigation

   Competing interests

   No competing interests declared

6. Francesco Paolo Ulloa Severino

   1. Department of Psychology and Neuroscience, Duke University, Durham, United States
   2. Department of Cell Biology, Duke University School of Medicine, Durham, United States

   Contribution

   Investigation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-3725-9713

7. Henry H Yin

   1. Department of Psychology and Neuroscience, Duke University, Durham, United States
   2. Department of Neurobiology, Duke University School of Medicine, Durham, United States

   Contribution

   Conceptualization, Resources, Formal analysis, Funding acquisition, Methodology, Writing – original draft, Project administration, Writing – review and editing

   For correspondence

   hy43@duke.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-1546-6850

• 1,146

  Page views

• 236

  Downloads

• 1

  Citations

Article citation count generated by polling the highest count across the following sources: PubMed Central, Crossref, Scopus.