Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Brief Communication
  • Published:

Fear generalization in the primate amygdala

Nature Neuroscience volume 18, pages 188–190 (2015)Cite this article

Subjects

Abstract

Broad generalization of negative memories is a potential etiology for anxiety disorders, yet the underlying mechanisms remain unknown. We developed a non-human primate model that replicates behavioral observations in humans and identifies specific changes in tuning properties of amygdala neurons: the width of auditory tuning increases with the distance of its center from the conditioned stimulus. This center-width relationship can account for better detection and at the same time explain the wide stimulus generalization.

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on Springer Link
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

Figure 1: Behavioral generalization.
Figure 2: Tuning width changes with distance from the CS.
Figure 3: Neural organization match behavioral generalization.

Similar content being viewed by others

References

  1. Pitman, R.K. et al. Nat. Rev. Neurosci. 13, 769–787 (2012).

    Article  CAS  Google Scholar 

  2. Lissek, S. Depress. Anxiety 29, 257–263 (2012).

    Article  Google Scholar 

  3. Jovanovic, T. & Ressler, K.J. Am. J. Psychiatry 167, 648–662 (2010).

    Article  Google Scholar 

  4. Aizenberg, M. & Geffen, M.N. Nat. Neurosci. 16, 994–996 (2013).

    Article  CAS  Google Scholar 

  5. Resnik, J., Sobel, N. & Paz, R. Nat. Neurosci. 14, 791–796 (2011).

    Article  CAS  Google Scholar 

  6. Laufer, O. & Paz, R. J. Neurosci. 32, 6304–6311 (2012).

    Article  CAS  Google Scholar 

  7. Duvarci, S. & Pare, D. Neuron 82, 966–980 (2014).

    Article  CAS  Google Scholar 

  8. Armony, J.L., Servan-Schreiber, D., Romanski, L.M., Cohen, J.D. & LeDoux, J.E. Cereb. Cortex 7, 157–165 (1997).

    Article  CAS  Google Scholar 

  9. Shaban, H. et al. Nat. Neurosci. 9, 1028–1035 (2006).

    Article  CAS  Google Scholar 

  10. Han, J.H. et al. Learn. Mem. 15, 443–453 (2008).

    Article  Google Scholar 

  11. Quirk, G.J., Armony, J.L. & LeDoux, J.E. Neuron 19, 613–624 (1997).

    Article  CAS  Google Scholar 

  12. David, S.V., Fritz, J.B. & Shamma, S.A. Proc. Natl. Acad. Sci. USA 109, 2144–2149 (2012).

    Article  CAS  Google Scholar 

  13. Likhtik, E. & Stujenske, J.M. Nat. Neurosci. 17, 106–113 (2014).

    Article  CAS  Google Scholar 

  14. Chavez, C.M., McGaugh, J.L. & Weinberger, N.M. Neurobiol. Learn. Mem. 91, 382–392 (2009).

    Article  Google Scholar 

  15. Dunsmoor, J.E., Prince, S.E., Murty, V.P., Kragel, P.A. & LaBar, K.S. Neuroimage 55, 1878–1888 (2011).

    Article  Google Scholar 

  16. Courtin, J. et al. Nature 505, 92–96 (2014).

    Article  Google Scholar 

  17. Xu, W. & Sudhof, T.C. Science 339, 1290–1295 (2013).

    Article  CAS  Google Scholar 

  18. Bergado-Acosta, J.R. et al. Learn. Mem. 15, 163–171 (2008).

    Article  CAS  Google Scholar 

  19. Kaouane, N. et al. Science 335, 1510–1513 (2012).

    Article  CAS  Google Scholar 

  20. Johansen, J.P., Tarpley, J.W., LeDoux, J.E. & Blair, H.T. Nat. Neurosci. 13, 979–986 (2010).

    Article  CAS  Google Scholar 

  21. Livneh, U. & Paz, R. J. Neurosci. Methods 192, 90–95 (2010).

    Article  Google Scholar 

  22. Livneh, U. & Paz, R. J. Neurosci. 32, 8598–8610 (2012).

    Article  CAS  Google Scholar 

  23. Livneh, U. & Paz, R. Neuron 75, 133–142 (2012).

    Article  CAS  Google Scholar 

