Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Oct;25(10):1300-1313.
doi: 10.1038/s41593-022-01169-4. Epub 2022 Sep 30.

Motor learning drives dynamic patterns of intermittent myelination on learning-activated axons

Clara M Bacmeister  1   2 , Rongchen Huang  3   4 , Lindsay A Osso  1 , Michael A Thornton  1 , Lauren Conant  1 , Anthony R Chavez  1 , Alon Poleg-Polsky  4 , Ethan G Hughes  5
Affiliations

Motor learning drives dynamic patterns of intermittent myelination on learning-activated axons

Clara M Bacmeister et al. Nat Neurosci. 2022 Oct.

Abstract

Myelin plasticity occurs when newly formed and pre-existing oligodendrocytes remodel existing patterns of myelination. Myelin remodeling occurs in response to changes in neuronal activity and is required for learning and memory. However, the link between behavior-induced neuronal activity and circuit-specific changes in myelination remains unclear. Using longitudinal in vivo two-photon imaging and targeted labeling of learning-activated neurons in mice, we explore how the pattern of intermittent myelination is altered on individual cortical axons during learning of a dexterous reach task. We show that behavior-induced myelin plasticity is targeted to learning-activated axons and occurs in a staged response across cortical layers in the mouse primary motor cortex. During learning, myelin sheaths retract, which results in lengthening of nodes of Ranvier. Following motor learning, addition of newly formed myelin sheaths increases the number of continuous stretches of myelination. Computational modeling suggests that motor learning-induced myelin plasticity initially slows and subsequently increases axonal conduction speed. Finally, we show that both the magnitude and timing of nodal and myelin dynamics correlate with improvement of behavioral performance during motor learning. Thus, learning-induced and circuit-specific myelination changes may contribute to information encoding in neural circuits during motor learning.

PubMed Disclaimer

Conflict of interest statement

Competing Interests

The authors declare no competing financial interests.

