The primary motor cortex (M1) possesses two intermediate layers upstream of the motor-output layer: layer 2/3 (L2/3) and layer 5a (L5a). Although repetitive training often improves motor performance and movement coding by M1 neuronal ensembles, it is unclear how neuronal activities in L2/3 and L5a are reorganized during motor task learning. We conducted two-photon calcium imaging in mouse M1 during 14 training sessions of a self-initiated lever-pull task. In L2/3, the accuracy of neuronal ensemble prediction of lever trajectory remained unchanged globally, with a subset of individual neurons retaining high prediction accuracy throughout the training period. However, in L5a, the ensemble prediction accuracy steadily improved, and one-third of neurons, including subcortical projection neurons, evolved to contribute substantially to ensemble prediction in the late stage of learning. The L2/3 network may represent coordination of signals from other areas throughout learning, whereas L5a may participate in the evolving network representing well-learned movements.
Subscribe to Journal
Get full journal access for 1 year
only $17.42 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
Weiler, N., Wood, L., Yu, J., Solla, S.A. & Shepherd, G.M.G. Top-down laminar organization of the excitatory network in motor cortex. Nat. Neurosci. 11, 360–366 (2008).
Anderson, C.T., Sheets, P.L., Kiritani, T. & Shepherd, G.M.G. Sublayer-specific microcircuits of corticospinal and corticostriatal neurons in motor cortex. Nat. Neurosci. 13, 739–744 (2010).
Thoroughman, K.A. & Shadmehr, R. Learning of action through adaptive combination of motor primitives. Nature 407, 742–747 (2000).
Todorov, E. & Jordan, M.I. Optimal feedback control as a theory of motor coordination. Nat. Neurosci. 5, 1226–1235 (2002).
Sing, G.C., Joiner, W.M., Nanayakkara, T., Brayanov, J.B. & Smith, M.A. Primitives for motor adaptation reflect correlated neural tuning to position and velocity. Neuron 64, 575–589 (2009).
Scott, S.H. The computational and neural basis of voluntary motor control and planning. Trends Cogn. Sci. 16, 541–549 (2012).
Mitz, A.R., Godschalk, M. & Wise, S.P. Learning-dependent neuronal activity in the premotor cortex: activity during the acquisition of conditional motor associations. J. Neurosci. 11, 1855–1872 (1991).
Nudo, R.J., Milliken, G.W., Jenkins, W.M. & Merzenich, M.M. Use-dependent alterations of movement representations in primary motor cortex of adult squirrel monkeys. J. Neurosci. 16, 785–807 (1996).
Sanes, J.N. & Donoghue, J.P. Plasticity and primary motor cortex. Annu. Rev. Neurosci. 23, 393–415 (2000).
Ungerleider, L.G., Doyon, J. & Karni, A. Imaging brain plasticity during motor skill learning. Neurobiol. Learn. Mem. 78, 553–564 (2002).
Kleim, J.A. et al. Cortical synaptogenesis and motor map reorganization occur during late, but not early, phase of motor skill learning. J. Neurosci. 24, 628–633 (2004).
Laubach, M., Wessberg, J. & Nicolelis, M.A.L. Cortical ensemble activity increasingly predicts behavior outcomes during learning of a motor task. Nature 405, 567–571 (2000).
Carmena, J.M. et al. Learning to control a brain-machine interface for reaching and grasping by primates. PLoS Biol. 1, e42 (2003).
Cohen, D. & Nicolelis, M.A.L. Reduction of single-neuron firing uncertainty by cortical ensembles during motor skill learning. J. Neurosci. 24, 3574–3582 (2004).
Paz, R. & Vaadia, E. Learning-induced improvement in encoding and decoding of specific movement directions by neurons in the primary motor cortex. PLoS Biol. 2, e45 (2004).
Huber, D. et al. Multiple dynamic representations in the motor cortex during sensorimotor learning. Nature 484, 473–478 (2012).
Rokni, U., Richardson, A.G., Bizzi, E. & Seung, H.S. Motor learning with unstable neural representations. Neuron 54, 653–666 (2007).
