Opposing thalamic and amygdalar projections to the dorsolateral striatum sculpt the grooming microstructure in stress and autism models

Opposing thalamic and amygdalar projections to the dorsolateral striatum sculpt the grooming microstructure in stress and autism models | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Opposing thalamic and amygdalar projections to the dorsolateral striatum sculpt the grooming microstructure in stress and autism models Wenting Wang, Xin Huang, Jinwei Xu, Zimeng Li, Baolin Guo, Honghui Mao, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9128688/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Self-grooming is an evolutionarily conserved behavior characterized by a stereotyped sequence that is often disrupted in neuropsychiatric disease models, such as obsessive‒compulsive disorder (OCD) and autism spectrum disorder (ASD). While the dorsolateral striatum (DLS) has been implicated in modulating the grooming microstructure, the underlying circuit mechanisms remain unclear. By combining automated behavioral analysis, whole-brain activity mapping, and circuit-specific manipulations, we identified two opposing DLS circuits governing this microstructure. The glutamatergic mediodorsal thalamus (MD)→DLS pathway facilitates the core motor program by specifically amplifying mid-sequence grooming phases. Conversely, the GABAergic basolateral amygdala (BLA)→DLS pathway couples grooming with anxiety states by modulating sequence initiation and termination. Relevant disease models show distinct alterations and treatment methods: in stressed mice, fragmented grooming with disrupted termination is rescued by inhibiting the BLA→DLS pathway, whereas in Shank3B knockout (KO) mice, hyperpersistent mid-sequence grooming is normalized by either the alleviation of MD→DLS hyperactivity or the activation of the BLA→DLS pathway. These findings reveal phase-specific circuits that coordinate stereotyped behavior, providing new insights into the pathophysiology of related disorders. Biological sciences/Neuroscience/Neural circuits Health sciences/Diseases/Psychiatric disorders/Autism spectrum disorders grooming microstructure dorsolateral striatum mediodorsal thalamic nucleus basolateral amygdala anxiety locomotion autism Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Introduction Stereotyped behaviors, characterized by repetitive, rhythmic actions involving complex routines and rituals, are observed both as uncommon patterns in healthy individuals and as core features of disorders such as autism spectrum disorder (ASD) 1 and obsessive-compulsive disorder (OCD) 2 . However, their pathophysiology remains incompletely understood. Self-grooming behavior in rodents offers a powerful and translatable model for investigating these behaviors, as it is a highly conserved, innate action that follows a rigid cephalocaudal sequence 3,4 . Importantly, both the quantity and the sequential microstructure of grooming are disrupted in models of ASD 5,6 and stress/anxiety 7,8 , mirroring clinical phenotypes. Thus, the grooming microstructure and the detailed organization and progression of its component stages collectively serve as a sensitive ethological measure for disentangling distinct neural mechanisms that may drive repetitive behaviors across different pathological contexts 9 . Recent studies employing an analysis of the grooming microstructure have begun to differentiate these mechanisms 10,11 . Our prior work demonstrated that Shank3B KO mice (a well-established ASD model) and acutely stressed mice both exhibit hypergrooming 5 but with fundamentally distinct microstructural patterns. Notably, the anxiolytic diazepam normalized the disrupted sequence in the anxiety model but only partially normalized it in the ASD model, suggesting that different neural pathways underlie distinct types of pathological grooming 5 . However, the specific circuit mechanisms regulating these distinct microstructural programs have not been fully elucidated. While numerous brain regions, including the limbic and hypothalamic areas 11-13 , regulate overall self-grooming behavior, researchers have increasingly focused on the specific brain circuits that generate and regulate these microstructural patterns. Corticobasal ganglia circuits 14,15 , particularly the dorsolateral striatum (DLS) 16,17 , have emerged as critical hubs for executing such sequential motor actions. Evidence has shown that lesions of the DLS disrupt the stereotyped cephalocaudal grooming chain 18 , confirming its indispensable role in organizing the grooming microstructure, which is consistent with its broader functions in action sequencing and habit formation 19,20 . Our previous work further established that dysfunction within the indirect pathway of the DLS contributes to the excessive, stereotyped grooming behavior observed in Shank3B KO mice 21 . Although the DLS is known to integrate convergent inputs from thalamic, cortical, and limbic regions 22,23 , the specific upstream circuits that shape distinct microstructural patterns and how these circuits become dysregulated in different disease states remain unclear. Thus, this study aimed to elucidate the DLS circuit mechanisms regulating the grooming microstructure and to define their distinct contributions to ASD-related stereotypy versus anxiety-driven repetitive grooming. We combined an automated grooming microstructure analysis using Janelia Automatic Animal Behavior Annotator (JAABA) software 24 , whole-brain input mapping, in vivo neuronal activity recordings, and optogenetic and chemogenetic manipulations. We investigated two principal upstream pathways to the DLS: the mediodorsal thalamus (MD)→DLS pathway, which promotes the maintenance phase of grooming behavior, and the basolateral amygdala (BLA)→DLS pathway, which shapes the initiation and termination of grooming behavior. These circuits embed distinct behavioral states, such as general locomotion and anxiety, into the grooming microstructure, explaining the divergent patterns of repetitive grooming in ASD-related stereotypy versus anxiety-driven pathology. This work provides a circuit-level framework for understanding how specific behavioral sequences are constructed and dysregulated in individuals with neuropsychiatric disorders. Results Shank3 KO and stress models display distinct grooming microstructure patterns We first used the JAABA GentleBoost algorithm 24 to train classifiers for segmenting grooming stages (Fig. 1A). Automated scoring strongly correlated with manual scoring (R 2 =0.875) and differed by less than 20% (Supplementary Fig. 1A–C). In 128 wild-type (WT) mice, the stage distributions were as follows: stage 1 (paw licking) 23.9%, stage 2 (head/face washing) 12.5%, stage 3 (body grooming) 43.1%, stage 4 (leg licking) 13%, and stage 5 (tail/genital grooming) 7.4% (Supplementary Fig. 1D). Two major microcycles emerged: 1–2/2–1 (initiation) and 3–4/4–3 (maintenance). Transitions such as 1–3 and 2–3 linked these phases, whereas 4–5 transitions marked termination (Supplementary Fig. 1D). We next applied this approach to a Shank3B KO mouse ASD model (Fig. 1B–D) and an acute restraint stress mouse model (Fig. 1G–I). Both groups presented increased total grooming duration, bout number, and mean bout length. Overlapping confidence intervals (Supplementary Fig. 1E) indicated that gross metrics alone could not distinguish these models. However, the grooming microstructure analysis revealed distinct profiles: KO mice exhibited a prolonged duration of stages 2–4 (Supplementary Fig. 1F), with a reduced percentage of stage 1 and increased percentage of stages 3–4 (Fig. 1E), consistent with perseveration of the maintenance sequence. In contrast, the stressed mice spent a prolonged time in stages 3–5, with a reduced percentage of stages 2 and 5 and an isolated increase in the percentage of stage 3 (Supplementary Fig. 1G and Fig. 1J), suggesting fragmented, anxiety-driven grooming with disrupted initiation and termination. These patterns are consistent with our previous manual observations and published reports 4,5,9 . The transition analysis further showed that the transitions of the KO mice were concentrated around stages 3–4, whereas the stressed mice displayed more transitions involving stage 5 and a globally disorganized transition architecture (Fig. 1F and Fig. 1K). In summary, although both models exhibit hypergrooming, KO mice exhibit stereotyped perseveration of the maintenance phase (stage 3–4), whereas stressed mice display a chaotic pattern with frequent shifts to the initiation and termination phases (stage 2 and 5). DLS regulation of the grooming microstructure may be linked to several upstream hub regions We systematically elucidated the broader network of the brain regions involved in grooming behavior by performing whole-brain c-Fos mapping. In Shank3B KO mice, c-Fos expression was significantly increased in 33 regions 1.5 hours after grooming onset compared with that in WT mice (Fig. 2A and Supplementary Fig. 2A). Graph theory analysis 25 of these differentially activated regions (Supplementary Table 2) revealed several hub regions with high connectivity, including the magnocellular reticular nucleus (MARN), anterior cingulate area (ACA), and somatomotor area (MO) (Fig. 2B). Similarly, compared with control conditions, acute restraint stress induced grooming-associated activation of 34 brain regions (Fig. 2C and Supplementary Fig. 2B), with hubs including the septofimbrial nucleus (SF), perirhinal area (PERI), agranular insular area (AI), BLA, and MD (Fig. 2D). A shared network across both conditions emerged, involving regions such as the MO, MD, BLA, somatosensory areas (SS), auditory areas (AUD), visual areas (VIS), nucleus accumbens (ACB), temporal association areas (TEa), ectorhinal area (ECT), PERI, cortical amygdalar area (COA), and medial amygdalar nucleus (MEA) (Fig. 2E and Supplementary Fig. 2A–D). Notably, and consistent with a previous report 26 , c-Fos activation in the striatum (including the DLS) itself was not very pronounced, likely because of the low sensitivity of c-Fos expression in this region. Although c-Fos activity in the DLS during grooming appears to be limited, the pivotal role of the DLS in governing the grooming microstructure is well documented 16,18,20 . Therefore, we employed optogenetics to directly assess its involvement in modulating the grooming microstructure. Bilateral injection of AAV-hSyn-ChR2 and sustained photostimulation of the DLS in WT mice did not alter the overall grooming behavior (Supplementary Fig. 2E–F). However, it specifically reduced the duration of grooming in stages 1 and 2 and decreased the proportional time spent grooming in stage 2 (Supplementary Fig. 1G–I); this pattern was distinct from that observed in both the Shank3 KO and restraint-stress models. These findings indicate that the DLS regulates the grooming microstructure via complex, circuit-specific mechanisms rather than through uniform activation. We subsequently mapped its upstream inputs using retrograde tracing (Fig. 2F). The densest inputs originated from the cortex, thalamus, and amygdala (Supplementary Fig. 3 and Supplementary Fig. 4A–B). Intersecting this input map with the c-Fos-activated network (Fig. 2E) pinpointed the primary sensory cortex (S1), secondary motor cortex (M2), MD, and BLA as key DLS-projecting regions that are functionally engaged during grooming behavior (Fig. 2G and Supplementary Fig. 2C–D). Detailed mapping of retrogradely labeled neurons revealed distinct spatial distributions: MD-projecting neurons clustered rostrally, whereas BLA-projecting neurons exhibited a slight rostrocaudal gradient (Fig. 2H, J). In contrast, M2 neurons were diffusely distributed, and S1 neurons clustered caudally (Supplementary Fig. 4C–D). Subsequent anterograde viral tracing from these precise locations confirmed robust terminal labeling in the DLS, establishing these regions as core hubs in the regulation of grooming (Fig. 2I, K and Supplementary Fig. 4E–F). The MD and BLA, two regions upstream of the DLS, are involved in regulating the grooming microstructure We investigated whether the identified DLS-projecting regions regulate the grooming microstructure. Fiber photometry revealed increased activity in the MD of WT mice during the grooming maintenance stage (stages 3–4) (Fig. 3A–B). Optogenetic activation confirmed that sustained MD activation increased the total duration of grooming (Fig. 3C–E and Supplementary Fig. 5A), driven by an increased duration in stages 2–4 (Supplementary Fig. 5B) and a proportional shift toward stage 4 (Fig. 3F). The transition analysis showed increased maintenance phase transitions (3–4/4–3 and 2–4/4–2; Fig. 3G). In contrast, calcium signals in the BLA increased selectively during the termination stage (stage 5; Fig. 3H–I). Further BLA activation suppressed grooming behavior (Fig. 3J–L and Supplementary Fig. 5C), reducing the time spent in stages 2–4 (Supplementary Fig. 5D) while proportionally increasing the time spent in stage 1 and decreasing the time spent in stage 4 (Fig. 3M). It also reduced transitions within the maintenance module (3–4/4–3) but increased transitions linking initiation to maintenance (Fig. 3N), creating a disorganized profile. Together, these results indicate that the MD promotes the maintenance and progression of the core sequence, whereas the BLA constrains maintenance and facilitates transitions into the initiation and termination phases. Further dynamic manipulation revealed that phasic MD activation (3 s after grooming onset, lasting 15 s) during an ongoing bout did not prolong grooming (Supplementary Fig. 6A–C), indicating that sustained activity is required for its maintenance effect. In contrast, even acute BLA activation upon grooming onset (3 s after grooming onset, lasting 5 s) robustly suppressed behavior and promoted termination, with grooming resuming after light offset (Supplementary Fig. 6D–J). These results further confirmed that the MD and BLA exert opposing, phase-specific control over the grooming microstructure. Calcium signals in M2 (Supplementary Fig. 7A) and S1 (Supplementary Fig. 7G) also increased during stages 3–4, which aligns with the observed increase in c-Fos expression. However, direct optogenetic activation of M2 (Supplementary Fig. 7B–F and Supplementary Fig. 8A) or S1 (Supplementary Fig. 7H–L and Supplementary Fig. 8B) did not alter the total grooming duration, stage distribution, or transition architecture. These findings suggest that M2 and S1 may be involved in overall motor execution 27 and somatosensory perception 28 during grooming but do not participate in the specific regulation of the grooming action sequence. MD→DLS and BLA→DLS circuits play antagonistic roles in regulating grooming behavior We next investigated whether the MD and BLA regulate grooming through the DLS. We injected a retrograde Cre virus (AAV/Retro-cre-mCherry) into the DLS and a Cre-dependent GCaMP7s (rAAV-CAG-FLEX-jGCaMP7s) into the MD to enable the specific expression of GCaMP in MD neurons projecting to the DLS. Afterward, we recorded the calcium activity of these projecting neurons. Their calcium signals increased specifically during stages 3–4 (Fig. 4A–B). Optogenetic activation of this pathway significantly increased the total duration of grooming in WT mice (Fig. 4C–E and Supplementary Fig. 9A), primarily by prolonging stages 2–4 (Supplementary Fig. 9C). An analysis of the proportions of the stages showed decreased proportions of stages 1 and 5 but increased proportions of stages 3 and 4 (Fig. 4F), indicating shortened initiation, prolonged maintenance, and accelerated termination. The transition analysis revealed enhanced transitions within the maintenance module (2–3, 3–2, 3–4, and 4–3) and suppressed initiation-linked transitions (1–2, 2–1, 1–3, and 3–1) (Fig. 4G), further confirming that MD→DLS activation specifically consolidates the core grooming sequence. Bidirectional chemogenetic manipulation confirmed that this pathway is both necessary and sufficient for sustaining the maintenance phase: inhibition shortened stages 3–4 and increased stage 1, whereas activation produced the opposite effects (Supplementary Fig. 10A–N). In contrast, BLA neurons projecting to the DLS (BLA→DLS) showed elevated calcium activity during stages 2–3 (Fig. 4H–I), in contrast to the stage 5-dominant activity observed in the general BLA population. Optogenetic activation of BLA terminals in the DLS (Fig. 4J–L and Supplementary Fig. 9B) suppressed grooming behavior, primarily by shortening stages 2–4 (Supplementary Fig. 9D). An analysis of the proportions of the stages revealed a decreased proportion of stage 4 and an increased proportion of stage 1 (Fig. 4M), indicating premature termination and a shift in behavior toward the initiation phase. Accordingly, transitions within the maintenance module (3–4 and 4–3) were reduced, whereas transitions linking initiation to maintenance (1–3 and 3–1) and within the initiation module (1–2 and 2–1) were increased (Fig. 4N), reflecting a fragmented and less consolidated grooming sequence. Chemogenetic inhibition of the BLA→DLS pathway increased total grooming and prolonged stages 3–4, consistent with a release from tonic suppression. Conversely, chemogenetic activation suppressed grooming, mirroring the optogenetic results (Supplementary Fig. 11A–M). Together, these results demonstrate that the BLA→DLS pathway exerts inhibitory control over grooming by constraining the maintenance phase (stages 3–4) and promoting transitions into the initiation and termination phases. MD→DLS and BLA→DLS pathways modulate the grooming microstructure via antagonistic synaptic transmission The antagonistic effects of the MD→DLS and BLA→DLS pathways on grooming behavior prompted us to examine their underlying projection properties. Retrograde tracing (AAV2/Retro-DIO-mCherry) into the DLS of CamkⅡ-cre mice and VGAT-cre mice revealed that the MD→DLS pathway lacks inhibitory projections (Supplementary Fig. 12A–B). In GAD67-GFP mice, AAV2/Retro-mCherry tracing showed that GABAergic neurons in the BLA→DLS pathway account for ~15% of all BLA→DLS neurons and ~14% of inhibitory neurons in the BLA (Fig. 5A). Further anterograde tracing confirmed that the MD→DLS pathway is purely glutamatergic (Fig. 5B). In vivo, optogenetic MD activation induced a transient increase in calcium levels in the DLS of both WT and Vglut2-cre mice (Fig. 5C, Supplementary Fig. 12C–D). Ex vivo optogenetic patch clamp results confirmed the presence of monosynaptic (in the presence of TTX and 4-AP), glutamatergic MD→DLS connectivity, as blue-light stimulation evoked NBQX-sensitive EPSCs (Fig. 5D). Anterograde tracing of the BLA→DLS pathway in Vglut2-cre and VGAT-cre mice confirmed that this pathway consists of both glutamatergic and GABAergic projections (Fig. 5E–F). BLA activation in WT mice evoked a biphasic DLS calcium response (an increase followed by a decrease) (Supplementary Fig. 12D–E). In Vglut2-cre mice, BLA activation caused a monophasic increase in the calcium response, whereas in VGAT-cre mice, BLA activation caused a decrease in the calcium response, which is consistent with the excitatory–inhibitory composition of this pathway (Fig. 5G–H). Ex vivo stimulation of BLA→DLS terminals evoked NBQX-sensitive EPSCs and bicuculline-sensitive IPSCs under different conditions, confirming direct glutamatergic and GABAergic projections, respectively (Fig. 5I–J). Together, these anatomical and functional results delineate a purely excitatory MD→DLS pathway and a dual-component BLA→DLS pathway that can exert opposing effects on DLS activity. We next asked how distinct neuronal populations within the BLA→DLS pathway regulate grooming. Only the GABAergic component of the BLA→DLS pathway suppressed grooming. Selective activation of GABAergic BLA VGAT →DLS projections significantly reduced the total grooming time, shortened stages 2–4 (Fig. 5K–L and Supplementary Fig. 12F), decreased the proportion of stage 4, and fragmented the sequence by reducing maintenance-phase transitions (e.g., 3–4/4–3) and increasing initiation-linked transitions (e.g., 1–3, 3–1, 2–3, and 3–2) (Fig. 5M–N), reflecting a fragmented sequence with impaired sustainment of the core grooming phase. In contrast, the selective activation of glutamatergic BLA Vglut2 →DLS projections had a minimal effect on grooming (Supplementary Fig. 12G–K). Similarly, in situ activation of BLA Vglut2 neurons did not alter grooming, whereas activating BLA VGAT neurons in situ reproduced the suppressive phenotype observed with pathway-specific stimulation (Supplementary Fig. 13A–H). Notably, in situ BLA VGAT activation increased the proportions of both stage 1 and stage 3 but decreased the proportion of stage 4, further underscoring its role in disrupting the maintenance–termination transition and promoting erratic, initiation-biased grooming. Together, these results establish that the grooming-suppressive function of the BLA→DLS pathway is specifically mediated by its GABAergic component, which constrains the maintenance phase and fragments the behavioral sequence. The MD→DLS and BLA→DLS pathways separately integrate the motor and affective components of grooming Previous studies suggest that grooming comprises multiple components, including pure motor execution, compulsive sequencing, emotional relief, and somatic sensation 5,8,28 . We further elucidated the potential regulatory roles of the MD→DLS and BLA→DLS circuits by analyzing the full behavioral repertoire of WT mice following the activation of these circuits using a high-resolution 3D pose-tracking system 29 . The behavioral sequences were clustered into 40 discrete modules (Fig. 6A–C and Supplementary Table 3). MD→DLS activation specifically increased the activity of modules associated with sustained, crouched grooming postures (modules 28 and 27) and grooming–maintenance movements (module 12) while reducing the activity of modules related to stationary inactivity, cessation-like behaviors (modules 37 and 9), and transitional actions (module 10: beginning to rear up) (Fig. 6D and Supplementary Fig. 14A). Conversely, BLA→DLS activation broadly suppressed exploration and orienting modules, including rearing (modules 30, 4, and 29), directed locomotion and turning (modules 25, 26, 36, 18, and 40), and sniffing/head investigation (modules 20, 19, and 14), while it also decreased stationary postures (module 37) and increased freezing-like pauses (module 2) (Fig. 6D and Supplementary Fig. 14B–C). These profiles indicate that MD→DLS activation promotes focused, sustained grooming postures and maintenance-related movements while reducing nongrooming static or transitional behaviors. In contrast, BLA→DLS activation increases behavioral arrest and suppresses locomotion, turning, and exploratory activity. Next, we assessed whether manipulations of these circuits influence general locomotion and anxiety-like states using an open field test (OFT) and an elevated plus maze (EPM). Activating the MD→DLS pathway significantly increased the total distance traveled in the OFT, which was positively correlated with the duration of grooming in the restricted chamber (Fig. 6E–F), indicating a coordinated enhancement of motor activity and grooming. However, it did not affect the time spent in the center of the open field or open arms of the EPM and was not correlated with anxiety-related behaviors (Fig. 6G–H), suggesting that this circuit modulates motor output without directly affecting anxiety-like behaviors. In contrast, the selective activation of the GABAergic BLA VGAT →DLS pathway did not alter locomotion but did significantly increase the time spent in the center zone of the open field and open arms of the EPM (Fig. 6I–K), and the reduction in anxiety was correlated negatively with the grooming duration, indicating concurrent grooming suppression and anxiety mitigation. Moreover, the activation of the glutamatergic BLA Vglut2 →DLS component reduced both locomotion and center/open-arm time without affecting grooming (Supplementary Fig. 14D–F), indicating that its role in anxiety is distinct from that of grooming control. Together, these results delineate a double dissociation: the MD→DLS pathway promotes grooming and locomotion without altering anxiety, whereas the GABAergic BLA VGAT →DLS pathway concurrently suppresses grooming and reduces anxiety-like behaviors, highlighting the circuit-specific integration of motor and affective components. The BLA→DLS circuit but not the MD→DLS circuit regulates the grooming microstructure in animals subjected to restraint stress We next examined the roles of the MD→DLS and BLA→DLS circuits in regulating grooming in mice under acute restraint stress. Calcium imaging revealed that the MD response remained unchanged except for elevated activity in stage 2, whereas BLA activity decreased in stages 1, 2, and 5 but increased during stages 3 and 4 (Fig. 7A–D). This pattern, with reduced engagement during initiation and termination but increased activity during maintenance, suggests that under acute stress, the BLA shifts from its typical role in structuring behavioral boundaries (initiation/termination) toward a more persistent, maintenance-associated signaling mode, which may contribute to the fragmented grooming sequence observed in stressed mice. Simultaneous calcium imaging in the DLS revealed no change (Supplementary Fig. 15A–B), indicating that the altered drive in stress-induced grooming originates upstream, likely from the BLA. Optogenetic inhibition of the MD→DLS circuit did not alter stress-induced grooming (Supplementary Fig. 15C–E, H–J). In contrast, activating the GABAergic BLA VGAT →DLS pathway rapidly suppressed this process, shortening stages 2–4, increasing stage 1, and fragmenting transitions by reducing the maintenance phase (3–4/4–3) and increasing the initiation-to-maintenance (1–3/3–1) shifts (Supplementary Fig. 15F–G, K‒M), reflecting fragmented progression characterized by short maintenance and inefficient sequence consolidation. Next, we used targeted recombination in active populations (TRAP) in Fos-cre ER mice to express the ChR2 protein in stress-activated neurons in the MD and BLA 26 (Fig. 7E). Following the restraint-stress protocol, 44.12% of TRAPed MD neurons were glutamatergic and grooming related (Supplementary Fig. 16A, C), whereas 29.44% of Trapped BLA cells were grooming related, with significant GABAergic enrichment (Supplementary Fig. 16B, D). Activating TRAPed BLA→DLS neurons suppressed stress-induced grooming, partially restoring the stage balance (increasing the characteristic deficit in stage 2) and restabilizing maintenance-phase transitions (e.g., increasing 2–3/3–2 transitions) (Fig. 7F–J and Supplementary Fig. 16E). In contrast, inhibiting TRAPed MD→DLS neurons had no significant effect (Supplementary Fig. 16F–J). Together, these findings highlight the critical role of the BLA→DLS circuit, particularly its GABAergic component, in regulating grooming under stress, whereas the MD→DLS circuit primarily supports motor execution without restructuring anxiety-driven behaviors. Decoupling of the grooming microstructure from motor and anxiety-like behaviors by the MD→DLS and BLA→DLS circuits in Shank3B KO mice In Shank3B KO mice, we investigated the roles of the MD→DLS and BLA→DLS circuits in stereotyped grooming. Calcium imaging revealed a stronger response in the MD and a weaker response in the BLA of the Shank3B KO mice than in the WT mice (Fig. 8A–B). The stage-resolved analysis showed that MD activity was increased in stages 1, 3, and 4 in KO mice, whereas BLA activity increased in stages 1, 2, and 5 but decreased in stages 3 and 4 (Fig. 8C–D). This pattern is consistent with the microstructural phenotype of the KO mice (Fig. 1E), suggesting that increased MD activity during the maintenance phases (stage 3–4) likely drives the prolonged core grooming sequence, whereas altered BLA engagement during initiation (stage 1) and termination (stage 5) may reflect disrupted anxiety-related modulation. Simultaneous calcium recordings in the DLS of the KO mice showed responses similar to those of the WT mice but revealed stage-specific shifts, with lower signals in stages 1 and 3 and a higher signal in stage 4 (Supplementary Fig. 17A–B), indicating a redistribution toward the maintenance phase. Optogenetic inhibition of the MD→DLS circuit reduced the total grooming time in the KO mice (Fig. 8E) by shortening stages 3 and 4 (Supplementary Fig. 17C), increasing stage 1 and decreasing stage 4 (Fig. 8F), indicating interrupted progression from initiation to sustained maintenance. The transition analysis revealed decreases in maintenance-phase transitions (3–4 and 4–3) and in shifts linking initiation to maintenance (1–3 and 3–1), as well as increases in initiation-phase transitions (1–2 and 2–1) and early maintenance transitions (2–3 and 3–2) (Fig. 8G). These shifts reflect a “forward-biased” sequence with more cephalic grooming and less sustained body grooming, ultimately leading to an overall reduction. Similarly, optogenetic activation of the GABAergic BLA VGAT →DLS pathway also reduced the total grooming time in KO mice, increasing stage 1, decreasing stages 2 and 5, and fragmenting maintenance-phase transitions while increasing erratic initiation-to-maintenance transitions (e.g., 1–3/3–1) (Fig. 8H–J, Supplementary Fig. 17D). Activating the general BLA→DLS pathway resulted in a similar disruption (Supplementary Fig. 17E, G–I), whereas activating its glutamatergic component had no effect, confirming that GABAergic projections mediated this corrective effect (Supplementary Fig. 17F, J–L). Inhibiting the MD→DLS circuit in the KO mice reduced general locomotion, which correlated with decreased grooming (Supplementary Fig. 18A–B), but did not affect anxiety (Supplementary Fig. 18C–E). Conversely, activating the BLA VGAT →DLS pathway reduced anxiety-like behaviors (increased center/open-arm times) and was negatively correlated with grooming but did not alter locomotion (Supplementary Fig. 18F–J). These findings indicate that in this autism model, the MD→DLS pathway coordinates grooming execution and motor activity, whereas the GABAergic BLA→DLS pathway links grooming modulation to the emotional state. Both pathways may serve as potential targets for addressing pathological stereotypy. Discussion Self-grooming is a crucial adaptive behavior for maintaining physiological homeostasis and coping with environmental stress, and its dysregulation is closely linked to anxiety and stereotypy 30,31 . Studying its microstructure is therefore essential for distinguishing distinct pathological states. Previous studies 16-18 , including our own 21 , have established the DLS as a key node regulating the stereotyped sequential syntax of grooming. Here, we identified two main upstream inputs to the DLS, namely, the glutamatergic MD→DLS pathway, which coregulates locomotion and sustains the grooming maintenance phase, and the GABAergic BLA→DLS pathway, which co-regulates anxiety, constrains maintenance, and promotes sequence initiation. Under pathological conditions, inhibiting the MD→DLS pathway alleviated stereotyped grooming in an autism model by shortening the hyperpersistent maintenance phase. Activating the BLA→DLS circuit ameliorated repetitive grooming in both the stress and autism models by normalizing the phase durations and correcting condition-specific transition abnormalities. These findings reveal how distinct circuit logics generate heterogeneous pathological grooming, providing a framework for dissecting related neuropsychiatric disorders. Self-grooming is a vital adaptive behavior, and its dysregulation manifests across various neuropsychiatric conditions 30 . Traditional analyses, which often rely on the total duration or coarse metrics such as “incorrect transitions” 3,32 , fail to capture the nuanced syntax of this behavior, thereby limiting their ability to distinguish distinct pathological states. Instead, we employed a detailed assessment of the phase durations, proportional distributions, and complete transition probability matrix. Building on the classic cephalocaudal sequence 4,9,33,34 , our analysis defines a tripartite organization during grooming behavior: the initiation phase (stages 1–2, paw and head grooming) 35 , the maintenance phase (stages 3–4, body and leg grooming) 3 , and the termination phase (stage 5, tail/genital grooming) 5 . The proportional time spent in each phase and the transition probabilities between them provide a framework for deconstructing the grooming microstructure. Applying this framework, we confirmed distinct microstructural signatures in pathological models. Acute restraint stress produced a pattern of truncated initiation and prolonged, oscillatory maintenance, characterized by a surge in reciprocal 3–4/4–3 transitions, which is indicative of a sequence “locked” in repetitive body cleaning with impaired progression to termination (reduced stage 5). This result aligns with the ethological concept of “displacement” grooming under stress, which is often described as fragmented and chaotic 36 . In stark contrast, Shank3B KO mice exhibited extended maintenance with altered initiation, persisting in a sustained body cleaning loop with reduced transitions to/from the cephalic phase (reduced 2–3/3–2 transitions). This pattern reflects a more rigid, stereotyped sequence, which is consistent with the behaviors observed in individuals with compulsive disorders 37 . These distinct signatures strongly support fundamentally different neural mechanisms, extending our prior finding that anxiolytic treatment normalizes the stress-induced but not ASD-related microstructure 5 . Our study elucidates the origins of the grooming microstructure by defining how the DLS, a key node for action sequencing 16,17 , integrates opposing commands from the MD 38 and BLA 39 to sculpt grooming behavior. We identified the MD→DLS and BLA→DLS pathways as critical antagonistic regulators. The MD→DLS circuit is purely glutamatergic and promotes the maintenance phase of grooming. Its activation accelerates and sustains grooming sequences, characterized by a shortened initiation phase (reduced proportion of stage 1) and an amplified maintenance phase (increased stage 3–4 activity and transitions). These findings are consistent with calcium activity being most elevated in MD neurons and MD→DLS projections during the maintenance phase (stages 3–4). This result aligns with the role of the MD in motor planning and compulsive behaviors 40,41 . Furthermore, recent studies have reported hyperactivity of the MD in adult Shank3B KO mice 42 . Conversely, the activation of the GABAergic component of the BLA→DLS pathway suppresses and fragments grooming sequences by prolonging the initiation phase (increased stage 1 proportion), prematurely aborting the maintenance phase (decreased proportion of stage 4), and shifting transition dynamics toward oscillations between initiation and early maintenance (e.g., increased 1–3/3–1 transitions). This shift effectively disrupted the fluent cephalocaudal progression of grooming. The functional profile of this pathway corresponds to BLA→DLS neuronal activity peaking during stages 2–3 and prominent BLA signals at termination, suggesting a role in monitoring and potentially curtailing ongoing sequences based on the internal state. This finding fits the established role of the BLA in assigning emotional valence and anxiety processing 43,44 and the comorbidity of anxiety with repetitive behaviors 45 . Thus, the DLS acts as an integrator, balancing a motor-sequencing “go” signal from the MD against an affective “check” signal from the BLA to generate fluent, context-appropriate grooming. The divergent dysfunctions of these circuits underlie distinct pathologies. The BLA→DLS circuit is critically involved in acute stress: BLA activity shifts, and activating its GABAergic projections to the DLS rapidly normalizes the chaotic microstructure. Inhibiting the MD→DLS pathway has no effect, indicating that anxiety-driven pathology primarily involves maladaptive affective regulation. In contrast, stereotypy in Shank3B KO mice is driven by hyperactivity of the MD→DLS pathway, which promotes perseverative maintenance. The inhibition of this motor-sequencing pathway is therapeutic. Activating the GABAergic BLA→DLS pathway also ameliorates stereotypy but by fragmenting the rigid sequence, highlighting anxiety as a separable, modulatory component of ASD. This circuit-based dissection provides a mechanistic framework for understanding how heterogeneous pathological grooming arises from imbalances between motor sequence facilitation and affective state modulation. Whole-brain c-Fos mapping confirmed that grooming engages a distributed network beyond these hubs, including motor regions (e.g., the MO and ACA) 27,46 , sensory areas (e.g., S1 and VIS) 47 , and limbic regions (e.g., the AI and MEA) 48 . These findings underscore grooming as a complex integration of motor execution, sensory feedback, and internal states. Notably, while M2 and S1 were active during grooming, particularly in the maintenance phase, their direct optogenetic manipulation did not alter the grooming microstructure. These findings suggest that these areas function as hierarchical executors and sensors within sensorimotor loops, which are necessary for movement generation and somatosensory feedback 49 but not for determining the sequential microstructure itself. Their influence may be mediated by other striatal territories, such as the DMS 50,51 , highlighting the specificity of the DLS for automated sequence execution. Several limitations point to future directions. First, while we focused on the net effect of DLS outputs, further investigation is needed to elucidate how MD and BLA inputs engage specific striatal pathways (D1/D2) 50 and interneurons 52 under pathological conditions. Second, our study examined anxiety-driven and intrinsic stereotyped grooming but did not address “comfort” grooming (triggered by mild pressure or physical disturbances such as sprays) 53 , which may engage distinct neural circuits 12 . Thus, comparative ethological studies are warranted. Third, all experiments were conducted using male mice. Given the well-known sex differences in stress 54 , anxiety 55 , ASD 56 , and grooming behavior 57 , including both sexes in the future is crucial for generalizability. In conclusion, we established that the fine-grained microstructure of grooming, rather than its overall quantity, is responsible for its underlying neural logic. The DLS integrates antagonistic commands from a glutamatergic MD-based motor circuit that sustains sequences and a GABAergic BLA-based affective circuit that constrains them. The hyperactivity of the former drives the perseveration of stereotypy, whereas altered engagement of the latter underlies anxiety-driven fragmentation. This framework advances our understanding of how specific behavioral sequences are constructed and pathologically distorted, identifying refined circuit targets for dissecting and potentially treating related neuropsychiatric symptoms. Methods Animals All the experimental protocols were approved by the Fourth Military Medical University Institutional Animal Care and Use Committee and were conducted in accordance with the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH). All the animals were 6- to 8-week-old (20 to 25 g) male mice housed at a constant temperature (23–25°C) on a 12 h light/dark cycle (lights on from 08:30 to 20:30) with ad libitum access to water and food. C57BL/6 mice were obtained from the Laboratory Animal Center of the Fourth Military Medical University. Shank3B KO mice were kindly provided by Prof. Guoping Feng 6 . All the mice used in this study were male and maintained on a pure C57BL/6J background. Shank3B KO::Vglut2-cre double-positive mice were generated by crossing Shank3B heterozygous mice with Vglut2-cre mice (strain #016963, The Jackson Laboratory). Similarly, Shank3B KO::VGAT-cre double-positive mice were obtained by crossing Shank3B heterozygous mice with VGAT-cre mice (strain #028862, The Jackson Laboratory). Fos-cre ER mice were purchased from The Jackson Laboratory (strain #021882). Grooming behaviors induced under different conditions Two methods were used to elicit grooming behaviors in mice. The first utilized Shank3B KO mice, which inherently display high levels of stereotypic grooming behavior 6 . The second method involved an acute restraint stress model 58 . Briefly, mice were placed into a modified 50 ml Falcon tube with breathing holes, where only the tails and tips of their noses were free, for 30 min; upon release, the mice exhibited extensive grooming behavior. Grooming behavior and grooming microstructure analysis Grooming behavior was recorded in custom-built, sound-attenuating chambers (20 cm × 20 cm × 25 cm) equipped with adjustable top-mounted LED lighting for behavioral recording. Experiments were conducted under low illumination (~40 lux) to capture 30-minute videos of freely moving mice for the analysis of grooming behavior. Three walls of the chamber were opaque, with only the wall facing the camera (30 fps) being transparent to minimize environmental stress. Prior to testing, the mice were acclimatized to the behavioral room for at least one hour before being gently transferred into the chamber using an opaque tray. After each session, the chamber was cleaned by removing waste, spraying with 75% ethanol to eliminate odors, and rinsing with water. The analysis of grooming behavior was performed using the open-source software JAABA (Janelia Automatic Animal Behavior Annotator) 59 (downloaded from GitHub https://github . com/kristinbranson/JAABA)). Mouse movement trajectories were first extracted using the Mouse Tracker toolbox 60 in MATLAB and then imported into JAABA. A supervised GentleBoost learning algorithm was trained on 115,268 manually annotated video frames to create a classifier for identifying grooming behavior. The trained classifier showed strong agreement with manual scoring, with an error margin of less than 20% for the grooming duration (Supplementary Fig. 1A–C), confirming its reliability for subsequent analyses under consistent experimental conditions. For the detailed bout analysis, the following criteria were applied: a grooming episode was analyzed only if it lasted longer than 3 seconds, and a pause in grooming exceeding 6 seconds was defined as the end of one bout and the start of a new bout 3,5 . Using the same training strategy, separate classifiers were developed for the distinct grooming stages: stage 1 (paw licking) trained on 93271 frames of video, stage 2 (nose/face/head wash) trained on 84014 frames, stage 3 (body grooming) trained on 83877 frames, stage 4 (leg licking) trained on 44697 frames, and stage 5 (tail/genitals grooming) trained on 42339 frames. By applying both the general grooming classifier and the stage-specific classifiers to the same video recordings, we extracted comprehensive quantitative metrics of the grooming microstructure 3 . These metrics included the absolute time spent in each grooming stage, the proportional distribution of grooming time across stages, the total number of stage transitions, and the frequency of transitions between each specific pair of stages 7 . c-Fos experiments and graph theory analysis Quantification of whole - brain c-Fos expression We used c-Fos, a transcription factor that is rapidly upregulated upon neuronal activation, as a marker of behavior-associated neural activity to identify brain regions upstream of the dorsolateral striatum (DLS) that were activated during grooming. Grooming behaviors were recorded and mice were sacrificed 1.5 hours later and processed for c-Fos staining. Briefly, the animals were anesthetized with isoflurane and transcardially perfused with 0.01 M PBS followed by 4% paraformaldehyde. The brains were postfixed overnight at 4°C, cryoprotected in 30% sucrose, and sagittally sectioned at 50 μm using a freezing microtome. Sections were incubated with a rabbit anti-c-Fos antibody (1:1500; Cell Signaling Technology #2250) and subsequently with a donkey anti-rabbit Alexa Fluor 488 antibody (1:500; Invitrogen A-21206). Nuclei were counterstained with Hoechst 33342 (1:1000; Abcam ab228551). High-resolution images of the brain sections were acquired at 150-μm intervals and analyzed using a three-step pipeline, as described previously 61 : (1) signal recognition—c-Fos-positive cells were automatically identified using a classifier trained in ImageJ (Fiji); (2) brain registration—raw images of the sections were aligned to the Allen Mouse Brain Atlas using MATLAB; and (3) brain segmentation—positive signals were assigned to specific subregions based on atlas-defined grayscale intensities in Imaris software. All the analyses were conducted independently by two observers. Graph theory and hub network identification We employed a graph theory-based approach to delineate functional networks among c-Fos-positive brain regions and to identify hub regions within these networks 62 . After the c-Fos-positive cell counts for each brain region were obtained, the raw data were normalized using log10 transformation to reduce variability. Pearson’s correlation coefficients were then calculated between every pair of brain regions. These interregional correlation coefficients were subsequently converted into pairwise Euclidean distances. Hierarchical cluster dendrograms were generated from the distance matrices to identify functional modules, with the tree-cutting threshold set at 50% of the tree height to split the dendrogram into distinct modules. In parallel, the correlation matrix was reorganized to construct the functional network, and a connection between two regions was included only if the Pearson correlation coefficient was ≥ 0.82 63 . The resulting network parameters were analyzed using the Brain Connectivity Toolbox (https://sites.google.com/site/bctnet/) to compute the following centrality metrics: (1) degree, which is defined as the number of links connected to a node; (2) normalized betweenness centrality, which is defined as the fraction of all shortest paths in the network that pass through a given node; and (3) eigenvector centrality, which is a measure of the influence of as node based on the centrality of its neighbors. Brain regions that ranked highly across all three centrality measures were defined as hub nodes (Supplementary Table 2). The functional network was visualized using Cytoscape software 64 . Stereotaxic injections and fiber optic implantation The mice were anesthetized with 1% isoflurane and aligned on a stereotactic frame (RWD Life Science Inc., China). All surgeries were performed under sterile conditions. The heads of the mice were shaved and cleaned with 75% alcohol, and their skulls were thinned with a dental drill and then removed. Stereotactic coordinates were based on the Paxinos and Franklin mouse brain atlas and adjusted appropriately. Each mouse was injected with viral vectors using a glass micropipette at a rate of 30–60 nL per min. The micropipette remained in position for 10 min after the injection and then was slowly withdrawn. The mice were rewarmed on a thermal blanket after surgery and allowed to recover for more than 3 weeks before they participated in the behavioral experiments. For optogenetic behavioral experiments, rAAV-hSyn-mCherry-WPRE-pA, rAAV-hSyn-EGFP-WPRE-pA, rAAV-hSyn-NpHR3.0-EGFP-WPRE-pA, or rAAV-hSyn-hChR2(H134R)-mCherry-WPRE-pA was injected bilaterally into M2 (+0.86 mm AP, ±0.75 mm ML, -1.30 mm DV; 300 nl per site), S1 (-1.34 mm AP, ±1.50 mm ML, -1.00 mm DV; 100 nl per site), MD (-1.34 mm AP, ±0.40 mm ML, -3.40 mm DV; 100 nl per site), or BLA (-1.40 mm AP, ±3.15 mm ML, -4.80 mm DV; 150 nl per site) using a glass micropipette. rAAV-EF1a-DIO-hChR2(H134R)-mCherry-WPRE-pA, rAAV-Ef1a-DIO-NpHR3.0-EGFP, or rAAV-EF1a-DIO-mCherry-WPRE-pA was injected bilaterally into the MD or BLA of Vglut2-Cre or VGAT-cre animals. A 200 mm core 0.37 NA optical fiber was implanted 0.5 mm above the injection site. For MD, bilateral ceramic ferrules were inserted at a 15° angle (-1.34 mm AP, ±1.26 mm ML, -3.00 mm DV). For optogenetic targeting of MD/BLA→DLS terminals, viruses similar to those described above were injected bilaterally into the MD or BLA. Optical fibers were then implanted in the DLS, specifically at an anterior site (+0.70 mm AP, ±2.25 mm ML, -2.5 mm DV) for MD→DLS manipulation and at a posterior site (+0.00 mm AP, ±2.5 mm ML, -2.3 mm DV) for BLA→DLS manipulation. For chemogenetic activation experiments, AAV2/retro-hSyn-NLS-Cre-P2A-mCherry was injected bilaterally into the DLS (+0.70 mm AP, ±2.25 mm ML, -3.00 mm DV; 300 nl per site), and rAAV-hSyn-DIO-hM3Dq (Gq)-eGFP-WPRE-pA was injected bilaterally into the MD (-1.46 mm AP, ±0.40 mm ML, -3.40 mm DV; 100 nl per site) or BLA (-1.40 mm AP, ±3.15 mm ML, -4.80 mm DV; 150 nl per site). For chemogenetic inhibition experiments, AAV2/retro-hSyn-NLS-Cre-P2A-mCherry was injected bilaterally into the DLS, and rAAV-EF1a-DIO-hM4D(Gi)-EGFP-WPRE was injected bilaterally into the MD or BLA. For chemogenetic control experiments, AAV2/retro-hSyn-NLS-Cre-P2A-mCherry was injected bilaterally into the DLS, and rAAV-EF1a-DIO-EGFP-WPRE-bGH pA virus was injected bilaterally into the MD or BLA. For input mapping experiments, AAV2/2Retro-hSyn-eGFP-3Flag-WPRE-SV40pA was injected unilaterally into the DLS (+0.70 mm AP, -2.25 mm ML, -3.00 mm DV; 200 nl per site) of WT mice; AAV2/2Retro-DIO-mCherry was injected unilaterally into the DLS of CamKⅡ-cre and VGAT-cre mice. For anterograde monosynaptic tracing, AAV-hSyn-mCherry was injected unilaterally into the MD and BLA of WT mice; AAV-EF1α-DIO-mRuby-T2A-Synaptophysin-EGFP was injected unilaterally into the MD and BLA of Vglut2-cre and VGAT-cre mice. The animals were sacrificed 3 weeks later for monosynaptic tracing. Fiber photometry Local calcium signals were recorded after rAAV-hSyn-jGCaMP8s was injected into the target brain regions (M2, S1, MD or BLA). For projection-specific calcium imaging, AAV2/retro-hSyn-NLS-Cre-P2A-mCherry was injected unilaterally into the DLS, while rAAV-CAG-FLEX-jGCaMP7s was injected ipsilaterally into either the MD or BLA. Ceramic ferrules were implanted 0.3 mm above the virus injection sites. After a 3-week recovery period, fiber photometry recordings were performed to simultaneously record calcium signals and grooming behavior evoked under different conditions over a 30-min session. The onset of calcium signals was aligned frame by frame with the behavioral start time, and signals were analyzed separately according to distinct grooming stages. A 2-s preevent window was used as the baseline, and Z scores during the 5-s postevent period were calculated to analyze the area under the curve. All analyses were conducted independently by two observers who were blinded to the group assignments. Modulation of behavior Optogenetic modulation of grooming behavior For the optogenetic manipulation of grooming, the mice were acclimated to the testing room for 1 h prior to recording. Grooming behavior was monitored over a 1.5-hour session, following a laser OFF‒ON‒OFF paradigm with 30-minute intervals. In ChR2-mediated activation experiments, blue laser light (473 nm) was delivered at 20 Hz (5-ms pulse width). For NpHR-mediated inhibition, continuous yellow light (589 nm) was applied. During the regional stimulation of M2, S1, BLA or MD, the laser power was calibrated before each session to ensure an output intensity of 5–7 mW per side at the fiber tip. For the terminal stimulation of MD/BLA-DLS projections, the output was set to 9–10 mW per side. Successive stimulation sessions for each mouse were separated by at least 24 h. We also performed acute optogenetic stimulation triggered upon the initiation of self-grooming to control for potential thermal effects induced by prolonged (30 min) laser exposure (Supplementary Fig. 6). A 5 s (BLA→DLS) or 15 s ( MD→DLS) light pulse (20 Hz) was delivered during the 30-min recording, and the following parameters were analyzed: grooming termination (percentage of grooming bouts that stopped during the 5 s or 30 s light-on window), offset latency (time from light onset to grooming cessation), grooming retrieval (percentage of terminated bouts in which the animals resumed grooming), retrieval grooming time (mean duration of resumed grooming bouts), and total grooming (cumulative grooming time during the 30-min session). Chemogenetic manipulation of grooming behavior For DREADD experiments, C21 or vehicle (saline) was intraperitoneally injected 30 min before behavioral testing to allow sufficient time for Gq- or Gi-mediated signaling modulation during the assay. The testing protocol consisted of two consecutive daily sessions: on day 1, the mice received an intraperitoneal injection of 0.9 % saline, and their grooming behavior was recorded for 30 min to establish a baseline; on day 2, C21 (2 mg/kg) was administered to examine how the activation or inhibition of MD→DLS or BLA→DLS circuits influenced grooming behavior. Three-dimensional behavioral recordings during optogenetic stimulation The 3D behavioral test was performed based on a hierarchical learning framework for spontaneous behavior mapping 29 . The mice were individually placed in a transparent, custom-built rectangular chamber (20 cm × 20 cm × 25 cm) and allowed to explore freely under dim, homogeneous lighting. Behavior was recorded in a 15-min session following a laser OFF–ON–OFF stimulation paradigm (5 min intervals). Optogenetic activation of MD/BLA→DLS terminals was achieved with blue light (473 nm, 20 Hz, 5 ms pulses, 9–10 mW output at the fiber tip). Four high-speed cameras (100 fps) were positioned at 90° intervals around the arena to capture synchronized multiview videos. The 3D posture of each animal was reconstructed by first detecting 2D key points in each view using DeepLabCut, followed by triangulation to establish 2D-to-3D correspondence. A hierarchical graph network model was then applied to extract continuous 3D kinematic features and cluster them into discrete, semantically meaningful behavioral units (e.g., rearing, turning, stretching, and grooming). This system enables the high-throughput, automated profiling of the spontaneous behavioral structure with minimal experimental interference. Targeted Recombination in Active Populations (TRAP) A TRAP strategy was employed in Fos-cre ER mice to label stress-responsive neurons. Cre-dependent viral vectors were stereotactically injected into either the MD or BLA, and optical fibers were bilaterally implanted above the DLS. Following a two-week post-operative recovery and viral expression period, the mice were subjected to a three-day restraint stress protocol. During this paradigm, 4-hydroxytamoxifen (4-OHT) (Sigma, H7904) was administered intraperitoneally (i.p.) at a dose of 100 mg/kg to induce recombination and permanently label neurons activated by stress. After an additional two-week period to ensure full transgene expression, circuit manipulation experiments were performed. Anxiety-related behavioral tests Open Field Test (OFT) The OFT was conducted as previously described to detect anxiety and locomotor activity 65 . The apparatus consisted of a 50 cm × 50 cm × 50 cm polystyrene arena . The mice were acclimated to the testing room for 1 h prior to the trial. Each mouse was then gently placed in the center of the arena, and its behavior was recorded for 10 min. The arena was cleaned with 75% ethanol between trials to remove olfactory cues. The total distance traveled over the 10-min session was analyzed using the Smart system (SMART 3.0, Panlab S.L.U., Spain) to assess general locomotor activity. The time spent in the central zone (defined as the central 25% of the arena area) was used as an index of anxiety-like behavior. Elevated Plus Maze (EPM) The EPM was constructed with two open arms (25 cm × 8 cm) and two enclosed arms (25 cm × 8 cm, with 12 cm high walls) connected by a central platform (8 cm × 8 cm) 66 . The maze was elevated 90 cm above the floor. The mice were habituated to the test room for 1 h before the trial. Each animal was placed on the central platform facing an open arm, and behavior was recorded for 5 min. The maze was cleaned with 75% ethanol between subjects. The time spent in the open arms was quantified to evaluate anxiety-like states. Immunohistochemistry For viral tracing, the mice were perfused 3 weeks after the virus injection. Sagittal brain sections (50 μm thick) were incubated with a rabbit anti-GFP antibody (1:1000; Invitrogen, A-11122) or, for site verification, a rabbit anti-RFP antibody (1:1000; Rockland, 600-401-379). Mice received 10 min of 20 Hz photostimulation and were perfused 90 min later to confirm neuronal activation in optogenetic experiments. The sections were stained with rabbit anti-c-Fos (1:1500; Cell Signaling Technology #2250) or rabbit anti-Arc (1:1000; Synaptic Systems 156003) antibodies. GABAergic neurons were labeled a mouse anti-GAD67 antibody (1:1000; Sigma‒Aldrich, MAB5406). For glutamatergic neuron labeling, a mouse anti-Vglut2 antibody (1:100; Sigma‒Aldrich, MAB5504) was used. For immunofluorescence staining of all sections, the following secondary antibodies were used: a donkey anti-rabbit antibody conjugated to Alexa Fluor 488 (1:500, Invitrogen, A-21206) or Alexa Fluor 594 (1:500, Invitrogen, A-21207). Nuclei were counterstained with Hoechst 33342 (1:1000; Abcam, ab228551). Injection and fiber placement sites (e.g., the DLS, M2, S1, MD, and BLA) were verified on coronal sections along the anteroposterior axis (Supplementary Figs. 6 and 12). Electrophysiology We characterized the functional properties of MD/BLA→DLS projections by combining optogenetic stimulation with whole-cell patch-clamp recordings to examine synaptic transmission. The upstream region was injected with AAV-hSyn-hChR2(H134R)-mCherry. Three weeks later, brain slices containing the DLS were prepared, and light-evoked postsynaptic currents were recorded from neurons near labeled terminals. Slice preparation: Mice were deeply anesthetized with isoflurane and transcardially perfused with ice-cold, oxygenated cutting solution (115 mM choline chloride, 2.5 mM KCl, 1.25 mM NaH 2 PO4, 0.5 mM CaCl 2 , 8 mM MgCl 2 , 26 mM NaHCO 3 , 10 mM D-(+)-glucose, 0.1 mM L-ascorbic acid, and 0.4 mM sodium pyruvate, pH 7.4, osmolarity of 295–300 mOsm/L). The brains were quickly removed, and 300-μm-thick slices containing the target regions were cut using a vibratome. Slices were recovered in oxygenated artificial cerebrospinal fluid (ACSF, 119 mM NaCl, 2.3 mM KCl, 1.0 mM NaH 2 PO 4 , 26 mM NaHCO 3 , 11 mM D-(+)-glucose, 1.3 mM MgCl 2 , and 2.5 mM CaCl 2 , pH 7.4, osmolarity of 295–300 mOsm/L) at 32°C for 30 min and then maintained at room temperature for ≥1 h before recording. Recording setup: The slices were transferred to a submerged recording chamber. Under fluorescence guidance, healthy neurons near mCherry-positive axons were selected for whole-cell recording. EPSC recording: Pipettes were filled with a potassium-based internal solution. ACSF was supplemented with bicuculline (a GABAA receptor antagonist, 100 μM), tetrodotoxin (TTX, 1 μM), and 4-AP (100 μM) to isolate monosynaptic currents. Neurons were voltage clamped at -60 mV. Blue light pulses (473 nm, 10 mW, 20 Hz, 200 ms) were delivered to evoke EPSCs. IPSC recording: Pipettes contained a high-chloride cesium-based internal solution. ACSF was supplemented with D-AP5 (an NMDA receptor antagonist, 50 μM), NBQX (10 μM), TTX (1 μM), and 4-AP (100 μM). Neurons were clamped at -60 mV, and identical light stimulation was applied to evoke IPSCs. Statistical analysis All the data are presented as the means ± standard errors of the means (SEMs). Statistical analysis and graphing were performed using SPSS 22.0. The normality of the distribution of the data was assessed using the Shapiro‒Wilk test, and the homogeneity of variance was evaluated with Levene’s test. Parametric tests were applied when the data met the assumptions of normality and homogeneity of variance. If normality was satisfied but homogeneity of variance was not met, a corrected unpaired t test was used for comparisons between two independent groups; otherwise, nonparametric tests were employed. Parametric tests included two-tailed paired and unpaired t tests, one-way ANOVA and repeated-measures ANOVA . The corresponding nonparametric tests included the Wilcoxon signed-rank test, Mann‒Whitney U test, Kruskal‒Wallis H test and Friedman test. Correlations between two variables were assessed by calculating Pearson’s correlation coefficient. A p value < 0.05 was considered to indicate statistical significance, with asterisks denoting the following levels of significance: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. All detailed behavior data are presented in Supplementary Table 1. Declarations Acknowledgments We thank Prof. Guoping Feng (Massachusetts Institute of Technology) for suggestions and discussions. This work was supported by grants from Brain Science and Brain-like Intelligence Technology–National Science and Technology Major Project (2021ZD0201005 to S.W. and 2025ZD0216200 to B.G.), the National Natural Science Foundation of China (82221001 to S.W., 82271577 to W.W., 32394032 to S.W., 82422030 to B.G., and 82401777 to X.H.), and Joint Founding Project of Innovation Research Institute, Xijing Hospital (LHJJ24JH05 to W.W.). We thank Springer Nature Editing Service for English language editing (certificate verification code: 2439-A7BA-9DCF-416C-75DP). Author contributions Xin Huang, Formal analysis, Investigation, Methodology, Writing—original draft; Jinwei Xu, Investigation, Methodology; Zimeng Li, Formal analysis, Investigation; Baolin Guo, Conceptualization, Funding acquisition; Honghui Mao, Formal analysis; Haiying Liu, Conceptualization, Formal analysis, Investigation, Methodology, Writing—original draft; Shengxi Wu, Conceptualization, Funding acquisition, Supervision, Writing—review and editing; Wenting Wang, Conceptualization, Funding acquisition, Supervision, Writing—review and editing. Data availability All data supporting the findings described in this manuscript are available in the article and in the supplementary information and the source data are provided with this paper. Competing interests The authors declare no competing interests. References Hodgetts, S., Magill-Evans, J. & Misiaszek, J. E. Weighted vests, stereotyped behaviors and arousal in children with autism. J Autism Dev Disord 41 , 805-814 (2011). https://doi.org/10.1007/s10803-010-1104-x O'Loghlen, J., McKenzie, M., Lang, C. & Paynter, J. Repetitive Behaviors in Autism and Obsessive-Compulsive Disorder: A Systematic Review. J Autism Dev Disord 55 , 2307-2321 (2025). https://doi.org/10.1007/s10803-024-06357-8 Kalueff, A. V., Aldridge, J. W., LaPorte, J. L., Murphy, D. L. & Tuohimaa, P. Analyzing grooming microstructure in neurobehavioral experiments. Nat Protoc 2 , 2538-2544 (2007). https://doi.org/10.1038/nprot.2007.367 Kalueff, A. V. & Tuohimaa, P. Mouse grooming microstructure is a reliable anxiety marker bidirectionally sensitive to GABAergic drugs. Eur J Pharmacol 508 , 147-153 (2005). https://doi.org/10.1016/j.ejphar.2004.11.054 Liu, H. et al. Dissection of the relationship between anxiety and stereotyped self-grooming using the Shank3B mutant autistic model, acute stress model and chronic pain model. Neurobiol Stress 15 , 100417 (2021). https://doi.org/10.1016/j.ynstr.2021.100417 Peca, J. et al. Shank3 mutant mice display autistic-like behaviours and striatal dysfunction. Nature 472 , 437-442 (2011). https://doi.org/10.1038/nature09965 Kalueff, A. V. & Tuohimaa, P. The grooming analysis algorithm discriminates between different levels of anxiety in rats: potential utility for neurobehavioural stress research. J Neurosci Methods 143 , 169-177 (2005). https://doi.org/10.1016/j.jneumeth.2004.10.001 Nin, M. S. et al. Anxiolytic effect of clonazepam in female rats: grooming microstructure and elevated plus maze tests. Eur J Pharmacol 684 , 95-101 (2012). https://doi.org/10.1016/j.ejphar.2012.03.038 Kalueff, A. V. & Tuohimaa, P. Grooming analysis algorithm for neurobehavioural stress research. Brain Res Brain Res Protoc 13 , 151-158 (2004). https://doi.org/10.1016/j.brainresprot.2004.04.002 Xu, Y. et al. Identification of a neurocircuit underlying regulation of feeding by stress-related emotional responses. Nat Commun 10 , 3446 (2019). https://doi.org/10.1038/s41467-019-11399-z Mu, M. D. et al. A limbic circuitry involved in emotional stress-induced grooming. Nat Commun 11 , 2261 (2020). https://doi.org/10.1038/s41467-020-16203-x Xie, Z. et al. A brain-to-spinal sensorimotor loop for repetitive self-grooming. Neuron 110 , 874-890 e877 (2022). https://doi.org/10.1016/j.neuron.2021.11.028 Sun, J. et al. Excitatory SST neurons in the medial paralemniscal nucleus control repetitive self-grooming and encode reward. Neuron 110 , 3356-3373 e3358 (2022). https://doi.org/10.1016/j.neuron.2022.08.010 Ahmari, S. E. et al. Repeated cortico-striatal stimulation generates persistent OCD-like behavior. Science 340 , 1234-1239 (2013). https://doi.org/10.1126/science.1234733 Burguiere, E., Monteiro, P., Feng, G. & Graybiel, A. M. Optogenetic stimulation of lateral orbitofronto-striatal pathway suppresses compulsive behaviors. Science 340 , 1243-1246 (2013). https://doi.org/10.1126/science.1232380 Tecuapetla, F., Matias, S., Dugue, G. P., Mainen, Z. F. & Costa, R. M. Balanced activity in basal ganglia projection pathways is critical for contraversive movements. Nat Commun 5 , 4315 (2014). https://doi.org/10.1038/ncomms5315 Aldridge, J. W. & Berridge, K. C. Coding of serial order by neostriatal neurons: a "natural action" approach to movement sequence. J Neurosci 18 , 2777-2787 (1998). https://doi.org/10.1523/JNEUROSCI.18-07-02777.1998 Cromwell, H. C. & Berridge, K. C. Implementation of action sequences by a neostriatal site: a lesion mapping study of grooming syntax. J Neurosci 16 , 3444-3458 (1996). https://doi.org/10.1523/JNEUROSCI.16-10-03444.1996 Corbit, L. H., Nie, H. & Janak, P. H. Habitual alcohol seeking: time course and the contribution of subregions of the dorsal striatum. Biol Psychiatry 72 , 389-395 (2012). https://doi.org/10.1016/j.biopsych.2012.02.024 Markowitz, J. E. et al. The Striatum Organizes 3D Behavior via Moment-to-Moment Action Selection. Cell 174 , 44-58 e17 (2018). https://doi.org/10.1016/j.cell.2018.04.019 Wang, W. et al. Striatopallidal dysfunction underlies repetitive behavior in Shank3-deficient model of autism. J Clin Invest 127 , 1978-1990 (2017). https://doi.org/10.1172/JCI87997 Lee, J., Wang, W. & Sabatini, B. L. Anatomically segregated basal ganglia pathways allow parallel behavioral modulation. Nat Neurosci 23 , 1388-1398 (2020). https://doi.org/10.1038/s41593-020-00712-5 Mandelbaum, G. et al. Distinct Cortical-Thalamic-Striatal Circuits through the Parafascicular Nucleus. Neuron 102 , 636-652 e637 (2019). https://doi.org/10.1016/j.neuron.2019.02.035 van den Boom, B. J. G., Pavlidi, P., Wolf, C. J. H., Mooij, A. H. & Willuhn, I. Automated classification of self-grooming in mice using open-source software. J Neurosci Methods 289 , 48-56 (2017). https://doi.org/10.1016/j.jneumeth.2017.05.026 Wang, Z. et al. Comparative analysis of functional network dynamics in high and low alcohol preference mice. Exp Neurol 389 , 115238 (2025). https://doi.org/10.1016/j.expneurol.2025.115238 Guenthner, C. J., Miyamichi, K., Yang, H. H., Heller, H. C. & Luo, L. Permanent genetic access to transiently active neurons via TRAP: targeted recombination in active populations. Neuron 78 , 773-784 (2013). https://doi.org/10.1016/j.neuron.2013.03.025 Corbit, V. L. et al. Dissociable roles of central striatum and anterior lateral motor area in initiating and sustaining naturalistic behavior. Cell Rep 44 , 115181 (2025). https://doi.org/10.1016/j.celrep.2024.115181 Ge, J. et al. Diverse Neuronal Activity in Barrel Cortex During Self-Grooming and Whisking in Mice. Neurosci Bull (2026). https://doi.org/10.1007/s12264-025-01583-1 Huang, K. et al. A hierarchical 3D-motion learning framework for animal spontaneous behavior mapping. Nat Commun 12 , 2784 (2021). https://doi.org/10.1038/s41467-021-22970-y Li, G. et al. Neural substrates for regulating self-grooming behavior in rodents. J Zhejiang Univ Sci B 25 , 841-856 (2024). https://doi.org/10.1631/jzus.B2300562 Kalueff, A. V. et al. Neurobiology of rodent self-grooming and its value for translational neuroscience. Nat Rev Neurosci 17 , 45-59 (2016). https://doi.org/10.1038/nrn.2015.8 Song, C., Berridge, K. C. & Kalueff, A. V. 'Stressing' rodent self-grooming for neuroscience research. Nat Rev Neurosci 17 , 591 (2016). https://doi.org/10.1038/nrn.2016.103 Kalueff, A. V., Lou, Y. R., Laaksi, I. & Tuohimaa, P. Abnormal behavioral organization of grooming in mice lacking the vitamin D receptor gene. J Neurogenet 19 , 1-24 (2005). https://doi.org/10.1080/01677060590949683 Kalueff, A. V. & Tuohimaa, P. Contrasting grooming phenotypes in three mouse strains markedly different in anxiety and activity (129S1, BALB/c and NMRI). Behav Brain Res 160 , 1-10 (2005). https://doi.org/10.1016/j.bbr.2004.11.010 Komorowska, J. & Pellis, S. M. Regulatory mechanisms underlying novelty-induced grooming in the laboratory rat. Behav Processes 67 , 287-293 (2004). https://doi.org/10.1016/j.beproc.2004.05.001 Pires, G. N., Tufik, S. & Andersen, M. L. Grooming analysis algorithm: use in the relationship between sleep deprivation and anxiety-like behavior. Prog Neuropsychopharmacol Biol Psychiatry 41 , 6-10 (2013). https://doi.org/10.1016/j.pnpbp.2012.11.006 Berridge, K. C., Aldridge, J. W., Houchard, K. R. & Zhuang, X. Sequential super-stereotypy of an instinctive fixed action pattern in hyper-dopaminergic mutant mice: a model of obsessive compulsive disorder and Tourette's. BMC Biol 3 , 4 (2005). https://doi.org/10.1186/1741-7007-3-4 Wang, C. C. & Shyu, B. C. Differential projections from the mediodorsal and centrolateral thalamic nuclei to the frontal cortex in rats. Brain Res 995 , 226-235 (2004). https://doi.org/10.1016/j.brainres.2003.10.006 Kita, H. & Kitai, S. T. Amygdaloid projections to the frontal cortex and the striatum in the rat. J Comp Neurol 298 , 40-49 (1990). https://doi.org/10.1002/cne.902980104 Inagaki, H. K. et al. A midbrain-thalamus-cortex circuit reorganizes cortical dynamics to initiate movement. Cell 185 , 1065-1081 e1023 (2022). https://doi.org/10.1016/j.cell.2022.02.006 Cassidy, R. M. & Gale, K. Mediodorsal thalamus plays a critical role in the development of limbic motor seizures. J Neurosci 18 , 9002-9009 (1998). https://doi.org/10.1523/JNEUROSCI.18-21-09002.1998 Devienne, G., Vantomme, G. & Huguenard, J. R. Disrupted Development of the mPFC-Thalamic Circuit in Shank3(-/-) mice, an autism-associated model. Mol Psychiatry (2025). https://doi.org/10.1038/s41380-025-03390-6 O'Neill, P. K., Gore, F. & Salzman, C. D. Basolateral amygdala circuitry in positive and negative valence. Curr Opin Neurobiol 49 , 175-183 (2018). https://doi.org/10.1016/j.conb.2018.02.012 Kim, J., Pignatelli, M., Xu, S., Itohara, S. & Tonegawa, S. Antagonistic negative and positive neurons of the basolateral amygdala. Nat Neurosci 19 , 1636-1646 (2016). https://doi.org/10.1038/nn.4414 Ferguson, E. F., Spackman, E., Cai, R. Y., Hardan, A. Y. & Uljarevic, M. Characterizing associations between emotion dysregulation, anxiety, and repetitive behaviors in autistic youth with intellectual disability. Autism Res 17 , 1810-1817 (2024). https://doi.org/10.1002/aur.3207 Cao, X. et al. Cadherin 11 regulates presynaptic vesicle trafficking and neuronal activity in autism-related brain circuit. Mol Psychiatry 31 , 192-208 (2026). https://doi.org/10.1038/s41380-025-03141-7 Mark, V. W. Acute versus chronic functional aspects of unilateral spatial neglect. Front Biosci 8 , e172-189 (2003). https://doi.org/10.2741/973 Grandi, L. C. & Gerbella, M. Single Neurons in the Insular Cortex of a Macaque Monkey Respond to Skin Brushing: Preliminary Data of the Possible Representation of Pleasant Touch. Front Behav Neurosci 10 , 90 (2016). https://doi.org/10.3389/fnbeh.2016.00090 Boyd, B. A., McBee, M., Holtzclaw, T., Baranek, G. T. & Bodfish, J. W. Relationships among Repetitive Behaviors, Sensory Features, and Executive Functions in High Functioning Autism. Res Autism Spectr Disord 3 , 959-966 (2009). https://doi.org/10.1016/j.rasd.2009.05.003 Ramirez-Armenta, K. I. et al. Optogenetic inhibition of indirect pathway neurons in the dorsomedial striatum reduces excessive grooming in Sapap3-knockout mice. Neuropsychopharmacology 47 , 477-487 (2022). https://doi.org/10.1038/s41386-021-01161-9 Givon, L., Edut, S. & Klavir, O. The role of fear and dopamine-striatal pathways in grooming. Neuropharmacology 269 , 110323 (2025). https://doi.org/10.1016/j.neuropharm.2025.110323 Mondragon-Gonzalez, S. L., Schreiweis, C. & Burguiere, E. Closed-loop recruitment of striatal interneurons prevents compulsive-like grooming behaviors. Nat Neurosci 27 , 1148-1156 (2024). https://doi.org/10.1038/s41593-024-01633-3 Merrick, C. & Filingeri, D. The evolution of wetness perception: A comparison of arachnid, insect and human models. J Therm Biol 85 , 102412 (2019). https://doi.org/10.1016/j.jtherbio.2019.102412 Oblong, J. E. Comparison of the impact of environmental stress on male and female skin. Br J Dermatol 166 Suppl 2 , 41-44 (2012). https://doi.org/10.1111/j.1365-2133.2012.10928.x Wiersielis, K. R. et al. Sex differences in corticotropin releasing factor-evoked behavior and activated networks. Psychoneuroendocrinology 73 , 204-216 (2016). https://doi.org/10.1016/j.psyneuen.2016.07.007 Jung, H. et al. Sexually dimorphic behavior, neuronal activity, and gene expression in Chd8-mutant mice. Nat Neurosci 21 , 1218-1228 (2018). https://doi.org/10.1038/s41593-018-0208-z Wilson, C. et al. Experience-dependent grooming microstructure alterations and gastrointestinal dysfunction in the SAPAP3 knockout mouse model of compulsive behaviour. J Affect Disord 363 , 520-531 (2024). https://doi.org/10.1016/j.jad.2024.07.143 Gammie, S. C. & Stevenson, S. A. Effects of daily and acute restraint stress during lactation on maternal aggression and behavior in mice. Stress 9 , 171-180 (2006). https://doi.org/10.1080/10253890600969106 Kabra, M., Robie, A. A., Rivera-Alba, M., Branson, S. & Branson, K. JAABA: interactive machine learning for automatic annotation of animal behavior. Nat Methods 10 , 64-67 (2013). https://doi.org/10.1038/nmeth.2281 Ohayon, S., Avni, O., Taylor, A. L., Perona, P. & Roian Egnor, S. E. Automated multi-day tracking of marked mice for the analysis of social behaviour. J Neurosci Methods 219 , 10-19 (2013). https://doi.org/10.1016/j.jneumeth.2013.05.013 Liu, H. et al. Comprehensive Brain-wide Mapping of Afferent and Efferent Nuclei Associated with the Heart in the Mouse. Neurosci Bull 41 , 1743-1760 (2025). https://doi.org/10.1007/s12264-025-01384-6 Yue, Y. et al. Mapping Brain-Wide Neural Activity of Murine Attentional Processing in the Five-Choice Serial Reaction Time Task. Neurosci Bull 41 , 741-758 (2025). https://doi.org/10.1007/s12264-025-01380-w Hu, Y. et al. Comparative brain-wide mapping of ketamine- and isoflurane-activated nuclei and functional networks in the mouse brain. Elife 12 (2024). https://doi.org/10.7554/eLife.88420 Shannon, P. et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res 13 , 2498-2504 (2003). https://doi.org/10.1101/gr.1239303 Kraeuter, A. K., Guest, P. C. & Sarnyai, Z. The Open Field Test for Measuring Locomotor Activity and Anxiety-Like Behavior. Methods Mol Biol 1916 , 99-103 (2019). https://doi.org/10.1007/978-1-4939-8994-2_9 Danduga, R. & Kola, P. K. Elevated Plus Maze for Assessment of Anxiety and Memory in Rodents. Methods Mol Biol 2761 , 93-96 (2024). https://doi.org/10.1007/978-1-0716-3662-6_8 Additional Declarations There is NO Competing Interest. Supplementary Files Supplymentarytable1X20260211.xlsx Supplymentary table 1 Supplymentarytable2X20260110.xlsx Supplymentary table 2 Supplymentarytable3X20251129.xlsx Supplymentary table 3 SupplementaryfiguresX20260315.docx Supplementary Figures Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9128688","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":611502588,"identity":"fff3e322-2a4e-4534-8997-34e36e60509b","order_by":0,"name":"Wenting Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAzUlEQVRIiWNgGAWjYJCCAx///JdjbwCzmYnRwMz4cGYDszHPARK0MBtzNjAn9hCtxeBG/jFpxh1s6T1ip9MkGCqsExvYzx4goCWZTbrwDE9uj3TuNgmGM+mJDTx5CXi1mN0GapnBJpG7H6SFse1wYoMEjwFhLTxsBuk8YC3/iNPCbMzblpAA0dJAhBb7+48NH844c8AQ6JfNFgnH0o3beHLwa5HsOfjgwIeKA/JAWzbe+FBjLdvPfga/FlSQAMRsJKgfBaNgFIyCUYADAACpSkSmFVmcWwAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0002-8344-0102","institution":"Fourth Military Medical University","correspondingAuthor":true,"prefix":"","firstName":"Wenting","middleName":"","lastName":"Wang","suffix":""},{"id":611502589,"identity":"f5591952-c4fa-419b-8c9f-62624b04b047","order_by":1,"name":"Xin Huang","email":"","orcid":"","institution":"Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xin","middleName":"","lastName":"Huang","suffix":""},{"id":611502590,"identity":"8f87cbea-b9cc-439f-9bbd-7489c1ace999","order_by":2,"name":"Jinwei Xu","email":"","orcid":"","institution":"Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jinwei","middleName":"","lastName":"Xu","suffix":""},{"id":611502591,"identity":"cd3b4d21-26a2-4646-bc86-fcfbf184bb96","order_by":3,"name":"Zimeng Li","email":"","orcid":"","institution":"Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Zimeng","middleName":"","lastName":"Li","suffix":""},{"id":611502592,"identity":"77c2feea-f0fb-4e6b-baba-ab747e16e904","order_by":4,"name":"Baolin Guo","email":"","orcid":"https://orcid.org/0000-0001-7309-5774","institution":"Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Baolin","middleName":"","lastName":"Guo","suffix":""},{"id":611502593,"identity":"0d3f87e2-f030-4f51-9949-ed3212f27f9f","order_by":5,"name":"Honghui Mao","email":"","orcid":"https://orcid.org/0000-0002-5625-3715","institution":"Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Honghui","middleName":"","lastName":"Mao","suffix":""},{"id":611502594,"identity":"e53ec437-d4f9-492c-9621-ca90133f6ad2","order_by":6,"name":"haiying Liu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"haiying","middleName":"","lastName":"Liu","suffix":""},{"id":611502595,"identity":"eb633032-0fa3-456c-807c-ef0deb14861a","order_by":7,"name":"Shengxi Wu","email":"","orcid":"https://orcid.org/0000-0002-3210-9567","institution":"Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Shengxi","middleName":"","lastName":"Wu","suffix":""}],"badges":[],"createdAt":"2026-03-15 12:50:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9128688/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9128688/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":105457299,"identity":"c6bb5445-43e6-44df-8f7e-b5d8a9afbff0","added_by":"auto","created_at":"2026-03-26 09:22:20","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":2390047,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAnalysis of the grooming microstructure under different conditions.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA.\u003c/strong\u003eSchematic diagrams of the JAABA interface for analyzing grooming behavior (left panel), classifier training (middle panel), and the method for obtaining the grooming microstructure based on different classifiers (right panel). \u003cstrong\u003eB.\u003c/strong\u003e Workflow for recording the grooming behaviors of \u003cem\u003eShank3B\u003c/em\u003e KO and littermate WT mice. \u003cstrong\u003eC–D.\u003c/strong\u003e KO mice showed significantly increased grooming duration, frequency, and mean duration compared with littermate WT mice (WT n=25 mice, KO n=36 mice). \u003cstrong\u003eE.\u003c/strong\u003e Comparison of the proportion of grooming duration across different stages between the KO and WT mice. \u003cstrong\u003eF.\u003c/strong\u003e Comparison of the total number of transitions and grooming transition proportions across different stagesbetween the KO and WT mice. \u003cstrong\u003eG.\u003c/strong\u003e Workflow for recording grooming behaviors in mice subjected to acute restraint stress and control mice. \u003cstrong\u003eH–I.\u003c/strong\u003eRestrained mice showed significantly increased grooming duration, frequency, and mean duration compared with control mice (n=28 control mice, n=37 restrained mice). \u003cstrong\u003eJ.\u003c/strong\u003e Comparison of the proportion of the grooming duration across different stages between the restrained and control groups. \u003cstrong\u003eK.\u003c/strong\u003e Comparison of the total number of transitions and grooming transition proportions across different stagesbetween the restrained and control groups. Statistical analyses were performed using the two-tailed unpaired t test (C, E, H, J); two-tailed unpaired separate variance estimation t test (H); and Mann‒WhitneyU test (C, E‒F, H, J‒K). The dataare presented as the means±SEMs. *p \u0026lt; 0.05, **p \u0026lt; 0.01, and ***\u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001. Detailed statistics are provided in Supplementary Table 1.\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-9128688/v1/7f11fe476db565fe156841d7.png"},{"id":105457301,"identity":"cb84dcce-f2ed-4d51-bc7e-f5b0923500c9","added_by":"auto","created_at":"2026-03-26 09:22:20","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":5506423,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMapping the DLS-related brain regions governing grooming behavior.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA.\u003c/strong\u003e \u0026nbsp;Workflow for recording grooming behaviors and c-Fos staining in \u003cem\u003eShank3B\u003c/em\u003e KO and littermate WT mice (n=5 WT mice, n=4 KO mice). \u003cstrong\u003eB.\u003c/strong\u003e Functional network diagrams in KO and WT mice with Pearson’s r \u0026gt; 0.82: nodes represent brain regions; the node size reflects the degree; the node color corresponds to the combined rank of degree, normalized betweenness centrality and eigenvector centrality; and the line thickness indicates the strength of the correlation between regions. Networks are clustered according to the results of the hierarchical clustering analysis and denoted by circle borders of different colors. \u003cstrong\u003eC.\u003c/strong\u003e Workflow for recording grooming behaviors and c-Fos staining in mice subjected to acute restraintstress and control mice (n=3 control mice, n=4 mice subjected to restrained stress). \u003cstrong\u003eD.\u003c/strong\u003e Functional network diagrams of restraint-stressed and control mice. \u003cstrong\u003eE.\u003c/strong\u003e Overlap of hub regions identified from functional network diagrams across the KO mousemodel and restraint stress mouse model. \u003cstrong\u003eF.\u003c/strong\u003e Schematic of virus (AAV2/Retro-hSyn-eGFP-3Flag-WPRE) injection into the DLS and subsequent retrograde tracing (n=4 mice). \u003cstrong\u003eG.\u003c/strong\u003e Gradient of the GFP-positive cell density in brain regions projecting to the DLS. \u003cstrong\u003eH.\u003c/strong\u003e Distribution of GFP-positive cells at different bregma points in the MD. \u003cstrong\u003eI.\u003c/strong\u003e Fiber projections observed in the DLS following AAV-hSyn-mCherry injection into the MD.\u003cstrong\u003e J.\u003c/strong\u003e Distribution of GFP-positive cells at different bregma points in the BLA. \u003cstrong\u003eK.\u003c/strong\u003e Fiber projections observed in the DLS following AAV-hSyn-mCherry injection into the BLA.\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-9128688/v1/b31ca470f011ae919fc38452.png"},{"id":105457304,"identity":"b2d85d13-2958-4f2d-af08-9f1b10649a81","added_by":"auto","created_at":"2026-03-26 09:22:21","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":5738646,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe MD and BLA play distinct roles in regulating the grooming microstructure\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA.\u003c/strong\u003eSchematic and representative image of calcium indicator virus injection into the MD (left panel); overall calcium signal responses initiated in all grooming stages(right panel). \u003cstrong\u003eB.\u003c/strong\u003e Calcium signal responses and area under the curve (AUC) for different grooming stages (n=8 mice). \u003cstrong\u003eC.\u003c/strong\u003e Schematic of the virus injection and optogenetic stimulation strategy for activating MD neurons in situ (left panel) and representative image of virus expression and increased c-Fos staining poststimulation, confirming MD activation (right panel) (n=9 control mice, n=6 ChR2 mice). \u003cstrong\u003eD–E.\u003c/strong\u003eActivation of MD neurons in the ChR2 group increased the total duration of grooming (n=9 control mice, n=11 ChR2 mice). \u003cstrong\u003eF.\u003c/strong\u003e MD activation decreased the proportion of stage 1 grooming and increased the proportion of stage 4 grooming. \u003cstrong\u003eG.\u003c/strong\u003eMD activation increased the total number of stage transitions, with the ChR2 and control groups exhibiting distinct transition patterns. \u003cstrong\u003eH.\u003c/strong\u003e Schematic and representative image of calcium indicator virus injection into the BLA (left panel) and overall calcium signal responses initiated during all grooming stages(right panel). \u003cstrong\u003eI. \u003c/strong\u003eCalcium signal responses and AUCs for different grooming stages(right panel) (n=10 mice). \u003cstrong\u003eJ.\u003c/strong\u003eSchematic of the virus injection and optogenetic stimulation strategy for activating BLA neurons in situ (left panel), and a representative image of virus expression and increased c-Fos staining poststimulation, confirming BLA activation (right panel) (n=9 control mice, n=6 ChR2mice). \u003cstrong\u003eK–L.\u003c/strong\u003eActivation of BLA neurons in the ChR2 group significantly decreased the total duration of grooming (n=9 control mice;n=8 ChR2 mice). \u003cstrong\u003eM.\u003c/strong\u003e BLA activation increased the proportion of stage 1 grooming and decreased the proportion of stage 4 grooming. \u003cstrong\u003eN.\u003c/strong\u003e BLA activation decreased the total number of stage transitions, with the ChR2 and control groups exhibiting distinct transition patterns. Statistical analyses were performed using two-tailed unpaired t test (D–G, M); two-tailed unpaired separate variance estimation t test (F, J–K); and the Mann‒WhitneyU test (C, F, M–N). The data are presented as the means±SEMs. *p \u0026lt; 0.05, **p \u0026lt; 0.01, and ***p \u0026lt; 0.001. Detailed statistics are provided in Supplementary Table 1.\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-9128688/v1/0c900660f34b59e199a3127e.png"},{"id":105457302,"identity":"2779960a-172a-43a9-add8-e7094f4b2040","added_by":"auto","created_at":"2026-03-26 09:22:20","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":6142905,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe MD→DLS and BLA→DLS circuits play antagonistic roles in regulating the grooming microstructure\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA.\u003c/strong\u003e Overall calcium signal responses initiated in all grooming stages for MD→DLS neurons. \u003cstrong\u003eB.\u003c/strong\u003eCalcium responses and AUCs of MD→DLS neurons across different grooming stages (n=6 mice). \u003cstrong\u003eC.\u003c/strong\u003e Schematic of virus injection and optogenetic stimulation strategy for activating the MD→DLS circuit (left panel), and a representative image of virus expression and increased c-Fos staining poststimulation, confirming MD→DLS activation (right panel) (n=7 control mice, n=9 ChR2 mice). \u003cstrong\u003eD-E. \u003c/strong\u003eOptogenetic activation of MD→DLS terminals significantly increased the duration of grooming in mice (n=7 control mice, n=10 ChR2 mice). \u003cstrong\u003eF.\u003c/strong\u003e Optogenetic activation of the MD→DLS circuit reduced the proportions of grooming stages1 and 5 while prolonging the proportions of stages3 and 4.\u003cstrong\u003e G.\u003c/strong\u003e Optogenetic activation of the MD→DLS circuit increased the total number of stage transitions, with distinct transition patterns between the ChR2 and control groups. \u003cstrong\u003eH.\u003c/strong\u003e Overall calcium signal responses initiated in all grooming stages for BLA→DLS neurons. \u003cstrong\u003eI.\u003c/strong\u003e Calcium responses and AUCs of BLA→DLS neurons across different grooming stages (n=5 mice). \u003cstrong\u003eJ.\u003c/strong\u003e Schematic of the virus injection and optogenetic stimulation strategy for activating the BLA→DLS circuit (left panel), and a representative image of virus expression and decreased c-Fos staining poststimulation, confirming the suppression of BLA output to the DLS (right panel) (n=6 control mice, n=8 ChR2 mice). \u003cstrong\u003eK–L.\u003c/strong\u003e Optogenetic activation of BLA→DLS terminals significantly decreased the duration of grooming in mice n=9 control mice, n=9 ChR2 mice). \u003cstrong\u003eM.\u003c/strong\u003e Optogenetic activation of the BLA→DLS circuit reduced the proportion of grooming stage 4 but increasedthat of stage 1. \u003cstrong\u003eN.\u003c/strong\u003e Optogenetic activation of the BLA→DLS circuit decreased the total number of stage transitions, with distinct transition patterns between the ChR2 and control groups. Statistical analyses were performed using two‑tailed unpaired t test (C, F, K, M); two‑tailed unpaired separate variance estimation t test (G, J, N); and the Mann–Whitney U test (D–F, M). The dataare presented as the means±SEMs. *p \u0026lt; 0.05, **p \u0026lt; 0.01, ***p \u0026lt; 0.001, and ****p \u0026lt; 0.0001. Detailed statistics are provided in Supplementary Table 1.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-9128688/v1/6c0be5d46e35c760ef6666e6.png"},{"id":105457306,"identity":"be83179c-0dda-4949-99d7-7e0a767814b8","added_by":"auto","created_at":"2026-03-26 09:22:21","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":7339408,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCharacterization of the MD→DLS and BLA→DLS circuits\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA. \u003c/strong\u003e The proportions of excitatory and inhibitory projection neurons from the MD and BLA to the DLS were quantified via retrograde tracing by injecting AAV2/Retro-mCherry into the DLS of GAD67-GFP mice (n = 4 mice). \u003cstrong\u003eB. \u003c/strong\u003eThe injection of AAV-EF1α-DIO-mRuby-T2A-Synaptophysin-EGFP into the MD of Vglut2-cre mice resulted in extensive mRuby/synaptophysin colabeling in the DLS, confirming that the MD→DLS pathway is glutamatergic. \u003cstrong\u003eC.\u003c/strong\u003e The activation of MD\u003csup\u003eVglut2\u003c/sup\u003e neurons led to a simultaneous increase in the recorded calcium signals in the DLS (n = 4 mice). \u003cstrong\u003eD.\u003c/strong\u003e Electrophysiological recordings confirmed the presence of monosynaptic excitatory projections from the MD to the DLS (n = 9 mice). \u003cstrong\u003eE.\u003c/strong\u003e Virus injection into the BLA of Vglut2-cre mice confirmed that the BLA→DLS pathway is glutamatergic. \u003cstrong\u003eF.\u003c/strong\u003e Virus injection into the BLA of VGAT-cre mice confirmed the presence of an additional GABAergic BLA→DLS projection. \u003cstrong\u003eG.\u003c/strong\u003e The activation of BLA\u003csup\u003eVglut2\u003c/sup\u003e neurons resulted in a simultaneous increase in calcium signals recorded in the DLS (n = 5 mice). \u003cstrong\u003eH.\u003c/strong\u003e The activation of BLA\u003csup\u003eVGAT\u003c/sup\u003e neurons resulted in a simultaneous decrease in calcium signals recorded in the DLS (n = 6 mice). \u003cstrong\u003eI.\u003c/strong\u003e Electrophysiological recordings confirmed the presence of monosynaptic excitatory projections from the BLA to the DLS (n = 7 mice). \u003cstrong\u003eJ.\u003c/strong\u003e Electrophysiological recordings confirmed the presence of monosynaptic inhibitory projections from the BLA to the DLS (n = 5 mice). \u003cstrong\u003eK–L.\u003c/strong\u003e Optogenetic excitation of the BLA\u003csup\u003eVGAT\u003c/sup\u003e→DLS circuit significantly reduced the duration of grooming in mice (n=8 control mice, n=10 ChR2 mice). \u003cstrong\u003eM.\u003c/strong\u003e Optogenetic activation of the BLA\u003csup\u003eVGAT\u003c/sup\u003e→DLS circuit reduced the proportion of grooming stage 4 but increased that of stage 1. \u003cstrong\u003eN.\u003c/strong\u003e Optogenetic activation of the BLA\u003csup\u003eVGAT\u003c/sup\u003e→DLS circuit decreased the total number of stage transitions, with distinct transition patterns observed between the ChR2 and control groups. Statistical analyses were performed using two-tailed unpaired t test (L–N); the Mann‒Whitney U test (M); two-tailed paired t test (C, G); and the Wilcoxon signed-rank test (D, H-J). The data are presented as the means±SEMs. *p \u0026lt; 0.05 and **p \u0026lt; 0.01. Detailed statistics are provided in Supplementary Table 1.\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-9128688/v1/1304eb14c484bab9ec99be34.png"},{"id":105566518,"identity":"88a036e1-6998-4a6e-a7da-0ab39fa5cbdd","added_by":"auto","created_at":"2026-03-27 12:56:36","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":3917994,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDivergent integration of motor and affective components in grooming by the MD→DLS and BLA→DLS circuits\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA.\u003c/strong\u003e Strategy for the optogenetic manipulation of the MD→DLS and BLA→DLS circuits (left panel), and a schematic of the 3D behavioral analysis system and data processing pipeline (right panel). \u003cstrong\u003eB–C.\u003c/strong\u003e Representative images of viral expression. \u003cstrong\u003eD.\u003c/strong\u003e Changes in the proportions of 40 behavioral modules recorded by the 3D behavioral analysis system following activation of the MD→DLS and BLA→DLS circuits (control, n=8 mice; MD→DLS, n=8 mice; BLA→DLS, n=8 mice). \u003cstrong\u003eE.\u003c/strong\u003e Optogenetic stimulation paradigm in the open field test (OFT) and elevated plus maze (EPM). \u003cstrong\u003eF.\u003c/strong\u003e Optogenetic activation of the MD→DLS circuit increased the total distance traveled in the OFT (left panel). The correlation analysis of individual mice revealed a significant positive correlation between the distance traveled in the OFT and the duration of grooming in the restricted chamber (right panel) (n=7 control mice, n=9 ChR2 mice). \u003cstrong\u003eG.\u003c/strong\u003e Optogenetic activation of the MD→DLS circuit had no effect on the center zone time in the OFT (left panel), and a significant correlation was not observed between the center zone time in the OFT and grooming duration in the grooming chamber for individual mice (right panel) (n=7 control mice, n=9 ChR2 mice). \u003cstrong\u003eH.\u003c/strong\u003e Optogenetic activation of the MD→DLS circuit had no effect on either the total distance traveled or the time spent in the open arms in the EPM (left panel); no correlation was observed between the open arm time and grooming duration (right panel) (n=7 control mice, n=9 ChR2 mice). \u003cstrong\u003eI.\u003c/strong\u003e Effect of optogenetic activation of the BLA \u003csup\u003eVGAT\u003c/sup\u003e→DLS circuit on locomotion in the OFT (left panel); no significant correlation was observed between the distance traveled and grooming duration for individual mice (right panel) (n=8 control mice, n=10 ChR2 mice). \u003cstrong\u003eJ.\u003c/strong\u003e Optogenetic activation of the BLA\u003csup\u003eVGAT\u003c/sup\u003e→DLS circuit increased the center zone time in the OFT (left panel); a negative correlation was observed between the center zone time and grooming duration for individual mice (right panel) (n=8 control mice, n=10 ChR2 mice). \u003cstrong\u003eK.\u003c/strong\u003e Optogenetic activation of the BLA\u003csup\u003eVGAT\u003c/sup\u003e→DLS circuit did not affect the total distance traveled but did increase the time spent in the open arms of the EPM (left panel); the time spent in the open arms negatively correlated with the duration of grooming (right panel) (n=8 control mice, n=10 ChR2 mice). Statistical analyses were performed using two-tailed unpaired t test (B–C, F, J) and the Mann‒Whitney U test (G–I). The data are presented as the means±SEMs. *p \u0026lt; 0.05. Detailed statistics are provided in Supplementary Table 1.\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-9128688/v1/17f8aece86b0c169212afcef.png"},{"id":105457303,"identity":"e33edeb5-2204-48b8-9ef5-177e58b212f1","added_by":"auto","created_at":"2026-03-26 09:22:20","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":3091180,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe BLA→DLS circuit but not the MD→DLS circuit regulates the grooming microstructure in mice subjected to restraint stress\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA. Simultaneous recording of calcium signals from the MD and DLS during grooming induced by restraint stress (left panel); no significant difference was observed in the overall calcium responses in the MD between the restraint and control groups (right panel) (n=4 mice in the control group, n=4 mice in the restraint group). \u003cstrong\u003eB.\u003c/strong\u003e Simultaneous recording of calcium signals from the BLA and DLS during grooming induced by restraint stress (left panel); the overall calcium response in the BLA was significantly increased in the restraint group (right panel) (n=4 mice in the control group, n=4 mice in the restraint group). \u003cstrong\u003eC.\u003c/strong\u003e Calcium responses in the MD across different grooming stages in the restraint and control groups during grooming behavior. \u003cstrong\u003eD.\u003c/strong\u003eCalcium responses in the BLA across different grooming stages in the restraint and control groups during grooming behavior. \u003cstrong\u003eE.\u003c/strong\u003e Experimental timeline for labeling stress-related neurons using targeted recombination in active populations (TRAP) and subsequent manipulation following restraint-induced grooming in mice. \u003cstrong\u003eF.\u003c/strong\u003eSchematic representation of TRAPed BLA neurons. \u003cstrong\u003eG–H.\u003c/strong\u003e The poststress activation of the TRAPed BLA→DLS circuit significantly reduced restraint-induced grooming (n=6 control mice, n=6 ChR2 mice). \u003cstrong\u003eI.\u003c/strong\u003e An analysis of the proportions of the stages showed that the activation of the TRAPed BLA→DLS circuit decreased during stages 1, 3, and 5 but increased during stage 2. \u003cstrong\u003eJ.\u003c/strong\u003eActivation of the TRAPed BLA→DLS circuit reduced the total number of grooming transitions, with distinct transition patterns observed between different groups. Statistical analyses were performed using two-tailed unpaired t test (A, C–D, G, I); two-tailed unpaired separate variance estimation t test (J); and the Mann‒Whitney U test (B, I). The data are presented as the means±SEMs. *p \u0026lt; 0.05. Detailed statistics are provided in Supplementary Table 1.\u003c/p\u003e","description":"","filename":"image7.png","url":"https://assets-eu.researchsquare.com/files/rs-9128688/v1/cc08858281d06d5816e42678.png"},{"id":105457297,"identity":"24cf02bd-467f-4794-a9bb-3e6ef6301bc9","added_by":"auto","created_at":"2026-03-26 09:22:20","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":2933057,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDecoupling of the grooming microstructure from motor and anxiety-like behaviors by the MD→DLS and BLA→DLS circuits in \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eShank3B\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e KO mice\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA.\u003c/strong\u003e Simultaneous recording of calcium signals from the MD and DLS during grooming in \u003cem\u003eShank3B\u003c/em\u003e KO mice (left panel); the overall calcium response in the MD was significantly increased in KO mice compared with WT mice (right panel) (n=4 control mice, n=4 KO mice). \u003cstrong\u003eB.\u003c/strong\u003e Simultaneous recording of calcium signals from the BLA and DLS in KO mice (left panel); the overall calcium response in the BLA was significantly decreased in KO mice (right panel) (n=4 control mice, n=5 KO mice). \u003cstrong\u003eC.\u003c/strong\u003e Calcium responses in the MD across different grooming stages in WT and KO mice during grooming behavior. \u003cstrong\u003eD.\u003c/strong\u003e Calcium responses in the BLA across different grooming stages in WT and KO mice during grooming behavior. \u003cstrong\u003eE.\u003c/strong\u003e Inhibition of the MD→DLS circuit significantly reduced the duration of grooming in the KO mice (n=8 control mice, n=9 NpHR mice). \u003cstrong\u003eF.\u003c/strong\u003e Inhibition of the MD\u003cstrong\u003e→\u003c/strong\u003eDLS circuit in KO mice increased the proportion of stage 1 grooming and decreased the proportion of stage 4 grooming. \u003cstrong\u003eG.\u003c/strong\u003e Inhibition of the MD\u003cstrong\u003e→\u003c/strong\u003eDLS circuit in KO mice decreased the total number of stage transitions, with the NpHR and control groups exhibiting distinct transition patterns. \u003cstrong\u003eH.\u003c/strong\u003e Activation of the BLA\u003csup\u003eVGAT\u003c/sup\u003e\u003cstrong\u003e→\u003c/strong\u003eDLS circuit significantly reduced the duration of grooming in KO mice (n=7 control mice, n=10 ChR2 mice). \u003cstrong\u003eI.\u003c/strong\u003e Activation of the BLA\u003csup\u003eVGAT\u003c/sup\u003e\u003cstrong\u003e→\u003c/strong\u003eDLS circuit in KO mice increased the proportion of stage 1 grooming and decreased the proportion of stage 2 and 5 grooming. \u003cstrong\u003eJ.\u003c/strong\u003e Activation of the BLA\u003csup\u003eVGAT\u003c/sup\u003e\u003cstrong\u003e→\u003c/strong\u003eDLS circuit in KO mice decreased the total number of stage transitions, with the ChR2 and control groups exhibiting distinct transition patterns. Statistical analyses were performed using two-tailed unpaired t test (A-I); two-tailed unpaired separate variance estimation t test (C, J); and the Mann‒Whitney U test (C, F, I). The data are presented as the means±SEMs. *p \u0026lt; 0.05, **p \u0026lt; 0.01, and ***p \u0026lt; 0.001. Detailed statistics are provided in Supplementary Table 1.\u003c/p\u003e","description":"","filename":"image8.png","url":"https://assets-eu.researchsquare.com/files/rs-9128688/v1/a96ed5636361876e6e5db969.png"},{"id":105570242,"identity":"4333c24c-44d8-47f9-91c8-1f2db99239a8","added_by":"auto","created_at":"2026-03-27 13:15:42","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":36977685,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9128688/v1/8f1cc617-b6f3-4d7b-8e5d-5b9d87e5ac9f.pdf"},{"id":105457296,"identity":"0acafbb2-267c-4fff-87f1-c2df18f16b94","added_by":"auto","created_at":"2026-03-26 09:22:20","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":100682,"visible":true,"origin":"","legend":"Supplymentary table 1","description":"","filename":"Supplymentarytable1X20260211.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9128688/v1/62cc961a3d5e6e4f2d294a02.xlsx"},{"id":105566303,"identity":"84c531af-d28f-4cfc-b094-d2bf01ae5fec","added_by":"auto","created_at":"2026-03-27 12:56:05","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":38734,"visible":true,"origin":"","legend":"Supplymentary table 2","description":"","filename":"Supplymentarytable2X20260110.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9128688/v1/cd8132e5ffee1ecab1f14040.xlsx"},{"id":105457295,"identity":"278150b8-30ad-485b-ab68-37fe217639fc","added_by":"auto","created_at":"2026-03-26 09:22:20","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":18528,"visible":true,"origin":"","legend":"Supplymentary table 3","description":"","filename":"Supplymentarytable3X20251129.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9128688/v1/750f54f185b85b01e734a6ef.xlsx"},{"id":105457307,"identity":"eb4912ac-3287-4ce9-ad57-1acdca868271","added_by":"auto","created_at":"2026-03-26 09:22:21","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":17095464,"visible":true,"origin":"","legend":"Supplementary Figures","description":"","filename":"SupplementaryfiguresX20260315.docx","url":"https://assets-eu.researchsquare.com/files/rs-9128688/v1/fe7a5ce9240d8dc8c92b38cd.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Opposing thalamic and amygdalar projections to the dorsolateral striatum sculpt the grooming microstructure in stress and autism models","fulltext":[{"header":"Introduction","content":"\u003cp\u003eStereotyped behaviors, characterized by repetitive, rhythmic actions involving complex routines and rituals, are observed both as uncommon patterns in healthy individuals and as core features of disorders such as autism spectrum disorder (ASD)\u003csup\u003e1\u003c/sup\u003e and obsessive-compulsive disorder (OCD)\u003csup\u003e2\u003c/sup\u003e. However, their pathophysiology remains incompletely understood. Self-grooming behavior in rodents offers a powerful and translatable model for investigating these behaviors, as it is a highly conserved, innate action that follows a rigid cephalocaudal sequence\u003csup\u003e3,4\u003c/sup\u003e. Importantly, both the quantity and the sequential microstructure of grooming are disrupted in models of ASD\u003csup\u003e5,6\u003c/sup\u003e and stress/anxiety\u003csup\u003e7,8\u003c/sup\u003e, mirroring clinical phenotypes. Thus, the grooming microstructure and the detailed organization and progression of its component stages collectively serve as a sensitive ethological measure for disentangling distinct neural mechanisms that may drive repetitive behaviors across different pathological contexts\u003csup\u003e9\u003c/sup\u003e. Recent studies employing an analysis of the grooming microstructure have begun to differentiate these mechanisms\u003csup\u003e10,11\u003c/sup\u003e. Our prior work demonstrated that \u003cem\u003eShank3B\u003c/em\u003e KO mice (a well-established ASD model) and acutely stressed mice both exhibit hypergrooming\u003csup\u003e5\u003c/sup\u003e but with fundamentally distinct microstructural patterns. Notably, the anxiolytic diazepam normalized the disrupted sequence in the anxiety model but only partially normalized it in the ASD model, suggesting that different neural pathways underlie distinct types of pathological grooming\u003csup\u003e5\u003c/sup\u003e. However, the specific circuit mechanisms regulating these distinct microstructural programs have not been fully elucidated.\u003c/p\u003e\n\u003cp\u003eWhile numerous brain regions, including the limbic and hypothalamic areas\u003csup\u003e11-13\u003c/sup\u003e, regulate overall self-grooming behavior, researchers have increasingly focused on the specific brain circuits that generate and regulate these microstructural patterns. Corticobasal ganglia circuits\u003csup\u003e14,15\u003c/sup\u003e, particularly the dorsolateral striatum (DLS)\u003csup\u003e16,17\u003c/sup\u003e, have emerged as critical hubs for executing such sequential motor actions. Evidence has shown that lesions of the DLS disrupt the stereotyped cephalocaudal grooming chain\u003csup\u003e18\u003c/sup\u003e, confirming its indispensable role in organizing the grooming microstructure, which is consistent with its broader functions in action sequencing and habit formation\u003csup\u003e19,20\u003c/sup\u003e. Our previous work further established that dysfunction within the indirect pathway of the DLS contributes to the excessive, stereotyped grooming behavior observed in \u003cem\u003eShank3B\u003c/em\u003e KO mice\u003csup\u003e21\u003c/sup\u003e. Although the DLS is known to integrate convergent inputs from thalamic, cortical, and limbic regions\u003csup\u003e22,23\u003c/sup\u003e, the specific upstream circuits that shape distinct microstructural patterns and how these circuits become dysregulated in different disease states remain unclear.\u003c/p\u003e\n\u003cp\u003eThus, this study aimed to elucidate the DLS circuit mechanisms regulating the grooming microstructure and to define their distinct contributions to ASD-related stereotypy versus anxiety-driven repetitive grooming. We combined an automated grooming microstructure analysis using Janelia Automatic Animal Behavior Annotator (JAABA) software\u003csup\u003e24\u003c/sup\u003e, whole-brain input mapping, in vivo neuronal activity recordings, and optogenetic and chemogenetic manipulations. We investigated two principal upstream pathways to the DLS: the mediodorsal thalamus (MD)\u0026rarr;DLS pathway, which promotes the maintenance phase of grooming behavior, and the basolateral amygdala (BLA)\u0026rarr;DLS pathway, which shapes the initiation and termination of grooming behavior. These circuits embed distinct behavioral states, such as general locomotion and anxiety, into the grooming microstructure, explaining the divergent patterns of repetitive grooming in ASD-related stereotypy versus anxiety-driven pathology. This work provides a circuit-level framework for understanding how specific behavioral sequences are constructed and dysregulated in individuals with neuropsychiatric disorders.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003e\u003cem\u003eShank3\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;KO and stress models display distinct grooming microstructure\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003epatterns\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe first used the JAABA GentleBoost algorithm\u003csup\u003e24\u003c/sup\u003e to train classifiers for segmenting grooming stages (Fig. 1A). Automated scoring strongly correlated with manual scoring (R\u003csup\u003e2\u003c/sup\u003e=0.875) and differed by less than 20% (Supplementary Fig. 1A\u0026ndash;C). In 128 wild-type (WT) mice, the stage distributions were as follows: stage 1 (paw licking) 23.9%, stage 2 (head/face washing) 12.5%, stage 3 (body grooming) 43.1%, stage 4 (leg licking) 13%, and stage 5 (tail/genital grooming) 7.4% (Supplementary Fig. 1D). Two major microcycles emerged: 1\u0026ndash;2/2\u0026ndash;1 (initiation) and 3\u0026ndash;4/4\u0026ndash;3 (maintenance). Transitions such as 1\u0026ndash;3 and 2\u0026ndash;3 linked these phases, whereas 4\u0026ndash;5 transitions marked termination (Supplementary Fig. 1D).\u003c/p\u003e\n\u003cp\u003eWe next applied this approach to a \u003cem\u003eShank3B\u003c/em\u003e KO mouse ASD model (Fig. 1B\u0026ndash;D) and an acute restraint stress mouse model (Fig. 1G\u0026ndash;I). Both groups presented increased total grooming duration, bout number, and mean bout length. Overlapping confidence intervals (Supplementary Fig. 1E) indicated that gross metrics alone could not distinguish these models. However, the grooming microstructure analysis revealed distinct profiles: KO mice exhibited a prolonged duration of stages 2\u0026ndash;4 (Supplementary Fig. 1F), with a reduced percentage of stage 1 and increased percentage of stages 3\u0026ndash;4 (Fig. 1E), consistent with perseveration of the maintenance sequence. In contrast, the stressed mice spent a prolonged time in stages 3\u0026ndash;5, with a reduced percentage of stages 2 and 5 and an isolated increase in the percentage of stage 3 (Supplementary Fig. 1G and Fig. 1J), suggesting fragmented, anxiety-driven grooming with disrupted initiation and termination. These patterns are consistent with our previous manual observations and published reports\u003csup\u003e4,5,9\u003c/sup\u003e. The transition analysis further showed that the transitions of the KO mice were\u0026nbsp;concentrated\u0026nbsp;around stages 3\u0026ndash;4, whereas the stressed mice displayed more transitions involving stage 5 and a globally disorganized transition architecture (Fig. 1F and Fig. 1K). In summary, although both models exhibit hypergrooming, KO mice exhibit stereotyped perseveration of the maintenance phase (stage 3\u0026ndash;4), whereas stressed mice display a chaotic pattern with frequent shifts to the initiation and termination phases (stage 2 and 5).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDLS regulation of the grooming microstructure may\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003ebe linked\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;to several upstream hub regions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe systematically elucidated the broader network of the brain regions involved in grooming behavior by performing whole-brain c-Fos mapping. In \u003cem\u003eShank3B\u003c/em\u003e KO mice, c-Fos expression was significantly increased in 33 regions 1.5 hours after grooming onset compared with that in WT mice (Fig. 2A and Supplementary Fig. 2A). Graph theory analysis\u003csup\u003e25\u003c/sup\u003e of these differentially activated regions (Supplementary Table 2) revealed several hub regions with high connectivity, including the magnocellular reticular nucleus (MARN), anterior cingulate area (ACA), and somatomotor area (MO) (Fig. 2B). Similarly, compared with control conditions, acute restraint stress induced grooming-associated activation of 34 brain regions (Fig. 2C and Supplementary Fig. 2B), with hubs including the septofimbrial nucleus (SF), perirhinal area (PERI), agranular insular area (AI), BLA, and MD (Fig. 2D). A shared network across both conditions emerged, involving regions such as the MO, MD, BLA, somatosensory areas (SS), auditory areas (AUD), visual areas (VIS), nucleus accumbens (ACB), temporal association areas (TEa), ectorhinal area (ECT), PERI, cortical amygdalar area (COA), and medial amygdalar nucleus (MEA) (Fig. 2E and Supplementary Fig. 2A\u0026ndash;D). Notably, and consistent with a previous report\u003csup\u003e26\u003c/sup\u003e, c-Fos activation in the striatum (including the DLS) itself was not very pronounced, likely because of the low sensitivity of c-Fos expression in this region.\u003c/p\u003e\n\u003cp\u003eAlthough c-Fos activity in the DLS during grooming appears to be limited, the pivotal role of the DLS in governing the grooming microstructure is well documented\u003csup\u003e16,18,20\u003c/sup\u003e. Therefore, we employed optogenetics to directly assess its involvement in modulating the grooming microstructure. Bilateral injection of AAV-hSyn-ChR2 and sustained photostimulation of the DLS in WT mice did not alter the overall grooming behavior (Supplementary Fig. 2E\u0026ndash;F). However, it specifically reduced the duration of grooming in stages 1 and 2 and decreased the proportional time spent grooming in stage 2 (Supplementary Fig. 1G\u0026ndash;I); this pattern was distinct from that observed in both the \u003cem\u003eShank3\u003c/em\u003e KO and restraint-stress models. These findings indicate that the DLS regulates the grooming microstructure via complex, circuit-specific mechanisms rather than through uniform activation. We subsequently mapped its upstream inputs using retrograde tracing (Fig. 2F). The densest inputs originated from the cortex, thalamus, and amygdala (Supplementary Fig. 3 and Supplementary Fig. 4A\u0026ndash;B). Intersecting this input map with the c-Fos-activated network (Fig. 2E) pinpointed the primary sensory cortex (S1), secondary motor cortex (M2), MD, and BLA as key DLS-projecting regions that are functionally engaged during grooming behavior (Fig. 2G and Supplementary Fig. 2C\u0026ndash;D). Detailed mapping of retrogradely labeled neurons revealed distinct spatial distributions: MD-projecting neurons clustered rostrally, whereas BLA-projecting neurons exhibited a slight rostrocaudal gradient (Fig. 2H, J). In contrast, M2 neurons were diffusely distributed, and S1 neurons clustered caudally (Supplementary Fig. 4C\u0026ndash;D). Subsequent anterograde viral tracing from these precise locations confirmed robust terminal labeling in the DLS, establishing these regions as core hubs in the regulation of grooming (Fig. 2I, K and Supplementary Fig. 4E\u0026ndash;F).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eThe\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eMD and BLA, two regions upstream of the DLS, are involved in regulating the grooming microstructure\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe investigated whether the identified DLS-projecting regions regulate the grooming microstructure. Fiber photometry revealed increased activity in the MD of WT mice during the grooming maintenance stage (stages 3\u0026ndash;4) (Fig. 3A\u0026ndash;B). Optogenetic activation confirmed that sustained MD activation increased the total duration of grooming (Fig. 3C\u0026ndash;E and Supplementary Fig. 5A), driven by an increased duration in stages 2\u0026ndash;4 (Supplementary Fig. 5B) and a proportional shift toward stage 4 (Fig. 3F). The transition analysis showed increased maintenance phase transitions (3\u0026ndash;4/4\u0026ndash;3 and 2\u0026ndash;4/4\u0026ndash;2; Fig. 3G). In contrast, calcium signals in the BLA increased selectively during the termination stage (stage 5; Fig. 3H\u0026ndash;I). Further BLA activation suppressed grooming behavior (Fig. 3J\u0026ndash;L and Supplementary Fig. 5C), reducing the time spent in stages 2\u0026ndash;4 (Supplementary Fig. 5D) while proportionally increasing the time spent in stage 1 and decreasing the time spent in stage 4 (Fig. 3M). It also reduced transitions within the maintenance module (3\u0026ndash;4/4\u0026ndash;3) but increased transitions linking initiation to maintenance (Fig. 3N), creating a disorganized profile. Together, these results indicate that the MD promotes the maintenance and progression of the core sequence, whereas the BLA constrains maintenance and facilitates transitions into the initiation and termination phases.\u003c/p\u003e\n\u003cp\u003eFurther dynamic manipulation revealed that phasic MD activation (3 s after grooming onset, lasting 15 s) during an ongoing bout did not prolong grooming (Supplementary Fig. 6A\u0026ndash;C), indicating that sustained activity is required for its maintenance effect. In contrast, even acute BLA activation upon grooming onset (3 s after grooming onset, lasting 5 s) robustly suppressed behavior and promoted termination, with grooming resuming after light offset (Supplementary Fig. 6D\u0026ndash;J). These results further confirmed that the MD and BLA exert opposing, phase-specific control over the grooming microstructure.\u003c/p\u003e\n\u003cp\u003eCalcium signals in M2 (Supplementary Fig. 7A) and S1 (Supplementary Fig. 7G) also increased during stages 3\u0026ndash;4, which aligns with the observed increase in c-Fos expression. However, direct optogenetic activation of M2 (Supplementary Fig. 7B\u0026ndash;F and Supplementary Fig. 8A) or S1 (Supplementary Fig. 7H\u0026ndash;L and Supplementary Fig. 8B) did not alter the total grooming duration, stage distribution, or transition architecture. These findings suggest that M2 and S1 may be involved in overall motor execution\u003csup\u003e27\u003c/sup\u003e and somatosensory perception\u003csup\u003e28\u003c/sup\u003e during grooming but do not participate in the specific regulation of the grooming action sequence.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMD\u0026rarr;DLS and BLA\u0026rarr;DLS circuits play antagonistic roles in regulating\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003egrooming behavior\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe next investigated whether the MD and BLA regulate grooming through the DLS. We injected a retrograde Cre virus (AAV/Retro-cre-mCherry) into the DLS and a Cre-dependent GCaMP7s (rAAV-CAG-FLEX-jGCaMP7s) into the MD to enable the specific expression of GCaMP in MD neurons projecting to the DLS. Afterward, we recorded the calcium activity of these projecting neurons. Their calcium signals increased specifically during stages 3\u0026ndash;4 (Fig. 4A\u0026ndash;B). Optogenetic activation of this pathway significantly increased the total duration of grooming in WT mice (Fig. 4C\u0026ndash;E and Supplementary Fig. 9A), primarily by prolonging stages 2\u0026ndash;4 (Supplementary Fig. 9C). An analysis of the proportions of the stages showed decreased proportions of stages 1 and 5 but increased proportions of stages 3 and 4 (Fig. 4F), indicating shortened initiation, prolonged maintenance, and accelerated termination. The transition analysis revealed enhanced transitions within the maintenance module (2\u0026ndash;3, 3\u0026ndash;2, 3\u0026ndash;4, and 4\u0026ndash;3) and suppressed initiation-linked transitions (1\u0026ndash;2, 2\u0026ndash;1, 1\u0026ndash;3, and 3\u0026ndash;1) (Fig. 4G), further confirming that MD\u0026rarr;DLS activation specifically consolidates the core grooming sequence. Bidirectional chemogenetic manipulation confirmed that this pathway is both necessary and sufficient for sustaining the maintenance phase: inhibition shortened stages 3\u0026ndash;4 and increased stage 1, whereas activation produced the opposite effects (Supplementary Fig. 10A\u0026ndash;N).\u003c/p\u003e\n\u003cp\u003eIn contrast, BLA neurons projecting to the DLS (BLA\u0026rarr;DLS) showed elevated calcium activity during stages 2\u0026ndash;3 (Fig. 4H\u0026ndash;I), in contrast to the stage 5-dominant activity observed in the general BLA population. Optogenetic activation of BLA terminals in the DLS (Fig. 4J\u0026ndash;L and Supplementary Fig. 9B) suppressed grooming behavior, primarily by shortening stages 2\u0026ndash;4 (Supplementary Fig. 9D). An analysis of the proportions of the stages revealed a decreased proportion of stage 4 and an increased proportion of stage 1 (Fig. 4M), indicating premature termination and a shift in behavior toward the initiation phase. Accordingly, transitions within the maintenance module (3\u0026ndash;4 and 4\u0026ndash;3) were reduced, whereas transitions linking initiation to maintenance (1\u0026ndash;3 and 3\u0026ndash;1) and within the initiation module (1\u0026ndash;2 and 2\u0026ndash;1) were increased (Fig. 4N), reflecting a fragmented and less consolidated grooming sequence. Chemogenetic inhibition of the BLA\u0026rarr;DLS pathway increased total grooming and prolonged stages 3\u0026ndash;4, consistent with a release from tonic suppression. Conversely, chemogenetic activation suppressed grooming, mirroring the optogenetic results (Supplementary Fig. 11A\u0026ndash;M). Together, these results demonstrate that the BLA\u0026rarr;DLS pathway exerts inhibitory control over grooming by constraining the maintenance phase (stages 3\u0026ndash;4) and promoting transitions into the initiation and termination phases.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMD\u0026rarr;DLS and BLA\u0026rarr;DLS pathways modulate the grooming microstructure via antagonistic synaptic transmission\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe antagonistic effects of the MD\u0026rarr;DLS and BLA\u0026rarr;DLS pathways on grooming behavior prompted us to examine their underlying projection properties. Retrograde tracing (AAV2/Retro-DIO-mCherry) into the DLS of CamkⅡ-cre mice and VGAT-cre mice revealed that the MD\u0026rarr;DLS pathway lacks inhibitory projections (Supplementary Fig. 12A\u0026ndash;B). In GAD67-GFP mice, AAV2/Retro-mCherry tracing showed that GABAergic neurons in the BLA\u0026rarr;DLS pathway account for ~15% of all BLA\u0026rarr;DLS neurons and ~14% of inhibitory neurons in the BLA (Fig. 5A). Further anterograde tracing confirmed that the MD\u0026rarr;DLS pathway is purely glutamatergic (Fig. 5B). In vivo, optogenetic MD activation induced a transient increase in calcium levels in the DLS of both WT and Vglut2-cre mice (Fig. 5C, Supplementary Fig. 12C\u0026ndash;D). Ex vivo optogenetic patch clamp results confirmed the presence of monosynaptic (in the presence of TTX and 4-AP), glutamatergic MD\u0026rarr;DLS connectivity, as blue-light stimulation evoked NBQX-sensitive EPSCs (Fig. 5D).\u003c/p\u003e\n\u003cp\u003eAnterograde tracing of the BLA\u0026rarr;DLS pathway in Vglut2-cre and VGAT-cre mice confirmed that this pathway\u0026nbsp;consists of\u0026nbsp;both glutamatergic and GABAergic projections (Fig. 5E\u0026ndash;F). BLA activation in WT mice evoked a biphasic DLS calcium response (an increase followed by a decrease) (Supplementary Fig. 12D\u0026ndash;E). In Vglut2-cre mice, BLA activation caused a monophasic increase in the calcium response, whereas in VGAT-cre mice, BLA activation caused a decrease in the calcium response, which is consistent with the excitatory\u0026ndash;inhibitory composition of this pathway (Fig. 5G\u0026ndash;H). Ex vivo stimulation of BLA\u0026rarr;DLS terminals evoked NBQX-sensitive EPSCs and bicuculline-sensitive IPSCs under different conditions, confirming direct glutamatergic and GABAergic projections, respectively (Fig. 5I\u0026ndash;J). Together, these anatomical and functional results delineate a purely excitatory MD\u0026rarr;DLS pathway and a dual-component BLA\u0026rarr;DLS pathway that can exert opposing effects on DLS activity.\u003c/p\u003e\n\u003cp\u003eWe next asked how distinct neuronal populations within the BLA\u0026rarr;DLS pathway regulate grooming. Only the GABAergic component of the BLA\u0026rarr;DLS pathway suppressed grooming. Selective activation of GABAergic BLA\u003csup\u003eVGAT\u003c/sup\u003e\u0026rarr;DLS projections significantly reduced the total grooming time, shortened stages 2\u0026ndash;4 (Fig. 5K\u0026ndash;L and Supplementary Fig. 12F), decreased the proportion of stage 4, and fragmented the sequence by reducing maintenance-phase transitions (e.g., 3\u0026ndash;4/4\u0026ndash;3) and increasing initiation-linked transitions (e.g., 1\u0026ndash;3, 3\u0026ndash;1, 2\u0026ndash;3, and 3\u0026ndash;2) (Fig. 5M\u0026ndash;N), reflecting a fragmented sequence with impaired sustainment of the core grooming phase. In contrast, the selective activation of glutamatergic BLA\u003csup\u003eVglut2\u003c/sup\u003e\u0026rarr;DLS projections had a minimal effect on grooming (Supplementary Fig. 12G\u0026ndash;K). Similarly, in situ activation of BLA\u003csup\u003eVglut2\u003c/sup\u003e neurons did not alter grooming, whereas activating BLA\u003csup\u003eVGAT\u0026nbsp;\u003c/sup\u003eneurons in situ reproduced the suppressive phenotype observed with pathway-specific stimulation (Supplementary Fig. 13A\u0026ndash;H). Notably, in situ BLA\u003csup\u003eVGAT\u003c/sup\u003e activation increased the proportions of both stage 1 and stage 3 but decreased the proportion of stage 4, further underscoring its role in disrupting the maintenance\u0026ndash;termination transition and promoting erratic, initiation-biased grooming. Together, these results establish that the grooming-suppressive function of the BLA\u0026rarr;DLS pathway is specifically mediated by its GABAergic component, which constrains the maintenance phase and fragments the behavioral sequence.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eThe MD\u0026rarr;DLS and BLA\u0026rarr;DLS pathways separately integrate the motor and affective components of grooming\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePrevious studies suggest that grooming comprises multiple components, including pure motor execution, compulsive sequencing, emotional relief, and somatic sensation\u003csup\u003e5,8,28\u003c/sup\u003e. We further elucidated the potential regulatory roles of the MD\u0026rarr;DLS and BLA\u0026rarr;DLS circuits by analyzing the full behavioral repertoire of WT mice following the activation of these circuits using a high-resolution 3D pose-tracking system\u003csup\u003e29\u003c/sup\u003e. The behavioral sequences were clustered into 40 discrete modules (Fig. 6A\u0026ndash;C and Supplementary Table 3). MD\u0026rarr;DLS activation specifically increased the activity of modules associated with sustained, crouched grooming postures (modules 28 and 27) and grooming\u0026ndash;maintenance movements (module 12) while reducing the activity of modules related to stationary inactivity, cessation-like behaviors (modules 37 and 9), and transitional actions (module 10: beginning to rear up) (Fig. 6D and Supplementary Fig. 14A). Conversely, BLA\u0026rarr;DLS activation broadly suppressed exploration and orienting modules, including rearing (modules 30, 4, and 29), directed locomotion and turning (modules 25, 26, 36, 18, and 40), and sniffing/head investigation (modules 20, 19, and 14), while it also decreased stationary postures (module 37) and increased freezing-like pauses (module 2) (Fig. 6D and Supplementary Fig. 14B\u0026ndash;C). These profiles indicate that MD\u0026rarr;DLS activation promotes focused, sustained grooming postures and maintenance-related movements while reducing nongrooming static or transitional behaviors. In contrast, BLA\u0026rarr;DLS activation increases behavioral arrest and suppresses locomotion, turning, and exploratory activity.\u003c/p\u003e\n\u003cp\u003eNext, we assessed whether manipulations of these circuits influence general locomotion and anxiety-like states using an open field test (OFT) and an elevated plus maze (EPM). Activating the MD\u0026rarr;DLS pathway significantly increased the total distance traveled in the OFT, which was positively correlated with the duration of grooming in the restricted chamber (Fig. 6E\u0026ndash;F), indicating a coordinated enhancement of motor activity and grooming. However, it did not affect the time spent in the center of the open field or open arms of the EPM and was not correlated with anxiety-related behaviors (Fig. 6G\u0026ndash;H), suggesting that this circuit modulates motor output without directly affecting anxiety-like behaviors. In contrast, the selective activation of the GABAergic BLA\u003csup\u003eVGAT\u003c/sup\u003e\u0026rarr;DLS pathway did not alter locomotion but did significantly increase the time spent in the center zone of the open field and open arms of the EPM (Fig. 6I\u0026ndash;K), and the reduction in anxiety was correlated negatively with the grooming duration, indicating concurrent grooming suppression and anxiety mitigation. Moreover, the activation of the glutamatergic BLA\u003csup\u003eVglut2\u003c/sup\u003e\u0026rarr;DLS component reduced both locomotion and center/open-arm time without affecting grooming (Supplementary Fig. 14D\u0026ndash;F), indicating that its role in anxiety is distinct from that of grooming control. Together, these results delineate a double dissociation: the MD\u0026rarr;DLS pathway promotes grooming and locomotion without altering anxiety, whereas the GABAergic BLA\u003csup\u003eVGAT\u003c/sup\u003e\u0026rarr;DLS pathway concurrently suppresses grooming and reduces anxiety-like behaviors, highlighting the circuit-specific integration of motor and affective components.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eThe BLA\u0026rarr;DLS circuit but not\u0026thinsp;the MD\u0026rarr;DLS circuit regulates the grooming microstructure in animals subjected to restraint stress\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe next examined the roles of the MD\u0026rarr;DLS and BLA\u0026rarr;DLS circuits in regulating grooming in mice under acute restraint stress. Calcium imaging revealed that the MD response remained unchanged except for elevated activity in stage 2, whereas BLA activity decreased in stages 1, 2, and 5 but increased during stages 3 and 4 (Fig. 7A\u0026ndash;D). This pattern, with reduced engagement during initiation and termination but increased activity during maintenance, suggests that under acute stress, the BLA shifts from its typical role in structuring behavioral boundaries (initiation/termination) toward a more persistent, maintenance-associated signaling mode, which may contribute to the fragmented grooming sequence observed in stressed mice. Simultaneous calcium imaging in the DLS revealed no change (Supplementary Fig. 15A\u0026ndash;B), indicating that the altered drive in stress-induced grooming originates upstream, likely from the BLA. Optogenetic inhibition of the MD\u0026rarr;DLS circuit did not alter stress-induced grooming (Supplementary Fig. 15C\u0026ndash;E, H\u0026ndash;J). In contrast, activating the GABAergic BLA\u003csup\u003eVGAT\u003c/sup\u003e\u0026rarr;DLS pathway rapidly suppressed this process, shortening stages 2\u0026ndash;4, increasing stage 1, and fragmenting transitions by reducing the maintenance phase (3\u0026ndash;4/4\u0026ndash;3) and increasing the initiation-to-maintenance (1\u0026ndash;3/3\u0026ndash;1) shifts (Supplementary Fig. 15F\u0026ndash;G, K‒M), reflecting fragmented progression characterized by short maintenance and inefficient sequence consolidation.\u003c/p\u003e\n\u003cp\u003eNext, we used targeted recombination in active populations (TRAP) in Fos-cre\u003csup\u003eER\u003c/sup\u003e mice to express the ChR2 protein in stress-activated neurons in the MD and BLA\u003csup\u003e26\u003c/sup\u003e\u003csup\u003e\u0026nbsp;\u003c/sup\u003e(Fig. 7E). Following the restraint-stress protocol, 44.12% of TRAPed MD neurons were glutamatergic and grooming related (Supplementary Fig. 16A, C), whereas 29.44% of Trapped BLA cells were grooming related, with significant GABAergic enrichment (Supplementary Fig. 16B, D). Activating TRAPed BLA\u0026rarr;DLS neurons suppressed stress-induced grooming, partially restoring the stage balance (increasing the characteristic deficit in stage 2) and restabilizing maintenance-phase transitions (e.g., increasing 2\u0026ndash;3/3\u0026ndash;2 transitions) (Fig. 7F\u0026ndash;J and Supplementary Fig. 16E). In contrast, inhibiting TRAPed MD\u0026rarr;DLS neurons had no significant effect (Supplementary Fig. 16F\u0026ndash;J). Together, these findings highlight the critical role of the BLA\u0026rarr;DLS circuit, particularly its GABAergic component, in regulating grooming under stress, whereas the MD\u0026rarr;DLS circuit primarily supports motor execution without restructuring anxiety-driven behaviors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDecoupling of the grooming microstructure from motor and anxiety-like behaviors by the\u0026thinsp;MD\u0026rarr;DLS and BLA\u0026rarr;DLS circuits in \u003cem\u003eShank3B\u003c/em\u003e KO mice\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn \u003cem\u003eShank3B\u003c/em\u003e KO mice, we investigated the roles of the MD\u0026rarr;DLS and BLA\u0026rarr;DLS circuits in stereotyped grooming. Calcium imaging revealed a stronger response in the MD and a weaker response in the BLA of the\u003cem\u003e\u0026nbsp;Shank3B\u003c/em\u003e KO mice than in the WT mice (Fig. 8A\u0026ndash;B). The stage-resolved analysis showed that MD activity was increased in stages 1, 3, and 4 in KO mice, whereas BLA activity increased in stages 1, 2, and 5 but decreased in stages 3 and 4 (Fig. 8C\u0026ndash;D). This pattern is consistent with the microstructural phenotype of the KO mice (Fig. 1E), suggesting that increased MD activity during the maintenance phases (stage 3\u0026ndash;4) likely drives the prolonged core grooming sequence, whereas altered BLA engagement during initiation (stage 1) and termination (stage 5) may reflect disrupted anxiety-related modulation. Simultaneous calcium recordings in the DLS of the KO mice showed responses similar to those of the WT mice but revealed stage-specific shifts, with lower signals in stages 1 and 3 and a higher signal in stage 4 (Supplementary Fig. 17A\u0026ndash;B), indicating a redistribution toward the maintenance phase.\u003c/p\u003e\n\u003cp\u003eOptogenetic inhibition of the MD\u0026rarr;DLS circuit reduced the total grooming time in the KO mice (Fig. 8E) by shortening stages 3 and 4 (Supplementary Fig. 17C), increasing stage 1 and decreasing stage 4 (Fig. 8F), indicating interrupted progression from initiation to sustained maintenance. The transition analysis revealed decreases in maintenance-phase transitions (3\u0026ndash;4 and 4\u0026ndash;3) and in shifts linking initiation to maintenance (1\u0026ndash;3 and 3\u0026ndash;1), as well as increases in initiation-phase transitions (1\u0026ndash;2 and 2\u0026ndash;1) and early maintenance transitions (2\u0026ndash;3 and 3\u0026ndash;2) (Fig. 8G). These shifts reflect a \u0026ldquo;forward-biased\u0026rdquo; sequence with more cephalic grooming and less sustained body grooming, ultimately leading to an overall reduction. Similarly, optogenetic activation of the GABAergic BLA\u003csup\u003eVGAT\u003c/sup\u003e\u0026rarr;DLS pathway also reduced the total grooming time in KO mice, increasing stage 1, decreasing stages 2 and 5, and fragmenting maintenance-phase transitions while increasing erratic initiation-to-maintenance transitions (e.g., 1\u0026ndash;3/3\u0026ndash;1) (Fig. 8H\u0026ndash;J, Supplementary Fig. 17D). Activating the general BLA\u0026rarr;DLS pathway resulted in a similar disruption (Supplementary Fig. 17E, G\u0026ndash;I), whereas activating its glutamatergic component had no effect, confirming that GABAergic projections mediated this corrective effect (Supplementary Fig. 17F, J\u0026ndash;L).\u003c/p\u003e\n\u003cp\u003eInhibiting the MD\u0026rarr;DLS circuit in the KO mice reduced general locomotion, which correlated with decreased grooming (Supplementary Fig. 18A\u0026ndash;B), but did not affect anxiety (Supplementary Fig. 18C\u0026ndash;E). Conversely, activating the BLA\u003csup\u003eVGAT\u003c/sup\u003e\u0026rarr;DLS pathway reduced anxiety-like behaviors (increased center/open-arm times) and was negatively correlated with grooming but did not alter locomotion (Supplementary Fig. 18F\u0026ndash;J). These findings indicate that in this autism model, the MD\u0026rarr;DLS pathway coordinates grooming execution and motor activity, whereas the GABAergic BLA\u0026rarr;DLS pathway links grooming modulation to the emotional state. Both pathways may serve as potential targets for addressing pathological stereotypy.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eSelf-grooming is a crucial adaptive behavior for maintaining physiological homeostasis and coping with environmental stress, and its dysregulation is closely linked to anxiety and stereotypy\u003csup\u003e30,31\u003c/sup\u003e. Studying its microstructure is therefore essential for distinguishing distinct pathological states. Previous studies\u003csup\u003e16-18\u003c/sup\u003e, including our own\u003csup\u003e21\u003c/sup\u003e, have established the DLS as a key node regulating the stereotyped sequential syntax of grooming. Here, we identified two main upstream inputs to the DLS, namely, the glutamatergic MD\u0026rarr;DLS pathway, which coregulates locomotion and sustains the grooming maintenance phase, and the GABAergic BLA\u0026rarr;DLS pathway, which co-regulates anxiety, constrains maintenance, and promotes sequence initiation. Under pathological conditions, inhibiting the MD\u0026rarr;DLS pathway alleviated stereotyped grooming in an autism model by shortening the hyperpersistent maintenance phase. Activating the BLA\u0026rarr;DLS circuit ameliorated repetitive grooming in both the stress and autism models by normalizing the phase durations and correcting condition-specific transition abnormalities. These findings reveal how distinct circuit logics generate heterogeneous pathological grooming, providing a framework for dissecting related neuropsychiatric disorders.\u003c/p\u003e\n\u003cp\u003eSelf-grooming is a vital adaptive behavior, and its dysregulation manifests across various neuropsychiatric conditions\u003csup\u003e30\u003c/sup\u003e. Traditional analyses, which often rely on the total duration or coarse metrics such as \u0026ldquo;incorrect transitions\u0026rdquo;\u003csup\u003e3,32\u003c/sup\u003e, fail to capture the nuanced syntax of this behavior, thereby limiting their ability to distinguish distinct pathological states. Instead, we employed a detailed assessment of the phase durations, proportional distributions, and complete transition probability matrix. Building on the classic cephalocaudal sequence\u003csup\u003e4,9,33,34\u003c/sup\u003e, our analysis defines a tripartite organization during grooming behavior: the initiation phase (stages 1\u0026ndash;2, paw and head grooming)\u003csup\u003e35\u003c/sup\u003e, the maintenance phase (stages 3\u0026ndash;4, body and leg grooming)\u003csup\u003e3\u003c/sup\u003e, and the termination phase (stage 5, tail/genital grooming)\u003csup\u003e5\u003c/sup\u003e. The proportional time spent in each phase and the transition probabilities between them provide a framework for deconstructing the grooming microstructure. Applying this framework, we confirmed distinct microstructural signatures in pathological models. Acute restraint stress produced a pattern of truncated initiation and prolonged, oscillatory maintenance, characterized by a surge in reciprocal 3\u0026ndash;4/4\u0026ndash;3 transitions, which is indicative of a sequence \u0026ldquo;locked\u0026rdquo; in repetitive body cleaning with impaired progression to termination (reduced stage 5). This result aligns with the ethological concept of \u0026ldquo;displacement\u0026rdquo; grooming under stress, which is often described as fragmented and chaotic\u003csup\u003e36\u003c/sup\u003e. In stark contrast, \u003cem\u003eShank3B\u003c/em\u003e KO mice exhibited extended maintenance with altered initiation, persisting in a sustained body cleaning loop with reduced transitions to/from the cephalic phase (reduced 2\u0026ndash;3/3\u0026ndash;2 transitions). This pattern reflects a more rigid, stereotyped sequence, which is consistent with the behaviors observed in individuals with compulsive disorders\u003csup\u003e37\u003c/sup\u003e. These distinct signatures strongly support fundamentally different neural mechanisms, extending our prior finding that anxiolytic treatment normalizes the stress-induced but not ASD-related microstructure\u003csup\u003e5\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eOur study elucidates the origins of the grooming microstructure by defining how the DLS, a key node for action sequencing\u003csup\u003e16,17\u003c/sup\u003e, integrates opposing commands from the MD\u003csup\u003e38\u003c/sup\u003e and BLA\u003csup\u003e39\u003c/sup\u003e to sculpt grooming behavior. We identified the MD\u0026rarr;DLS and BLA\u0026rarr;DLS pathways as critical antagonistic regulators. The MD\u0026rarr;DLS circuit is purely glutamatergic and promotes the maintenance phase of grooming. Its activation\u0026nbsp;accelerates and sustains grooming sequences, characterized by a shortened initiation phase (reduced proportion of stage 1) and an amplified maintenance phase (increased stage 3\u0026ndash;4 activity and transitions). These findings are consistent with calcium activity being most elevated in MD neurons and MD\u0026rarr;DLS projections during the maintenance phase (stages 3\u0026ndash;4). This result aligns with the role of the MD in motor planning and compulsive behaviors\u003csup\u003e40,41\u003c/sup\u003e.\u0026nbsp;Furthermore, recent studies have reported hyperactivity of the MD in adult \u003cem\u003eShank3B\u003c/em\u003e KO mice\u003csup\u003e42\u003c/sup\u003e. Conversely, the activation of the GABAergic component of the BLA\u0026rarr;DLS pathway suppresses and fragments grooming sequences by prolonging the initiation phase (increased stage 1 proportion), prematurely aborting the maintenance phase (decreased proportion of stage 4), and shifting transition dynamics toward oscillations between initiation and early maintenance (e.g., increased 1\u0026ndash;3/3\u0026ndash;1 transitions). This shift effectively disrupted the fluent cephalocaudal progression of grooming. The functional profile of this pathway corresponds to BLA\u0026rarr;DLS neuronal activity peaking during stages 2\u0026ndash;3 and prominent BLA signals at termination, suggesting a role in monitoring and potentially curtailing ongoing sequences based on the internal state. This finding fits the established role of the BLA in assigning emotional valence and anxiety processing\u003csup\u003e43,44\u003c/sup\u003e and the comorbidity of anxiety with repetitive behaviors\u003csup\u003e45\u003c/sup\u003e. Thus, the DLS acts as an integrator, balancing a motor-sequencing \u0026ldquo;go\u0026rdquo; signal from the MD against an affective \u0026ldquo;check\u0026rdquo; signal from the BLA to generate fluent, context-appropriate grooming.\u003c/p\u003e\n\u003cp\u003eThe\u0026nbsp;divergent dysfunctions of these circuits underlie distinct pathologies. The BLA\u0026rarr;DLS circuit is critically involved in acute stress: BLA activity shifts, and activating its GABAergic projections to the DLS rapidly normalizes the chaotic microstructure. Inhibiting the MD\u0026rarr;DLS pathway has no effect, indicating that anxiety-driven pathology primarily involves maladaptive affective regulation. In contrast, stereotypy in \u003cem\u003eShank3B\u003c/em\u003e KO mice is driven by hyperactivity of the MD\u0026rarr;DLS pathway, which promotes perseverative maintenance. The inhibition of this motor-sequencing pathway is therapeutic. Activating the GABAergic BLA\u0026rarr;DLS pathway also ameliorates stereotypy but by fragmenting the rigid sequence, highlighting anxiety as a separable, modulatory component of ASD. This circuit-based dissection provides a mechanistic framework for understanding how heterogeneous pathological grooming arises from imbalances between motor sequence facilitation and affective state modulation.\u003c/p\u003e\n\u003cp\u003eWhole-brain c-Fos mapping confirmed that grooming engages a distributed network beyond these hubs, including motor regions (e.g., the MO and ACA)\u003csup\u003e27,46\u003c/sup\u003e, sensory areas (e.g., S1 and VIS)\u003csup\u003e47\u003c/sup\u003e, and limbic regions (e.g., the AI and MEA)\u003csup\u003e48\u003c/sup\u003e. These findings underscore grooming as a complex integration of motor execution, sensory feedback, and internal states. Notably, while M2 and S1 were active during grooming, particularly in the maintenance phase, their direct optogenetic manipulation did not alter the grooming microstructure. These findings suggest that these areas function as hierarchical executors and sensors within sensorimotor loops, which are necessary for movement generation and somatosensory feedback\u003csup\u003e49\u003c/sup\u003e but not for determining the sequential microstructure itself. Their influence may be mediated by other striatal territories, such as the DMS\u003csup\u003e50,51\u003c/sup\u003e, highlighting the specificity of the DLS for automated sequence execution.\u003c/p\u003e\n\u003cp\u003eSeveral limitations point to future directions. First, while we focused on the net effect of DLS outputs, further investigation is needed to elucidate how MD and BLA inputs engage specific striatal pathways (D1/D2)\u003csup\u003e50\u003c/sup\u003e and interneurons\u003csup\u003e52\u003c/sup\u003e under pathological conditions. Second, our study examined anxiety-driven and intrinsic stereotyped grooming but did not address \u0026ldquo;comfort\u0026rdquo; grooming (triggered by mild pressure or physical disturbances such as sprays)\u003csup\u003e53\u003c/sup\u003e, which may engage distinct neural circuits\u003csup\u003e12\u003c/sup\u003e. Thus, comparative ethological studies are warranted. Third, all experiments were conducted using male mice. Given the well-known sex differences in stress\u003csup\u003e54\u003c/sup\u003e, anxiety\u003csup\u003e55\u003c/sup\u003e, ASD\u003csup\u003e56\u003c/sup\u003e, and grooming behavior\u003csup\u003e57\u003c/sup\u003e, including both sexes in the future is crucial for generalizability.\u003c/p\u003e\n\u003cp\u003eIn conclusion, we established that the fine-grained microstructure of grooming, rather than its overall quantity, is responsible for its underlying neural logic. The DLS integrates antagonistic commands from a glutamatergic MD-based motor circuit that sustains sequences and a GABAergic BLA-based affective circuit that constrains them. The hyperactivity of the former drives the perseveration of stereotypy, whereas altered engagement of the latter underlies anxiety-driven fragmentation. This framework advances our understanding of how specific behavioral sequences are constructed and pathologically distorted, identifying refined circuit targets for dissecting and potentially treating related neuropsychiatric symptoms.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eAnimals\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll the\u0026nbsp;experimental\u0026nbsp;protocols were approved by the Fourth Military Medical University Institutional Animal Care and Use Committee and\u0026nbsp;were\u0026nbsp;conducted\u0026nbsp;in accordance with\u0026nbsp;the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH). All\u0026nbsp;the\u0026nbsp;animals were 6-\u0026nbsp;to 8-week-old (20 to 25 g) male mice housed\u0026nbsp;at\u0026nbsp;a constant temperature (23\u0026ndash;25\u0026deg;C) on a 12 h light/dark cycle (lights on from 08:30 to 20:30) with ad libitum access to water and food. C57BL/6 mice were obtained from the Laboratory Animal Center of the Fourth Military Medical University. \u003cem\u003eShank3B\u003c/em\u003e KO mice were kindly provided by Prof. Guoping Feng\u003csup\u003e6\u003c/sup\u003e. All\u0026nbsp;the\u0026nbsp;mice used in this study were male and maintained on a pure C57BL/6J background. \u003cem\u003eShank3B\u003c/em\u003e KO::Vglut2-cre double-positive mice were generated by crossing \u003cem\u003eShank3B\u0026nbsp;\u003c/em\u003eheterozygous mice with Vglut2-cre mice (strain #016963, The Jackson Laboratory). Similarly, \u003cem\u003eShank3B\u003c/em\u003e KO::VGAT-cre double-positive mice were obtained by crossing \u003cem\u003eShank3B\u0026nbsp;\u003c/em\u003eheterozygous mice with VGAT-cre mice (strain #028862, The Jackson Laboratory). Fos-cre\u003csup\u003eER\u003c/sup\u003e mice were purchased from The Jackson Laboratory (strain #021882).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGrooming behaviors induced under different conditions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTwo\u0026nbsp;methods\u0026nbsp;were used to elicit grooming behaviors in mice. The first utilized \u003cem\u003eShank3B\u003c/em\u003e KO mice, which inherently display high levels of stereotypic grooming behavior\u003csup\u003e6\u003c/sup\u003e. The second method involved an acute restraint stress model\u003csup\u003e58\u003c/sup\u003e. Briefly, mice were placed into a modified 50 ml Falcon tube with breathing holes, where only the tails and tips of their noses were free, for 30 min; upon release, the mice exhibited extensive grooming behavior.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGrooming behavior and grooming microstructure analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGrooming behavior was recorded in custom-built, sound-attenuating chambers (20 cm \u0026times; 20 cm \u0026times; 25 cm) equipped with adjustable top-mounted LED lighting for behavioral recording. Experiments were conducted\u0026nbsp;under\u0026nbsp;low illumination (~40 lux) to capture 30-minute videos of freely moving mice for the analysis of grooming behavior. Three walls of the chamber were opaque, with only the wall facing the camera (30 fps) being transparent to minimize environmental stress. Prior to testing, the mice were acclimatized to the behavioral room for at least one hour before being gently transferred into the chamber using an opaque tray. After each session, the chamber was cleaned by removing waste, spraying with 75% ethanol to eliminate odors, and rinsing with water. The analysis of grooming behavior was performed using the open-source software JAABA (Janelia Automatic Animal Behavior Annotator)\u003csup\u003e59\u003c/sup\u003e (downloaded from GitHub https://github\u003cstrong\u003e.\u003c/strong\u003ecom/kristinbranson/JAABA)). Mouse movement trajectories were first extracted using the Mouse Tracker toolbox\u003csup\u003e60\u003c/sup\u003e in MATLAB and then imported into JAABA. A supervised GentleBoost learning algorithm was trained on 115,268 manually annotated video frames to create a classifier for identifying grooming behavior. The trained classifier showed strong agreement with manual scoring, with an error margin of less than 20%\u0026nbsp;for\u0026nbsp;the grooming duration (Supplementary\u0026nbsp;Fig.\u0026nbsp;1A\u0026ndash;C), confirming its reliability for subsequent analyses under consistent experimental conditions. For the detailed bout analysis, the following criteria were applied: a grooming episode was analyzed\u0026nbsp;only\u0026nbsp;if it lasted longer than 3 seconds, and a pause in grooming exceeding 6 seconds was defined as the end of one bout and the start of a new\u0026nbsp;bout\u003csup\u003e3,5\u003c/sup\u003e. Using the same training strategy, separate classifiers were developed for\u0026nbsp;the\u0026nbsp;distinct grooming stages: stage 1 (paw licking) trained on 93271 frames of video, stage 2 (nose/face/head wash) trained on 84014 frames, stage 3 (body grooming) trained on 83877 frames, stage 4 (leg licking) trained on 44697 frames,\u0026nbsp;and\u0026nbsp;stage 5 (tail/genitals grooming) trained on 42339 frames. By applying both the general grooming classifier and the stage-specific classifiers to the same video recordings, we extracted comprehensive quantitative metrics of\u0026nbsp;the\u0026nbsp;grooming microstructure\u003csup\u003e3\u003c/sup\u003e. These metrics included the absolute time spent in each grooming stage, the proportional distribution of grooming time across\u0026nbsp;stages, the total number of stage transitions, and the frequency of transitions between each specific pair of\u0026nbsp;stages\u003csup\u003e7\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec-Fos experiments and graph\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003etheory\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eQuantification of whole\u003c/strong\u003e\u003cstrong\u003e-\u003c/strong\u003e\u003cstrong\u003ebrain c-Fos expression\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe used c-Fos, a transcription factor that is rapidly upregulated upon neuronal activation, as a marker of behavior-associated neural activity to identify brain regions upstream of the dorsolateral striatum (DLS) that were activated during grooming. Grooming behaviors were recorded and mice were sacrificed 1.5 hours later and processed for c-Fos staining. Briefly, the animals were anesthetized with isoflurane and\u0026nbsp;transcardially\u0026nbsp;perfused with 0.01 M PBS followed by 4% paraformaldehyde.\u0026nbsp;The brains\u0026nbsp;were postfixed overnight at\u0026nbsp;4\u0026deg;C, cryoprotected in 30% sucrose, and sagittally sectioned at 50 \u0026mu;m using a freezing microtome. Sections were incubated with\u0026nbsp;a\u0026nbsp;rabbit anti-c-Fos\u0026nbsp;antibody\u0026nbsp;(1:1500;\u0026nbsp;Cell Signaling Technology #2250) and subsequently with\u0026nbsp;a\u0026nbsp;donkey anti-rabbit Alexa Fluor 488\u0026nbsp;antibody\u0026nbsp;(1:500;\u0026nbsp;Invitrogen A-21206). Nuclei were counterstained with Hoechst 33342 (1:1000;\u0026nbsp;Abcam ab228551). High-resolution images of\u0026nbsp;the\u0026nbsp;brain sections were acquired at 150-\u0026mu;m intervals and analyzed using a three-step pipeline, as described previously\u003csup\u003e61\u003c/sup\u003e: (1) signal recognition\u0026mdash;c-Fos-positive cells were automatically identified using a classifier trained in ImageJ (Fiji); (2) brain registration\u0026mdash;raw images of the sections were aligned to the Allen Mouse Brain Atlas using MATLAB; and (3) brain segmentation\u0026mdash;positive signals were assigned to specific subregions based on atlas-defined grayscale intensities in Imaris software. All\u0026nbsp;the\u0026nbsp;analyses were conducted independently by two observers.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGraph\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003etheory\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;and hub network identification\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe employed a graph theory-based approach to delineate functional networks among c-Fos-positive brain\u0026nbsp;regions\u0026nbsp;and to identify hub regions within these networks\u003csup\u003e62\u003c/sup\u003e. After the c-Fos-positive cell counts for each brain region\u0026nbsp;were obtained, the raw data were normalized using log10 transformation to reduce variability. Pearson\u0026rsquo;s correlation coefficients were then calculated between every pair of brain regions. These interregional correlation coefficients were subsequently converted into pairwise Euclidean distances. Hierarchical cluster dendrograms were generated from the distance matrices to identify functional modules, with the tree-cutting threshold set at 50% of the tree height to split the dendrogram into distinct modules. In parallel, the correlation matrix was reorganized to construct the functional network, and a connection between two regions was included only if the Pearson correlation coefficient was \u0026ge; 0.82\u003csup\u003e63\u003c/sup\u003e. The resulting network parameters were analyzed using the Brain Connectivity Toolbox (https://sites.google.com/site/bctnet/) to compute the following centrality metrics: (1) degree, which is defined as the number of links connected to a node; (2) normalized betweenness centrality, which is defined as the fraction of all shortest paths in the network that pass through a given node; and (3) eigenvector centrality, which is a measure of the influence of as node based on the centrality of its neighbors. Brain regions that ranked highly across all three centrality measures were defined as hub nodes (Supplementary Table 2). The functional network was visualized using Cytoscape software\u003csup\u003e64\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStereotaxic injections and fiber optic implantation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe mice were anesthetized with 1% isoflurane and aligned on a stereotactic frame (RWD Life Science Inc., China). All surgeries were performed under sterile conditions. The heads of the mice were shaved and\u0026nbsp;cleaned\u0026nbsp;with 75% alcohol,\u0026nbsp;and\u0026nbsp;their skulls were thinned with a dental drill and then removed. Stereotactic coordinates were based on\u0026nbsp;the\u0026nbsp;Paxinos and Franklin mouse brain atlas and adjusted appropriately. Each mouse was injected with viral vectors using a glass micropipette at a rate of 30\u0026ndash;60\u0026nbsp;nL per min. The micropipette remained in position for 10 min\u0026nbsp;after the injection\u0026nbsp;and then was slowly withdrawn.\u0026nbsp;The mice\u0026nbsp;were rewarmed on a thermal blanket after surgery and allowed to recover for\u0026nbsp;more than\u0026nbsp;3 weeks before\u0026nbsp;they participated in the\u0026nbsp;behavioral experiments.\u003c/p\u003e\n\u003cp\u003eFor optogenetic behavioral experiments, rAAV-hSyn-mCherry-WPRE-pA, rAAV-hSyn-EGFP-WPRE-pA, rAAV-hSyn-NpHR3.0-EGFP-WPRE-pA, or rAAV-hSyn-hChR2(H134R)-mCherry-WPRE-pA was injected bilaterally into M2 (+0.86 mm AP, \u0026plusmn;0.75 mm ML, -1.30 mm DV; 300 nl per site), S1 (-1.34 mm AP, \u0026plusmn;1.50 mm ML, -1.00 mm DV; 100 nl per site), MD (-1.34 mm AP, \u0026plusmn;0.40 mm ML, -3.40 mm DV; 100 nl per site), or BLA (-1.40 mm AP, \u0026plusmn;3.15 mm ML, -4.80 mm DV; 150 nl per site) using a glass micropipette. rAAV-EF1a-DIO-hChR2(H134R)-mCherry-WPRE-pA, rAAV-Ef1a-DIO-NpHR3.0-EGFP, or rAAV-EF1a-DIO-mCherry-WPRE-pA was injected bilaterally into the MD or BLA of Vglut2-Cre or VGAT-cre animals. A 200 mm core 0.37 NA optical fiber was implanted 0.5 mm above the injection site. For MD, bilateral ceramic ferrules were inserted at a 15\u0026deg; angle (-1.34 mm AP, \u0026plusmn;1.26 mm ML, -3.00 mm DV).\u003c/p\u003e\n\u003cp\u003eFor\u0026nbsp;optogenetic\u0026nbsp;targeting of MD/BLA\u0026rarr;DLS terminals, viruses similar to those described above were injected bilaterally into the MD or BLA. Optical fibers were then implanted in the DLS, specifically at an anterior site (+0.70 mm AP, \u0026plusmn;2.25 mm ML, -2.5 mm DV) for MD\u0026rarr;DLS manipulation and at a posterior site (+0.00 mm AP, \u0026plusmn;2.5 mm ML, -2.3 mm DV) for BLA\u0026rarr;DLS manipulation.\u003c/p\u003e\n\u003cp\u003eFor chemogenetic activation experiments, AAV2/retro-hSyn-NLS-Cre-P2A-mCherry was injected bilaterally into the DLS (+0.70 mm AP, \u0026plusmn;2.25 mm ML, -3.00 mm DV; 300 nl per site), and rAAV-hSyn-DIO-hM3Dq\u0026nbsp;(Gq)-eGFP-WPRE-pA was injected bilaterally into\u0026nbsp;the\u0026nbsp;MD (-1.46 mm AP, \u0026plusmn;0.40 mm ML, -3.40 mm DV; 100 nl per site) or BLA (-1.40 mm AP, \u0026plusmn;3.15 mm ML, -4.80 mm DV; 150 nl per site). For chemogenetic inhibition experiments, AAV2/retro-hSyn-NLS-Cre-P2A-mCherry was injected bilaterally\u0026nbsp;into the\u0026nbsp;DLS,\u0026nbsp;and\u0026nbsp;rAAV-EF1a-DIO-hM4D(Gi)-EGFP-WPRE\u0026nbsp;was injected bilaterally\u0026nbsp;into\u0026nbsp;the\u0026nbsp;MD or BLA. For chemogenetic control experiments, AAV2/retro-hSyn-NLS-Cre-P2A-mCherry was injected bilaterally\u0026nbsp;into the\u0026nbsp;DLS,\u0026nbsp;and\u0026nbsp;rAAV-EF1a-DIO-EGFP-WPRE-bGH pA virus\u0026nbsp;was\u0026nbsp;injected bilaterally\u0026nbsp;into the MD\u0026nbsp;or BLA.