Figures

Extended Data Fig. 1
Extended Data Fig. 1. Near-infrared branding identifies the same oligodendrocytes and myelin sheaths in longitudinally, in vivo imaged areas and post-hoc stained tissue.
a, The same field of view imaged in vivo (“Live Tissue”, left) and fixed, sectioned, and stained tissue (“Fixed Tissue”, right). Patterns of cell bodies (examples outlined in white dotted lines) were maintained across live and processed tissue. Note the new oligodendrocyte generated at 60d and delineated with a white arrowhead. b, A newly generated oligodendrocyte in vivo (“Live Tissue”, left) and fixed, sectioned tissue (“Fixed Tissue”, right) stained for oligodendrocytes and myelin (GFP, green) and axons (NFH, magenta). Note the same T-junction across live and fixed samples is marked with the white arrowhead.
Extended Data Fig. 2
Extended Data Fig. 2. Learning trajectory of mice engaging in forelimb reach training.
a, Illustration of forelimb reach training box environment. Mice learn to reach by extending their left hand through a slit in a plexiglass box to grab a pellet and return it to their mouth. b, Imaging and training timelines for untrained (top) and learning (bottom) mice. c, The majority of mice learn to perform the forelimb reach task (learners, green). Learners improve their success rate gradually over the course of seven days of training. In contrast, non-learners maintain a low success rate across time and do not attain higher than a 15% success rate at the end of the training regimen (purple). d, Successful learners of the task perform significantly better on the last day of training (Paired Student’s t-test, t(15)=11.72, p<0.0001) and achieve higher than 15% success rate, in contrast to non-learners which do not improve significantly and do not attain above 15% success on the last day of training. Mice were excluded from data analysis if they did not succeed in at least 10% of reaches across the seven days of training (n = 2 mice). e, f, Breakdown of all mice included in each of the figures, where each dot is a unique color that represents a single mouse and is consistent across graphs. Neither success rate (e) nor change in success rate (f) differ across figures (REML with post-hoc Tukey’s HSD). Bars and error bars represent mean ± s.e.m unless otherwise noted. For detailed statistics, see Supplementary Table 3.
Extended Data Fig. 3
Extended Data Fig. 3. Identifying and tracing nodes and axons across live and fixed tissue.
a, The same node in live imaged (“Live Tissue”, left) and fixed, sectioned tissue (“Fixed Tissue”, right) stained for oligodendrocytes and myelin (GFP, green), sodium channels (NaV, magenta), and axons (NFH, yellow). White arrow points to node across time and in fixed tissue. b,c, 90.90±3.73% of nodes visualized in vivo possess NaV staining characteristic of nodes of Ranvier (b), comparable to values of nodes identified fixed tissue of age-matched controls (c; random sample, nmice = 6; longitudinally imaged, nmice = 3). d, No difference in the proportion of gaps in myelination (larger than 3 microns) with and without appreciable sodium channel distributions in fixed tissue of age-matched controls (nmice = 8; Paired Student’s t-test). e, Representative in vivo imaging and post-hoc immunostaining of a lengthening node with sodium channels (left) and without appreciable sodium channels (right). f, The same node confirmed with NaV staining in (a) is identified in live imaging by a marked decrease in autofluorescence between two GFP-labeled myelin paranodes. g, Axon morphology is accurately identified in vivo as confirmed by post-hoc immunostaining of the same tissue. A lengthening node in vivo (“Live Tissue”, left) and fixed, sectioned tissue (“Fixed Tissue”, middle). Axon trajectory is identified using the shape of the myelin sheath prior to node lengthening (“Live axon morphology”, magenta) and confirmed using post-hoc immunostaining of the same sheaths for axons (NFH, magenta) and myelin (GFP, green) and reconstructions of the axons (“Fixed axon morphology”, magenta). In vivo node morphology (“In vivo nodal morphology”, top right) is reconstructed and confirmed using post-hoc immunostaining and morphological reconstruction (“Fixed nodal morphology”, bottom right). h, Sheath length and change in sheath length is determined by aligning fields of view using fiduciary marks (e.g. cell soma) which extend across the entire duration of the study, including pre-existing oligodendrocytes, which maintain their position throughout the course of imaging (left). Change in sheath length is mirrored by a change in paranode position (right). To determine the change in sheath length, the distance between the initial and final paranodal position is traced using Simple Neurite Tracer and overlayed stacks. Bars and error bars represent mean ± s.e.m unless otherwise noted. For detailed statistics, see Supplementary Table 3.
Extended Data Fig. 4
Extended Data Fig. 4. Sheath retraction, but not sheath growth, is affected by learning a new skill.