  24. Paton, J.J., Belova, M.A., Morrison, S.E. & Salzman, C.D. Nature 439, 865–870 (2006).

    Article  CAS  Google Scholar 

  25. Balleine, B.W. & Killcross, S. Trends Neurosci. 29, 272–279 (2006).

    Article  CAS  Google Scholar 

  26. Herry, C. et al. Nature 454, 600–606 (2008).

    Article  CAS  Google Scholar 

  27. Sangha, S., Chadick, J.Z. & Janak, P.H. J. Neurosci. 33, 3744–3751 (2013).

    Article  CAS  Google Scholar 

  28. Genud-Gabai, R., Klavir, O. & Paz, R. J. Neurosci. 33, 17986–17994 (2013).

    Article  CAS  Google Scholar 

  29. Moore, A.K. & Wehr, M. J. Neurosci. 33, 13713–13723 (2013).

    Article  CAS  Google Scholar 

  30. Bordi, F. & LeDoux, J. J. Neurosci. 12, 2493–2503 (1992).

    Article  CAS  Google Scholar 

  31. Ciocchi, S. et al. Nature 468, 277–282 (2010).

    Article  CAS  Google Scholar 

  32. Bartlett, E.L., Sadagopan, S. & Wang, X. J. Neurophysiol. 106, 849–859 (2011).

    Article  Google Scholar 

  33. Cover, T.T., Thomas, J.A. Elements of Information Theory (Wiley-Interscience, 1991).

  34. Seung, H.S. & Sompolinsky, H. Proc. Natl. Acad. Sci. USA 90, 10749–10753 (1993).

    Article  CAS  Google Scholar 

  35. Salinas, E. & Abbott, L.F. J. Comput. Neurosci. 1, 89–107 (1994).

    Article  CAS  Google Scholar 

Download references

Acknowledgements

We are grateful to H. Sompolinsky for discussions and suggestions. We thank Y. Shohat for major contribution to the work and welfare of the animals, E. Kahana and G. Hecht for help with medical and surgical procedures, E. Furman-Haran and N. Stern for magnetic resonance imaging procedures. This work was supported by the Minerva Foundation and the Israel Science Foundation (#26613 and ERC-FP7-StG #281171, R.P.).

Author information

Authors and Affiliations

  1. Department of Neurobiology, Weizmann Institute of Science, Rehovot, Israel

    Jennifer Resnik & Rony Paz

Authors
  1. Jennifer Resnik
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Rony Paz
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

J.R. and R.P. designed the research. J.R. performed the experiments. J.R. analyzed the data. J.R. and R.P. wrote the paper.

Corresponding author

Correspondence to Rony Paz.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Integrated supplementary information

Supplementary Figure 1 Behavioral generalization.

a. Behavioral generalization curves in four Individual sessions. Shown is the conditioned response (CR, mean ± SEM), as a function of absolute (main) or real (inset) distance from CS.

The preparatory conditioned-response (CR) is the change in the inhale that triggers the CS tone, i.e. before the inhale that triggers odor release – hence a preparatory response, and quantified as the half-height-time of the inhale in ms. It is then normalized within session to account for differences across days.

b. The width of behavioral generalization (i.e. the half-height width from Fig.1e), but computed separately for the first third of recording sessions, the 2nd, and the last 3rd (mean±SEM). This shows that the behavioral generalization was similar along the experimental period (p>0.1, ANOVA).

c. Behavioral generalization, similar to main Fig.1e but here the CR is quantified as the peak of the preparatory inhale (mean ± SEM).

d. Same as Fig.1g when the CR is quantified as the peak as in (c) (significant difference between aversive and pleasant response).

e. Same as Fig.1h when the CR is quantified as the peak as in (c). The just-noticeable-difference (JND, 70.7% of the maximal response) computed here was 6.9%.

