Skip to main content

Thank you for visiting 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.

Atallah et al. reply



replying to S.-H. Lee, A. C. Kwan & Y. Dan Nature508,

Solving discrepancies in the literature is critical for the advancement of science, and the Comment by Lee et al.1 is thus welcome. It clarifies that there is no contradiction between the earlier study of Lee et al.2 and our study3. The disagreement is in the interpretation of the results and in the model used to fit the data.


Both studies2,3 optogenetically perturbed the activity of inhibitory parvalbumin-expressing (PV+) cells in mouse visual cortex and measured the resulting impact on the orientation tuning width of pyramidal cells. Whereas Lee et al.2 reported a narrowing of the tuning width, our study3 did not observe any systematic change.

The most obvious effect on photoactivation of PV+ cells is the reduction in the firing of pyramidal cells. Accordingly, we are glad that when Lee et al.1 (see accompanying Comment) consider the same range of pyramidal cell firing reduction as that described in our study3 (that is, up to 50% reduction, for a ratio of −0.34), there is no narrowing of the tuning curve. This narrowing, on the other hand, is present when exploring larger reductions in pyramidal cell firing, consistent with the findings of Lee et al.2.

These effects can be explained by the simple linear model with threshold proposed by our study3. In this model, the impact of PV+ cells is to subtract and scale orientation tuning curves, unless the firing rate is 0. The model captures not only the data in our study3, but also the data points of the example cell shown in figure 1 of the Comment1. As illustrated in our Fig. 1, this model fits the data very well, so all these results can be explained by a simple ‘iceberg effect’.

Figure 1: The linear-threshold model captures the effects of increasing PV activation.

a, Responses of a pyramidal cell to stimuli of different orientations, in control conditions (black) or in the presence of increasing PV cell activation (blue, moderate PV cell activation; red, stronger PV cell activation). The data are the same of those in figure 1 of Lee et al.1, and were obtained with the Matlab function ‘grabit.m’. We did not consider a fourth set of points, with the lowest firing rate, as there are barely any data points above zero. The curves indicate the fits of the linear-threshold model introduced by our study3. b, The same curves as in a, rescaled to peak at 1, to illustrate a mild but progressive narrowing of tuning curves with increasing PV+ cell activation.

PowerPoint slide

Moreover, the model explains additional data obtained by our study3, which are not mentioned in the Comment1. Our study3 performed the reverse experiment, namely the optogenetic suppression of PV+ cells to increase pyramidal cell firing rate up to 250%, and again found that there was no systematic increase in tuning width. As described in our study3, the model explains this finding because once the iceberg is out of the water it cannot get wider by rising further.

Therefore, one can fully reconcile Lee et al.2 and our study3 by pointing out (1) that when one explores both intermediate and large reductions in pyramidal cell firing rates one sees both effects (invariance and narrowing of tuning width, respectively, as the Comment does1); and (2) that when one uses the linear-threshold model3 one explains all of these effects. We believe that interpreting the data in the context of such a model is superior to comparing Gaussian functions fit to responses obtained with and without stimulation, as was done in Lee et al.2. It is closer to the biological reality of a spike threshold, more parsimonious, and therefore more informative as to the functional effect of PV+ cells on pyramidal cells.


  1. 1

    Lee, S.-H., Kwan, A. C. & Dan, Y. Interneuron subtypes and orientation tuning. Nature 508, (2014)

  2. 2

    Lee, S.-H. et al. Activation of specific interneurons improves V1 feature selectivity and visual perception. Nature 488, 379–383 (2012)

    ADS  PubMed  PubMed Central  CAS  Article  Google Scholar 

  3. 3

    Atallah, B. V., Bruns, W., Carandini, M. & Scanziani, M. Parvalbumin-expressing interneurons linearly transform cortical responses to visual stimuli. Neuron 73, 159–170 (2012)

    PubMed  PubMed Central  CAS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Matteo Carandini.

PowerPoint slides

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Atallah, B., Scanziani, M. & Carandini, M. Atallah et al. reply. Nature 508, E3 (2014).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links