a, Sheath length is similar across sheaths with 2, 1, or 0 neighbors. b, Sheaths of many ages initiate sheath dynamics in young adult mice. c, Proportion of dynamic sheaths engaging in growth three weeks before learning, during learning (one week), and in the two weeks after learning. d, Proportion of dynamic sheaths engaging in retraction before, during and after learning. Learning modulates sheath retraction (F2,14 = 6.76, p = 0.0088). During learning, more sheaths retract relative to untrained mice (p = 0.0095; Tukey’s HSD) and relative to pre-learning values in trained mice (p = 0.016; Tukey’s HSD). Two weeks after learning, more sheaths retract relative to pre-learning values in trained mice (p = 0.0358; Tukey’s HSD). e, In learning mice, there are fewer stable sheaths with 2 (p=0.013, t(5.91)=3.49), 1 (p=0.0089, t(2.47)=7.62), and 0 neighbors (p=0.043, t(2.04)=4.55). f, In learning mice, there are more growing sheaths with 1 neighbor (p=0.0071, t(6.00)=−4.00). g, In learning mice, there are more retracting sheaths with 2 (p=0.024, t(4.00)=−3.62) and 1 neighbors (p=0.0050, t(6.00)=−4.31). h, Nodes and gaps in myelin that shorten are not modulated by learning (REML with post-hoc Tukey’s HSD). Bars and error bars represent mean ± s.e.m unless otherwise noted. For detailed statistics, see Supplementary Table 3.
Extended Data Fig. 5
Extended Data Fig. 5. Tracking virally labeled axons across time in vivo.
a, Maximum projection of a cFos+ neuron (AAv8-cFos-ER, red) and surrounding oligodendrocytes and myelin (in vivo, green) (top). Single slice at 159 microns below the pial surface, with white arrow identifying unmyelinated region of a cFos+ axon (middle). Single slice at 185 microns below the pial surface, with pink arrow identifying a myelinated region of a cFos+ axon (bottom). b, 3D projection of the cFos+ neuron and associated myelin sheaths from (a), with dashed lines corresponding to the single slices in (a). c, The same neuron as in (a), reconstructed using Simple Neurite Tracer. d, 3D projection of a longitudinally imaged axon at the outset of the imaging experiment (0d) and at the final imaging timepoint (88d).
Extended Data Fig. 6
Extended Data Fig. 6. Biological correlates of computational modeling data.
a, Average pre-existing myelin sheath length per mouse (nmice = 4). b, Mean node length of lengthening nodes per mouse, measured across time in pre-learning, learning, and post-learning stages (nmice = 3). c, Distribution of the number of consecutive lengthening nodes along a single axon. d-g, Modeled change in conduction speed as a function of the number of consecutively remodeled nodes at different sodium:leak conductance ratios (node with = 35 microns, pink exponential fit; Chi-square goodness of fit test). Conductance ratios were generated using either 0.4 S/cm2 (low) or 3.4 S/cm2 (high) for sodium conductance (gNa) and either 0.01 S/cm2 (low) or 0.08S/cm2 (high) for leak conductance (gL). h, Length of unmyelinated axon in that received new sheath addition following learning, measured from one sheath to the nearest neighboring sheath (nmice = 6; box plot, white bar = median, bounds of rectangle are Q1 and Q3; whiskers reflect minima and maxima). i, Proportion of gaps filled by 1, 2, or 3 sheaths following learning. j, Modeled change in conduction speed as a function of proportion of a 184 micron gap filled in by new myelin (pink exponential fit; Chi-square goodness of fit test). Conductance ratios as in d-g. Bars and error bars represent mean ± s.e.m unless otherwise noted.
Extended Data Fig. 7
Extended Data Fig. 7. Characterizing AAv8-cFos-ER virus.
a, Maximum projection of motor cortex injected with the cFos virus 1 day before learning (−1d), one week after learning (7d), three weeks after learning (22d), and 5 weeks after learning (42d). Tamoxifen was injected 3 hours after the final day of learning (7d). b, Fit curve for modeled change in cFos+ cell gain (calculated as % of tamoxifen-independent cFos+ neurons, i.e. the number of neurons labeled at 60 days) in mice that learn and receive tamoxifen (pink), untrained mice (grey), and mice that learn and are injected with sunflower oil (red). Each dot represents proportion of labeled neurons per mouse at a given timepoint. Shading represents 95% confidence interval. c, Rate of cFos+ cell gain in mice that learn and receive tamoxifen (pink), untrained mice (grey), and mice that learn and are injected with sunflower oil (red). Mice that learn have a heightened rate of cFos+ cell gain in the two weeks following learning, while the percentage of cFos+ neurons did not change in untrained mice and oil injected mice over the course of the experiment. Lines and shading represent mean ± s.e.m. d, Significantly more cFos+ cells appear in L2/3 relative to L1 following learning and injection of tamoxifen. e, The majority of learning-activated cells are putative L2/3 pyramidal neurons (determined by morphology). f, Distribution of traced axon lengths for cFos+ neurons (nmice = 5; box plot, white bar = median, bounds of rectangle are Q1 and Q3; whiskers reflect minima and maxima). Bars and error bars represent mean ± s.e.m unless otherwise noted. For detailed statistics, see Supplementary Table 3.
Figure 1 |