Supplementary Figure 2 Examples of tuning curves in the amygdala.

a. Two examples of single-cells’ tuning curves. The middle panel shows the firing rate as a function of the nine presented tones in the generalization phase (mean ± SEM). Outer panels show raster plots overlaid by PSTH of individual responses to each tone.

b. Three additional examples of single-cells’ tuning curves: firing rate as a function of the nine presented tones in the generalization phase (mean ± SEM, dots indicate FR in individual trials). Insets show the same generalization curves but as a function of absolute distance from the CS (mean ± SEM).

Supplementary Figure 3 Tuning-width changes with preferred-stimulus post-conditioning, but not pre-conditioning.

a. Change in tuning width (width in post- minus pre-conditioning, mean±SEM) as a function of PS distance from CS (p=0.03, F=2.96, ANOVA). This suggests that neurons generally had wide tuning pre-conditioning and narrowed/widened their tuning with direct relationship to the distance of their post-conditioning PS from the CS.

b. The same analysis using PS from pre-conditioning showed no relationship (p=0.92 F=0.16 ANOVA).

c. Tuning width during post-conditioning generalization as a function of the PS in the pre-conditioning habituation (phase/tone/interaction p=0.25/0.76/0.59 F=1.33/0.39/0.63, 2-way ANOVA).

d. Same as in Fig.2e, but with width of tuning from pre-conditioning habituation (r=0.17, p=0.95, Pearson).

Supplementary Figure 4 There were no changes in tonic activity of BLA neurons.

Firing rate during the pre-conditioning (habituation) plotted against the firing rate in the post-conditioning (generalization) phase for each neuron. Insets show the overall average in both phases (mean ± SEM). Left: activity taken from 1 sec before the tones and represents baseline activity. Right: activity taken from 1 sec after the tone and represents evoked response. There was no change in tonic responses (p=0.829 for left and 0.724 for right, paired t-test).

Supplementary Figure 5 Additional characteristics of tuning in the post-learning phase.

a. Correlation between Gaussian fit (R2) and the absolute distance between the PS and the CS during generalization (r=-0.137, p=0.046; during habituation: r=-0.01, p=0.92). The slightly better fit for neurons with PS closer to the CS can reflect the narrower tuning described.

b. Distribution of PS in pre- and post-conditioning periods. There was no change in variance of the distributions (right, variance ± CI, p>0.1), indicating that in the current paradigm, neurons did not shift their PS selectively towards the CS.

c. Daily behavioral PS (center of Gaussian fit to the behavioral responses) varied with the neural PS recorded in the same session, but only during post-conditioning phase (left, r=0.38 p=0.005; pre-conditioning: right, r=-0.04 p=0.8).

Supplementary Figure 6 Neurons increase their tuning width rather than adding a CS response.

a. Normalized firing rates for all tones as a function of absolute distance from the CS. Each panel depicts the average over all neurons with PS distance from the CS in a specific range (indicated in the color legend). All curves have a single peak at the PS, rather than a double peak at the PS and at the CS.

b. Difference in (normalized) firing rate between the PS segment and the segment next to it, for the four groups of neurons as in (a). Inset shows the slope of these lines. The closer the PS to the CS, the steeper the slope (p<0.001, F=7.7, ANOVA). This strengthens the main finding, of narrower curves next to the CS.

Supplementary Figure 7 Basic auditory tuning properties of neurons in the primate amygdala.

Pseudorandomized presentation of 40 pure-tones, ranging 300–4800 Hz (4 octaves) in equal logarithmic steps. Each tone was presented 30 times for 200ms (10ms rise/fall ramp; ITI:200ms).

a. Rater plots overlaid with PSTH for 4 BLA neurons, aligned on tone onset.

b. Auditory STRF for 4 BLA neurons (40 tones spanning 4 octaves in equal logarithmic distance)

c. Distribution of number of peaks showing that most neurons had one best-frequency (BF, n=126). Zero-peaks was defined for broadly un-tuned units that responded equally to the entire tested spectrum. Inset shows the mean firing rate for the three different groups (mean ± SEM)

d. BF distribution from tuned units was homogenous (logarithmic scale, χ2=25.7, p>0.1, n=99)

e. Five examples of single-cells’ tuning. STRFs were averaged over the first 200 ms, and shown either in color scale or as a graph (upper three are 1-peak, fourth is 2-peak, lower is 0-peak).

f. Distribution of response bandwidths and its cumulative distribution. Median=0.76

g. Latency distribution of response to each unit’s PS. Median=76ms. Notice the relatively large group of neurons with latency of <50ms.

