The development of cortical circuits for motion discrimination

Journal name:
Nature Neuroscience
Volume:
18,
Pages:
252–261
Year published:
DOI:
doi:10.1038/nn.3921
Received
Accepted
Published online

Abstract

Stimulus discrimination depends on the selectivity and variability of neural responses, as well as the size and correlation structure of the responsive population. For direction discrimination in visual cortex, only the selectivity of neurons has been well characterized across development. Here we show in ferrets that at eye opening, the cortical response to visual stimulation exhibits several immaturities, including a high density of active neurons that display prominent wave-like activity, a high degree of variability and strong noise correlations. Over the next three weeks, the population response becomes increasingly sparse, wave-like activity disappears, and variability and noise correlations are markedly reduced. Similar changes were observed in identified neuronal populations imaged repeatedly over days. Furthermore, experience with a moving stimulus was capable of driving a reduction in noise correlations over a matter of hours. These changes in variability and correlation contribute significantly to a marked improvement in direction discriminability over development.

At a glance

Figures

  1. Response properties change dramatically following eye opening.
    Figure 1: Response properties change dramatically following eye opening.

    (a) Experimental timeline. GCaMP3expressing AAV was delivered intracortically via microinjection and two-photon imaging was performed 7–14 d later. Animals were imaged at either P29–32 (naive), P32–36 (immature) or P48–50 (mature). (b) Representative responses from animals in each age group. Left, baseline image. Scale bars, 50 μm. Middle, single-trial response to preferred stimulus (maximum projection across stimulus duration). Right, maximum response across all stimuli and all trials. EO, days after eye opening. (c) Responses for individual neurons highlighted in b. Left, response to eight directional stimuli, averaged across trials. Right, tuning curves fit with a two-peaked Gaussian. Horizontal line indicates the mean response to a blank stimulus.

  2. Stimulus selectivity increases and population response density decreases with age and experience.
    Figure 2: Stimulus selectivity increases and population response density decreases with age and experience.

    Naive, P29–32 (n = 9 animals); immature, P33–36 (n = 5 animals); mature, P48–50 (n = 5 animals). (a) Orientation selectivity (OSI) increased significantly with age (n = 1,134, 679 and 315 neurons for naive, immature and mature, respectively). (b) Direction selectivity (DSI) increases significantly with age (n = 924, 633 and 252 for naive, immature and mature, respectively). (c) Response variability (shown as s.d. across trials) to the preferred stimulus decreases with age. (d) Response amplitude to the preferred stimulus does not change across age (mean response across trials). (e) Response density (fraction of active neurons on a given trial out of all identified neuronal ROIs with at least one response) declines significantly from the naive and immature groups to the mature group. (f) Response density is stimulus specific. In all age groups, the fraction of neurons active on a given trial is significantly greater for the dominant (Dom) stimulus (the stimulus producing activity in the largest fraction of neurons, aligned across animals) than for a stimulus with opposite direction of motion (null, +180°) or an orthogonal orientation. Dashed line indicates mean fraction of neurons active during blank stimuli. Error bars are mean ± s.e.m. across animals.

  3. Wave-like responses to visual stimulation in young animals.
    Figure 3: Wave-like responses to visual stimulation in young animals.

    (a) Pseudocolored time course of response to single stimulus. Propagating activity appears as a gradient from blue to orange. Scale bar, 50 μm; applies to ad. (b) Single-trial example responses. Top row, responses to grating drifting down and to the left; bottom row, responses to stimulus of same orientation but opposite direction of motion. (c) Average response across all trials with strong wave-like activity. (d) Example response well fit by a linear traveling wave (wave index (WI) = 100). (e) Frequency of linear waves declines significantly with age (n = 3 FOV from 3 animals, 6 FOV from 3 animals and 4 FOV from 2 animals for naive, immature and mature, respectively). Circles indicate individual FOVs; squares indicate mean ± s.e.m. (f) Velocity of linear waves declines significantly with age (mean ± s.e.m.: naive: 256.8 ± 23.7 μm/s, n = 115 waves; immature: 91.3 ± 10.6 μm/s, n = 60 waves; mature: 44.2 ± 7.1 μm/s n = 6 waves). Dashed lines indicate geometric means. (g) Occurrence of linear waves is stimulus specific. P, preferred stimulus; N, null stimulus with opposite direction of motion from preferred; B, blank stimulus. Error bars are mean ± s.e.m. across animals. (h) Wave direction is largely consistent across all stimuli within an animal but differs across animals. Each histogram shows propagation directions for one animal in the naive group. Bar color indicates stimulus identity.