Figure 1 |. Learning-induced sheath retraction modifies node length.
a, All types of sheaths exhibit three behaviors: stable (top), growing (middle), and retracting (bottom). Note newly-generated oligodendrocyte (middle-right pink asterisk). b, Retraction of sheaths with two neighbors is specifically modulated by learning (F3,48 = 9.04, p < 0.0001) and continues for two weeks after learning (0d vs. 21d, p < 0.0001; 9d vs. 21d, p = 0.021; Tukey’s HSD). c, Retraction of sheaths with one neighbor is specifically modulated by learning (F3,40 = 9.93, p < 0.0001) and finishes within the learning period (p = 0.029; Tukey’s HSD). d, Dynamics of sheaths with 0 neighbors are not specifically modulated by learning. e, A longitudinally-imaged, lengthened node before (0d) and after learning (18d), and that same node in fixed and sectioned tissue stained for GFP (green) and neurofilament (NFH, red). The bottom panel shows a reconstruction of the lengthened node and axon. Images manually resliced for clarity. Sheaths of interest pseudo-colored. f, Learning modulates proportion of nodes lengthening (F1,9 = 38.85, p = 0.0002). More nodes lengthen in the two weeks following learning than before learning in mice that receive motor learning (p < 0.0001; Tukey’s HSD). In the two weeks following learning, more nodes lengthen in learning mice than in untrained mice (p < 0.0001; Tukey’s HSD). g, Learning modulates node length (F3,83.38=4.88, p=0.0036). Nodes increase significantly in length during learning (p = 0.039; Tukey’s HSD) and continue to lengthen following learning (p = 0.03; Tukey’s HSD). h, Changes in node length are significantly correlated to changes in sheath length of associated myelin sheaths (Standard least squares regression, p < 0.0001; shading represents 95% confidence interval). *p < 0.05, **p < 0.01, ***p < 0.0001, NS, not significant; bars and error bars represent mean ± s.e.m. For detailed statistics, see Supplementary Table 3, Figure 1.
Figure 2 |
Figure 2 |. Myelin and nodal plasticity occur in distinct phases in response to learning.
a, An entire arbor of an individual oligodendrocyte in vivo (left) and in reconstruction (right), in untrained (top) and learning (bottom) MOBP-EGFP mice. For clarity, only sheaths belonging to the relevant oligodendrocyte are shown in these images. b, No difference in sheath number per new oligodendrocyte in untrained and learning mice. c, No difference in sheath length on new oligodendrocytes in untrained and learning mice (box plot, white bar = median, bounds of rectangle are Q1 and Q3; whiskers reflect I.Q.R.). d, Learning modulates the types of sheaths produced by new oligodendrocytes (F5,15=15.03, p<0.0001). Continuous sheath production is heightened after learning (p=0.048; Tukey’s HSD), while new oligodendrocytes produce fewer isolated sheaths (p=0.044; Tukey’s HSD). e, A new sheath (blue, 37d) fills in a gap in pre-existing myelination (red, 0d) after learning (37d). f, After learning, more sheaths per new oligodendrocyte fill in pre-existing gaps in myelination (Student’s t-test, t(6.85)=−2.78, p=0.028). g, Following learning, the majority of newly filled gaps in myelination are pre-existing gaps in myelination rather than nodes that lengthened (Paired Student’s t-test, t(5)=6.99, p=0.0009). h, The majority of pre-existing gaps in myelination are filled in by one sheath, but up to three new sheaths can be responsible for filling in a gap in pre-existing myelination. i, A larger proportion of axons possess myelin dynamics in learning mice (Student’s t-test, t(6.58)=−3.73, p=0.0082). j, Cumulative change in myelin coverage on individual axons from baseline to two weeks post-learning separated by myelin and node dynamics. k, Learning-induced sheath retraction precedes the post-learning burst in oligodendrogenesis and new sheath addition (Student’s t-test, t(13.43)=−2.46, p=0.028; lines and shading represent mean±s.e.m.). *p<0.05, **p<0.01, ***p<0.0001, NS, not significant; bars and error bars represent mean±s.e.m. Sheaths of interest pseudo-colored. For detailed statistics, see Supplementary Table 3, Figure 2.
Figure 3 |
Figure 3 |. Learning-induced changes to myelin occur at specific locations along axons.
a, Multiple nodes along the same axonal segment lengthen in response to learning (in vivo data shown on top, reconstructions shown on bottom). b, Individual oligodendrocytes in learning and untrained mice exhibit multiple sheath behaviors. c, Learning modulates the proportion of axonal segments with majority dynamic nodes (F3,12 = 14.17, p = 0.0003). In learning mice, significantly more axonal segments with dynamic myelin possess majority dynamic nodes than in untrained mice (p=0.01; Tukey’s HSD). d, Axons can be myelinated by multiple new sheaths. Sheaths of interest are pseudo-colored. e, Learning modulates the proportion of axonal segments receiving multiple new myelin sheaths (F1,9.46 = 14.17, p = 0.0027). In the two weeks following learning, axons are more often myelinated by multiple new sheaths than before learning (p = 0.0055; Tukey’s HSD). The percent of axonal segments receiving increased myelin sheaths is heightened in learning compared to untrained mice (p = 0.006; Tukey’s HSD). f, Proportion of axonal segments receiving new myelin from multiple oligodendrocytes, multiple sheaths from a single oligodendrocyte, or only one new sheath in untrained mice (gray) and before and after learning (blue). *p<0.05, **p < 0.01, ***p < 0.0001, NS, not significant; bars and error bars represent mean ± s.e.m. For detailed statistics, see Supplementary Table 3, Figure 3.