Supplementary Figure 8 Tuning curves with dense homogenous sampling and following 24 h.

a. Ten examples of single-cells’ tuning curves, evaluated by presentation of 40 different tones spanning 4 octaves (300 Hz to 4800 Hz) with equal spacing in logarithmic scale, and each tone is presented 10 times for 200ms in pseudorandom order.

Shown is the firing rate, either in a colormap (bottom) or on the y-axis, as a function of distance (%) from the CS+ of the previous day.

From top to bottom: the further the PS is from the CS+ of the previous day, the wider the tuning.

Hence, even with dense homogenous sampling, the characterization of tuning curve show the same attribute of width as a function of distance from the CS+. Moreover, this is evident even in the following day before a new session begins, indicating correlates of overnight memory.

b. Correlation between the width of the tuning curves (in octaves) and the distance between the neuron’s PS and the CS+ from the previous day (solid line, p<0.01, r=0.5). As control, we repeated this analysis using random tones as the CS+ (1000 repetitions, dashed line, mean ± SEM).

c. Distribution of PS was homogenous.

Supplementary Figure 9 Robustness of relationship between width and PS distance from CS.

a. Linear regression between tuning-width and PS distance from CS for neurons with different tuning fits. From left to right: for neurons with tuning-fit of R2 > 0.1, >0.2, >0.3, >0.5, >0.6; R2>0.4 is shown in main Fig.2e. All regressions were significant with similar coefficients.

b. The slope between the PS and the tuning-width, separately for the 1st third of the recording epoch, the 2nd, and the final third; showing similar neural changes following conditioning along the recording period.

Supplementary Figure 10 Behavioral and neural dynamics during conditioning and post-learning.

a. Behavioral conditioned response (CR, the half-height-time of the inhale that triggers the CS, mean±SEM) during the first third of the acquisition trials, the 2nd, and the last third (F=5.16, p=0.002, ANOVA).

b. Neural response (firing rate evoked by the CS, mean±SEM) during the first third of the acquisition trials, the 2nd, and the last third (F=2.76, p=0.04, ANOVA).

c. Comparing rate of change for the neural response vs. the behavioral response on an individual basis (each dot is a neuron in one conditioning session). Upper: by fitting an exponential function to each and plotting λs; Lower: by fitting linear regression and plotting the slope.

Solid lines represent the identity (upper: p=0.3, χ2=1.0; lower: p=0.04, χ2=4.5).

d. Larger response for the CS that decays with distance from it, when counting spikes (left) or number of responding neurons (right). Insets show the same for the surrogate data.

Supplementary Figure 11 Tone discrimination based on neural activity.

a. d’ between the CS and other tones, averaged over all neurons (top: experimental data; bottom: surrogate data), during habituation (grey) and generalization-test (black). This is the same figure as main Fig.3b – inserted here for clarity.

b. Change in d’ for all tone combinations. Color scale represents the difference in d’ from post-conditioning minus the d’ from pre-conditioning. The narrow diagonal band indicates no or negative improvement between near-by tones.

c. To see how this result emerges from the population data, one can calculate the Fisher-Information (FI) for the CS when including neurons with PS farther from the CS and wide tuning curves (here, FI is the normalized average of the tuning-curves’ slope). This validated that information is high when considering only neurons with PS right next to the CS - because they have sharp tuning, hence their slope at the CS is large; but information decays when taking into account neurons with wide generalization curves – because the information they provide is low due to a shallower slope around the CS.

Supplementary information

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Resnik, J., Paz, R. Fear generalization in the primate amygdala. Nat Neurosci 18, 188–190 (2015). https://doi.org/10.1038/nn.3900

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/nn.3900

This article is cited by

Search

Advanced search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing