State-specific gating of salient cues by midbrain dopaminergic input to basal amygdala

Abstract

Basal amygdala (BA) neurons guide associative learning via acquisition of responses to stimuli that predict salient appetitive or aversive outcomes. We examined the learning- and state-dependent dynamics of BA neurons and ventral tegmental area (VTA) dopamine (DA) axons that innervate BA (VTADA→BA) using two-photon imaging and photometry in behaving mice. BA neurons did not respond to arbitrary visual stimuli, but acquired responses to stimuli that predicted either rewards or punishments. Most VTADA→BA axons were activated by both rewards and punishments, and they acquired responses to cues predicting these outcomes during learning. Responses to cues predicting food rewards in VTADA→BA axons and BA neurons in hungry mice were strongly attenuated following satiation, while responses to cues predicting unavoidable punishments persisted or increased. Therefore, VTADA→BA axons may provide a reinforcement signal of motivational salience that invigorates adaptive behaviors by promoting learned responses to appetitive or aversive cues in distinct, intermingled sets of BA excitatory neurons.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Fig. 1: Mouse BA neurons acquire hunger-dependent responses to food-predicting cues.
Fig. 2: VTA dopamine axons in BA are activated by cues predicting food reward availability in hungry mice.
Fig. 3: VTA dopamine axons in BA are activated by cues predicting unavoidable aversive outcomes.
Fig. 4: Opposite responses to aversive cues in BA and NAc in simultaneous recordings of VTA dopamine axons or dopamine release.
Fig. 5: Individual VTADABA axons are activated by both cues predicting reward and cues predicting unavoidable aversive outcomes.
Fig. 6: Distinct, intermingled BA neurons acquire responses to cues predicting either reward or unavoidable aversive outcomes.
Fig. 7: VTADABA axons release glutamate.

Data availability

The data that support these findings are available from the authors upon reasonable request.

Code availability

The code used for these analyses is available from the authors upon reasonable request.

References

  1. 1.

    Burgess, C. R., Livneh, Y., Ramesh, R. N. & Andermann, M. L. Gating of visual processing by physiological need. Curr. Opin. Neurobiol. 49, 16–23 (2018).

  2. 2.

    Johnson, A. W., Gallagher, M. & Holland, P. C. The basolateral amygdala is critical to the expression of pavlovian and instrumental outcome-specific reinforcer devaluation effects. J. Neurosci. 29, 696–704 (2009).

  3. 3.

    Tye, K. M. & Janak, P. H. Amygdala neurons differentially encode motivation and reinforcement. J. Neurosci. 27, 3937–3945 (2007).

  4. 4.

    O’Neill, P.-K., Gore, F. & Salzman, C. D. Basolateral amygdala circuitry in positive and negative valence. Curr. Opin. Neurobiol. 49, 175–183 (2018).

  5. 5.

    Grewe, B. F. et al. Neural ensemble dynamics underlying a long-term associative memory. Nature 543, 670–675 (2017).

  6. 6.

    Johansen, J. P. et al. Hebbian and neuromodulatory mechanisms interact to trigger associative memory formation. Proc. Natl Acad. Sci. USA 111, E5584–E5592 (2014).

  7. 7.

    de Oliveira, A. R. et al. Conditioned fear is modulated by D2 receptor pathway connecting the ventral tegmental area and basolateral amygdala. Neurobiol. Learn. Mem. 95, 37–45 (2011).

  8. 8.

    Fadok, J. P., Dickerson, T. M. K. & Palmiter, R. D. Dopamine is necessary for cue-dependent fear conditioning. J. Neurosci. 29, 11089–11097 (2009).

  9. 9.

    Esber, G. R. et al. Attention-related Pearce–Kaye–Hall signals in basolateral amygdala require the midbrain dopaminergic system. Biol. Psychiatry 72, 1012–1019 (2012).

  10. 10.

    Bissière, S., Humeau, Y. & Lüthi, A. Dopamine gates LTP induction in lateral amygdala by suppressing feedforward inhibition. Nat. Neurosci. 6, 587–592 (2003).

  11. 11.

    Yu, K. et al. The central amygdala controls learning in the lateral amygdala. Nat. Neurosci. 20, 1680–1685 (2017).

  12. 12.

    Tye, K. M. et al. Methylphenidate facilitates learning-induced amygdala plasticity. Nat. Neurosci. 13, 475–481 (2010).

  13. 13.

    Menegas, W., Akiti, K., Amo, R., Uchida, N. & Watabe-Uchida, M. Dopamine neurons projecting to the posterior striatum reinforce avoidance of threatening stimuli. Nat. Neurosci. 1, 1421–1430 (2018).

  14. 14.

    Menegas, W., Babayan, B. M., Uchida, N. & Watabe-Uchida, M. Opposite initialization to novel cues in dopamine signaling in ventral and posterior striatum in mice. eLife 6, e21886 (2017).

  15. 15.

    Groessl, F. et al. Dorsal tegmental dopamine neurons gate associative learning of fear. Nat. Neurosci. 21, 952–962 (2018).

  16. 16.

    Lammel, S., Lim, B. K. & Malenka, R. C. Reward and aversion in a heterogeneous midbrain dopamine system. Neuropharmacology 76, 351–359 (2014). Pt B.

  17. 17.

    Ehrlich, I. et al. Amygdala inhibitory circuits and the control of fear memory. Neuron 62, 757–771 (2009).

  18. 18.

    Calhoon, G. G. et al. Acute food deprivation rapidly modifies valence-coding microcircuits in the amygdala. Preprint at bioRxiv https://doi.org/10.1101/285189 (2018).

  19. 19.

    Zhang, X. & Li, B. Population coding of valence in the basolateral amygdala. Nature Commun. 9, 5195 (2018).

  20. 20.

    Burgess, C. R. et al. Hunger-dependent enhancement of food cue responses in mouse postrhinal cortex and lateral amygdala. Neuron 91, 1154–1169 (2016).

  21. 21.

    Ramesh, R. N., Burgess, C. R., Sugden, A. U., Gyetvan, M. & Andermann, M. L. Intermingled ensembles in visual association cortex encode stimulus identity or predicted outcome. Neuron 100, 900–915.e9 (2018).

  22. 22.

    Livneh, Y. et al. Homeostatic circuits selectively gate food cue responses in insular cortex. Nature 546, 611–616 (2017).

  23. 23.

    Scibilia, R. J., Lachowicz, J. E. & Kilts, C. D. Topographic nonoverlapping distribution of D1 and D2 dopamine receptors in the amygdaloid nuclear complex of the rat brain. Synapse 11, 146–154 (1992).

  24. 24.

    Lein, E. S. et al. Genome-wide atlas of gene expression in the adult mouse brain. Nature 445, 168–176 (2006).

  25. 25.

    Schultz, W., Dayan, P. & Montague, P. R. A neural substrate of prediction and reward. Science 275, 1593–1599 (1997).

  26. 26.

    Cohen, J. Y., Haesler, S., Vong, L., Lowell, B. B. & Uchida, N. Neuron-type-specific signals for reward and punishment in the ventral tegmental area. Nature 482, 85–88 (2012).

  27. 27.

    Pearce, J. M. & Hall, G. A model for Pavlovian learning: variations in the effectiveness of conditioned but not of unconditioned stimuli. Psychol. Rev. 87, 532–552 (1980).

  28. 28.

    Kim, J., Pignatelli, M., Xu, S., Itohara, S. & Tonegawa, S. Antagonistic negative and positive neurons of the basolateral amygdala. Nat. Neurosci. 19, 1636–1646 (2016).

  29. 29.

    Beyeler, A. et al. Organization of valence-encoding and projection-defined neurons in the basolateral amygdala. Cell Rep. 22, 905–918 (2018).

  30. 30.

    Wang, L. et al. The coding of valence and identity in the mammalian taste system. Nature 558, 127–131 (2018).

  31. 31.

    Parker, N. F. et al. Reward and choice encoding in terminals of midbrain dopamine neurons depends on striatal target. Nat. Neurosci. 19, 845–854 (2016).

  32. 32.

    Poulin, J.-F. et al. Mapping projections of molecularly defined dopamine neuron subtypes using intersectional genetic approaches. Nat. Neurosci. 21, 1260–1271 (2018).

  33. 33.

    Patriarchi, T. et al. Ultrafast neuronal imaging of dopamine dynamics with designed genetically encoded sensors. Science 360, eaat4422 (2018).

  34. 34.

    Howe, M. W. & Dombeck, D. A. Rapid signalling in distinct dopaminergic axons during locomotion and reward. Nature 535, 505–510 (2016).

  35. 35.

    Beyeler, A. et al. Divergent routing of positive and negative information from the amygdala during memory retrieval. Neuron 90, 348–361 (2016).

  36. 36.

    Roesch, M. R., Esber, G. R., Li, J., Daw, N. D. & Schoenbaum, G. Surprise! Neural correlates of Pearce–Hall and Rescorla–Wagner coexist within the brain. Eur. J. Neurosci. 35, 1190–1200 (2012).

  37. 37.

    Roesch, M. R., Calu, D. J., Esber, G. R. & Schoenbaum, G. Neural correlates of variations in event processing during learning in basolateral amygdala. J. Neurosci. 30, 2464–2471 (2010).

  38. 38.

    Tye, K. M., Cone, J. J., Schairer, W. W. & Janak, P. H. Amygdala neural encoding of the absence of reward during extinction. J. Neurosci. 30, 116–125 (2010).

  39. 39.

    Herry, C. et al. Switching on and off fear by distinct neuronal circuits. Nature 454, 600–606 (2008).

  40. 40.

    Young, A. M. & Rees, K. R. Dopamine release in the amygdaloid complex of the rat, studied by brain microdialysis. Neurosci. Lett. 249, 49–52 (1998).

  41. 41.

    Manassero, E., Renna, A., Milano, L. & Sacchetti, B. Lateral and basal amygdala account for opposite behavioral responses during the long-term expression of fearful memories. Sci. Rep. 8, 518 (2018).

  42. 42.

    Lee, S.-C. et al. Basolateral amygdala nucleus responses to appetitive conditioned stimuli correlate with variations in conditioned behaviour. Nat. Commun. 7, 12275 (2016).

  43. 43.

    Jikomes, N., Ramesh, R. N., Mandelblat-Cerf, Y. & Andermann, M. L. Preemptive stimulation of AgRP neurons in fed mice enables conditioned food seeking under threat. Curr. Biol. 26, 2500–2507 (2016).

  44. 44.

    Alhadeff, A. L. et al. A neural circuit for the suppression of pain by a competing need state. Cell 173, 140–152.e15 (2018).

  45. 45.

    Kröner, S., Rosenkranz, J. A., Grace, A. A. & Barrionuevo, G. Dopamine modulates excitability of basolateral amygdala neurons in vitro. J. Neurophysiol. 93, 1598–1610 (2005).

  46. 46.

    Kroener, S., Chandler, L. J., Phillips, P. E. M. & Seamans, J. K. Dopamine modulates persistent synaptic activity and enhances the signal-to-noise ratio in the prefrontal cortex. PLoS One 4, e6507 (2009).

  47. 47.

    Shabel, S. J. & Janak, P. H. Substantial similarity in amygdala neuronal activity during conditioned appetitive and aversive emotional arousal. Proc. Natl Acad. Sci. USA 106, 15031–15036 (2009).

  48. 48.

    Gore, F. et al. Neural representations of unconditioned stimuli in basolateral amygdala mediate innate and learned responses. Cell 162, 134–145 (2015).

  49. 49.

    Johansen, J. P., Tarpley, J. W., LeDoux, J. E. & Blair, H. T. Neural substrates for expectation-modulated fear learning in the amygdala and periaqueductal gray. Nat. Neurosci. 13, 979–986 (2010).

  50. 50.

    Madisen, L. et al. Transgenic mice for intersectional targeting of neural sensors and effectors with high specificity and performance. Neuron 85, 942–958 (2015).

  51. 51.

    Bäckman, C. M. et al. Characterization of a mouse strain expressing Cre recombinase from the 3’ untranslated region of the dopamine transporter locus. Genesis 44, 383–390 (2006).

  52. 52.

    Asaad, W. F. & Eskandar, E. N. A flexible software tool for temporally-precise behavioral control in Matlab. J. Neurosci. Methods 174, 245–258 (2008).

  53. 53.

    Klapoetke, N. C. et al. Independent optical excitation of distinct neural populations. Nat. Methods 11, 338–346 (2014).

  54. 54.

    Broussard, G. J. et al. In vivo measurement of afferent activity with axon-specific calcium imaging. Nat. Neurosci. 21, 1272–1280 (2018).

  55. 55.

    Dimidschstein, J. et al. A viral strategy for targeting and manipulating interneurons across vertebrate species. Nat. Neurosci. 19, 1743–1749 (2016).

  56. 56.

    Bocarsly, M. E. et al. Minimally invasive microendoscopy system for in vivo functional imaging of deep nuclei in the mouse brain. Biomed. Opt. Express 6, 4546–4556 (2015).

  57. 57.

    Tervo, D. G. R. et al. A designer AAV variant permits efficient retrograde access to projection neurons. Neuron 92, 372–382 (2016).

  58. 58.

    Resendez, S. L. et al. Visualization of cortical, subcortical and deep brain neural circuit dynamics during naturalistic mammalian behavior with head-mounted microscopes and chronically implanted lenses. Nat. Protoc. 11, 566–597 (2016).

  59. 59.

    Ting, J. T., Daigle, T. L., Chen, Q. & Feng, G. Acute brain slice methods for adult and aging animals: application of targeted patch clamp analysis and optogenetics. Methods Mol. Biol. 1183, 221–242 (2014).

  60. 60.

    Petreanu, L., Huber, D., Sobczyk, A. & Svoboda, K. Channelrhodopsin-2–assisted circuit mapping of long-range callosal projections. Nat. Neurosci. 10, 663–668 (2007).

  61. 61.

    Garfield, A. S. et al. A neural basis for melanocortin-4 receptor–regulated appetite. Nat. Neurosci. 18, 863–871 (2015).

  62. 62.

    Bonin, V., Histed, M. H., Yurgenson, S. & Reid, R. C. Local diversity and fine-scale organization of receptive fields in mouse visual cortex. J. Neurosci. 31, 18506–18521 (2011).

  63. 63.

    Mukamel, E. A., Nimmerjahn, A. & Schnitzer, M. J. Automated analysis of cellular signals from large-scale calcium imaging data. Neuron 63, 747–760 (2009).

  64. 64.

    Ziv, Y. et al. Long-term dynamics of CA1 hippocampal place codes. Nat. Neurosci. 16, 264–266 (2013).

  65. 65.

    Petreanu, L. et al. Activity in motor-sensory projections reveals distributed coding in somatosensation. Nature 489, 299–303 (2012).

  66. 66.

    Driscoll, L. N., Pettit, N. L., Minderer, M., Chettih, S. N. & Harvey, C. D. Dynamic reorganization of neuronal activity patterns in parietal cortex. Cell 170, 986–999.e16 (2017).

  67. 67.

    Matsumoto, H., Tian, J., Uchida, N. & Watabe-Uchida, M. Midbrain dopamine neurons signal aversion in a reward-context-dependent manner. eLife 5, e17328 (2016).

  68. 68.

    Washburn, M. & Moises, H. Electrophysiological and morphological properties of rat basolateral amygdaloid neurons in vitro. J. Neurosci. 12, 4066–4079 (1992).

  69. 69.

    Paxinos, G. & Franklin, K. B. J. The Mouse Brain in Stereotaxic Coordinates. 4th ed. (Academic Press, 2012).

