Introduction

Even before sensory inputs are fully established, immature animals can respond to environmental cues and generate diverse outputs, implying that complex neural networks are pre-organized. This organization relies on robust developmental processes, among which spontaneous activity plays a central role in shaping neural circuits. During early postnatal development, spontaneous activity is highly correlated across neural populations in various species and systems (Kirkby et al., 2013; Ackman & Crair, 2014). The widespread occurrence of correlated spontaneous activity suggests a universal principle of brain maturation, facilitating the formation of precise sensory maps and large-scale networks required for sensory processing and behavioral output.

In the developing central nervous system, correlated spontaneous activity initially manifests as synchronized patterns in localized or extensive regions (Adelsberger et al., 2005; Good et al., 2017; Babola et al., 2018; Mizuno et al., 2018; Murakami et al., 2022; Fujimoto et al., 2023). These patterns gradually transition to desynchronized states, as observed in the cortex, cerebellum, and olfactory bulb (Good et al., 2017; Mizuno et al., 2018; Fujimoto et al., 2023). Despite extensive documentation of these patterns, the mechanisms that generate synchronized activity and their functional significance remain poorly understood. Oligodendrocytes are strong candidates for regulating synchronized spontaneous activity because they control myelination and conduction velocity (Pajevic et al., 2014; Pajevic et al., 2023). Myelin formation continues throughout life, but the majority occurs during postnatal development in many brain regions (de Faria et al., 2021). However, how developmental oligodendrocytes contribute to brain maturation remains largely unknown.

Here we tried to determine the causal relationships among developmental oligodendrocytes, synchronized spontaneous activity, and brain function in the mouse cerebellum. The developing cerebellum is a suitable model, as the developmental trajectory of synchronized activity is well characterized (Good et al., 2017). To manipulate oligodendrocytes during development, we used an adeno-associated virus (AAV) vector with high oligodendrocyte specificity, enabling precise spatiotemporal reduction of these cells in the developing cerebellum. We found that reducing oligodendrocytes during the second to third postnatal week disrupted Purkinje cell (PC) synchrony. Behavioral analyses further revealed that perturbing developmental oligodendrocytes and synchronized activity led to lasting impairments in adult cerebellar function, including anxiety-like behaviors, reduced social interactions, and impaired motor coordination. Optogenetic re-synchronization in adulthood restored motor and social functions but not anxiety-like behavior. These results suggest that oligodendrocytes act as key regulators of synchronized spontaneous activity and are essential for the proper development of brain function.

Materials and methods

Experimental animals

This study was conducted under the recommendations of the Regulations for Animal Experiments and Related Activities at Tokyo Medical and Dental University. All animal experiment were approved by the Committees on Gene Recombination Experiments and Animal Experiments of Tokyo Medical and Dental University. All animals were group housed and maintained on a 12hr:12hr light dark cycle. All efforts were made to reduce the number of animals used and to minimize the suffering and pain of animals. WT mice (C57BL6N, 6 days −16 weeks, Sankyo labo service) were used in this study.

Plasmids

The human MAG gene promoter sequences were amplified by PCR from human genomic DNA using primer pairs (Table. 1) and inserted into sites of hSyn promoter in AAV-U6-sgRNA-hSyn-mCherry (a gift from Alex Hewitt, Addgene plasmid # 87916) (AAV-hMAG-mcherry). hSyn promoter was removed by ApaI and BaMH1. The promoter was cloned using the Gibson Cloning Assembly Kit (New England BioLabs) following standard procedures. For AAV-hMAG-DTA, we amplified the DTA coding sequence from the pAAV-mCherry-flex-dtA (a gift from Naoshige Uchida, Addgene plasmid # 58536) using primer pairs (Table. 1) and cloned it into AAV-MAG-mcherry. The sequences of jGCaMP7f gene and jGCaMP7s were amplified by PCR from pGP-AAV-syn-jGCaMP7f-WPRE and pGP-AAV-syn-jGCaMP7s-WPRE (a gift from Douglas Kim & GENIE Project, Addgene plasmid #104488 and #104487) and inserted into sites of GFP in pAAV/L7-6-GFP-WPRE (a gift from Hirokazu Hirai, Addgene plasmid # 126462) (pAAV-L7-6-jGCaMP7f and pAAV-L7-6-jGCaMP7s) and into sites of GFP in pAAV-TRE-EGFP (a gift from Hyungbae Kwon, Addgene plasmid # 89875) (pAAV-TRE-jGCaMP7f and pAAV-TRe-jGCaMP7s). For AAV-TRE-Kir2.1-T2A-GFP, we amplified the sequences of Kir2.1-T2A-GFP from the pAAV hSyn FLEx-loxp Kir2.1-2A-GFP (a gift from Kevin Beier & Robert Malenka, Addgene plasmid # 161574), and cloned it into sites of GFP in pAAV-TRE-EGFP. For AAV-TRE-jGCaMP7f-T2A-Kir2.1, we amplified the sequences of jGCaMP7f and Kir2.1 from the pGP-AAV-syn-jGCaMP7f-WPRE and pAAV hSyn FLEx-loxp Kir2.1-2A-GFP, and cloned it into sites of GFP in pAAV-TRE-EGFP. For pAAV-TRE-rsChRmine-oScarlet, we amplified the sequences of rsChRmine-oScarlet from the pAAV -CaMKIIa-rsChRmine-oScarlet (a gift from Karl Deisseroth, Addgene plasmid # 183522) (Kishi et al., 2022), and cloned it into sites of GFP in pAAV-TRE-EGFP. For pAAV-L7-6-tTA, tTA was inserted into sites of GFP in pAAV/L7-6-GFP-WPRE.