  4. Noise correlations decline with age and experience.
    Figure 4: Noise correlations decline with age and experience.

    (a) Pairwise noise correlation as a function of intra-pair spatial distance. There is a significant decrease in noise correlation across age groups, as well as a significant decrease in noise correlation as a function of distance within each group. (b) Pairwise noise correlation as a function of the intra-pair difference in preferred direction. For all age groups, noise correlations were significantly higher for pairs with similar orientation preferences than for those with orthogonal preferences. Likewise, all age groups exhibited significantly higher correlations for pairs with similar as opposed to opposite direction preferences. Data in a,b are represented as the mean ± s.e.m. across animals after taking the mean across all pairs within an animal. (c) Noise correlations decrease and DSI increases with age. Circles, mean ± s.e.m. for each FOV; squares, mean ± s.e.m. across all animals in each age group. There is a significant correlation (R = −0.59; P < 0.01) between noise correlation and DSI across all experiments. (df) Maps of pairwise noise correlation as a function of intra-pair spatial distance and difference in preferred direction (local average based on Gaussian kernel, σx = 15 μm, σy = 15 deg). Early, noise correlation is large for small distances, in particular in pairs with similar preferred direction. In the mature cortex, this bias is reduced and noise correlations are small irrespective of distance and tuning difference. Black indicates regions with insufficient data (less than 20 pairs per 30-μm × 22.5-deg region).

  5. Longitudinal two-photon imaging reveals emergence of direction selectivity in identified neurons.
    Figure 5: Longitudinal two-photon imaging reveals emergence of direction selectivity in identified neurons.

    (a) Schematic of chronically implanted headplate. (b) Time course of imaging experiments. (c,d) Longitudinal imaging across 5 d following eye opening. (c) Top, cortical blood vessel pattern from day 0 to day 5. Blue box indicates area for functional imaging. Bottom, imaging field over days. Yellow boxes outline examples of corresponding neurons. 126 neurons (63% of day 0) remained visible and visually responsive across all imaging sessions. Images were aligned across days with an affine transform. Purple symbols indicate example neurons shown in g. Scale bars: top, 500 μm; bottom, 50 μm. (d) Direction selectivity (DSI) increases in a population of neurons imaged from a single animal over 5 d (n = 126 neurons). Cum. prob., cumulative probability. (e) Direction preference is stable in the majority of neurons, whereas a subset of cells exhibit 180-degree reversals. Red dashed lines indicate 180-degree shift from day 0. Pref. dir., preferred direction. (f) DSI increases in neurons with stable preferences (blue symbols; red circle indicates mean, with s.e.m. smaller than marker size) from the initial to final imaging session. Neurons that reverse preference (black; green indicates mean, with s.e.m. smaller than marker size) do not show an increase in selectivity. (g,h) Example (g) and average (h) tuning curves showing neurons that maintained a preferred direction (top), reversed preferred direction (middle) and developed a clear preferred direction (bottom). Purple symbols in c,eg indicate corresponding neurons. Error bars in g are s.e.m. across trials (n = 12). Error bars in h are s.e.m. across neurons (n = 115, 24, 98 for maintained, reversed and developed a preferred direction, respectively). For ease of comparison, tuning curves were shifted to align preferred stimuli on day 5.

  6. Correlation between decrease in noise correlation and direction selectivity.
    Figure 6: Correlation between decrease in noise correlation and direction selectivity.

    (a) Pairwise noise correlations decreased significantly from the initial to the final imaging session. Red dot indicates mean across all pairs and animals. (b) Relationship between change in pairwise noise correlation and direction selectivity (relative to day 0) for 4 animals in which chronic imaging was performed. In 3 of 4 cases, noise correlations exhibited a significant decrease by the final imaging session, whereas direction selectivity significantly increased (n = 33, 147 and 126 neurons per experiment). In one experiment (F1319, orange, n = 13 neurons), changes in neither DSI nor correlations were significant. Data are shown as mean ± s.e.m. across all neurons. Averaging across all neurons and pairs on the final imaging session (black) reveals a significant decrease in noise correlation and a significant rise in direction selectivity. (c) Cartoon depicting possible relationships in pairwise angular preference across imaging sessions. (d) Initial pairwise noise correlations were higher for pairs that will maintain similar preferences (S–S pairs) than those that adopt opposite preferences (S–O pairs). (e) Initial correlations were higher for pairs with opposite preferences if that pair will adopt matching preferred directions on the final imaging session (O–S) than for pairs that maintain opposite preferences (O–O).

  7. Changes in variance and correlation structure contribute to increased discriminability over development.
    Figure 7: Changes in variance and correlation structure contribute to increased discriminability over development.

    (a) Single-cell direction discriminability increases significantly between the naive and immature ages (black versus dark blue). Combining immature levels of variance with naive levels of selectivity (gray) results in near immature levels of discriminability, whereas the gain in selectivity alone between the naive and immature ages (light blue) has a much smaller effect on discriminability (gray versus light blue). (b) Effects of variance and structure of noise correlations on discriminability by groups of cells. The overall change in variance and correlation structure increases discriminability substantially between the naive and immature ages. A considerable fraction (increasing with group size) is due to a change in the structure of noise correlations alone (black dashed line). (c) Effect of eliminating correlations by trial shuffling on discriminability by groups of cells. (d) Differences between solid and dashed lines from c. Shaded region shows ± 1 s.e.m.

  8. Motion training induces decrease in noise correlations and increase in direction selectivity.
    Figure 8: Motion training induces decrease in noise correlations and increase in direction selectivity.

    (a) Experimental timeline. Shortly after eye opening, animals received 4–6 h of motion training. PND, postnatal day. (b) Example (from point with black error bars in d) of direction selectivity increase following motion training. (c) Example of decrease in pairwise noise correlations following motion training. (d) Following motion training, noise correlations decreased and direction selectivity increased, whereas direction selectivity did not change and noise correlations increased following flash training. In control animals with no training, noise correlations and direction selectivity remained unchanged. Pref. dir., preferred direction; corr. coeff., correlation coefficient. (e) Change in noise correlation as a function of training type, pooled across all animals. Error bars in d,e are mean ± s.e.m. across neurons. (f) Before training, noise correlation was highest in pairs for which the difference in preferred direction and cortical distance was small (map generated as in Fig. 4d–f using a Gaussian smoothing kernel, σx = 15 μm, σy = 15 deg). (g) The correlation in this group decreased after training, primarily owing to the subset of pairs maintaining a small difference in preferred direction, as opposed to those preferring opposite directions after training. (hk) Discriminability measures in training data set. (h) Discriminability of the trained orientation increases with training. Switching the pre- and post-training variances (light green) did not change discriminability. (i) In larger groups of cells, the change in variance, but not noise correlation, had a detectable effect on discriminability (discrim.). (j) Effect of eliminating noise correlations by trial shuffling on discriminability in training data set. (k) Difference between solid and dashed lines from j. Shaded regions show ± 1 s.e.m.

  9. Responses in identified neurons imaged repeatedly over days.
    Supplementary Fig. 1: Responses in identified neurons imaged repeatedly over days.

    (a) Six representative neurons imaged on Day 0 (P30), Day 3 (P33), and Day 5 (P35). Scale bar is 50 µm. (b) Responses for individual neurons highlighted in a. Left column: Response to 8 directional stimuli, averaged across trials. Right column: Tuning curves fit with a 2-peaked Gaussian. Horizontal line indicates the mean response to a blank stimulus.

  10. Development of orientation and direction selective responses in identified neurons over days.
    Supplementary Fig. 2: Development of orientation and direction selective responses in identified neurons over days.

    (a,c) Orientation preference is stable over days in longitudinally imaged animals. (b,d) Direction preference is stable in the majority of neurons, whereas a subset of cells exhibit 180 degree reversals. Red dashed lines indicate 180 degree shift from d0. (e) Orientation selectivity increases over days in identified neurons. Red dot indicates mean across neurons (s.e.m. error bars are smaller than marker). (f) DSI increases from the initial to final imaging session. Neurons with negative initial DSI (black symbols, mean indicated by green dot) reversed their direction preference over imaging sessions, but on average do not show an increase in selectivity. Neurons that maintain a preferred direction (positive initial DSI, blue symbols, mean indicated by red dot) exhibit significantly increased selectivity over days. Error bars indicating s.e.m. are smaller than markers for mean. Panels d & f are re-plotted from Figure 5 to facilitate comparison with orientation preference and selectivity.

  11. Evolution of noise correlation matrix and relative preferred direction over five days of development.
    Supplementary Fig. 3: Evolution of noise correlation matrix and relative preferred direction over five days of development.

    Cells are sorted by their preferred direction on the last day. (a-c): Noise correlation on day 0, day 3, and day 5. (d-f): Angular difference between cell pairs on day 0, day 3, and day 5. Correlations generally decrease from day 0 to day 5, with a prominent decrease in sets of cells sharing a preferred direction (cells 1 to 40).

  12. Initial noise correlations are higher for pairs that will ultimately adopt similar preferred directions.
    Supplementary Fig. 4: Initial noise correlations are higher for pairs that will ultimately adopt similar preferred directions.

    (a) Cartoon depicting possible relationships in pairwise angular preference across imaging sessions, re-plotted from Figure 6. (b) Pairwise noise correlations during initial imaging as function of pairwise angular difference in preferred direction. Colored squares indicate regions of plot classified as S-S, S-O, O-O, and O-S, and quantified in Figure 6d,e.

  13. Possible scenarios for the maturation of direction discriminability.
    Supplementary Fig. 5: Possible scenarios for the maturation of direction discriminability.

    (a) Possible scenarios for the maturation of single cell discriminability. Stimuli to be discriminated are moving gratings with opposite directions of motion. Early cortex (left): Response distribution of a single cell for a left moving grating (blue) and a right moving grating (red). Mean responses are indicated by vertical lines. Direction selectivity is commonly defined as the difference of mean responses divided by their sum. Single trial discriminability of motion direction critically depends on the fraction of overlap between the two distributions. If the overlap is close to 100%, discriminability is close to 0. If the overlap is small, discriminability is close to 1. During development, discriminability can improve by shifting the mean responses apart, i.e. by increasing the cell’s direction selectivity (right, upper). However, discriminability can also improve by reducing the response variance (without changing the selectivity; right, lower). (b) Possible scenarios for the maturation of multi-cell discriminability, here for a population of two neurons. Early cortex (left): Response distribution for a left (blue) and a right moving grating (red). Assuming response fluctuations follow a Gaussian statistics, the bivariate distributions have elliptic shapes (marking the 95% confidence region for the response distributions). Mean response vectors are marked by dots. In the shown example both cells are tuned for the same stimulus direction and their noise correlation is positive. As in the single cell case, discriminability depends on the fraction of overlap between the two distributions. This fractional overlap can decrease over development as cells become more selective (right, upper), as the overall magnitude of fluctuations decreases (right, middle) or as noise correlations become smaller (right, lower).

  14. Individual neurons increase discriminability despite reduced selectivity due to decreased variability.
    Supplementary Fig. 6: Individual neurons increase discriminability despite reduced selectivity due to decreased variability.

    (a) Discriminability increases significantly over days in identified neurons. (b) The majority of neurons exhibit increased discriminability over days (light blue). These neurons are analyzed further in panels c-i. (c) Of the neurons exhibiting increased discriminability over days, the majority also display increased direction selectivity (HS cells, red), whereas a subset show decreased selectivity over days (LS cells, blue). (d) In LS cells, neither the preferred nor the null response changes significantly. (e) HS neurons achieve enhanced selectivity both through an increase in the preferred response and a decrease in the null response. (f) Variability decreases significantly for both LS and HS groups, however variability decreases to a greater extent in cells without increased selectivity (blue) than in those where selectivity increases (red). (g) For LS cells, only the change in variance contributes positively to the increase in discriminability. (h) For HS cells, both the decreased variance and the increased selectivity contribute significantly to increased discriminability, with the change in mean response providing a significantly greater contribution. (i) Changes in variance provide a significantly greater contribution to improved discriminability in LS vs. HS neurons.

  15. Noise correlations decrease asymmetrically in longitudinally imaged animals.
    Supplementary Fig. 7: Noise correlations decrease asymmetrically in longitudinally imaged animals.

    Correlations decrease over imaging sessions to a larger extent in S-S vs. S-O pairs. Likewise, the decrease in correlations is larger in O-S vs. O-O pairs. Change in correlations are normalized within animal to the change exhibited by S-S pairs. (*) indicates MW test, p<0.01. Error bars indicate s.e.m.

  16. Change in response variance vs. change in selectivity.
    Supplementary Fig. 8: Change in response variance vs. change in selectivity.

    (a) Distribution of standard deviation of response to preferred stimulus for all cells, pre- and post-training. Variance of response does not change significantly over motion training. (b) Response variance is calculated across the preferred orientation before and after training. Each point represents the average change for a single field of view. (c) Overall change in standard deviation for each training type. (d) Same as (b), but for the orientation orthogonal to the preferred orientation. (e) Same as (c), but for the orthogonal orientation. Error bars represent +/- 1 SEM.

  17. Single neuron response distributions across age groups.
    Supplementary Fig. 9: Single neuron response distributions across age groups.

    Distribution of activity levels over all stimulus conditions with stimulus means subtracted (see Eqn. 8, 9, Methods) for 10 representative and randomly selected cells from the naive (a), immature (b), and mature (c) populations of cells.

Videos

  1. Supplementary Video 1: Wave-like responses to drifting grating stimuli at eye opening.
    Video 1: Supplementary Video 1: Wave-like responses to drifting grating stimuli at eye opening.
    Response to two presentations of drifting grating stimuli with the same orientation, drifting in opposite directions. Movie is shown at 2x real-time. Stimulus presentation and direction are indicated by arrow in upper left. Scale bar is 50 μm. 512 × 256 images with a 1:2 aspect ratio were acquired at 60 Hz, downsampled to 15 Hz and resized to 512 × 512 by bilinear interpolation. Images were then filtered in x-y by a 2×2 pixel radius mean filter, Z-scored relative to the mean and standard deviation of the blank stimulus, and clipped at 0 to 4 Z-units.

References

  1. Tomko, G.J. & Crapper, D.R. Neuronal variability: non-stationary responses to identical visual stimuli. Brain Res. 79, 405418 (1974).
  2. Tolhurst, D.J., Movshon, J.A. & Dean, A.F. The statistical reliability of signals in single neurons in cat and monkey visual cortex. Vision Res. 23, 775785 (1983).
  3. Tolhurst, D.J. The amount of information transmitted about contrast by neurones in the cat's visual cortex. Vis. Neurosci. 2, 409413 (1989).
  4. Willmore, B. & Tolhurst, D.J. Characterizing the sparseness of neural codes. Network 12, 255270 (2001).
  5. Yen, S.C., Baker, J. & Gray, C.M. Heterogeneity in the responses of adjacent neurons to natural stimuli in cat striate cortex. J. Neurophysiol. 97, 13261341 (2007).
  6. Weliky, M., Fiser, J., Hunt, R.H., Wagner, D.N. & York, N. Coding of natural scenes in primary visual cortex. Neuron 37, 703718 (2003).
  7. Olshausen, B.A. & Field, D.J. Sparse coding of sensory inputs. Curr. Opin. Neurobiol. 14, 481487 (2004).
  8. Shadlen, M.N. & Newsome, W.T. Noise, neural codes and cortical organization. Curr. Opin. Neurobiol. 4, 569579 (1994).
  9. Abbott, L.F. & Dayan, P. The effect of correlated variability on the accuracy of a population code. Neural Comput. 11, 91101 (1999).
  10. Averbeck, B.B., Latham, P.E. & Pouget, A. Neural correlations, population coding and computation. Nat. Rev. Neurosci. 7, 358366 (2006).
  11. Chapman, B., Stryker, M.P. & Bonhoeffer, T. Development of orientation preference maps in ferret primary visual cortex. J. Neurosci. 16, 64436453 (1996).
  12. Li, Y., Fitzpatrick, D. & White, L.E. The development of direction selectivity in ferret visual cortex requires early visual experience. Nat. Neurosci. 9, 676681 (2006).
  13. Chapman, B. & Stryker, M.P. Development of orientation selectivity in ferret visual cortex and effects of deprivation. J. Neurosci. 13, 52515262 (1993).
  14. Krug, K., Akerman, C.J. & Thompson, I.D. Responses of neurons in neonatal cortex and thalamus to patterned visual stimulation through the naturally closed lids. J. Neurophysiol. 85, 14361443 (2001).
  15. Rochefort, N.L. et al. Sparsification of neuronal activity in the visual cortex at eye-opening. Proc. Natl. Acad. Sci. USA 106, 1504915054 (2009).
  16. Ikezoe, K., Tamura, H., Kimura, F. & Fujita, I. Decorrelation of sensory-evoked neuronal responses in rat barrel cortex during postnatal development. Neurosci. Res. 73, 312320 (2012).
  17. Li, Y., Van Hooser, S.D., Mazurek, M., White, L.E. & Fitzpatrick, D. Experience with moving visual stimuli drives the early development of cortical direction selectivity. Nature 456, 952956 (2008).
  18. Derrington, A.M. & Fuchs, A.F. The development of spatial-frequency selectivity in kitten striate cortex. J. Physiol. (Lond.) 316, 110 (1981).
  19. Berens, P. CircStat: a MATLAB toolbox for circular statistics. J. Stat. Softw. 31, 121 (2009).
  20. Bair, W., Zohary, E. & Newsome, W.T. Correlated firing in macaque visual area MT: time scales and relationship to behavior. J. Neurosci. 21, 16761697 (2001).
  21. Ecker, A.S. et al. Decorrelated neuronal firing in cortical microcircuits. Science 327, 584587 (2010).
  22. Smith, M.A. & Kohn, A. Spatial and temporal scales of neuronal correlation in primary visual cortex. J. Neurosci. 28, 1259112603 (2008).
  23. Clemens, J.M., Ritter, N.J., Roy, A., Miller, J.M. & Van Hooser, S.D. The laminar development of direction selectivity in ferret visual cortex. J. Neurosci. 32, 1817718185 (2012).
  24. Shadlen, M.N. & Newsome, W.T. The variable discharge of cortical neurons: implications for connectivity, computation, and information coding. J. Neurosci. 18, 38703896 (1998).
  25. Kriener, B., Tetzlaff, T., Aertsen, A., Diesmann, M. & Rotter, S. Correlations and population dynamics in cortical networks. Neural Comput. 20, 21852226 (2008).
  26. Van Hooser, S.D. et al. Initial neighborhood biases and the quality of motion stimulation jointly influence the rapid emergence of direction preference in visual cortex. J. Neurosci. 32, 72587266 (2012).
  27. Jeanne, J.M., Sharpee, T.O. & Gentner, T.Q. Associative learning enhances population coding by inverting interneuronal correlation patterns. Neuron 78, 352363 (2013).
  28. Blankenship, A.G. & Feller, M.B. Mechanisms underlying spontaneous patterned activity in developing neural circuits. Nat. Rev. Neurosci. 11, 1829 (2010).
  29. Huberman, A.D., Feller, M.B. & Chapman, B. Mechanisms underlying development of visual maps and receptive fields. Annu. Rev. Neurosci. 31, 479509 (2008).
  30. Ackman, J.B. & Crair, M.C. Role of emergent neural activity in visual map development. Curr. Opin. Neurobiol. 24, 166175 (2014).
  31. Wu, J.Y., Xiaoying, H. & Chuan, Z. Propagating waves of activity in the neocortex: what they are, what they do. Neuroscientist 14, 487502 (2008).
  32. Sato, T.K., Nauhaus, I. & Carandini, M. Traveling waves in visual cortex. Neuron 75, 218229 (2012).
  33. Ackman, J.B., Burbridge, T.J. & Crair, M.C. Retinal waves coordinate patterned activity throughout the developing visual system. Nature 490, 219225 (2012).
  34. Siegel, F., Heimel, J.A., Peters, J. & Lohmann, C. Peripheral and central inputs shape network dynamics in the developing visual cortex in vivo. Curr. Biol. 22, 253258 (2012).
  35. Stroh, A. et al. Making waves: initiation and propagation of corticothalamic Ca2+ waves in vivo. Neuron 77, 11361150 (2013).
  36. Wong, R.O., Meister, M. & Shatz, C.J. Transient period of correlated bursting activity during development of the mammalian retina. Neuron 11, 923938 (1993).
  37. Hanganu, I.L., Ben-Ari, Y. & Khazipov, R. Retinal waves trigger spindle bursts in the neonatal rat visual cortex. J. Neurosci. 26, 67286736 (2006).
  38. Colonnese, M.T. et al. A conserved switch in sensory processing prepares developing neocortex for vision. Neuron 67, 480498 (2010).
  39. Ko, H. et al. Functional specificity of local synaptic connections in neocortical networks. Nature 473, 8791 (2011).
  40. Ko, H. et al. The emergence of functional microcircuits in visual cortex. Nature 496, 96100 (2013).
  41. Le Magueresse, C. & Monyer, H. GABAergic interneurons shape the functional maturation of the cortex. Neuron 77, 388405 (2013).
  42. Gao, W.J., Newman, D.E., Wormington, A.B. & Pallas, S.L. Development of inhibitory circuitry in visual and auditory cortex of postnatal ferrets: immunocytochemical localization of GABAergic neurons. J. Comp. Neurol. 409, 261273 (1999).
  43. Van Hooser, S.D., Escobar, G.M., Maffei, A. & Miller, P. Emerging feed-forward inhibition allows the robust formation of direction selectivity in the developing ferret visual cortex. J. Neurophysiol. 111, 23552373 (2014).
  44. King, P.D., Zylberberg, J. & DeWeese, M.R. Inhibitory interneurons decorrelate excitatory cells to drive sparse code formation in a spiking model of V1. J. Neurosci. 33, 54755485 (2013).
  45. Bernacchia, A. & Wang, X.J. Decorrelation by recurrent inhibition in heterogeneous neural circuits. Neural Comput. 25, 17321767 (2013).
  46. Tetzlaff, T., Helias, M., Einevoll, G.T. & Diesmann, M. Decorrelation of neural-network activity by inhibitory feedback. PLOS Comput. Biol. 8, e1002596 (2012).
  47. van Vreeswijk, C. & Sompolinsky, H. Chaotic balanced state in a model of cortical circuits. Neural Comput. 10, 13211371 (1998).
  48. Renart, A. et al. The asynchronous state in cortical circuits. Science 327, 587590 (2010).
  49. Renart, A. & van Rossum, M.C. Transmission of population-coded information. Neural Comput. 24, 391407 (2012).
  50. Tian, L. et al. Imaging neural activity in worms, flies and mice with improved GCaMP calcium indicators. Nat. Methods 6, 875881 (2009).
  51. Brainard, D.H. The Psychophysics Toolbox. Spat. Vis. 10, 433436 (1997).
  52. Pelli, D.G. The VideoToolbox software for visual psychophysics: transforming numbers into movies. Spat. Vis. 10, 437442 (1997).
  53. Peirce, J.W. PsychoPy–Psychophysics software in Python. J. Neurosci. Methods 162, 813 (2007).
  54. Ohki, K., Chung, S., Ch'ng, Y.H., Kara, P. & Reid, R.C. Functional imaging with cellular resolution reveals precise micro-architecture in visual cortex. Nature 433, 597603 (2005).
  55. Kara, P. & Boyd, J.D. A micro-architecture for binocular disparity and ocular dominance in visual cortex. Nature 458, 627631 (2009).
  56. Hendel, T. et al. Fluorescence changes of genetic calcium indicators and OGB-1 correlated with neural activity and calcium in vivo and in vitro. J. Neurosci. 28, 73997411 (2008).
  57. Chen, T.W. et al. Ultrasensitive fluorescent proteins for imaging neuronal activity. Nature 499, 295300 (2013).
  58. Chen, L.M. et al. A chamber and artificial dura method for long-term optical imaging in the monkey. J. Neurosci. Methods 113, 4149 (2002).
  59. Averbeck, B.B. & Lee, D. Effects of noise correlations on information encoding and decoding. J. Neurophysiol. 95, 36333644 (2006).

Download references

Author information

  1. Present address: Department of Organismal Biology and Anatomy and Department of Neurobiology, University of Chicago, Chicago, Illinois, USA.

    • Audrey Sederberg
  2. These authors contributed equally to this work.

    • Gordon B Smith &
    • Audrey Sederberg
  3. These authors jointly directed this work.

    • Matthias Kaschube &
    • David Fitzpatrick

Affiliations

  1. Department of Functional Architecture and Development of Cerebral Cortex, Max Planck Florida Institute for Neuroscience, Jupiter, Florida, USA.

    • Gordon B Smith,
    • Yishai M Elyada &
    • David Fitzpatrick
  2. Department of Physics, Princeton University, Princeton, New Jersey, USA.

    • Audrey Sederberg
  3. Department of Biology, Brandeis University, Waltham, Massachusetts, USA.

    • Stephen D Van Hooser
  4. Frankfurt Institute for Advanced Studies, Frankfurt, Germany.

    • Matthias Kaschube
  5. Faculty of Computer Science and Mathematics, Goethe University, Frankfurt, Germany.

  6. Bernstein Focus: Neurotechnology Frankfurt, Frankfurt, Germany.

    • Matthias Kaschube

Contributions

G.B.S., A.S., M.K. and D.F. designed the study, analyzed the results and wrote the paper. G.B.S. performed the acute and longitudinal GCaMP imaging. Y.M.E. developed the method for longitudinal imaging. S.D.V.H. originally acquired the motion training data and prepared these data for the additional analyses reported here.

Competing financial interests

The authors declare no competing financial interests.

Corresponding authors

Correspondence to:

Author details

Supplementary information

Supplementary Figures

  1. Supplementary Figure 1: Responses in identified neurons imaged repeatedly over days. (388 KB)

    (a) Six representative neurons imaged on Day 0 (P30), Day 3 (P33), and Day 5 (P35). Scale bar is 50 µm. (b) Responses for individual neurons highlighted in a. Left column: Response to 8 directional stimuli, averaged across trials. Right column: Tuning curves fit with a 2-peaked Gaussian. Horizontal line indicates the mean response to a blank stimulus.

  2. Supplementary Figure 2: Development of orientation and direction selective responses in identified neurons over days. (285 KB)

    (a,c) Orientation preference is stable over days in longitudinally imaged animals. (b,d) Direction preference is stable in the majority of neurons, whereas a subset of cells exhibit 180 degree reversals. Red dashed lines indicate 180 degree shift from d0. (e) Orientation selectivity increases over days in identified neurons. Red dot indicates mean across neurons (s.e.m. error bars are smaller than marker). (f) DSI increases from the initial to final imaging session. Neurons with negative initial DSI (black symbols, mean indicated by green dot) reversed their direction preference over imaging sessions, but on average do not show an increase in selectivity. Neurons that maintain a preferred direction (positive initial DSI, blue symbols, mean indicated by red dot) exhibit significantly increased selectivity over days. Error bars indicating s.e.m. are smaller than markers for mean. Panels d & f are re-plotted from Figure 5 to facilitate comparison with orientation preference and selectivity.

  3. Supplementary Figure 3: Evolution of noise correlation matrix and relative preferred direction over five days of development. (634 KB)

    Cells are sorted by their preferred direction on the last day. (a-c): Noise correlation on day 0, day 3, and day 5. (d-f): Angular difference between cell pairs on day 0, day 3, and day 5. Correlations generally decrease from day 0 to day 5, with a prominent decrease in sets of cells sharing a preferred direction (cells 1 to 40).

  4. Supplementary Figure 4: Initial noise correlations are higher for pairs that will ultimately adopt similar preferred directions. (640 KB)

    (a) Cartoon depicting possible relationships in pairwise angular preference across imaging sessions, re-plotted from Figure 6. (b) Pairwise noise correlations during initial imaging as function of pairwise angular difference in preferred direction. Colored squares indicate regions of plot classified as S-S, S-O, O-O, and O-S, and quantified in Figure 6d,e.

  5. Supplementary Figure 5: Possible scenarios for the maturation of direction discriminability. (84 KB)

    (a) Possible scenarios for the maturation of single cell discriminability. Stimuli to be discriminated are moving gratings with opposite directions of motion. Early cortex (left): Response distribution of a single cell for a left moving grating (blue) and a right moving grating (red). Mean responses are indicated by vertical lines. Direction selectivity is commonly defined as the difference of mean responses divided by their sum. Single trial discriminability of motion direction critically depends on the fraction of overlap between the two distributions. If the overlap is close to 100%, discriminability is close to 0. If the overlap is small, discriminability is close to 1. During development, discriminability can improve by shifting the mean responses apart, i.e. by increasing the cell’s direction selectivity (right, upper). However, discriminability can also improve by reducing the response variance (without changing the selectivity; right, lower). (b) Possible scenarios for the maturation of multi-cell discriminability, here for a population of two neurons. Early cortex (left): Response distribution for a left (blue) and a right moving grating (red). Assuming response fluctuations follow a Gaussian statistics, the bivariate distributions have elliptic shapes (marking the 95% confidence region for the response distributions). Mean response vectors are marked by dots. In the shown example both cells are tuned for the same stimulus direction and their noise correlation is positive. As in the single cell case, discriminability depends on the fraction of overlap between the two distributions. This fractional overlap can decrease over development as cells become more selective (right, upper), as the overall magnitude of fluctuations decreases (right, middle) or as noise correlations become smaller (right, lower).

  6. Supplementary Figure 6: Individual neurons increase discriminability despite reduced selectivity due to decreased variability. (273 KB)

    (a) Discriminability increases significantly over days in identified neurons. (b) The majority of neurons exhibit increased discriminability over days (light blue). These neurons are analyzed further in panels c-i. (c) Of the neurons exhibiting increased discriminability over days, the majority also display increased direction selectivity (HS cells, red), whereas a subset show decreased selectivity over days (LS cells, blue). (d) In LS cells, neither the preferred nor the null response changes significantly. (e) HS neurons achieve enhanced selectivity both through an increase in the preferred response and a decrease in the null response. (f) Variability decreases significantly for both LS and HS groups, however variability decreases to a greater extent in cells without increased selectivity (blue) than in those where selectivity increases (red). (g) For LS cells, only the change in variance contributes positively to the increase in discriminability. (h) For HS cells, both the decreased variance and the increased selectivity contribute significantly to increased discriminability, with the change in mean response providing a significantly greater contribution. (i) Changes in variance provide a significantly greater contribution to improved discriminability in LS vs. HS neurons.

  7. Supplementary Figure 7: Noise correlations decrease asymmetrically in longitudinally imaged animals. (55 KB)

    Correlations decrease over imaging sessions to a larger extent in S-S vs. S-O pairs. Likewise, the decrease in correlations is larger in O-S vs. O-O pairs. Change in correlations are normalized within animal to the change exhibited by S-S pairs. (*) indicates MW test, p<0.01. Error bars indicate s.e.m.

  8. Supplementary Figure 8: Change in response variance vs. change in selectivity. (131 KB)

    (a) Distribution of standard deviation of response to preferred stimulus for all cells, pre- and post-training. Variance of response does not change significantly over motion training. (b) Response variance is calculated across the preferred orientation before and after training. Each point represents the average change for a single field of view. (c) Overall change in standard deviation for each training type. (d) Same as (b), but for the orientation orthogonal to the preferred orientation. (e) Same as (c), but for the orthogonal orientation. Error bars represent +/- 1 SEM.

  9. Supplementary Figure 9: Single neuron response distributions across age groups. (372 KB)

    Distribution of activity levels over all stimulus conditions with stimulus means subtracted (see Eqn. 8, 9, Methods) for 10 representative and randomly selected cells from the naive (a), immature (b), and mature (c) populations of cells.

Video

  1. Video 1: Supplementary Video 1: Wave-like responses to drifting grating stimuli at eye opening. (27.86 MB, Download)
    Response to two presentations of drifting grating stimuli with the same orientation, drifting in opposite directions. Movie is shown at 2x real-time. Stimulus presentation and direction are indicated by arrow in upper left. Scale bar is 50 μm. 512 × 256 images with a 1:2 aspect ratio were acquired at 60 Hz, downsampled to 15 Hz and resized to 512 × 512 by bilinear interpolation. Images were then filtered in x-y by a 2×2 pixel radius mean filter, Z-scored relative to the mean and standard deviation of the blank stimulus, and clipped at 0 to 4 Z-units.

PDF files

  1. Supplementary Text and Figures (5,878 KB)

    Supplementary Figures 1–9 and Supplementary Tables 1–6

  2. Supplementary Methods Checklist (649 KB)

Additional data