Download references

Acknowledgements

We thank S. J. Lee, K. McGuire, J. Reggiani, A. Fratzl, M. Uchida, N. Uchida, J. Assad, L. Liang, Y. Livneh, J. Zaremba, S. Zhang and other members of the Andermann laboratory for useful feedback. We also thank N. Pettit and Y. Aponte for helpful advice regarding GRIN lenses. We thank R. Jozanovic, E. Bamberg, F. Finkel, T. Pottala, I. Shurnayte and L. Rupert for help with mouse training and histology. We thank J. Madara and H. Fenselau for assistance with brain slice electrophysiology. We thank L. Tian for providing dLight1.1 DNA plasmid, the Boston Children’s Hospital Viral Core for virus packaging and the GENIE Project, HHMI, for use of GCaMP6. Authors were supported by the NIH (grant nos. F32 DK112589 and 5T32NS007484-15 to A.L. and T32 5T32DK007516 to A.U.S.); NIH New Innovator Award grants (nos. DP2 DK105570 and R01 DK109930); a McKnight Scholar Award; a Pew Scholar Award; a Smith Family Foundation Award; grants from the Klarman Family Foundation and the American Federation for Aging Research; the Boston Nutrition and Obesity Research Center (grant no. P30 DK046200); and a Harvard Brain Science Initiative Bipolar Disorder Seed Grant, supported by Kent and Liz Dauten (M.L.A.).

Author information

A.L. and M.L.A. designed experiments and analyses and wrote the manuscript. A.L., O.A., H.K. and C.C. performed two-photon imaging experiments. A.L., C.C., V.D. and K.F. performed photometry experiments. A.L. performed brain slice electrophysiology experiments. A.L., O.A., C.C., V.F.-M. and K.F. performed surgical procedures. A.L. analyzed two-photon imaging data with assistance from H.K. and A.U.S. A.L. analyzed photometry data and electrophysiology data.

Correspondence to Mark L. Andermann.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Nature Neuroscience thanks Kay Tye and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Integrated supplementary information

Supplementary Figure 1 Go/NoGo behavior and GRIN lens locations.

a, Behavioral performance of trained mice (n = 15 sessions from 7 mice) used for recordings in Fig. 1. Left: percentage of correct responses across reward cue trials (‘hit rate’). Right: percentage of false alarms (incorrect licking after neutral or aversive cues) across all neutral and aversive cue trials. Mean ± s.e.m. b, Confirmation of locations of the surface of the GRIN lens within the BLA (horizontal colored lines) for recordings related to Figs. 1 and 6, determined from post-hoc histology (lenses are depicted at scale). The two-photon imaging fields of view were between 100 and 300 μm ventral to the lens surface. Coronal sections and coordinates relative to Bregma are based on Paxinos and Franklin’s mouse brain atlas (4th edition).

Supplementary Figure 2 Changes in pupil area and locomotion across states do not account for observed changes in BA activity.

a, Mean pupil area in the 2 s preceding a cue during hungry and sated states (normalized to maximum pupil area during the entire session across both states). Mean ± s.e.m. across 15 sessions from 7 mice. ** p = 0.0012, two-sided Wilcoxon sign-rank. b, Mean speed of locomotion in the 2 s preceding a cue during hungry and sated states (normalized to maximum locomotion during the entire session across both states). Mean ± s.e.m. across 15 sessions from 7 mice. *** p = 0.0006, two-sided Wilcoxon sign-rank. c, Histograms of pupil areas in the 2 s preceding onset of each cue during hungry and sated states from an example session before (left) and after (right) matching trials for pupil area (see Methods). Note that following matching of trials, distributions of pupil areas are now completely overlapping. Insets show moments of small (left) and large (right) pupil diameter (bright ellipse against a dark background, as infra-red laser light emitted from the pupil was used for tracking, see Methods). d, Histograms of relative locomotion speed in the 2 s preceding cues during hungry and sated states from an example session before (left) and after (right) matching trials for locomotion. e, Comparison of cue responses (mean ± s.e.m.) of BA neurons across hungry and sated states before and after matching for pupil areas. The finding of attenuated cue responses following satiation persisted even when matching pupil area distributions across states. Sample size for RC: n = 66 activated neurons and n = 58 suppressed neurons; for AC-Av: n = 13 activated neurons and n = 26 suppressed neurons; for NC: n = 21 activated neurons and n = 25 suppressed neurons. Original AC-Av activated response magnitude, hungry vs. sated: ** p = 0.0089. Pupil-matched AC-Av activated response magnitude hungry vs. sated: ** p = 0.0061. For every other comparison, *** p < 0.0001, two-sided Wilcoxon sign-rank. Mean response magnitudes were analyzed separately for neurons that were significantly activated (red) or suppressed (blue) by cue presentation. f, Comparison of cue responses (mean ± s.e.m.) of BA neurons across hungry and sated states before and after matching for locomotion. The finding of attenuated cue responses following satiation persisted even when matching locomotion distributions across states. Sample sizes are same as in panel e. Original AC-Av activated response magnitude, hungry vs. sated: ** p = 0.0012. Locomotion-matched AC-Av activated response magnitude hungry vs. sated: * p = 0.019. Original NC activated response magnitude hungry vs. sated: *** p < 0.0001. Locomotion-matched NC activated response magnitude, hungry vs. sated: ** p = 0.0026. Original NC suppressed response magnitude, hungry vs. sated: *** p < 0.0001. Locomotion-matched NC suppressed response magnitude, hungry vs. sated: * p = 0.027. For every other comparison, *** p < 0.0001, two-sided Wilcoxon sign-rank.

Supplementary Figure 3 D1 receptor expression in basal amygdala.

a, A cross of a red fluorescent reporter mouse with a D1 receptor Cre line (left) or a D2 receptor Cre line (right) provided initial evidence of dense D1 receptor mRNA expression and sparse D2 receptor mRNA expression across both anterior basal amygdala (aBA) and posterior basal amygdala (pBA) neurons, in sagittal sections (data modified from Allen Institute Brain Atlas; D1 receptor: image 3 [http://connectivity.brain-map.org/transgenic/experiment/81034295]; D2 receptor: image 1 [http://connectivity.brain-map.org/transgenic/experiment/100146223]). Because reporter fluorescence could reflect transient receptor expression during development, we also assessed mRNA expression levels in BA neurons using the Allen Brain Atlas (panel b). Scale bar = 0.5 mm. b, D1 receptor mRNA expression (left) and D2 receptor mRNA expression (right) in sagittal sections of BA. Note that expression of the D1 receptor was evident throughout anterior and posterior regions of BA, while D2 receptor expression was largely not detectable in BA (Allen Institute Brain Atlas; D1 receptor: image 3 [http://mouse.brain-map.org/experiment/show/71307280]; D2 receptor: image 3 [http://mouse.brain-map.org/experiment/show/81790728]). Scale bar = 0.5 mm. c, Left: mCherry-labeled VTA dopaminergic axons target BA but not lateral amygdala (LA) (for details, see Fig. 2a). Middle and right: coronal sections demonstrating detectable expression of D1 receptor mRNA (middle, arrow) but not of D2 receptor mRNA (right, arrow) in BA (Allen Institute Brain Atlas; D1 receptor: image 24 [http://mouse.brain-map.org/experiment/show/352]; D2 receptor: image 26 [http://mouse.brain-map.org/experiment/show/357]). Scale bar = 0.5 mm.

Supplementary Figure 4 Implant locations for VTADABA fiber photometry experiments.

a, Location of photometry fiber implants from all GCaMP6s photometry mice (horizontal colored lines) with basal amygdala (BA) implants, determined using post-hoc histology (lenses are depicted at scale): 8 mice with bilateral BA implants targeted either to anterior BA (aBA) or posterior BA (pBA) regions, and mice used for two GFP control experiments. Photometry data from these mice are shown in Fig. 2c–e, Fig. 3c–i, Supplementary Figs. 5–9. Coronal sections and coordinates relative to Bregma are based on Paxinos and Franklin’s mouse brain atlas (4th edition). ‘X’ marks a fiber implant excluded from subsequent analyses due to mis-targeting. b, Location of photometry fiber implants from all BA implanted axon-targeted GCaMP6s photometry mice for sham-satiation and home cage satiation experiments (horizontal colored lines) determined from post-hoc histology (lenses are depicted at scale): 8 mice. Photometry data from these mice are shown in Figs. 2f–h and 3j–l. Coronal sections and coordinates relative to Bregma are based on Paxinos and Franklin’s mouse brain atlas (4th edition). c, Location of photometry fiber implants from all GCaMP6s photometry mice for comparing nucleus accumbens (NAc) vs. BA responses (horizontal colored lines) determined from post-hoc histology (lenses are depicted at scale): 4 mice with ipsilateral BA and NAc implants. Photometry data from these mice are shown in Fig. 4b–d. Coronal sections and coordinates relative to Bregma are based on Paxinos and Franklin’s mouse brain atlas (4th edition). d, Location of photometry fiber implants from all dLight1.1 photometry mice for comparing NAc vs. BA responses (horizontal colored lines) determined from post-hoc histology (lenses are depicted at scale): 4 mice with ipsilateral BA and NAc implants. Photometry data from these mice are shown in Fig. 4e–g. Coronal sections and coordinates relative to Bregma are based on Paxinos and Franklin’s mouse brain atlas (4th edition). e, Location of photometry fiber implants from all BA implanted dLight1.1 photometry mice for Chrimson optogenetic stimulation of VTA dopamine axon terminals (horizontal colored lines) determined from post-hoc histology (lenses are depicted at scale): 4 mice. Photometry data from these mice are shown in Fig. 4i. Coronal sections and coordinates relative to Bregma are based on Paxinos and Franklin’s mouse brain atlas (4th edition).

Supplementary Figure 5 VTADABA axon responses during training and following completion of training.

a, Left: all single-trial responses (rows) from 8 mice during Pavlovian training of RC paired with Ensure delivery (“During training” refers to sessions during the first 3 days of training before introducing aversive trials), sorted by onset of first lick following RC onset (blue ticks). Green ticks: time of Ensure delivery. Responses were mostly evident during Ensure consumption. This was especially pronounced when Ensure delivery was unexpected, as seen in rows with longer latencies to first lick. Right: behavioral performance of mice during training (n = 8 sessions, one session/mouse) used for recordings in panels c-e, below. Percentage of reward cue trials with licking during the cue (‘hit rate’) and percentage of false alarms (incorrect licking after neutral cues or during blank trials) were both high early in learning, indicative of poor discrimination between cues. Mean ± s.e.m. b, Left: all single-trial responses (rows) from 10 mice following completion of training, sorted by onset of first lick following RC onset (blue ticks). Green ticks: time of Ensure delivery. Responses were locked to cue onset, not to motor response onset. Right: behavioral performance of trained mice (n = 10 sessions, one session/mouse) used for recordings in Fig. 2d, e and in panels c-e, below. Left: percentage of correct responses across reward cue trials (‘hit rate’). Right: percentage of false alarms (incorrect licking after neutral or aversive cues) across all neutral or aversive trials. Mean ± s.e.m. c, Mean response timecourses of VTADABA axons during training and following completion of training on task. Response timecourse of Ensure delivery related activity was obtained by subtracting a monoexponential fit of the cue response. The window used for analysis of Ensure delivery responses is indicated by a blue bar. Error bars: s.e.m. across mice. Z: Z-score. d, Comparison of cue responses and Ensure delivery responses during training vs. following training. Mean ± s.e.m. *** p < 0.0001, two-sided Wilcoxon rank-sum. e, The outcome response bias index shifted from positive to negative following training, indicating a shift in response from the time of Ensure delivery to the time of cue onset. Mean ± s.e.m. *** p < 0.0001, two-sided Wilcoxon rank-sum. f, Population response, averaged across those AC-Av trials early in training that involved licking during the behavioral response window (“false alarm”, thereby eliciting quinine delivery) or across trials without licking (“correct reject”, thereby eliciting no outcome). A monoexponential fit was subtracted to isolate the response magnitude to quinine delivery or no outcome. The window used for analysis of outcome responses is indicated by a blue bar. Mean ± s.e.m., n = 10 mice. Z: Z-score. g, Comparison of AC-Av cue and outcome responses for false alarm trials vs. correct reject trials. Mean ± s.e.m., n = 10 mice, * p = 0.04, two-sided Wilcoxon sign-rank. h, Population response, averaged across those NC trials early in training that involved licking during the behavioral response window (“false alarm”) or across trials without licking (“correct reject”). In both cases, no outcome occurs following cue offset. A monoexponential fit was subtracted to isolate the response magnitude during the outcome window. The window used for analysis of outcome responses is indicated by a blue bar. Mean ± s.e.m., n = 10 mice. Z: Z-score. i, Comparison of NC cue and outcome responses during false alarm trials vs. correct reject trials. Mean ± s.e.m., n = 10 mice.

Supplementary Figure 6 VTADABA axon responses during training with cues predicting air puff or tail shock.

a, Left: schematic of delivery of air puff aversive outcome during the head-fixed Go/NoGo visual discrimination task. Behavioral performance of trained mice (n = 6 sessions, 1 session/mouse) remained high on the day preceding air puff training when mice were already trained on avoidable quinine task (day = −1) and during 4 days of training with air puff. Middle: percentage of correct responses across RC trials (‘hit rate’) during hungry and sated states. Right: percentage of false alarms (incorrect licking after neutral or aversive cues) across all NC or AC-Un trials during hungry and sated states. Mean ± s.e.m. b, Left: schematic of delivery of tail shock aversive outcome during the head-fixed Go/NoGo visual discrimination task. Behavioral performance of trained mice (n = 6 sessions, 1 session/mouse) during 5 days of training with tail shock remained high. Middle: percentage of correct responses across RC trials (‘hit rate’) during hungry and sated states. Right: percentage of false alarms (incorrect licking after neutral or aversive cues) across all NC or AC-Un trials during hungry and sated states. Mean ± s.e.m. c, Mean response timecourse on first sessions with aversive cue pairing with air puff (top) or tail shock (bottom) (s.e.m. across 6 mice). Trials in which we omitted the air puff (top) or tail shock (bottom) delivery did not have outcome responses and were subtracted from the trials with delivery of air puff (top) or tail shock (bottom) in order to isolate the response specific to the outcome. d, Outcome and cue response magnitudes (mean ± s.e.m., n = 6 mice) on the first day of training with air puff delivery (left) or tail shock delivery (right) separated in 6 consecutive bins (~ 10 trials per bin). Within the first day of training, outcome response magnitudes decreased whereas cue responses increased. Note that training with tail shock delivery followed training with air puff and therefore cue responses began at an already elevated magnitude. e, Outcome and cue response magnitudes (mean ± s.e.m., n = 6 mice) across days of training with air puff (left) or tail shock (right). Across days, outcome responses continued to decrease while cue responses increased. f, An outcome bias was initially present on the first day of training with air puff delivery, but this bias showed a significant decrease following days of training. Mean ± s.e.m., n = 6 mice, * p = 0.03, two-sided Wilcoxon sign-rank. g, Across all days of training with air puff (left) and tail shock (right) delivery, the RC response magnitude remained stable. Mean ± s.e.m., n = 6 mice.

Supplementary Figure 7 Behavior during tasks involving unavoidable aversive air puff or tail shock and related photometry controls.

a, Percentage change in eye closure behavior for task involving unavoidable air puff, relative to pre-cue baseline period for reward cue (RC), neutral cue (NC), and aversive cue (AC-Un) trials. Error bars: s.e.m. across 6 mice. See Methods for details. b, Left: percentage change in eye closure during each cue period. Right: percentage change in eye closure after air puff delivery (n = 6, *** p = 0.0008, ** p = 0.0055, two-sided t-test against null distribution with mean of zero). Mean ± s.e.m. across 6 mice. c, Percentage change in eye closure for air puff experiments, relative to pre-cue baseline period from hungry trials. Note that persistent defensive eye closure increased with satiety. Note that VTADABA axon responses to unavoidable aversive outcome predicting cue (AC-Un) did not decrease with satiety, indicating that vision was not impaired in the contralateral eye receiving visual stimuli. Mean ± s.e.m., n = 6 mice. d, Percentage increase in pre-cue persistent eye closure during sated trials relative to hungry trials. (n = 6, * p = 0.011, two-sided t-test again null distribution with mean of zero). e, Mean running behavior for task involving unavoidable tail shock during second session of training for reward cue (RC), neutral cue (NC), and aversive cue (AC-Un) trials. Error bars: s.e.m. across 8 mice (1 session/mouse). f, Left: change in speed during cue period relative to pre-cue baseline (n = 8 mice, * p = 0.028, two-sided t-test again null distribution with mean of zero). Right: change in speed during tail shock delivery relative to pre-cue baseline (n = 8 mice, * p = 0.011, two-sided t-test again null distribution with mean of zero). Mean ± s.e.m. g, Left: mean cue responses in fiber photometry recordings from VTADABA axons (4th and final day of air puff exposure) in hungry mice and sated mice. Error bars: s.e.m. across 6 mice. Note that response to AC-Un progressively decreased across days (see also Supplementary Fig. 6; we speculate that this may reflect a decrease in motivational salience of the cue as mice develop learned helplessness), but continues to increase in sated sessions. Right: mean VTADABA cue responses (first session of tail shock exposure) in hungry mice. Error bars: s.e.m. across 6 mice. Z: Z-score. Note that the same mice previously learned the AC-Un association with unavoidable air puff and that the AC-Un response is rapidly restored upon exposure to the same visual cue now predicting unavoidable tail shock. h, Top: example GCaMP6s fiber photometry recording from VTADABA axons showing mean response and single-trial responses to reward cue (RC; left, mean ± s.e.m., n = 56 trials from 1 mouse) and to cues predicting unavoidable air puff delivery (AC-Un; right, mean ± s.e.m., n = 32 trials from 1 mouse). Note that Ensure reward delivery or air puff delivery occur following the second vertical dashed line. Bottom: mean and single-trial photometry traces from GFP controls during reward cues (RC; left, mean ± s.e.m., n = 69 trials from 1 mouse) and during cues predicting unavoidable air puff delivery (AC-Un; right; mean ± s.e.m., n = 30 trials from 1 mouse) do not show the same transient events observed in GCaMP6s recordings. Images at right: example histological reconstructions of fiber placements over BA and fluorescence of VTADABA axons for GCaMP6s (top) and GFP (bottom) experiments.

Supplementary Figure 8 VTADABA axons show an increase in tonic inter-trial activity with reward expectation and a decrease in activity upon reward/shock omission, but only when omission is preceded by a previously rewarded/shocked trial, respectively.

a, Left: mean timecourse for tail shock omission trials that were not preceded by trials with tail shock delivery in the hungry state. Right: mean timecourse for tail shock omission trials that were preceded by trials with tail shock delivery in the hungry state. Error bars: s.e.m. across 13 sessions from 8 mice. The window used for analysis of outcome responses is indicated by a blue bar. b, Left: cue response magnitude for tail shock omission trials that were not preceded (light gray) or were preceded (dark gray) by tail shock were not significantly different (two-sided Wilcoxon sign-rank). Right: mean activity after tail shock omission was not different between trials that were or were not preceded by tail shock in hungry mice. Error bars: s.e.m. across 13 sessions from 8 mice. n.s. = not significant. c, Left: mean timecourse for tail shock omission trials that were not preceded by trials with tail shock delivery in the sated state. Right: mean timecourse for tail shock omission trials that were preceded by trials with tail shock delivery in the sated state. Error bars: s.e.m. across 13 sessions from 8 mice. The window used for analysis of outcome responses is indicated by a blue bar. d, Left: cue response magnitude for tail shock omission trials that were not preceded by tail shock (light gray) were larger than those that were preceded by tail shock (dark gray) in sated mice. Right: mean decrease in activity after tail shock omission was significantly greater when the prior trial contained a tail shock in sated mice, potentially due to increased expectation of additional tail shocks following tail shock receipt in the sated state when mice may be in a more defensive state. Error bars: s.e.m. across 13 sessions from 8 mice. Comparisons between cue responses: ** p = 0.0012; comparison between outcome window responses: ** p = 0.0081, two-sided Wilcoxon sign-rank. e, Modified task design for assessing effects of reward expectation and omission on activity of VTADABA axons. For experiments presented in this Supplementary Figure 8e-h only, we used a modified task that only involved presentation of a single type of cue – the reward cue (RC). Reward cues were followed either by reward delivery (67% of trials) or by reward omission (33% of trials). f, Left: responses were persistently elevated for many seconds following termination of reward consumption (well beyond the ~1-s timescale associated with decay of GCaMP6s fluorescence) for trials that followed a non-rewarded trial. Right: reward trials following a previously rewarded trial did not show persistently elevated activity following cue offset (right of the shaded box) relative to pre-cue baseline. Error bars: s.e.m. across 8 mice. Note that response timecourses were re-zeroed to pre-cue baseline. g, Left: mean timecourse of reward omission trials that were not preceded by a rewarded trial. Right: mean timecourse of reward omission trials that were preceded by a rewarded trial. Error bars: s.e.m. across 8 mice. Note that response timecourses were re-zeroed to pre-cue baseline. The window used for analysis of outcome responses is indicated by a blue bar. h, Left: mean cue response for reward omission trials that were not preceded (light gray) or were preceded (dark gray) by reward were not significantly different (p > 0.05, Wilcoxon sign-rank). Right: mean decrease in activity after reward omission was significantly greater when the prior trial was rewarded, potentially due to increased expectation of additional rewards following reward receipt. Error bars: s.e.m. across 8 mice. * p = 0.015, two-sided Wilcoxon sign-rank.

Supplementary Figure 9 VTADABA axon cue responses are similar across anterior BA and posterior BA.

a, Mean timecourse during trials involving a reward cue (RC), a neutral cue (NC), or an aversive cue predicting unavoidable air puff (AC-Un), for anterior BA (aBA) and posterior BA (pBA) recordings. Error bars: s.e.m. across 10 sessions from 6 mice. b, Cue response magnitude of VTADABA recordings in aBA or pBA (mean ± s.e.m., n = 10 sessions from 6 mice). **p = 0.002, two-sided Wilcoxon sign-rank. c, Mean timecourse during trials involving a reward cue (RC), a neutral cue (NC), or an aversive cue predicting unavoidable tail shock (AC-Un), for aBA and pBA recordings. Error bars: s.e.m. across 11 sessions from 8 mice. d, Cue response magnitude of VTADABA axon recordings from anterior vs. posterior BA (mean ± s.e.m., n = 11 sessions from 8 mice). *** p = 0.0009, two-sided Wilcoxon sign-rank. e, Mean response timecourses of VTADABA axons during training and following completion of training on avoidable quinine task, from aBA (left) or pBA (right). Response timecourse of Ensure delivery related activity was obtained by subtracting a monoexponential fit of the cue response. The window used for analysis of Ensure delivery responses is indicated by a blue bar. Error bars: s.e.m. across 8 “during training” and 6 “trained” mice. Z: Z-score. f, Cue and outcome response magnitudes of VTADABA axon recordings from aBA or pBA during training (mean ± s.e.m., n = 8 mice) vs. following training (n = 6). During training, no differences were found between aBA and pBA (p > 0.05, Wilcoxon rank-sum). Following training, similar to Supplementary Fig 9b,d, RC response magnitudes were greater in aBA. * p = 0.03, two-sided Wilcoxon rank-sum. g, Mean response timecourses of VTADABA axons during first sessions of training with air puff (top) or tail shock (bottom) in aBA (left) and pBA (right). Response timecourse of aversive outcome delivery-related activity was obtained by subtracting a monoexponential fit of the cue response. The window used for analysis of air puff or tail shock delivery responses is indicated by a blue bar. Error bars: s.e.m. across 6 mice. h, Outcome response magnitudes of VTADABA axon recordings from aBA or pBA during first day of air puff training (top) or first day of tail shock training (bottom). Responses to delivery of air puff or tail shock were not different between aBA and pBA (mean ± s.e.m., n = 6 mice), two-sided Wilcoxon rank-sum.

Supplementary Figure 10 Individual VTADABA axon cue responses are restricted to the reward cue during a task involving aversive cues predicting passively avoidable quinine.

a, Left: location of GRIN lens implants in medial BA (mBA) from VTADABA axon imaging experiments (horizontal colored lines; lenses are depicted at scale), determined from post-hoc histology (n = 4 mice). Right: images of injection site showing GCaMP6s-labeled dopamine neurons in VTA from all 4 mice. b, Top: mean response of individual axons (rows) to presentation of the reward cue (RC), aversive cue predicting avoidable quinine (AC-Av), and neutral cue (NC) (n = 38 axons, 7 fields of view from 4 mice). Bottom: population responses (mean ± s.e.m. across 38 axons). c, Top: percent of all axons with significant cue responses during hungry runs (RC: 16/38 neurons; AC-Av: 0/38; NC: 1/38). Bottom: percent of cue responsive axons preferring a given cue. All data are from hungry runs. d, Mean RC response for activated axons. Lines: individual axon responses across hunger and satiety (mean ± s.e.m., n = 16 axons, *** p = 0.0004, two-sided Wilcoxon sign-rank). e, Response bias of individual VTADABA axons activated by RC and/or AC from 4 mice following training with avoidable quinine (left, mouse 1: n = 18 axons, mouse 2: n = 6 axons, mouse 3: n = 5 axons, mouse 4: n = 9 axons), following training with air puff (middle, mouse 1: n = 14 axons, mouse 2: n = 6 axons, mouse 3: n = 2 axons, mouse 4: n = 8 axons), and following training with tail shock (right, mouse 1: n = 16 axons, mouse 2: n = 7 axons, mouse 3: n = 8 axons, mouse 4: n = 10 axons). Note that axons recorded from all four mice were initially biased to the RC and shifted towards equal preference of RC and AC following training with air puff and tail shock. Mean ± s.e.m. f, RC (left, n = 8 axons, ** p = 0.0078) and AC-Un (right, n = 35 axons, ** p = 0.0026) response magnitudes of VTADABA axons during hungry vs. sated states for the task involving unavoidable tail shocks, after subtraction of NC responses. Mean ± s.e.m., two-sided Wilcoxon sign-rank.

Supplementary Figure 11 Additional analyses of BA cell body responses during the task involving cues predicting unavoidable tail shock.

a, Left: correlation between magnitude of responses to the reward cue (RC) and to the aversive cue predicting unavoidable tail shock (AC-Un) across basal amygdala (BA) neurons (n = 186 neurons; Pearson’s r: −0.20, p = 0.0055). Middle: correlation between RC and NC response magnitude (n = 186 neurons; Pearson’s r: 0.36, p < 0.0001). Right: correlation between NC and AC-Un response magnitude (n = 186 neurons; Pearson’s r: 0.40, p < 0.0001). b, Heatmap showing mean cue response timecourses of BA neurons (rows) from sated mice. Rows are sorted by response magnitude during earlier runs in the same session, in hungry mice (see Fig. 5c; n = 482 neurons, 9 fields of view from 4 mice). Vertical dashed lines demarcate visual stimulus onsets and offsets. Horizontal lines demarcate sorting of axons by preferred cue (cue with the largest absolute value response). c, Left: response magnitude of neurons that were significantly activated by the RC during hunger or satiety (mean ± s.e.m., n = 98 neurons in hungry mice, n = 79 neurons in sated mice, *** p < 0.0001, two-sided Wilcoxon rank sum). Right: response magnitude of neurons that were significantly suppressed by the RC during hunger or satiety (mean ± s.e.m., n = 83 neurons in hungry mice, n = 21 neurons in sated mice, *** p < 0.0001, two-sided Wilcoxon rank sum). d, Left: response magnitude of neurons that were significantly activated by the AC-Un during hunger or satiety (mean ± s.e.m., n = 136 neurons in hungry mice, n = 108 neurons in sated mice, n.s. = not significant, p = 0.2, two-sided Wilcoxon rank sum). Right: response magnitude of neurons that were significantly suppressed by AC-Un during hunger or satiety (mean ± s.e.m., n = 82 neurons in hungry mice, n = 38 neurons in sated mice, * p = 0.013, two-sided Wilcoxon rank sum). e, Left: percent of BA neurons that were significantly activated by cues in hungry (RC: 98/482 neurons; AC-Un: 136/482 neurons; NC: 98/482 neurons) and sated mice (RC: 79/482 neurons; AC-Un: 108/482 neurons; NC: 89/482 neurons). There was a higher percentage of neurons significantly activated by AC-Un in hungry vs. sated mice (* p = 0.019, one-tailed binomial proportion test). Right: percent of BA neurons with a given cue preference (activated) in hungry (RC: 81/287 neurons; AC-Un: 92/287 neurons; NC: 8/287 neurons) and sated states (RC: 50/205 neurons; AC-Un: 93/205 neurons; NC: 7/205 neurons). There was a higher percentage of BA neurons preferentially activated by the AC-Un in the sated state (** p = 0.0013, one-tailed binomial proportion test). f, Left: percent of BA neurons that were significantly suppressed by cues in hungry (RC: 83/482 neurons; AC-Un: 82/482 neurons; NC: 69/482 neurons) and sated states (RC: 21/482 neurons; AC-Un: 38/482 neurons; NC: 26/482 neurons). There was a higher percentage of suppressed BA neurons in hungry mice (*** p < 0.0001, one-tailed binomial proportion test). Right: percent of BA neurons with a given cue preference (suppressed) in hungry (RC: 48/287 neurons; AC-Un: 53/287 neurons; NC: 5/287 neurons) and sated states (RC: 10/205 neurons; AC-Un: 34/205 neurons; NC: 11/205 neurons). There was a higher percentage of BA neurons preferentially suppressed by the RC in the hungry state (*** p < 0.0001) and NC in the sated state (* p = 0.012, one-tailed binomial proportion test). g, Percent of BA neurons significantly activated (left) or suppressed (right) by any cue in hungry (n = 181 activated neurons; n = 106 suppressed neurons) vs. sated states (n = 150 activated neurons; n = 55 suppressed neurons). * p = 0.017, *** p < 0.0001, one-sided binomial proportion test.

Supplementary Figure 12 Changes in pupil area and locomotion across states do not account for observed changes in BA activity in mice trained with cue predicting rewards and cues predicting unavoidable tail shock.

a, Histograms of pupil areas in the 2 s preceding cues during hungry and sated states from an example session, before (left) and after (right) matching trials for pupil area (see Methods). Right: mean pupil area in the 2 s preceding a cue during hungry and sated states (normalized to maximum pupil area during that entire session across both states). Mean ± s.e.m. across 9 sessions from 4 mice. n.s.= not significant, p = 0.054, two-sided Wilcoxon sign-rank. b, Histograms of locomotion in the 2 s preceding cues during hungry and sated states from an example session before (left) and after (right) matching trials for locomotion. Right: mean locomotion in the 2 s preceding a cue during hungry and sated states (normalized to maximum locomotion during that entire session across both states). Mean ± s.e.m. across 9 sessions from 4 mice. ** p = 0.0039, two-sided Wilcoxon sign-rank. c, Comparison of cue responses (mean ± s.e.m.) of BA neurons that were significantly responsive in the hungry state. Attenuation of cue responses following satiation persisted after matching pupil area distributions across states. Sample size for RC: n = 98 activated neurons and n = 83 suppressed neurons; for AC-Un: n = 136 activated neurons and n = 82 suppressed neurons; for NC: n = 98 activated neurons and n = 69 suppressed neurons. For all comparisons: *** p < 0.0001, two-sided Wilcoxon sign-rank. d, Comparison of cue responses (mean ± s.e.m.) of BA neurons that were significantly responsive in the hungry state. Attenuation of cue responses following satiation persisted after matching locomotion distributions across states. Sample sizes are identical to panel c. For all comparisons: *** p < 0.0001, two-sided Wilcoxon sign-rank. e, Comparison of cue responses (mean ± s.e.m.) of BA neurons that were significantly responsive in the sated state. Enhancement of cue responses following satiation persisted after matching pupil area distributions across states. Sample size for RC: n = 79 activated neurons and n = 21 suppressed neurons, for AC-Un: n = 108 activated neurons and n = 38 suppressed neurons, for NC: n = 89 activated neurons and n = 26 suppressed neurons. Original AC-Un activated response magnitude, hungry vs. sated: ** p = 0.0031. Pupil-matched AC-Un activated response magnitude hungry vs. sated: *** p < 0.0001. Original AC-Un suppressed response magnitude, hungry vs. sated: * p = 0.013. Pupil-matched AC-Un suppressed response magnitude hungry vs. sated: ** p = 0.0029. Original NC activated response magnitude hungry vs. sated: *** p < 0.0001. Pupil-matched NC activated response magnitude,, hungry vs. sated: *** p < 0.0001. Two-sided Wilcoxon sign-rank. f, Comparison of cue responses (mean ± s.e.m.) of BA neurons that were significantly responsive in the sated state. Enhancements of cue responses following satiation persisted after matching locomotion distributions across states. Sample sizes are identical to panel e. Original AC-Un activated response magnitude, hungry vs. sated: ** p = 0.0031. Locomotion-matched AC-Un activated response magnitude hungry vs. sated: *** p < 0.0001. Original AC-Un suppressed response magnitude, hungry vs. sated: * p = 0.013. Locomotion-matched AC-Un suppressed response magnitude, hungry vs. sated: * p = 0.017. Original NC activated response magnitude, hungry vs. sated: *** p < 0.0001. Locomotion-matched NC activated response magnitude, hungry vs. sated: p = 0.1. Two-sided Wilcoxon sign-rank.

Supplementary Figure 13 Similar findings in NAc-projecting BA neurons as in nearby unlabeled BA neurons.

a, Schematic depicting retrograde labeling of nucleus accumbens (NAc)-projecting basal amygdala (BA) neurons and two-photon imaging of these labeled neurons (see Methods for additional details). b, Left top: mean image of GCaMP6s fluorescence expression in excitatory BA neurons in an example field of view. Left bottom: mean image from the same field of view as above, showing mCherry fluorescence in the nuclei of retrogradely labeled NAc-projecting BA neurons. Right: overlaying the GCaMP6s and mCherry fluorescence within the yellow box allows identification of neurons with and without red-labeled nuclei. These representative images from a single animal were similar to those observed in the 10 other mice included in the analyses in this figure. c, Top left: heatmap with rows depicting mean responses of NAc-projecting BA neurons (n = 182 neurons, 9 fields of view from 4 mice) from the task involving an aversive cue predicting unavoidable tail shock (AC-Un). Top right: responses of the same neurons as in top left, but during sated trials. In this panel, neurons (rows) are sorted by response magnitude during the sated state, to facilitate comparison of the two subpopulations of BA neurons. Bottom left: heatmap with rows depicting mean responses of BA neurons that are not retrogradely labeled with nuclear-localized mCherry (n = 300 neurons, 9 fields of view from 4 mice), from the task involving unavoidable tail shock. Bottom right: responses of the same neurons as in bottom left but during sated trials and sorted by response magnitude in the sated state. Note the similarity in response profiles across the two subpopulations. These data, combined across NAc-projecting BA neurons and other BA neurons, are also shown in Fig. 6c. d, Top left: Response magnitude was not significantly different in NAc-projecting BA neurons activated by the RC (‘1’, n = 39) vs. non-labeled BA neurons activated by the RC (‘2’, n = 59). Top right: similarly, response magnitude was not significantly different in NAc-projecting BA neurons suppressed by the RC (‘1’, n = 43) vs. non-labeled BA neurons suppressed by the RC (‘2’, n = 40). Bottom left: response magnitude was not significantly different in NAc-projecting BA neurons activated by the AC-Un (‘1’, n = 55) vs. non-labeled BA neurons activated by the AC-Un (‘2’, n = 81). Bottom right: similarly, response magnitude was not significantly different in NAc-projecting BA neurons suppressed by the AC-Un (‘1’, n = 33) vs. non-labeled BA neurons suppressed by the AC-Un (‘2’, n = 49). Mean ± s.e.m., p > 0.05, two-sided Wilcoxon rank sum. e, Top left: percent of significantly responsive neurons from either NAc-projecting BA neurons (‘1’) and for non-labeled BA neurons (‘2’) during hungry trials. Top right: percent of significantly responsive neurons from either NAc-projecting BA neurons (‘1’) or from non-labeled neurons (‘2’) during sated trials. Bottom left: percent of neurons with a given cue preference of NAc-projecting BA neurons (‘1’) or from non-labeled neurons (‘2’) during hungry trials. Bottom right: percent of neurons with a given cue preference for NAc-projecting BA neurons (‘1’) vs. non-labeled BA neurons (‘2’) during hungry trials. Red: activated neurons. Blue: suppressed neurons. f, Top left: heatmap with rows depicting mean responses of NAc-projecting BA neurons (n = 155 neurons, 15 fields of view from 7 mice) from task using avoidable quinine (AC-Av). Top right: responses of the same neurons with same sorting as in top left, but during sated trials. Bottom left: rows depicting mean responses of BA neurons that lack nuclear mCherry labeling (n = 205 neurons, 15 fields of view from 7 mice), from the task involving avoidable quinine. Bottom right: responses of the same neurons with the same sorting as in bottom left but during sated trials. Note: unsorted data combined from NAc-projecting and other BA neurons also shown in Fig. 1h. g, Left: for the above task involving reward cues (RC) and cues predicting passively avoidable quinine (AC-Av), RC response magnitude for NAc-projecting BA neurons (‘1’, n = 34) that were activated by cue presentation (red) showed significantly weaker response magnitudes compared to non-labeled BA neurons (‘2’, n = 32). Mean ± s.e.m., ** p = 0.0084, two-sided Wilcoxon rank sum. Right: RC response magnitude for NAc-projecting BA neurons (‘1’, n = 32) suppressed by cue presentation (blue) showed significantly weaker response magnitudes compared to non-labeled BA neurons (‘2’, n = 26). Mean ± s.e.m., ** p = 0.0077, two-sided Wilcoxon rank sum. h, Top: percent of significantly responsive NAc-projecting BA neurons (‘1’) or significantly responsive non-labeled BA neurons (‘2’). Bottom: cue preference of significantly responsive NAc-projecting BA neurons (‘1’) and of significantly responsive, non-labeled BA neurons (‘2’). Red: activated neurons. Blue: suppressed neurons.

Supplementary Figure 14 Analyses of BA cell body cue responses in anterior and posterior subregions of BA.

a, Left: comparison of percentage of BA neurons in anterior BA (aBA; n = 321 total neurons) vs. posterior BA (pBA; n = 161 total neurons) that were significantly activated (top left; RC: n = 76 aBA neurons vs. 22 pBA neurons, ** p = 0.005; AC-Un: n = 100 aBA neurons vs. 36 pBA neurons, * p = 0.021; NC: n = 78 aBA neurons vs. 20 pBA neurons, ** p = 0.001) or suppressed (bottom left; RC: n = 58 aBA neurons vs. 25 pBA neurons, p = 0.24; AC-Un: n = 38 aBA neurons vs. 44 pBA neurons, *** p < 0.0001; NC: n = 44 aBA neurons vs. 25 pBA neurons, p = 0.29) in hungry mice. Right: comparison of percentage of BA neurons in aBA (n = 196 total neurons) vs. pBA (n = 91 total neurons) that were preferentially activated (top right; RC: n = 67 aBA neurons vs. 14 pBA neurons, *** p = 0.0005; AC-Un: n = 62 aBA neurons vs. 30 pBA neurons, p = 0.41; NC: n = 7 aBA neurons vs. 1 pBA neurons, p = 0.1) or suppressed (bottom right; RC: n = 39 aBA neurons vs. 9 pBA neurons, * p = 0.017; AC-Un: n = 19 aBA neurons vs. 34 pBA neurons, *** p < 0.0001; NC: n = 2 aBA neurons vs. 3 pBA neurons, p = 0.09) by cues in hungry mice. One-sided binomial proportion test. b, Left: comparison of percentage of BA neurons in anterior BA (aBA; n = 321 total neurons) vs. posterior BA (pBA; n = 161 total neurons) that were significantly activated (top left; RC: n = 74 aBA neurons vs. 5 pBA neurons, *** p < 0.0001; AC-Un: n = 67 aBA neurons vs. 41 pBA neurons, p = 0.13; NC: n = 61 aBA neurons vs. 28 pBA neurons, p = 0.33) or suppressed (bottom left; RC: n = 21 aBA neurons vs. 0 pBA neurons, *** p = 0.0005; AC-Un: n = 27 aBA neurons vs. 11 pBA neurons, p = 0.27; NC: n = 21 aBA neurons vs. 5 pBA neurons, p = 0.06) in sated mice. Right: comparison of percentage of BA neurons in aBA (n = 147 total neurons) vs. pBA (n = 58 total neurons) that were preferentially activated (top right; RC: n = 49 aBA neurons vs. 1 pBA neurons, *** p < 0.0001; AC-Un: n = 53 aBA neurons vs. 40 pBA neurons, *** p < 0.0001; NC: n = 3 aBA neurons vs. 4 pBA neurons, * p = 0.042) or suppressed (bottom right; RC: n = 10 aBA neurons vs. 0 pBA neurons, * p = 0.02; AC-Un: n = 23 aBA neurons vs. 11 pBA neurons, p = 0.28; NC: n = 9 aBA neurons vs. 2 pBA neurons, p = 0.22) by cues in sated mice. One-sided binomial proportion test. c, Comparison of hungry vs. sated cue responses of aBA and pBA neurons that were significantly responsive in the hungry state. Attenuation of cue response magnitudes following satiation was observed in both aBA and pBA. Sample sizes are the same as reported for cue responsive neurons in panel a, * p = 0.01, *** p < 0.0001, two-sided Wilcoxon sign-rank. Mean ± s.e.m. d, Comparison of hungry vs. sated cue responses of aBA and pBA neurons that were significantly responsive in the sated state. Enhancement of the AC-Un and NC response magnitudes following satiation was observed in pBA, but not in aBA. Sample sizes are the same as reported for cue responsive neurons in panel b, * p = 0.027, *** p < 0.0001, two-sided Wilcoxon sign-rank. Mean ± s.e.m. e, Comparison of cue response magnitudes of aBA vs. pBA (left: activated; right: suppressed) neurons which were found to be responsive in the hungry state. The RC response was greater in aBA in hungry mice. Sample sizes are the same as reported for cue responsive neurons in panel a, * p = 0.017, *** p < 0.0001, two-sided Wilcoxon sign-rank. Mean ± s.e.m. f, Comparison of cue response magnitudes of aBA vs. pBA (left: activated; right: suppressed) neurons which were found to be responsive in the sated state. The AC-Un (*** p < 0.0001) and NC (*** p < 0.0001) activated responses were greater in aBA in hungry mice whereas the AC-Un (* p = 0.02) and NC (** p = 0.005) response was greater in pBA in sated mice. The AC-Un (** p = 0.001) and NC (* p = 0.022) suppressed responses were also greater in pBA in sated mice. Sample sizes are the same as reported for cue responsive neurons in panel a, two-sided Wilcoxon sign-rank. Mean ± s.e.m.

Supplementary Figure 15 Analyses of cell body outcome responses in anterior and posterior subregions of BA using a generalized linear model to estimate distinct components of overall activity.

a, Heatmap showing mean cue response timecourses of BA neurons (rows) from anterior BA (aBA; top) and from posterior BA (pBA; bottom) during hungry sessions, sorted by response magnitude in the 2 s following cue offset (n = 321 aBA neurons, 5 fields of view from 2 mice; n = 161 pBA neurons, 4 fields of view from 2 mice), for the task involving unavoidable tail shock. Vertical dashed lines demarcate visual stimulus onsets and offsets. b, Left: heatmap showing mean generalized linear model (GLM) predicted cue-related activity for RC and AC-Un from aBA (top) and pBA (bottom) neurons, sorted by response magnitude during 2 s cue period (see Methods). Right: heatmap showing mean GLM Ensure/tail shock-related activity from aBA (top) and pBA (bottom) neurons, sorted by response magnitude during 2 s following cue offset. c, Left: example neurons from aBA (top) and pBA (bottom) which had significant (see Methods) Ensure-related activity. Right: example neurons from aBA (top) and pBA (bottom) which had significant tail shock-related activity. Actual deconvolved neuronal activity (black lines), Ensure-related activity (red lines) (RC trials)/tail shock(AC-Un trials)/offset(NC trials), cue-related activity (green: RC; purple: AC-Un; gray: NC), running-related activity (cyan lines), and lick bout-related activity (blue lines). d, Mean predicted activity timecourses for cues (top) and outcomes (bottom) using all GLM components, from all cells with significant Ensure (left) or tail shock (right) components. Note that cells with Ensure components preferentially exhibited RC-related responses and cells with tail shock components preferentially exhibited AC-Un-related responses. e, Percent of all aBA vs. pBA neurons that have significant predicted activity related to Ensure (left), tail shock (middle), or licking (right). Top: neurons with positive beta-coefficients indicative of predicted increases in activity. Bottom: neurons with negative beta-coefficients. There was a larger percentage of neurons in pBA with significant predicted increases in activity to Ensure (n = 19/161 pBA neurons vs. n = 21/321 aBA neurons, * p = 0.02) while there was a larger percentage of cells with Ensure-related (n = 10/161 pBA neurons vs. n = 51/321 aBA neurons, ** p = 0.0013) and Lick-related (n = 14/161 pBA neurons vs. n = 62/321 aBA neurons, ** p = 0.001) decreases in activity in aBA vs. pBA. One-tailed binomial proportion test. f, Percent of all aBA vs. pBA neurons that have significant activity related to presentation of the RC (left), AC-Un (middle), or NC (right). Similar to results in Supplementary Fig. 14, there was a larger percentage of neurons in aBA with significant predicted increases (n = 6/161 pBA neurons vs. n = 36/361 aBA neurons, ** p = 0.003) and predicted suppression in activity related to the RC (n = 6/161 pBA neurons vs. n = 55/361 aBA neurons, *** p < 0.0001), and with significant predicted increases in activity related to the NC (n = 6/161 pBA neurons vs. n = 32/361 aBA neurons, ** p = 0.0082). One-tailed binomial proportion test.

Supplementary information

Supplementary Figs. 1–15.

Reporting Summary

Supplementary Video 1

Videos of average VTADA→BA axon responses to salient and nonsalient cues. Left, movie of cue-evoked responses, averaged across all reward cue (RC) trials (n = 58 trials) from a single session. Middle, movie of cue-evoked responses, averaged across all presentations of the aversive cue predicting unavoidable air puff (AC-Un) (n = 53 trials). Right, movie of cue-evoked responses, averaged across all presentations of the neutral cue (NC) (n = 56 trials). Movies have been spatially downsampled by 2× and temporally binned by four frames. Frame rate, 7 frames s–1 (that is, movie is playing at 1.8× real-time). Stimulus duration (indicated by white squares) was 2 s. White circles indicate perimeter of field of view. Percentage change in GCaMP6s fluorescence ranges from −15% (black) to +15% (white). Gray, no change in activity. See also Fig. 5b.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Lutas, A., Kucukdereli, H., Alturkistani, O. et al. State-specific gating of salient cues by midbrain dopaminergic input to basal amygdala. Nat Neurosci 22, 1820–1833 (2019). https://doi.org/10.1038/s41593-019-0506-0

Download citation

Further reading