Chestek, C.A. et al. Single-neuron stability during repeated reaching in macaque premotor cortex. J. Neurosci. 27, 10742–10750 (2007).
Lütcke, H., Margolis, D.J. & Helmchen, F. Steady or changing? Long-term monitoring of neuronal population activity. Trends Neurosci. 36, 375–384 (2013).
Carmena, J.M., Lebedev, M.A., Henriquez, C.S. & Nicolelis, M.A.L. Stable ensemble performance with single-neuron variability during reaching movements in primates. J. Neurosci. 25, 10712–10716 (2005).
Hira, R. et al. Spatiotemporal dynamics of functional clusters of neurons in the mouse motor cortex during a voluntary movement. J. Neurosci. 33, 1377–1390 (2013).
Tennant, K.A. et al. The organization of the forelimb representation of the C57BL/6 mouse motor cortex as defined by intracortical microstimulation and cytoarchitecture. Cereb. Cortex 21, 865–876 (2011).
Isomura, Y., Harukuni, R., Takekawa, T., Aizawa, H. & Fukai, T. Microcircuitry coordination of cortical motor information in self-initiation of voluntary movements. Nat. Neurosci. 12, 1586–1593 (2009).
Tian, L. et al. Imaging neural activity in worms, flies and mice with improved GCaMP calcium indicators. Nat. Methods 6, 875–881 (2009).
Thévenaz, P., Ruttimann, U.E. & Unser, M. A pyramid approach to subpixel registration based on intensity. IEEE Trans. Image Process. 7, 27–41 (1998).
Dombeck, D.A., Khabbaz, A.N., Collman, F., Adelman, T.L. & Tank, D.W. Imaging large-scale neural activity with cellular resolution in awake, mobile mice. Neuron 56, 43–57 (2007).
Vogelstein, J.T. et al. Spike inference from calcium imaging using sequential monte carlo methods. Biophys. J. 97, 636–655 (2009).
Drucker, H., Burges, C.J.C., Kaufman, L., Smola, A. & Vapnik, V. Support vector regression machines. Adv. Neural Inf. Process. Syst. 9, 155–161 (1997).
Nelson, R.B. An Introduction to Copulas, 2nd edn. (Springer, New York, 2006).
Davy, M. & Doucet, A. Copulas: a new insight into positive time-frequency distributions. IEEE Signal Process. Lett. 10, 215–218 (2003).
Wessberg, J. et al. Real-time prediction of hand trajectory by ensembles of cortical neurons in primates. Nature 408, 361–365 (2000).
Isomura, Y. et al. Reward-modulated motor information in identified striatum neurons. J. Neurosci. 33, 10209–10220 (2013).
Yin, H.H. et al. Dynamic reorganization of striatal circuits during the acquisition and consolidation of a skill. Nat. Neurosci. 12, 333–341 (2009).
Kiritani, T., Wickersham, I.R., Seung, H.S. & Shepherd, G.M. Hierarchical connectivity and connection-specific dynamics in the corticospinal-corticostriatal microcircuit in mouse motor cortex. J. Neurosci. 32, 4992–5001 (2012).
Masamizu, Y. et al. Local and retrograde gene transfer into primate neuronal pathways via adeno-associated virus serotype 8 and 9. Neuroscience 193, 249–258 (2011).
Abbott, L.F. & Nelson, S.B. Synaptic plasticity: taming the beast. Nat. Neurosci. 3, 1178–1183 (2000).
Mao, T. et al. Long-range neuronal circuits underlying the interaction between sensory and motor cortex. Neuron 72, 111–123 (2011).
Thorn, C.A., Atallah, H., Howe, M. & Graybiel, A.M. Differential dynamics of activity changes in dorsolateral and dorsomedial striatal loops during learning. Neuron 66, 781–795 (2010).
Yin, H.H. & Knowlton, B.J. The role of the basal ganglia in habit formation. Nat. Rev. Neurosci. 7, 464–476 (2006).
Iwaniuk, A.N. & Whishaw, I.Q. On the origin of skilled forelimb movements. Trends Neurosci. 23, 372–376 (2000).
Kimura, R. et al. Reinforcing operandum: rapid and reliable learning of skilled forelimb movements by head-fixed rodents. J. Neurophysiol. 108, 1781–1792 (2012).
Morishima, M. & Kawaguchi, Y. Recurrent connection patterns of corticostriatal pyramidal cells in frontal cortex. J. Neurosci. 26, 4394–4405 (2006).
Hira, R. et al. In vivo optogenetic tracing of functional corticocortical connections between motor forelimb areas. Front. Neural Circuits 7, 55 (2013).
Xiao, W. et al. Gene therapy vectors based on adeno-associated virus type 1. J. Virol. 73, 3994–4003 (1999).
Okada, T. et al. Scalable purification of adeno-associated virus serotype 1 (AAV1) and AAV8 vectors, using dual ion-exchange adsorptive membranes. Hum. Gene Ther. 20, 1013–1021 (2009).
Holtmaat, A. et al. Long-term, high-resolution imaging in the mouse neocortex through a chronic cranial window. Nat. Protoc. 4, 1128–1144 (2009).
Mittmann, W. et al. Two-photon calcium imaging of evoked activity from L5 somatosensory neurons in vivo. Nat. Neurosci. 14, 1089–1093 (2011).
Glickfeld, L.L., Andermann, M.L., Bonin, V. & Reid, R.C. Cortico-cortical projections in mouse visual cortex are functionally target specific. Nat. Neurosci. 16, 219–226 (2013).
Keck, T. et al. Synaptic scaling and homeostatic plasticity in the mouse visual cortex in vivo. Neuron 80, 327–334 (2013).
Helmchen, F. & Denk, W. Deep tissue two-photon microscopy. Nat. Methods 2, 932–940 (2005).
Oheim, M., Beaurepaire, E., Chaigneau, E., Mertz, J. & Charpak, S. Two-photon microscopy in brain tissue: parameters influencing the imaging depth. J. Neurosci. Methods 111, 29–37 (2001).
Kitamura, K., Judkewitz, B., Kano, M., Denk, W. & Häusser, M. Targeted patch-clamp recordings and single-cell electroporation of unlabeled neurons in vivo. Nat. Methods 5, 61–67 (2008).
Zariwala, H.A. et al. A Cre-dependent GCaMP3 reporter mouse for neuronal imaging in vivo. J. Neurosci. 32, 3131–3141 (2012).
Tanaka, Y., Tanaka, Y., Furuta, T., Yanagawa, Y. & Kaneko, T. The effects of cutting solutions on the viability of GABAergic interneurons in cerebral cortical slices of adult mice. J. Neurosci. Methods 171, 118–125 (2008).
O'Connor, D.H., Peron, S.P., Huber, D. & Svoboda, K. Neural activity in barrel cortex underlying vibrissa-based object localization in mice. Neuron 67, 1048–1061 (2010).
Gonchar, Y., Wang, Q. & Burkhalter, A. Multiple distinct subtypes of GABAergic neurons in mouse visual cortex identified by triple immunostaining. Front. Neuroanat. 1, 3 (2007).
Zhao, Y. et al. An expanded palette of genetically encoded Ca2+ indicators. Science 333, 1888–1891 (2011).
We thank M. Himeno, J. Saito, H. Sugiura, T. Sugiyama and K. Ozawa for technical assistance, T. Chiyo for AAV purification assistance, J. Noguchi, K. Sohya and A. Nagaoka for technical advice, K. Ohara for help with the development of the task device, M. Kimura and Y. Sakai for comments on the manuscript, and S. Ozawa for discussion. We are grateful to the Functional Genomics and the Spectrography and Bioimaging Facility at National Institute for Basic Biology for allowing the use of their equipment. We thank L.L. Looger (Howard Hughes Medical Institute) for providing GCaMP3 vector (Addgene plasmid 22692), rAAV2/9-Syn-GCaMP3 and rAAV2/9-Syn-Flex-GCaMP3, K. Deisseroth (Stanford University) for providing pAAV (Addgene plasmid 26973), and J.M. Wilson (University of Pennsylvania) for providing helper plasmids pAAV2-1 and pAAV2-9. This work was supported by Grants-in-Aid for Young Scientists (no. 19680020 to M.M., 26830020 to Y.M., and 22680031 to K.K.), Scientific Research on Innovative Areas 'Mesoscopic Neurocircuitry' (no. 22115005 to M.M. and 23115504 to K.K.) and 'Neural Creativity for Communication' (no. 22120520 to Y.I.), Challenging Exploratory Research (no. 22650083 to K.K.), Scientific Research (no. 23300148 to M.M.), the Japan Society for the Promotion of Science Research Fellowships for Young Scientists (no. 268449 to Y.R.T., 253960 to Y.H.T. and 22136 to R.H.), Research Activity Start-up (no. 23800071 to Y.M.), the Strategic Research Program for Brain Sciences (to M.M.) from the Ministry of Education, Culture, Sports, Science, and Technology, grants from the Mitsubishi Foundation, Takeda Foundation and Toyoaki Foundation to M.M., and by the Uehara Memorial Foundation and Brain Science Foundation to Y.I.
The authors declare no competing financial interests.
Integrated supplementary information
Supplementary Figure 1 Spatial distribution of all reconstructed ROIs over sessions, representative ROI selections, and ROI areas.
(a, b) Contours of all ROIs in the seven L2/3 fields (a; A–G) and seven L5a fields (b; H–N) in each of the analyzed sessions. The horizontal location and size of each field are shown in Figure 2b.The session number is stated on the left of each image. Each field indicates the field in which ROIs were selected after motion correction was applied in each session. In some sequential sessions there was a very high correlation between the images after correction for lateral displacement, and the same ROI morphologies were assigned. Scale bars, 100 μm. (c) Top, two representative mean images showing cell morphology and their corresponding 2-pixel cluster images delineated by assessing correlations between pixels. Cluster 1 was cell-shaped in the left image, but was not in the right image. Bottom, traces of fluorescence intensity averaged within the numbered areas in top panels. In the left image, fluorescent changes in cell-shaped cluster 1 are distinct from the background in cluster 2; thus, cluster 1 was accepted as a reconstructed neuron. In the right image, no distinct peak was found in the fluorescent intensity trace in cluster 1; thus, it was not accepted. Scale bars, 5 μm. (d) Cumulative histograms of areas of all ROIs determined by the ROI selection method (black) and ROIs manually determined from the soma morphology (gray) in L2/3 (left) and L5a (right). One-hundred manually-determined ROIs were obtained from three fields in each layer. Dotted vertical lines denote the median. Note that L5a neurons are larger than L2/3 neurons.
(a) Contours of all ROIs for the reconstructed neurons overlaid in the L2/3 field images in training sessions 2 (left) and 14 (right), as shown in Figure 2d. The numbered ROIs (1–13) correspond to the neurons that were monitored in more than seven training sessions, including sessions 2 and 14. Scale bar, 50 μm. Motion-corrected calcium transients of ROIs 1–10 are shown in Figure 2f. (b) Contours of all ROIs for the reconstructed neurons overlaid in the L5a field images in training sessions 2 (left) and 11 (right), as shown in Figure 2e. The numbered ROIs (1–26) correspond to the neurons that were monitored in more than seven training sessions, including the sessions 2 and 11. Scale bar, 50 μm. Motion-corrected calcium transients of ROIs 1–10 are shown in Figure 2g. (c) Motion-corrected calcium transients of all reconstructed neurons in sessions 2 and 14 that are not shown in Figure 2f. Traces 11–13 are from ROIs 11–13 in both sessions. The other traces are from unnumbered ROIs in each session. The lever trajectory, reward timing, and occurrence of licking behaviors are shown in the bottom three traces. The trace period is the same as in Figure 2f. (d) Motion-corrected calcium transients of all reconstructed neurons in sessions 2 and 11 that are not shown in Figure 2g. Traces 11–26 are from ROIs 11–26 in both sessions. The other traces are from unnumbered ROIs in each session. The lever trajectory, reward timing, and occurrence of licking behaviors are shown in the bottom three traces. The trace period is the same as in Figure 2g.
Supplementary Figure 3 The relationship between calcium transients and action potentials, and the changes in the baseline fluctuation of ΔF/F and the signal-to-noise ratio of ΔF/F from early to late sessions.
(a) Left, an example of in vivo two-photon images of cell-attached L2/3 neurons that expressed GCaMP3. Pipette containing Alexa Fluor 594 is colored red. Scale bar, 10 μm. Right, mean traces of ΔF/F of four L2/3 neurons from two mice when 1, 2, 3, 5, and 10 action potentials (APs) naturally occurred. (b) Mean ΔF/F as a function of the number of recorded APs. The number of APs between two adjacent imaging time points and the peak ΔF/F at the latter imaging time point were averaged for every sample. Data were obtained from four L2/3 neurons in two mice. Each color indicates a different neuron. Error bars indicate s.e.m. (c) Left, an example of simultaneous two-photon imaging and whole-cell recording from a GCaMP3-expressing neuron in L5 in the cortical slice. Pipette containing Alexa Fluor 594 is colored red. Scale bar, 10 μm. Right, mean traces of ΔF/F of four L5 neurons from two mice for trials exhibiting 1–5, 6–11, and 20–21 APs that were induced by current injections through the patch pipette (3–10 nA, 2–5 ms) at 83 Hz. (d) Mean ΔF/F as a function of the number of recorded APs. The number of APs between two adjacent imaging time points and the peak ΔF/F at the latter imaging time point were averaged for every sample. Data were obtained from four L5 neurons in two mice. (e) Histograms of baseline fluctuation of ΔF/F of all neurons during non-lever-pull periods (from 1 s after the end of a lever pull to 1 s before the onset of the next lever pull) in early (left top; n = 993 neurons from 7 fields of 6 mice) and late sessions (left bottom; n = 1005 neurons from 7 fields of 6 mice) in L2/3 and in early (right top; n = 1269 neurons from 7 fields of 6 mice) and late sessions (right bottom; n = 1193 neurons from 7 fields of 6 mice) in L5a. For each neuron, the baseline fluctuation of ΔF/F was defined as the standard deviation of values that were less than the median during non-lever-pull periods in each session. (f) Histograms of signal-to-noise ratios of ΔF/F for all neurons during non-lever-pull periods in early (left top) and late sessions (left bottom) in L2/3 and in early (right top) and late sessions (right bottom) in L5a. For each neuron, the signal-to-noise ratio of ΔF/F was defined as the 95 percentile value during non-lever-pull periods divided by the baseline fluctuation of ΔF/F during non-lever-pull periods in each session.
Supplementary Figure 4 Lever movement represented forepaw movement throughout the training sessions.
(a) Trajectory of the right forepaw (magenta lines) during a 5 minute period of two-photon imaging in session 1 (left), session 2 (middle), and session 14 (right) of the same animal overlaid on single frames from video recordings. The video rate was 30 Hz. The center of the forepaw was tracked by a particle filter-based method and confirmed by visual inspection. (b) Two-dimensional (x and y) forepaw trajectories in the middle in a (top two rows), the lever trajectory linearly fitted from the two-dimensional forepaw trajectories (third row), and the lever trajectory measured using a magnetic sensor (bottom row). (c) Time course of the correlation coefficient between the recorded lever trajectory and the fitted lever trajectory in each imaging session (open circle, L2/3, n = 6 fields from 5 mice; closed circle, L5a, n = 6 fields from 5 mice). Error bars indicate s.e.m. In the other two fields, the video camera was directed from the front of the mouse face so that the forepaw trajectory when the lever was pulled could not be sufficiently resolved. In the correlation coefficients, there was statistically non-significant difference throughout training or between L2/3 and L5a imaging sessions (P = 0.19, F[13, 113] = 1.37, and P = 0.94, F[1,113] = 0.79, respectively, by two-way ANOVA). This indicates that it is unlikely that a change in the number of non-lever-related forepaw movements during learning had a large effect on the change in neuronal coding of lever movement in L2/3 and L5a. (d) Time course of the mean correlation coefficients during periods starting 0.2 s before and ending 0.9 s after the start of each successful lever pull between the recorded lever trajectory and the fitted lever trajectory in each session (open circle, L2/3, n = 6 fields from 5 mice; closed circle, L5a, n = 6 fields from 5 mice). Error bars indicate s.e.m. The correlation coefficients around and during lever pulling were high throughout training and showed statistically non-significant difference throughout training or between L2/3 and L5a imaging sessions (P = 0.87, F[13. 113] = 0.58, and P = 0.38, F[1, 113] = 0.79, respectively, by two-way ANOVA).
Supplementary Figure 5 Estimation of mutual information (Î) between bivariate distributions using copula function and Akaike information criterion (AIC).
(a) Representative distribution of the experimentally-obtained lever trajectory (x; n = 1654 frame points) and the corresponding predicted lever trajectory (y). Bottom and left histograms approximate the corresponding marginal density of x and y. (b) The empirical copula frequency for the distribution in panel a. (c) Distribution of random variables (n = 1654 points) generated according to beta distributions (α = 0.06, β = 0.265 for x; α = 1.7, β = 12.6 for y) coupled by a Gaussian copula whose parameter, r, was set at 0.5. (d) The empirical copula frequency for the distribution in panel c. (e,f) The AIC of the estimated copula density (e) and Î (f) for the distribution in panel c, according to m, the number of histogram bins in the interval [0,1]. In this case, the AIC is minimized at m = 6 (green line) and the corresponding Î is 0.205 bit, which is close to 0.208 bit, the theoretical mutual information for this distribution (red line). Inset in panel e is the estimated copula density. White areas denote high-density binned areas. (g) The estimated mutual information based on the AIC (blue circle; 200 iterations) had a relatively small bias from its theoretical value (red). Green lines show the estimation of mutual information using a fixed bin number (m = 10, diamond; m = 15, square; m = 20, triangle; 200 iterations each) with Miller–Madow bias correction. Error bars indicate s.d. (h) The bias of Î with the copula function and the AIC (blue; 200 iterations) was relatively small for distributions with various theoretical values of mutual information. The number of random variables was set at 1024. Error bars indicate s.d. (i) The mutual information calculated from the copula entropy (the predictive information, Iensemble) relative to R2 (squared correlation coefficient) from all sessions in all imaged fields. Iensemble and R2 were calculated from the predicted lever trajectory when using all available frames and 20 reconstructed neurons in each session in each field. (j) Iensemble calculated from 768 randomly chosen frames relative to Iensemble calculated from all available frames (768–4447) in each session in each field. The number of neurons that were used was from one to the maximum number of reconstructed neurons in each session in each field. The red line shows equality. (k) Iensemble averaged over early and late sessions for each field (L2/3, P = 1, n = 7 fields from 6 mice; L5a, P = 0.016, n = 7 fields from 6 mice, sign test), where Iensemble was directly calculated according to the formula I(x,y) = H(x) + H(y) – H(x,y), where H(x) and H(y) are the entropy of x and y, respectively, and H(x, y) is the joint entropy of x and y. The bin number in x was fixed to 6 and the bin number in y was determined with AIC calculation. *P < 0.05.
Supplementary Figure 6 Iensemble changes did not depend on the horizontal location of the imaged field or the fluorescence intensity.
(a) The ratio of Iensemble from 20 neurons averaged over late sessions to Iensemble from 20 neurons averaged over early sessions (late/early Iensemble ratio) plotted according to the anterior-posterior (AP) location of the center of the field in each of the six L2/3 fields from 5 mice (top) and seven L5a fields from 6 mice (bottom) that had ≥20 neurons in at least one early session and one late session. The P-value of the Spearman's rank correlation was 0.46 for L2/3 and 0.58 for L5a. (b) Late/early Iensemble ratio plotted according to the medial-lateral location of the center of the field for each of the six L2/3 fields from 5 mice (top) and seven L5a fields from 6 mice (bottom) that had ≥20 neurons in at least one early session and one late session. The P-value of the Spearman's rank correlation was 0.65 for L2/3 and 0.09 for L5a. (c) The relation between the mean baseline fluorescence intensity of a group of 20 ROIs (neurons) and their Iensemble in each session in L2/3 (top) and L5a (bottom) fields. Data are summed across all sessions and fields in which ≥20 neurons were reconstructed (L2/3, n = 72 sessions from 7 fields from 6 mice; L5a, n = 82 sessions from 7 fields from 6 mice). The regression lines are shown. The P-value of the Pearson's correlation was 0.59 for L2/3 and 0.80 for L5a. (d) The relation between the mean coefficient of variation (CV) of baseline fluorescence intensity of a group of 20 ROIs (neurons) and their Iensemble in each session in L2/3 (top) and L5a (bottom) fields. The regression lines are shown. The P-value of the Pearson's correlation was 0.58 for L2/3 (n = 72 sessions from 7 fields from 6 mice) and 2.5 × 10–4 for L5a (n = 82 sessions from 7 fields from 6 mice).
Supplementary Figure 7 Temporal correlations between the lever trajectory and licking, and Iensemble for licking.
(a) A single frame of a video recording made during an imaging session. The yellow area indicates the ROI that includes the tip of the mouth. (b) The intensity of the yellow ROI in panel a during part of an imaging session (5 min). (c) The amplitude of the 5–9 Hz component of the trace shown in panel b, quantified from a Fourier transform of the data. The 5–9 Hz frequency band was examined because licking behaviors are repeated at a frequency of 6–8 Hz (Isomura et al., Nat. Neurosci., 2009; Kimura et al., J. Neurophysiol., 2012). (d) The number of licking per imaging frame (4.5 Hz). (e) The lever trajectory during the same period as panel d measured using a magnetic sensor. (f) An example of a cross-correlogram between the lever trajectory and recorded lick-rate trace. Recorded lick-rate was obtained by convolution of the number of licking with a Gaussian function (width parameter was set to the single frame). Negative time lag means that the lever movement preceded the lick. (g) The peak amplitude of the cross-correlogram between lever trajectory and the lick-rate trace in early and late sessions in which L2/3 fields (left) and L5a fields (right) were imaged (L2/3, P = 0.016, n = 7 fields from 6 mice; L5a, P = 0.031, n = 7 fields from 6 mice, Wilcoxon signed-rank test). Each line indicates a different field. *P < 0.05. (h) Iensemble between the recorded and predicted lick-rate traces averaged over early and late sessions in each field. (L2/3, P = 0.45, n = 7 fields from 6 mice; L5a, P = 0.016, n = 7 fields from 6 mice, sign test). *P < 0.05. (i) The time lag of the peak correlogram between the lever trajectory and lick-rate trace in early and late sessions in each field. (L2/3, P = 0.41, n = 7 fields from 6 mice; L5a, P = 0.38, n = 7 fields from 6 mice, Wilcoxon signed-rank test).
Inferred spike event rate of individual neurons in L2/3 (left) and L5a (right) during non-lever-pull periods averaged over early and late sessions. Inferred spike event rate for each neuron was defined as the sum of the inferred spike events during the non-lever-pull periods divided by the total duration of the non-lever-pull periods during that session. Each line indicates a pursued neuron. Thin red, blue, and gray lines indicate increase-, decrease-, and other-neurons respectively. Thick red, blue, and gray lines indicate the averaged values across increase-, decrease-, and other-neurons respectively. In each layer, no significant difference was detected between sessions in any neuron group or between any pair of neuron groups in any sessions. The smallest P-value observed was 0.086 for L2/3 and 0.44 for L5a (Wilcoxon rank sum test with Bonferroni correction).
Supplementary Figure 1–8, Supplementary Table 1 and 2, and Supplementary Modeling (PDF 11812 kb)
Real-speed video of the mouse behavior in session 2 (20 s duration). (AVI 9569 kb)
Real-speed video of the mouse behavior in session 14 (20 s duration). The mouse is the same as in Supplementary Video 1. (AVI 9426 kb)
About this article
Cite this article
Masamizu, Y., Tanaka, Y., Tanaka, Y. et al. Two distinct layer-specific dynamics of cortical ensembles during learning of a motor task. Nat Neurosci 17, 987–994 (2014). https://doi.org/10.1038/nn.3739
Current Opinion in Neurobiology (2021)
Frontiers in Behavioral Neuroscience (2020)
Structural dynamics and stability of corticocortical and thalamocortical axon terminals during motor learning
PLOS ONE (2020)