Figure 4 |
Figure 4 |. Learning induces myelin and nodal dynamics broadly across cortical layers.
a, Max projections (top) and reconstructions (bottom) of axons with multiple nodes lengthening (left) and heightened sheath addition (right) following learning in L1. b, Learning modulates node lengthening in L1 (F1,1 4=49.33, p<0.0001). Two weeks after learning, more nodes lengthen than before learning in L1 (p<0.0001; Tukey’s HSD) and compared to untrained mice in L1 (p<0.0001; Tukey’s HSD. and L2/3 (p = 0.0014; Tukey’s HSD). c, Learning modulates the proportion of axons with multiple dynamic nodes in L1 (F3,24=25.12, p<0.0001). In learning mice, significantly more axons with dynamic myelin possess multiple dynamic nodes than in untrained mice in L1 (p<0.0001; Tukey’s HSD). In L1, learning mice have significantly more axons with multiple dynamic nodes than a single dynamic node (p=0.0052; Tukey’s HSD). d, Learning modulates the proportion of axons receiving multiple new myelin sheaths in L1 (F1,11.95=12.65, p=0.0040). Two weeks after learning, more axons receive multiple new myelin sheaths relative to before learning (p=0.0065; Tukey’s HSD) (d) and relative to axons in untrained mice (p=0.0087; Tukey’s HSD). e, Learning-induced sheath retraction precedes the post-learning sheath addition in L1 (Student’s t-test, t(4.00) = −4.51, p = 0.011; lines and shading represent mean ± s.e.m.). f, Raw data in XY plane corresponding to the same axon in 3D projections (left) and reconstructions (right) with multiple nodes lengthening. g, 3D projection of an axon with sheath addition following learning in L2/3. h, Learning modulates node lengthening in L2/3 (F1,9=15.85, p=0.0038). More nodes lengthen after than before learning in L2/3 (p=0.0014; Tukey’s HSD) and compared to untrained mice in L2/3 (p=0.0009; Tukey’s HSD). i, Learning modulates the proportion of axons with multiple dynamic nodes in L2/3 (F3,12=9.87, p=0.0015). In learning mice, significantly more axons with dynamic myelin possess multiple dynamic nodes than in untrained mice in L2/3 (p=0.015; Tukey’s HSD) in the two weeks after learning. j, Learning modulates the proportion of axons receiving heightened myelination in L2/3 (F1,8.92=7.87, p=0.021). Two weeks after learning, more axons receive new myelin sheaths relative to before learning (p=0.0063) and relative to axons in untrained mice (p=0.048; Tukey’s HSD). k Learning-induced sheath retraction precedes the post-learning sheath addition in L2/3 (Student’s t-test, t(4.8)=−2.91, p=0.035; lines and shading represent mean±s.e.m.). *p<0.05, **p<0.01, ***p<0.0001, NS, not significant; bars and error bars represent mean±s.e.m. For detailed statistics, see Supplementary Table 3, Figure 4.
Figure 5 |
Figure 5 |. Modeled effect of learning-induced sheath and nodal dynamics on conduction.
a, Schematic of myelinated axon model. Control trace and example traces of three categories of propagation along the modified region of axon: no change in conduction speed (“Success”), slowed conduction speed (“Delayed”), and failure to propagate (“Failure”). Note the arrival time of the control spike (dark pink) vs. the arrival time in each condition indicated (light pink). b, Matrix of conduction speeds across node relative to the ratio of sodium:leak conductance. Node lengths are characteristic of lengths before learning (“Pre-learn”, 1 micron), directly after learning (“Learn”, 20 microns), two weeks following learning (“Post-learn”, 35 microns), and of a large unmyelinated gap filled in by sheath addition (“Maximum observed”, 184 microns). c, Proportion of events that result in successful (lime), delayed (teal), or failures (navy) in propagation at 1, 20, 35, and 184 micron lengths. d, Modeled conduction speed as a function of sodium:leak conductance ratio at 1, 20, 35, and 184 micron lengths. e, Schematization of model used to predict the effect of coupled node lengthening and sheath retraction on conduction speed. Nodes were lengthened incrementally from 1 to 40 microns and associated sheaths were retracted a reciprocal amount. Asterisks indicate node retraction, while blue sheaths indicate sheath retraction. Nodes and sheaths were modified sequentially until 7 adjacent nodes were being lengthened at once. Four conductance ratios were modeled separately, using either 0.4 S/cm2 or 3.4 S/cm2 for sodium conductance (gNa) and either 0.01 S/cm2 or 0.08S/cm2 for leak conductance (gL). f, Modeled change in conduction speed as a function of node length at each of the four gNa:gL ratios with 1 (left) or 7 (right) nodes modified. g, Modeled conduction speed at a node length of 35 microns (dotted line in (f)) as a function of number of nodes modified. Learning modulates modeled conduction speed (F7,21=17.33, p<0.0001). Modification of 1 node significantly reduces conduction speed (0 vs. 1 nodes modified, p=0.0008; Tukey’s HSD), as does modification of 6 nodes (1 vs. 6 nodes modified, p=0.042; Tukey’s HSD). h, Schematization of model used to predict the effect of adjacent sheath addition on conduction velocity. Sheaths of 60 microns in length were added sequentially to a 184 micron unmyelinated gap. i, Modeled change in conduction speed as a function of sheaths added. j, Modeled conduction speed as a function of coverage of the unmyelinated gap, averaged across the four sodium:leak conditions. Sheath addition modulates conduction speed (F3,9=14.39, p=0.0009). The addition of a sheath that completely closes the unmyelinated gap results in a significant increase in conduction speed (p=0.01; Tukey’s HSD; shading represents 95% confidence interval). *p<0.05, **p<0.01, ***p<0.0001, NS, not significant; bars and error bars represent mean±s.e.m. For detailed statistics, see Supplementary Table 3, Figure 5.
Figure 6 |
Figure 6 |. Learning-activated axons exhibit heightened myelin dynamics.
a, Time course of experiment and appearance of tdTom+ labeling in experimental groups. b, Representative images of baseline-labeled and learning-activated neurons in untrained mice (top), learning mice (middle), and neurons labeled following learning and administration of 4-OHT in learning mice (bottom). c, Reconstructions of Layer 2/3 axons exhibiting coupled node lengthening and sheath retraction (top) and sheath addition (bottom) following learning. See Supp. Fig. 8 for raw data associated with reconstructions. d, Learning modulates number of retracting sheaths per axon (F2,26.19=6.65, p=0.0007). One month after learning, learning-activated axons have significantly more sheath retraction relative to axons labeled at baseline in learning mice (p=0.0007; Tukey’s HSD) and in untrained mice (p=0.0062; Tukey’s HSD). Dot and error bars represent median and I.Q.R. e, Learning modulates dynamics of sheath retraction (F6,7.14=28.07, p<0.0001). During learning, sheaths on behaviorally-activated axons retract significantly (p=0.0065; Tukey’s HSD), and one month after learning, sheaths are significantly shorter than sheaths on baseline-labeled axons in learning mice (p<0.0026; Tukey’s HSD) and in untrained mice (p<0.0001; Tukey’s HSD). f, Learning modulates the proportion axons with lengthening nodes (F2,14.68=2.96, p=0.029). One month after learning, a significantly higher proportion of learning-activated axons have node lengthening relative to axons labeled at baseline in learning mice (p=0.030; Tukey’s HSD) and in untrained mice (p=0.044; Tukey’s HSD). g, Learning modulates dynamics of node lengthening (F6,4.76=24.32, p=0.0025). During learning, nodes on learning-activated axons retract significantly (p=0.022; Tukey’s HSD), and one month after learning, nodes are significantly longer than nodes on baseline-labeled axons in learning mice (p=0.047; Tukey’s HSD) and in untrained mice (p=0.020; Tukey’s HSD). h, Learning modulates the proportion axons with new sheath addition (F2,25.02=8.70, p=0.0002). One month after learning, a significantly higher proportion of learning-activated axons have new sheath addition relative to axons labeled at baseline in learning mice (p=0.0006; Tukey’s HSD) and in untrained mice (p=0.0004; Tukey’s HSD). i, The majority of learning-activated axons receive 1 new sheath following training. j, Proportion of axons showing no dynamics (grey), only sheath addition (pink), only node lengthening (red), or both sheath addition and node lengthening (dark red). k, Cumulative change in myelin coverage on individual axons from baseline to four weeks post-learning. *p<0.05, **p<0.01, ***p<0.0001, NS, not significant; bars and error bars represent mean±s.e.m unless otherwise noted. For detailed statistics, see Supplementary Table 3, Figure 6.
Figure 7 |
Figure 7 |. The extent and timing of myelin dynamics is correlated to behavioral performance.
a, The proportion of nodes lengthening per mouse is correlated to degree of behavioral success only during learning, but not before or in the two weeks after learning (Standard least squares regression, shading represents 95% confidence interval). b, The proportion of axons receiving new myelin sheaths per mouse is correlated to degree of behavioral success only in the two weeks following learning, but not before or during learning (Standard least squares regression, shading represents 95% confidence interval). c, In high performing animals (defined as animals which attain 3 or more days above 20% successful reaches), node lengthening occurs significantly earlier than sheath addition (Paired Student’s t-test; t(5)=3.99, p=0.11). d, In low performing animals (defined as animals which attain less than 3 days at above 20% successful reaches), node lengthening and sheath addition occur concurrently. *p < 0.05, **p < 0.01, ***p < 0.0001, NS, not significant; lines and shaded areas represent mean±s.e.m; dots represent half maximum. For detailed statistics, see Supplementary Table 3, Figure 7.
See this image and copyright information in PMC