AAV packaging and injection

The AAVs were produced using standard production methods. HEK293 cells were transfected with Polyethylenimine. AAV9, PHPeB (Chan et al., 2017) and Olig001 (Powell et al., 2016) were used. Virus was collected after 120 h from both cell lysates and media. Purification Kit (Takara) was used for viral particle purification. All batches produced were in the range 10¹⁰ to 10¹¹ viral genomes per milliliter. The procedure for viral vector injection has been modified from the previous lentivirus protocol(Uesaka et al., 2014). 1-1.5 µL of viral solution were injected into the cerebellum of C57BL/6N mice at 100 nl / min.

Immunohistochemistry

Mice were perfused with 4% paraformaldehyde in 0.1 M phosphate buffer, and processed for parasagittal microslicer sections (100 μm in thickness). After permeabilization and blockade of nonspecific binding, the following antibodies were applied for 2 days at 4 °C: an antibody against Car8 (Car8-GP-Af500 or Car8-Go-Af780, diluted 1:300, Nittobo Medical) to visualize PCs; an antibody against S100b (S100b-GP-Af630, 1:300, Nittobo Medical); an antibody against Iba (Wako 019-19741, 1:1000), an antibody against RFP (390004, 1:1000, Synaptic Systems or PM005, 1:500, MBL), an antibody against MBP (NBP2-50035, 1:1000, Novus), an antibody against ASPA (ABN16986, 1:1000, Merck), an antibody against vGluT2 (MSFR106290, 1:300, Nittobo Medical), and/or an antibody against GFP (#06083-05, 1:1000, Nacalai Tesque). After incubation with secondary antibodies (an anti-rat Alexa Fluor 488, an anti-guinea pig Cy3, an anti-goat Alexa Fluor 488, an anti-goat Alexa Fluor 647, an anti-mouse Cy5, an anti-rabbit Cy3, an anti-rabbit Alexa Fluor 488), the immunolabeled sections were washed and then examined under a fluorescent microscope (BZ-X700 or BZ-X800, Keyence).

In vivo calcium imaging

Animals were put on a warm blanket and anesthetized with a mixture of midazolam (4 mg/kg body weight (BW)), butorphanol (5 mg/kg BW), and medetomidine (0.3 mg/kg BW). The depth of anesthesia was constantly monitored by observing pinch-induced forelimb withdrawal reflex. The skin and muscles over the skull were removed and a metal plate was fixed with dental acrylic cement over the lobule 5-7 of the cerebellar vermis. A craniotomy of 3 mm diameter was made and the dura mater was carefully removed. A sterile circular 3mm diameter glass coverslip (#CS01078 3mm diameter, Matsunami) was put directly on the dura mater and was secured in place with surgical adhesive (Alon-Alpha A, Sankyo). A custom-made metal headplate with a 5 mm circular imaging well was fixed to the skull over the cranial window with dental cement (Super-Bond, Sun-medical, Japan). The head of the mouse was immobilized by attaching the head plate to a custom-made stage. We then administered atipamezole (0.3mg/kg, anti-sedan) (100 μg/g of body weight, intraperitoneal injection) for anesthesia reversal. After a 10 - 15 minutes waiting period, in vivo calcium imaging was performed using an Olympus MVX10 fluorescence microscope or a two-photon microscope (Nikon, AX R MP) equipped with a ×16 objective lens (Nikon, CFI75 LWD 16X W) and an ultrafast laser (Axon 920-2 TPC). During the recordings, the animals were laid on a warm blanket to keep the body temperature at 37°C. Laser power for the two-photon microscope was maintained <20 mW at the sample. Images were obtained in a 2D plane (2048 x 2048 pixels) using Andor SOLIS software (Oxford Instruments, frame rate 5 Hz) or in a 2D plane (512 × 512 pixels) using Nikon NIS-Elements software (frame rate, 1 Hz) for the two photon microscope. Data were saved as TIFF files.

We resized raw calcium imaging videos (TIFF format) to a resolution of 512 x 512 pixels with an 8-bit depth using ImageJ (JAVA-based processing program, USA). We cropped out the unwanted areas. Regions of interest (ROIs) corresponding to the active PC dendrites were detected using the Suite2p software (Marius Pachitariu, GitHub). This software executed a two-dimensional phase-correlation-based non-rigid image registration in the XY plane, extracted the ROIs considered as PC dendrites, and calculated the fluorescence intensity changes in each ROI. ROIs were extracted using the parameters shown in Tanigawa et al., 2024 (Tanigawa et al., 2024). We exported these data as MAT files. The MAT files were imported into MATLAB R2023b (MathWorks, USA) for further analysis using custom-made MATLAB routines. The fluorescence intensity (F) was converted to ΔF/F0 wave, expressed as ΔF/F0 = (F - F0)/(F0), where F0 is the baseline fluorescence in the absence of calcium transients. F0 was defined as follows. First, we calculated a threshold as mean + 1 SD of the F wave of each ROI, and then the F wave below the threshold was averaged to obtain F0. Parameters such as amplitude, frequency, half-width, and area of each calcium transient were quantified by the custom-made MATLAB routines. The correlation coefficient among all ROI pairs was calculated as previously described (Tsutsumi et al., 2015). The resulting correlation matrix captured relationships among all ROI pairs in the dataset.

Behaviors

Male mice from 2 to 3 months of age 6-7 days after AAV injection were used for behavioral tests. Before the experiments, the mice were habituated in the testing areas for at least 30 min. Mice behavior was recorded through the video camera set in the experimental room. The open field test was performed in a 50 cm x 50 cm x 40 cm (W x D x H) open field box for 10 min. Total distance mice traveled and the time in the center and peripheral regions were automatically analyzed by the ImageJ software (MouBeAT) (Bello-Arroyo et al., 2018).

The sociality test was performed in the open field apparatus (50 x 50 x 40 cm, W x D x H). The test consisted of a 10-min habituation session and a 10-min test session. In the habituation session, mice were allowed to explore the open field freely. In the test session, two quadrangular cages with slits (8 cm x 8 cm, 18 cm high) were placed in two adjacent corners. One cage contained a novel mouse (8-week-old male C57BL6N) and the other was a mouse doll, and a test mouse was first placed in the outer regions of open field arena at the longest distance from these cages. The times spent around the two cages areas (12 cm in diameter) were analyzed automatically with the ImageJ software (MouBeAT). Social preference was assessed by the following equation: Sociality index = (T_mouse)/(T_doll) where T_mouse and T_doll are the staying time around the mouse cage and that around the doll cage, respectively.

To assess motor coordination and motor learning, mice were subjected to a rotarod test. Mice were placed on a resting state rotarod (model LE8205, Panlab) for three consecutive days with two trials per day spaced 10 min apart (30 min and 40 min after CNO injection). For each trial, the rotarod was accelerated linearly from 4 rpm to 40 rpm over 300 s. Latency to fall from the start of rotation was measured.

Optogenetic stimulation

To manipulate PC synchrony, we expressed a red-shifted channelrhodopsin variant selectively in PCs of control and DTA mice by stereotaxic injection of pAAV-TRE-rsChRmine-oScarlet and pAAV-L7-6-tTA into the cerebellar vermis 2 weeks before behavior tests. Mice were subjected to behavioral assays at P60-70. For light delivery, we used a wireless optogenetic system (Teleopto; Bio Research Center Co., Ltd., Aichi, Japan) equipped with a brain-surface LED probe (TeleLP-c, 590 nm). After making a small midline skin incision, the LED probe was placed directly above the cerebellar vermis on the intact skull and secured with dental cement (Super-Bond C&B). The cement was coated with black nail polish to minimize light leakage. This approach allowed stable illumination of the cerebellar surface without penetrating the tissue. Probe placement was verified after the completion of behavioral experiments. During behavioral testing, 590-nm light pulses (10 ms, 1 Hz) were delivered throughout the task sessions using a Teleopto Remote Controller via an infrared emitter. Stimulation was applied during open-field, three-chamber social interaction, and rotarod assays. Illumination parameters were selected to evoke population-wide synchronous PC spiking while minimizing nonspecific effects on locomotor activity.

Statistical analysis

All data are presented as mean ± SEM. Statistical significance was assessed by Mann-Whitney U test for the comparison of two independent samples. To compare the two different categorical independent samples on one dependent variable, Two-way ANOVA with post-hoc test (Bonferroni correction) was used as indicated in the text. For datasets in which multiple measurements were obtained from the same animal (e.g., multiple fields of view or recordings), statistical dependence within animals was accounted for by performing nested ANOVA or nested t-tests, treating animal identity as the nesting factor. This ensured that repeated measurements within the same animal were not treated as independent replicates. Because the nested models yielded comparable effect sizes to the Mann–Whitney tests, we have retained the mean ± SEM for ease of comparison with prior literature but now also report all values for each mouse in Table 1. Statistical analysis was conducted with GraphPad Prism program. Differences between groups were judged to be significant when p values were smaller than 0.05. *, **, ***, **** and ns represents p < 0.05, p < 0.01, p < 0.001, p < 0.0001 and not significant, respectively.

Results

Developmental oligodendrocytes regulate synchronized spontaneous activity

We tested the role of developmental oligodendrocytes in regulating synchronized spontaneous activity and brain function. We generated an oligodendrocyte-specific AAV (OL-AAV) containing a human myelin associated glycoprotein (hMAG) promoter and an oligodendrocyte-tropic capsid (Figure 1A). We focused on the developing mouse cerebellum in which synchronized spontaneous activity and its transition were observed during postnatal development (Good et al., 2017). Myelination in the cerebellum appeared by postnatal day 5 (P5) and P7 (Figure supplement 1A-B).

Figure 1

Injection of OL-AAVs expressing mCherry into the cerebellum of P7 mice resulted in robust and selective labeling of cerebellar oligodendrocytes by P21, with >98% of mCherry-positive cells co-expressing the mature oligodendrocyte marker ASPA (Figure 1B-D). In contrast, mCherry was not detected in Olig2-positive but ASPA-negative immature oligodendrocyte-lineage cells, confirming specific targeting of mature oligodendrocytes (Figure supplement 1C-D). mCherry expression was restricted to the cerebellum and was not detected in extracerebellar regions (Figure 1E-G), confirming the regional specificity of this approach.

To selectively ablate oligodendrocytes, we injected OL-AAVs carrying Diphtheria toxin A (DTA) into the cerebellum of P7 mice. This manipulation resulted in a significant reduction of oligodendrocyte density (ASPA⁺ cells) and myelin coverage (MBP staining) by P10 and P14 (Figure 1H-K and Figure supplement 1E). The density and coverage returned to control levels by P21 and remained normal in adulthood, indicating a transient but robust disruption of oligodendrocytes and myelination during a defined developmental window. This temporal profile provided the opportunity to test how developmental oligodendrocytes regulate PC activity synchrony and its maturation. Despite the transient loss of mature oligodendrocytes, gross cerebellar morphology remained indistinguishable from controls (Figure supplement 1F-G).

To determine whether other cell types were affected (Mathis et al., 2003), we examined microglia, astrocytes, PCs, and molecular layer interneurons in DTA-ablated mice. IBA1 staining showed no differences in microglial density and IBA1 intensity between DTA and control mice at both of P14 and P21 (Figure supplement Fig. 2A-D). S100β staining showed no differences in Bergmann glia density and S100β intensity (Figure supplement Fig. 2E-H). Parvalbumin (PV) immunostaining demonstrated that PC density and dendritic arborization were preserved in DTA mice (Figure supplement Fig. 2I-K). PV-positive interneurons in the molecular layer were present at normal density in DTA mice (Figure supplement Fig. 2L). The same results were obtained at P21 (Figure supplement Fig. 2M-P). Together, these results indicate that transient developmental ablation of oligodendrocytes does not induce compensatory gliosis or neuronal loss.

We next tested whether oligodendrocyte deficit affected synchronized spontaneous activity in the cerebellar PC population (Figure 2A-B). Here, spontaneous activity refers to neural events occurring in the absence of external sensory stimulation. Developmental PCs display highly synchronized CF–driven activity in newborn mice and undergo progressive desynchronization as development advances (Good et al., 2017). In our imaging paradigm, regions of interest were restricted to the dendritic arbor of PCs where calcium signals are dominated by climbing-fiber (CF) synaptic inputs (Ramirez & Stell, 2016; Good et al., 2017). In vivo calcium imaging of PCs revealed a marked reduction in synchrony of spontaneous activity among PCs at P13-15 following oligodendrocyte reduction in DTA mice (Figure 2C-F). This effect was observed across both proximal and distal PC populations (Figure 2E-F), indicating a widespread disruption of neural activity synchronization. Notably, other parameters such as frequency, amplitude, and area of calcium transients remained unchanged (Figure 2G-I), suggesting a specific effect on PC activity synchrony rather than general neural activity.

Figure 2

Having found that developmental oligodendrocyte loss decreases PC synchrony, we next examined whether this effect originated from altered CF activity. We imaged CF population activity by expressing jGCaMP7f in inferior olive neurons (Figure 3A). Correlation analysis of CFs revealed that CF–CF synchrony, but not amplitude, frequency, area, width, was significantly reduced in DTA mice at P13-15 (Figure 3B–D), indicating that impaired CF coordination contributes to the reduction of PC synchrony. Electrophysiological analyses showed that parallel-fiber EPSCs and inhibitory synaptic inputs onto PCs were not significantly altered (Figure 3E–H), making it unlikely that reduced synchrony results from altered excitatory or inhibitory drive other than CF inputs. Together, these findings demonstrate that developmental oligodendrocytes are required to maintain CF-driven PC synchronization during a critical window of cerebellar maturation.

Figure 3

Developmental oligodendrocyte affects brain function

We next examined whether developmental oligodendrocytes and synchronized spontaneous activity affect adult brain function. We first investigated the long-term effects of oligodendrocyte reduction during cerebellar development on the synchrony of PC activity in the adult cerebellum. We selectively ablated oligodendrocytes in early postnatal mice and monitored PC activity as the animals matured. Developmental reduction of oligodendrocytes caused persistent hyposynchrony of PC activity in adulthood, whereas other parameters of PC activity remain unaffected (Figure 4). These findings suggest that disturbances in oligodendrocyte function during development have lasting consequences for brain function, underscoring the role of developmental oligodendrocytes in sustaining neuronal synchrony across the lifespan.

Figure 4

Because developmental oligodendrocyte loss disrupted PC synchrony in adulthood, we next asked how this perturbation impacts brain function at the behavioral level. The cerebellum contributes not only to sensory-motor functions, including chewing, postural control, tone, movement, the coordination of balance, eye movements, and reflexes (Stoodley & Schmahmann, 2018; Algin et al., 2023), but also to cognitive, social, and emotional processes (Depping et al., 2018; Schmahmann, 2019; Van Overwalle et al., 2020). To examine roles of developmental oligodendrocytes in these brain function, we injected OL-AAVs with hMAG-mCherry or hMAG-DTA at P6-7 into distinct cerebellar compartments (left, center, right) (Figure supplement Fig. 3) corresponding to functional domains (Manni & Petrosini, 2004; Guell et al., 2018). Behavioral assessments in adulthood unveiled that oligodendrocyte reduction during development induced anxiety-like behavior specifically when targeted to the central cerebellar region, without altering distance travelled, underscoring the compartment-specific contribution to anxiety regulation (Figure 5A-C). Across all regions, oligodendrocyte reduction led to reduced social interaction and impaired motor coordination, as shown by reduced social approach to novel mouse and shorter latency to fall in the Rotarod test, respectively (Figure 5D-G). These data indicate that developmental oligodendrocytes are essential for mental states, sociality, and motor function in adult stages.

Figure 5

Developmental oligodendrocytes during specific time windows affect brain functions

We next examined whether oligodendrocytes influence cerebellar function within specific developmental time windows. To address this, we administrated hMAG-DTA to the central cerebellum at postnatal week 3 (3W_DTA mice), shortly after weaning. In these mice, oligodendrocytes in the cerebellum were significantly reduced at 4 weeks (1 week after AAV injection) compared with control mice with only hMAG-mCherry, but returned to near-control levels by 2 months old (Figure 6A-C), indicating that the ablation was temporary. Calcium imaging of 3W_DTA mice at 4weeks old revealed no significant alteration for correlation coefficient and activity (Figure 6D-J). These results suggest that oligodendrocytes before and after the weaning exert distinct influences on PC activity. Behavioral assessments in adulthood of these mice uncovered no significant changes in anxiety or sociality (Figure 6K-O). However, 3W_DTA mice showed a subtle motor impairment as evidenced by shorter latency to fall at day3 in the Rotarod test (Figure 6P). These data suggest that oligodendrocytes before and after weaning make temporally distinct contributions to neural activity and brain function.

Figure 6

Synchronized spontaneous activity during development regulates brain functions

To test the causal link between synchronized PC activity during developmental stage and cerebellar function, we used an experimental strategy that selectively manipulated CF activity, which drives PC calcium transients. We reduced synchronized spontaneous activity of PC population by expressing Kir2.1, a potassium channel that reduces neuronal excitability, in a subset of CFs projecting to PCs. Co-injection of AAVs with EGFP or Kir2.1-T2A-GFP under the control of TRE promoter and AAVs with tetracycline transactivator (tTA) under the control of Htr5b promoter (Dorgans et al., 2022) into the inferior olive at P7 labeled a subset of CFs in the cerebellum at P22 (Figure 7A-C). We quantified the proportion of PCs that received labeled CFs and found that PCs with GFP-labeled CFs were detected in 19.3 ± 11.9 % (mean ± S.D.) for control and 26.2 ± 11.8 % (mean ± S.D.) for Kir2.1 groups (Figure 7D). In vivo calcium imaging of CFs showed that activity of CFs with Kir2.1-T2A-jGCaMP7 was lower than that for CFs with only jGCaMP7 at P13-15 (Figure supplement 4A-D). In vivo calcium imaging of PC populations revealed that synchrony among distal PCs was unchanged, whereas synchrony among proximal PCs receiving Kir2.1-expressing CF inputs was significantly reduced at P13–15 compared with controls (Figure 7E-H). The frequency, amplitude and area of calcium transients were not significantly altered by Kir2.1 expression in a subset of CFs (Figure 7I-K). These results indicate that kir2.1 expression in a subset of CFs leads to reduced synchrony of proximal PC population activity.

Figure 7.

Because Kir2.1 expression reduced PC synchrony during development, we next asked whether this manipulation also affected adult behavior. Behavioral testing at 2 months revealed that Kir2.1-expressing mice (injection of AAVs with Kir2.1 at P7) showed a trend toward decreased distance travelled and percentage time spent in center (Figure 7L-N). Kir2.1-expressing mice also exhibited reduced social interaction and impaired motor coordination, as shown by reduced social approach to a novel mouse and shorter latency to fall in the Rotarod test, respectively (Figure 7O-Q). These results indicate that the behavioral consequences of developmental oligodendrocyte ablation are phenocopied by experimentally reducing synchrony during development.

To directly test whether restoring PC synchrony in adulthood can rescue behavior, we expressed a red-shifted opsin (AAV-L7-rsChRmine) selectively in PCs of control and DTA mice and applied 590-nm light pulses (10 ms, 1 Hz) to the cerebellar vermis during behavioral testing (Figure 8A–B). Periodic optogenetic re-synchronization did not alter locomotor activity and anxiety-like behavior, as distance traveled and open-field center time remained low in DTA mice (Figure 8C-D). Unexpectedly, in control mice the same stimulation reduced center time (Figure 8E), indicating that artificially elevating activity synchrony and firing rate of PCs can enhance anxiety-like responses. In contrast, motor and social deficits in DTA mice were significantly improved. Optogenetic stimulation rescued sociability, as measured by the time spent interacting with a novel mouse (Figure 8F–G). The same stimulation restored rotarod performance to control levels (Fig. 8H). On day 1 without light stimulation, DTA mice with rsChrrmine expression performed similarly to DTA mice without rsChrmine and showed reduced latency to fall compared with control mice with rsChrmine. On day 2 and day 3 with light stimulation, DTA mice with rsChrmine showed longer latency to fall than DTA mice without rsChrmine, reaching values comparable to controls. These results demonstrate that reduced PC synchrony is the primary driver of the observed motor and social impairments, whereas anxiety-like behavior likely involves additional mechanisms. Together, these findings establish a causal link between PC synchrony and specific behavioral domains and show that periodic re-synchronization can normalize motor and social functions in adult mice that experienced developmental oligodendrocyte loss.

Figure 8

Discussion

This study provides compelling evidence that oligodendrocytes play a crucial role beyond their traditional functions in myelination, extending to the regulation of synchronized spontaneous activity, and emphasizing the complexity of glial-neuronal interactions in the developing brain. Moreover, our research illuminates the significance of brain development following a proper course in brain function integrity The interpretation of our synchrony measurements depends critically on the cellular sources of PC calcium signals. Our calcium imaging approach was designed to capture dendritic transients in PCs, which are overwhelmingly driven by CF complex spikes (Good et al., 2017). Previous work has shown that simple-spike (SS) bursts can evoke modest Ca²⁺ rises, but these are restricted to the soma and proximal dendrites of adult PCs and do not propagate into the distal dendritic arbor where our regions of interest were defined (Ramirez & Stell, 2016). Consistent with these, in vivo imaging showed that oligodendrocyte ablation reduced CF–CF synchrony at P14, directly linking oligodendrocytes to impaired CF coordination and PC hyposynchrony. While SS-dependent Ca²⁺ signals may still influence long-term circuit function, they likely do not contribute to the synchrony observed here. Future somatic imaging and electrophysiology will clarify whether oligodendrocytes also modulate SS activity.

In adult brain, adaptive myelination can induce and maintains synchrony of spontaneous activity among neural populations (Kato et al., 2020; Bonetto et al., 2021; Pajevic et al., 2023). Developmental oligodendrocytes have been also speculated to regulate the synchrony of neuronal activities (Jang et al., 2019), and our study now provides direct evidence for this role. Oligodendrocytes regulate synchrony in part through myelin formation and plasticity, which modulate axonal conduction velocity, while reciprocal neuronal signals drive myelination (Fields, 2015). Thus, bidirectional oligodendrocyte–neuron interactions can shape the timing of circuit activity, and alterations in these interactions may lead to desynchronization. Mechanistically, oligodendrocytes can regulate synchrony through at least two complementary routes. First, multilamellar myelin shortens membrane time constants and accelerates conduction velocity, thereby reducing temporal dispersion of CF inputs and sharpening coincidence detection in PCs. Second, mature oligodendrocytes provide metabolic support—including lactate, cholesterol, and mitochondrial intermediates—that sustains high-frequency axonal firing. Loss of this trophic support can prolong refractory periods and promote excessive desynchronization, even in the absence of overt demyelination.

Our findings reveal that the timing of PC desynchronization coincides with the initial appearance of oligodendrocytes and myelin in the cerebellum, with synchrony declining from P3–P5 (Good et al., 2017), just as myelin sheaths begin to emerge (Figure supplement Fig. 1A-B). However, we found that genetic ablation of oligodendrocytes and myelin during P7-P21 declined PC synchrony, indicating that myelination could be not the trigger for desynchronization. Instead, these results suggest that the loss of synchrony from P3-P5 is mediated by other mechanisms, such as transient gap junction coupling or shared afferent drive, and that the onset of myelination later serves to counterbalance these processes. In this view, oligodendrocytes are not required to initiate the desynchronization of PC activity but rather become essential afterward, once axons are ensheathed, to stabilize and maintain the mature low-synchrony state. Based on these observations, we propose a two-phase model. Phase I (P3–P7): High early synchrony is generated by non-myelin mechanisms, including transient gap junctions and shared climbing-fiber input. Phase II (P7 onward): As oligodendrocytes proliferate and ensheath axons, they fine-tune conduction velocity and provide metabolic support, thereby stabilizing the network and maintaining the mature low-synchrony state.

Correlated spontaneous activity has been shown to play a pivotal role in neural development and the establishment of functional neural circuits especially in the developing visual system (Kirkby et al., 2013; Ackman & Crair, 2014). Retinal waves observed in various species, play a fundamental role in developing interconnections within the visual system including topographic maps. Recent study demonstrates that retinal wave is not merely random but possesses an intrinsic directionality that prefig the optic flow patterns encountered after eye opening. However, the effects of correlated spontaneous activity during development on adult brain function remain largely unknown. Our study suggests that synchronized spontaneous activity during development play a crucial role in the subsequent brain function. These studies provide new insights into specific information content carried by correlated spontaneous activity and their influence on the development of brain functions including higher-order function, extending beyond the previously understood role of these activity in neural circuit maturation.

Our findings demonstrate that the impact of oligodendrocyte loss on cerebellar synchronized activity and behavior is not merely transient but is gated by a narrow developmental window. Ablating oligodendrocytes during the second postnatal week— when myelination is beginning—produced a reduction in PC synchrony that was already evident at P14 and persisted unchanged into adulthood (P60–P70), even though oligodendrocyte density and MBP-positive myelin had fully recovered. The same early manipulation also led to lasting anxiety-like behavior, impaired motor coordination, and reduced sociability. By contrast, an equally robust ablation initiated after the third postnatal week (from P21) caused only a transient fall in oligodendrocyte numbers and had no measurable effect on synchrony or behavior. These results indicate that developmental myelination and/or the associated metabolic support provided by oligodendrocytes is indispensable for locking in the low-synchrony, high-fidelity state of mature cerebellar circuits, whereas the disruption outside this critical period confers little impact. Thus, the lasting network and behavioral consequences we observe stem from a time-limited developmental requirement for oligodendrocyte function rather than from chronic hypomyelination per se.

There is growing evidence that oligodendrocytes and myelin affect learning, memory, and cognitive function (de Faria et al., 2021; Munyeshyaka & Fields, 2022). Moreover, numerous studies have reported the changes of white matter, oligodendrocyte, and myelin in schizophrenia, autism spectrum disorders, depression, and anxiety, suggesting that the deficit of oligodendrocytes and myelin may underlie multiple brain disorders (Edgar & Sibille, 2012; de Faria et al., 2021). However, it is unclear when, where, and how oligodendrocytes and myelin are involved in the functionalization of brain and the pathogenesis of these brain disorders. Our data showing that synchronized spontaneous activity during development, controlled by oligodendrocytes, is a precursor to the maturation of functional neural circuits and brain functions provides a novel lens through which to examine the etiology of disorders characterized by synaptic imbalances and dysregulation of neural activity. Supporting this view, we found that reduced PC synchrony persists into adulthood after developmental oligodendrocyte loss, and that optogenetic restoration of synchrony rescues social and motor behaviors. Importantly, this effect cannot be explained by a mere increase in firing frequency. PC calcium event rates did not differ between control and DTA mice, and optogenetic stimulation of control mice did not further enhance social interaction or rotarod performance. Thus, oligodendrocyte and myelination abnormalities, and the resulting disruption of synchronized activity in neuronal populations, may be new therapeutic targets for multiple brain disorders.

Mounting evidence indicates that even subtle shifts in the temporal coordination of PC ensembles can cascade through cerebello–thalamo–cortical pathways and reshape distributed brain dynamics underlying motor planning, emotional regulation, and social cognition (Guell et al., 2018; Stoodley & Schmahmann, 2018; Lindeman et al., 2021; McAfee et al., 2021). Complex-spike synchrony is organized by aldolase-C module boundaries, generating modular timing channels within the cerebellar cortex (Tsutsumi et al., 2015), and disrupting synchrony in a single module is sufficient to alter behavior in real time (Tsutsumi et al., 2019). During reinforcement learning, these modules reorganize their synchrony into a handful of low-dimensional components that track motor initiation, error prediction, and reward valuation (Hoang et al., 2023), highlighting flexible coupling between cerebellar timing and higher-order cognitive circuits. To test the behavioral relevance of this persistent hyposynchrony we used optogenetics to transiently resynchronize PC ensembles in adult mice with DTA during development. This manipulation restored sociability and motor coordination but did not normalize anxiety-like behavior. These findings indicate that different behavioral domains vary in their dependence on cerebellar synchrony, with motor and social functions strongly reliant on precise Purkinje cell activity timing while anxiety involves additional circuit mechanisms. More broadly, neural synchrony does not operate in isolation within a single structure but propagates through cerebello–thalamo– cortical loops to influence prefrontal, limbic and parietal networks. Thus, even localized perturbations in Purkinje synchrony can acquire system-level significance and shape distributed brain states that support cognition, affect and social interaction.

Given the pivotal role of oligodendrocytes in neural circuit maturation, future research should aim to further dissect the molecular and cellular mechanisms underpinning their influence on synchronized spontaneous activity. Understanding the signaling pathways and interaction networks involved could open new therapeutic avenues for addressing developmental brain disorders and enhancing neural repair after injury.

Data availability

Requests for behavioral and/or calcium imaging data reported in this paper will be shared by the lead contact upon reasonable request. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon reasonable request.

Supplementary figures

Fig. S1.

Fig. S2.

Fig. S3.

Fig. S4.

Acknowledgements

We thank Dr. Daisuke Tanaka, Dr. Shizuki Inaba, Dr. Ayaka Nakai, and other lab members of Uesaka lab for helpful comments and discussions.

Additional information

Funding

This work is supported by MEXT Grants-in-Aid for Scientific Research on transformative research area B “Design-build of brain through spontaneous brain multimodal activity” (KAKENHI 22H05092, 22H05093), the Asahi Glass Foundation, Toray Foundation, Takeda Science Foundation, Tokumori Yasumoto Memorial Trust, and Uehara Memorial Foundation to N.U.

Author contributions

Initiation and conceptualization: N.U. Investigation: R.M., K.G., M.S., and N.U. Writing - original draft: N.U. Writing - editing: R.M., K.G., N.U., with inputs from all authors. Visualization: R.M., and N.U. Supervision: N.U. Funding acquisition: N.U.

Additional files

Supplementary table