Comment in

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Lillie RS FACTORS AFFECTING TRANSMISSION AND RECOVERY IN THE PASSIVE IRON NERVE MODEL. Journal of General Physiology 7, 473–507 (1925). - PMC - PubMed
    1. Tomassy GS et al. Distinct Profiles of Myelin Distribution Along Single Axons of Pyramidal Neurons in the Neocortex. Science 344, 319–324 (2014). - PMC - PubMed
    1. Yakovlev P & Lecours A The myelogenetic cycles of regional maturation of the brain. Regional development of the brain in early life. Edited by: Minkowski A 1967. (Oxford: Blackwell Scientific Publications, 1967).
    1. Gibson EM et al. Neuronal Activity Promotes Oligodendrogenesis and Adaptive Myelination in the Mammalian Brain. Science 344, 1252304 (2014). - PMC - PubMed
    1. Hughes EG, Orthmann-Murphy JL, Langseth AJ & Bergles DE Myelin remodeling through experience-dependent oligodendrogenesis in the adult somatosensory cortex. Nature Neuroscience 21, 696–706 (2018). - PMC - PubMed

Methods-only references

    1. Hughes EG, Kang SH, Fukaya M & Bergles DE Oligodendrocyte progenitors balance growth with self-repulsion to achieve homeostasis in the adult brain. Nat Neurosci 16, 668–676 (2013). - PMC - PubMed
    1. Xu T et al. Rapid formation and selective stabilization of synapses for enduring motor memories. Nature 462, 915–919 (2009). - PMC - PubMed
    1. Cohen CCH et al. Saltatory Conduction along Myelinated Axons Involves a Periaxonal Nanocircuit. Cell 180, 311–322.e15 (2020). - PMC - PubMed
    1. Hu W et al. Distinct contributions of Nav1.6 and Nav1.2 in action potential initiation and backpropagation. Nat Neurosci 12, 996–1002 (2009). - PubMed
    1. Mainen ZF, Joerges J, Huguenard JR & Sejnowski TJ A model of spike initiation in neocortical pyramidal neurons. Neuron 15, 1427–1439 (1995). - PubMed

Publication types

LinkOut - more resources

-