\u003c/p\u003e\n\u003cp\u003eFor input\u0026nbsp;mapping\u0026nbsp;experiments, AAV2/2Retro-hSyn-eGFP-3Flag-WPRE-SV40pA was injected unilaterally into the DLS (+0.70 mm AP, -2.25 mm ML, -3.00 mm DV; 200 nl per site) of WT mice; AAV2/2Retro-DIO-mCherry was injected unilaterally into the DLS of CamKⅡ-cre and VGAT-cre mice. For anterograde monosynaptic tracing, AAV-hSyn-mCherry was injected unilaterally into the MD and BLA of WT mice; AAV-EF1\u0026alpha;-DIO-mRuby-T2A-Synaptophysin-EGFP was injected unilaterally into the MD and BLA of Vglut2-cre and VGAT-cre mice. The animals were sacrificed 3 weeks later for monosynaptic tracing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFiber\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003ephotometry\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLocal calcium signals were recorded after rAAV-hSyn-jGCaMP8s was injected into the target brain regions (M2, S1, MD or BLA). For projection-specific calcium imaging, AAV2/retro-hSyn-NLS-Cre-P2A-mCherry was injected unilaterally into the DLS, while rAAV-CAG-FLEX-jGCaMP7s was injected ipsilaterally into either the MD or BLA. Ceramic ferrules were implanted 0.3 mm above the virus injection sites. After a 3-week recovery period, fiber photometry recordings were performed to simultaneously record calcium signals and grooming behavior evoked under different conditions over a 30-min session. The onset of calcium signals was aligned frame by frame with the behavioral start time, and signals were analyzed separately according to distinct grooming stages. A 2-s preevent window was used as the baseline, and Z scores during the 5-s postevent period were calculated to analyze the area under the curve. All analyses were conducted independently by two observers who were blinded to the group assignments.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eModulation of behavior\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOptogenetic modulation of grooming behavior\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor the optogenetic manipulation of grooming, the mice were acclimated to the testing room for 1 h prior to recording. Grooming behavior was monitored over a 1.5-hour session, following a laser OFF‒ON‒OFF paradigm with 30-minute intervals. In ChR2-mediated activation experiments, blue laser light (473 nm) was delivered at 20 Hz (5-ms pulse width). For NpHR-mediated inhibition, continuous yellow light (589 nm)\u0026nbsp;was applied. During the regional stimulation of M2, S1, BLA or MD, the\u0026nbsp;laser power was calibrated before each session to ensure an output intensity of\u0026nbsp;5\u0026ndash;7 mW per side at the fiber tip. For the terminal stimulation of MD/BLA-DLS projections, the output was set to\u0026nbsp;9\u0026ndash;10 mW per side. Successive stimulation sessions for each mouse were separated by at least 24 h.\u003c/p\u003e\n\u003cp\u003eWe also\u0026nbsp;performed acute optogenetic stimulation triggered upon the initiation of self-grooming to control for potential thermal effects induced by prolonged (30 min) laser exposure (Supplementary Fig.\u0026nbsp;6). A 5 s (BLA\u0026rarr;DLS) or 15 s (\u0026thinsp;MD\u0026rarr;DLS) light pulse (20 Hz) was delivered during the 30-min recording, and the following parameters were analyzed: grooming termination (percentage of grooming bouts that stopped during the 5 s or 30 s light-on window), offset latency (time from light onset to grooming cessation), grooming retrieval (percentage of terminated bouts in which the animals resumed grooming), retrieval grooming time (mean duration of resumed grooming bouts), and total grooming (cumulative grooming time during the 30-min session).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eChemogenetic manipulation of grooming behavior\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor\u0026nbsp;DREADD experiments, C21 or vehicle (saline) was intraperitoneally injected 30 min before behavioral testing to allow sufficient time for Gq- or Gi-mediated signaling modulation during the assay. The testing protocol consisted of two consecutive daily sessions: on day 1, the mice received an intraperitoneal injection of 0.9 % saline, and their grooming behavior was recorded for 30 min to establish a baseline; on day 2, C21 (2 mg/kg) was administered to examine how the activation or inhibition of MD\u0026rarr;DLS or BLA\u0026rarr;DLS circuits influenced grooming behavior.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eThree-dimensional behavioral recordings during optogenetic stimulation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe 3D behavioral test was performed based on a hierarchical learning framework for spontaneous behavior mapping\u003csup\u003e29\u003c/sup\u003e. The mice were individually placed in a transparent, custom-built rectangular chamber (20 cm \u0026times; 20 cm \u0026times; 25 cm) and allowed to explore freely under dim, homogeneous lighting. Behavior was recorded in a 15-min session following a laser OFF\u0026ndash;ON\u0026ndash;OFF stimulation paradigm (5 min intervals). Optogenetic activation of MD/BLA\u0026rarr;DLS terminals was achieved with blue light (473 nm, 20 Hz, 5 ms pulses, 9\u0026ndash;10 mW output at the fiber tip). Four high-speed cameras (100 fps) were positioned at 90\u0026deg; intervals around the arena to capture synchronized multiview videos. The 3D posture of each animal was reconstructed by first detecting 2D key points in each view using DeepLabCut, followed by triangulation to establish 2D-to-3D correspondence. A hierarchical graph network model was then applied to extract continuous 3D kinematic features and cluster them into discrete, semantically meaningful behavioral units (e.g., rearing, turning, stretching, and grooming). This system enables the high-throughput, automated profiling of the spontaneous behavioral structure with minimal experimental interference.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTargeted Recombination in Active Populations (TRAP)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA TRAP strategy was employed in Fos-cre\u003csup\u003eER\u003c/sup\u003e mice to label stress-responsive neurons. Cre-dependent viral vectors were stereotactically injected into either the MD or BLA, and optical fibers were bilaterally implanted above the DLS. Following a two-week post-operative recovery and viral expression period, the mice were subjected to a three-day restraint stress protocol. During this paradigm, 4-hydroxytamoxifen (4-OHT) (Sigma, H7904) was administered intraperitoneally (i.p.) at a dose of 100 mg/kg to induce recombination and permanently label neurons activated by stress. After an additional two-week period to ensure full transgene expression, circuit manipulation experiments were performed.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAnxiety-related behavioral tests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOpen Field Test (OFT)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe OFT was conducted as previously described to detect anxiety and locomotor activity\u003csup\u003e65\u003c/sup\u003e. The apparatus consisted of a 50 cm \u0026times; 50 cm \u0026times; 50 cm polystyrene arena\u003cstrong\u003e.\u003c/strong\u003e The mice were acclimated to the testing room for 1 h prior to the trial. Each mouse was then gently placed in the center of the arena, and its behavior was recorded for 10 min. The arena was cleaned with 75% ethanol between trials to remove olfactory cues. The total distance traveled over the 10-min session was analyzed using the Smart system (SMART 3.0, Panlab S.L.U., Spain) to assess general locomotor activity. The time spent in the central zone (defined as the central 25% of the arena area) was used as an index of anxiety-like behavior.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eElevated Plus Maze (EPM)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe EPM was constructed with two open arms (25 cm \u0026times; 8 cm) and two enclosed arms (25 cm \u0026times; 8 cm, with 12 cm high walls) connected by a central platform (8 cm \u0026times; 8 cm)\u003csup\u003e66\u003c/sup\u003e. The maze was elevated 90 cm above the floor. The mice were habituated to the test room for 1 h before the trial. Each animal was placed on the central platform facing an open arm, and behavior was recorded for 5 min. The maze was cleaned with 75% ethanol between subjects. The time spent in the open arms was quantified to evaluate anxiety-like states.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eImmunohistochemistry\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor viral\u0026nbsp;tracing, the mice were perfused 3 weeks after the virus injection. Sagittal brain sections (50 \u0026mu;m thick) were incubated with a rabbit anti-GFP antibody (1:1000; Invitrogen, A-11122) or, for site verification, a rabbit anti-RFP antibody (1:1000; Rockland, 600-401-379). Mice received 10 min of 20 Hz photostimulation and were perfused 90 min later to confirm neuronal activation in optogenetic experiments. The sections were stained with rabbit anti-c-Fos (1:1500; Cell Signaling Technology #2250) or rabbit anti-Arc (1:1000; Synaptic Systems 156003) antibodies. GABAergic neurons were labeled a mouse anti-GAD67 antibody (1:1000; Sigma‒Aldrich, MAB5406). For glutamatergic neuron labeling, a mouse anti-Vglut2 antibody (1:100; Sigma‒Aldrich, MAB5504) was used. For immunofluorescence staining of all sections, the following secondary antibodies were used: a donkey anti-rabbit antibody conjugated to Alexa Fluor 488 (1:500, Invitrogen, A-21206) or Alexa Fluor 594 (1:500, Invitrogen, A-21207). Nuclei were counterstained with Hoechst 33342 (1:1000; Abcam, ab228551). Injection and fiber placement sites (e.g., the DLS, M2, S1, MD, and BLA) were verified on coronal sections along the anteroposterior axis (Supplementary Figs. 6 and 12).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eElectrophysiology\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe characterized the functional properties of MD/BLA\u0026rarr;DLS projections by combining optogenetic stimulation with whole-cell patch-clamp recordings to examine synaptic transmission. The upstream region was injected with AAV-hSyn-hChR2(H134R)-mCherry. Three weeks later, brain slices containing the DLS were prepared, and light-evoked postsynaptic currents were recorded from neurons near labeled terminals.\u003c/p\u003e\n\u003cp\u003eSlice preparation: Mice were deeply anesthetized with isoflurane and transcardially perfused with ice-cold, oxygenated cutting solution (115\u0026thinsp;mM choline chloride, 2.5\u0026thinsp;mM KCl, 1.25\u0026thinsp;mM NaH\u003csub\u003e2\u003c/sub\u003ePO4, 0.5\u0026thinsp;mM CaCl\u003csub\u003e2\u003c/sub\u003e, 8\u0026thinsp;mM MgCl\u003csub\u003e2\u003c/sub\u003e, 26\u0026thinsp;mM NaHCO\u003csub\u003e3\u003c/sub\u003e, 10\u0026thinsp;mM D-(+)-glucose, 0.1\u0026thinsp;mM L-ascorbic acid, and 0.4\u0026thinsp;mM sodium pyruvate, pH 7.4, osmolarity of 295\u0026ndash;300\u0026thinsp;mOsm/L). The brains were quickly removed, and 300-\u0026mu;m-thick slices containing the target regions were cut using a vibratome. Slices were recovered in oxygenated artificial cerebrospinal fluid (ACSF, 119\u0026thinsp;mM NaCl, 2.3\u0026thinsp;mM KCl, 1.0\u0026thinsp;mM NaH\u003csub\u003e2\u003c/sub\u003ePO\u003csub\u003e4\u003c/sub\u003e, 26\u0026thinsp;mM NaHCO\u003csub\u003e3\u003c/sub\u003e, 11\u0026thinsp;mM D-(+)-glucose, 1.3\u0026thinsp;mM MgCl\u003csub\u003e2\u003c/sub\u003e, and 2.5\u0026thinsp;mM CaCl\u003csub\u003e2\u003c/sub\u003e, pH 7.4, osmolarity of 295\u0026ndash;300\u0026thinsp;mOsm/L) at 32\u0026deg;C for 30 min and then maintained at room temperature for \u0026ge;1 h before recording.\u003c/p\u003e\n\u003cp\u003eRecording setup: The slices were transferred to a submerged recording chamber. Under fluorescence guidance,\u0026nbsp;healthy\u0026nbsp;neurons near mCherry-positive axons were selected for whole-cell recording.\u003c/p\u003e\n\u003cp\u003eEPSC recording: Pipettes were filled with a potassium-based internal solution. ACSF was supplemented with bicuculline (a GABAA receptor antagonist, 100 \u0026mu;M), tetrodotoxin (TTX, 1 \u0026mu;M), and 4-AP (100 \u0026mu;M) to isolate monosynaptic currents. Neurons were voltage clamped at -60 mV. Blue light pulses (473 nm, 10 mW, 20 Hz, 200 ms) were delivered to evoke EPSCs.\u003c/p\u003e\n\u003cp\u003eIPSC recording: Pipettes contained a high-chloride cesium-based internal solution. ACSF was supplemented with D-AP5 (an NMDA receptor antagonist, 50 \u0026mu;M), NBQX (10 \u0026mu;M), TTX (1 \u0026mu;M), and 4-AP (100 \u0026mu;M). Neurons were clamped at -60 mV, and identical light stimulation was applied to evoke IPSCs.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll the data are presented as the means \u0026plusmn; standard errors of the means (SEMs). Statistical analysis and graphing were performed using SPSS 22.0. The normality of the distribution of the data was assessed using the Shapiro‒Wilk test, and the homogeneity of variance was evaluated with Levene\u0026rsquo;s test. Parametric tests were applied when the data met the assumptions of normality and homogeneity of variance. If normality was satisfied but\u0026nbsp;homogeneity\u0026nbsp;of variance was not\u0026nbsp;met, a corrected unpaired t\u0026nbsp;test was used for comparisons between two independent groups; otherwise,\u0026nbsp;nonparametric\u0026nbsp;tests were employed. Parametric tests included two-tailed\u0026nbsp;paired\u0026nbsp;and unpaired t\u0026nbsp;tests, one-way ANOVA and repeated-measures ANOVA\u003cstrong\u003e.\u003c/strong\u003e The corresponding nonparametric tests included the Wilcoxon signed-rank test, Mann‒Whitney U test, Kruskal‒Wallis H test and Friedman test. Correlations between two variables were assessed by calculating Pearson\u0026rsquo;s correlation coefficient. A p value \u0026lt; 0.05 was considered to indicate statistical significance, with asterisks denoting the following levels of significance: *p \u0026lt; 0.05, **p \u0026lt; 0.01, ***p \u0026lt; 0.001, and ****p \u0026lt; 0.0001. All detailed behavior data are presented in Supplementary Table 1.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank Prof. Guoping Feng (Massachusetts Institute of Technology) for suggestions and discussions. This work was supported by grants from Brain Science and Brain-like Intelligence Technology\u0026ndash;National Science and Technology Major Project (2021ZD0201005 to S.W. and 2025ZD0216200 to B.G.), the National Natural Science Foundation of China (82221001 to S.W., 82271577 to W.W., 32394032 to S.W., 82422030 to B.G., and 82401777 to X.H.), and Joint Founding Project of Innovation Research Institute, Xijing Hospital (LHJJ24JH05 to W.W.). We thank Springer Nature Editing Service for English language editing (certificate verification code: 2439-A7BA-9DCF-416C-75DP).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eXin Huang, Formal analysis, Investigation, Methodology, Writing\u0026mdash;original draft; Jinwei Xu, Investigation, Methodology; Zimeng Li, Formal analysis, Investigation; Baolin Guo, Conceptualization, Funding acquisition; Honghui Mao, Formal analysis; Haiying Liu, Conceptualization, Formal analysis, Investigation, Methodology, Writing\u0026mdash;original draft; Shengxi Wu, Conceptualization, Funding acquisition, Supervision, Writing\u0026mdash;review and editing; Wenting Wang, Conceptualization, Funding acquisition, Supervision, Writing\u0026mdash;review and editing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data supporting the findings described in this manuscript are available in the article and in the supplementary information and the source data are provided with this paper.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eHodgetts, S., Magill-Evans, J. \u0026amp; Misiaszek, J. E. Weighted vests, stereotyped behaviors and arousal in children with autism. \u003cem\u003eJ Autism Dev Disord\u003c/em\u003e \u003cstrong\u003e41\u003c/strong\u003e, 805-814 (2011). https://doi.org/10.1007/s10803-010-1104-x\u003c/li\u003e\n\u003cli\u003eO\u0026apos;Loghlen, J., McKenzie, M., Lang, C. \u0026amp; Paynter, J. Repetitive Behaviors in Autism and Obsessive-Compulsive Disorder: A Systematic Review. \u003cem\u003eJ Autism Dev Disord\u003c/em\u003e \u003cstrong\u003e55\u003c/strong\u003e, 2307-2321 (2025). https://doi.org/10.1007/s10803-024-06357-8\u003c/li\u003e\n\u003cli\u003eKalueff, A. V., Aldridge, J. W., LaPorte, J. L., Murphy, D. L. \u0026amp; Tuohimaa, P. Analyzing grooming microstructure in neurobehavioral experiments. \u003cem\u003eNat Protoc\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, 2538-2544 (2007). https://doi.org/10.1038/nprot.2007.367\u003c/li\u003e\n\u003cli\u003eKalueff, A. V. \u0026amp; Tuohimaa, P. Mouse grooming microstructure is a reliable anxiety marker bidirectionally sensitive to GABAergic drugs. \u003cem\u003eEur J Pharmacol\u003c/em\u003e \u003cstrong\u003e508\u003c/strong\u003e, 147-153 (2005). https://doi.org/10.1016/j.ejphar.2004.11.054\u003c/li\u003e\n\u003cli\u003eLiu, H.\u003cem\u003e et al.\u003c/em\u003e Dissection of the relationship between anxiety and stereotyped self-grooming using the Shank3B mutant autistic model, acute stress model and chronic pain model. \u003cem\u003eNeurobiol Stress\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 100417 (2021). https://doi.org/10.1016/j.ynstr.2021.100417\u003c/li\u003e\n\u003cli\u003ePeca, J.\u003cem\u003e et al.\u003c/em\u003e Shank3 mutant mice display autistic-like behaviours and striatal dysfunction. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e472\u003c/strong\u003e, 437-442 (2011). https://doi.org/10.1038/nature09965\u003c/li\u003e\n\u003cli\u003eKalueff, A. V. \u0026amp; Tuohimaa, P. The grooming analysis algorithm discriminates between different levels of anxiety in rats: potential utility for neurobehavioural stress research. \u003cem\u003eJ Neurosci Methods\u003c/em\u003e \u003cstrong\u003e143\u003c/strong\u003e, 169-177 (2005). https://doi.org/10.1016/j.jneumeth.2004.10.001\u003c/li\u003e\n\u003cli\u003eNin, M. S.\u003cem\u003e et al.\u003c/em\u003e Anxiolytic effect of clonazepam in female rats: grooming microstructure and elevated plus maze tests. \u003cem\u003eEur J Pharmacol\u003c/em\u003e \u003cstrong\u003e684\u003c/strong\u003e, 95-101 (2012). https://doi.org/10.1016/j.ejphar.2012.03.038\u003c/li\u003e\n\u003cli\u003eKalueff, A. V. \u0026amp; Tuohimaa, P. Grooming analysis algorithm for neurobehavioural stress research. \u003cem\u003eBrain Res Brain Res Protoc\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 151-158 (2004). https://doi.org/10.1016/j.brainresprot.2004.04.002\u003c/li\u003e\n\u003cli\u003eXu, Y.\u003cem\u003e et al.\u003c/em\u003e Identification of a neurocircuit underlying regulation of feeding by stress-related emotional responses. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 3446 (2019). https://doi.org/10.1038/s41467-019-11399-z\u003c/li\u003e\n\u003cli\u003eMu, M. D.\u003cem\u003e et al.\u003c/em\u003e A limbic circuitry involved in emotional stress-induced grooming. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 2261 (2020). https://doi.org/10.1038/s41467-020-16203-x\u003c/li\u003e\n\u003cli\u003eXie, Z.\u003cem\u003e et al.\u003c/em\u003e A brain-to-spinal sensorimotor loop for repetitive self-grooming. \u003cem\u003eNeuron\u003c/em\u003e \u003cstrong\u003e110\u003c/strong\u003e, 874-890 e877 (2022). https://doi.org/10.1016/j.neuron.2021.11.028\u003c/li\u003e\n\u003cli\u003eSun, J.\u003cem\u003e et al.\u003c/em\u003e Excitatory SST neurons in the medial paralemniscal nucleus control repetitive self-grooming and encode reward. \u003cem\u003eNeuron\u003c/em\u003e \u003cstrong\u003e110\u003c/strong\u003e, 3356-3373 e3358 (2022). https://doi.org/10.1016/j.neuron.2022.08.010\u003c/li\u003e\n\u003cli\u003eAhmari, S. E.\u003cem\u003e et al.\u003c/em\u003e Repeated cortico-striatal stimulation generates persistent OCD-like behavior. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e340\u003c/strong\u003e, 1234-1239 (2013). https://doi.org/10.1126/science.1234733\u003c/li\u003e\n\u003cli\u003eBurguiere, E., Monteiro, P., Feng, G. \u0026amp; Graybiel, A. M. Optogenetic stimulation of lateral orbitofronto-striatal pathway suppresses compulsive behaviors. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e340\u003c/strong\u003e, 1243-1246 (2013). https://doi.org/10.1126/science.1232380\u003c/li\u003e\n\u003cli\u003eTecuapetla, F., Matias, S., Dugue, G. P., Mainen, Z. F. \u0026amp; Costa, R. M. Balanced activity in basal ganglia projection pathways is critical for contraversive movements. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, 4315 (2014). https://doi.org/10.1038/ncomms5315\u003c/li\u003e\n\u003cli\u003eAldridge, J. W. \u0026amp; Berridge, K. C. Coding of serial order by neostriatal neurons: a \u0026quot;natural action\u0026quot; approach to movement sequence. \u003cem\u003eJ Neurosci\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e, 2777-2787 (1998). https://doi.org/10.1523/JNEUROSCI.18-07-02777.1998\u003c/li\u003e\n\u003cli\u003eCromwell, H. C. \u0026amp; Berridge, K. C. Implementation of action sequences by a neostriatal site: a lesion mapping study of grooming syntax. \u003cem\u003eJ Neurosci\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 3444-3458 (1996). https://doi.org/10.1523/JNEUROSCI.16-10-03444.1996\u003c/li\u003e\n\u003cli\u003eCorbit, L. H., Nie, H. \u0026amp; Janak, P. H. Habitual alcohol seeking: time course and the contribution of subregions of the dorsal striatum. \u003cem\u003eBiol Psychiatry\u003c/em\u003e \u003cstrong\u003e72\u003c/strong\u003e, 389-395 (2012). https://doi.org/10.1016/j.biopsych.2012.02.024\u003c/li\u003e\n\u003cli\u003eMarkowitz, J. E.\u003cem\u003e et al.\u003c/em\u003e The Striatum Organizes 3D Behavior via Moment-to-Moment Action Selection. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e174\u003c/strong\u003e, 44-58 e17 (2018). https://doi.org/10.1016/j.cell.2018.04.019\u003c/li\u003e\n\u003cli\u003eWang, W.\u003cem\u003e et al.\u003c/em\u003e Striatopallidal dysfunction underlies repetitive behavior in Shank3-deficient model of autism. \u003cem\u003eJ Clin Invest\u003c/em\u003e \u003cstrong\u003e127\u003c/strong\u003e, 1978-1990 (2017). https://doi.org/10.1172/JCI87997\u003c/li\u003e\n\u003cli\u003eLee, J., Wang, W. \u0026amp; Sabatini, B. L. Anatomically segregated basal ganglia pathways allow parallel behavioral modulation. \u003cem\u003eNat Neurosci\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 1388-1398 (2020). https://doi.org/10.1038/s41593-020-00712-5\u003c/li\u003e\n\u003cli\u003eMandelbaum, G.\u003cem\u003e et al.\u003c/em\u003e Distinct Cortical-Thalamic-Striatal Circuits through the Parafascicular Nucleus. \u003cem\u003eNeuron\u003c/em\u003e \u003cstrong\u003e102\u003c/strong\u003e, 636-652 e637 (2019). https://doi.org/10.1016/j.neuron.2019.02.035\u003c/li\u003e\n\u003cli\u003evan den Boom, B. J. G., Pavlidi, P., Wolf, C. J. H., Mooij, A. H. \u0026amp; Willuhn, I. Automated classification of self-grooming in mice using open-source software. \u003cem\u003eJ Neurosci Methods\u003c/em\u003e \u003cstrong\u003e289\u003c/strong\u003e, 48-56 (2017). https://doi.org/10.1016/j.jneumeth.2017.05.026\u003c/li\u003e\n\u003cli\u003eWang, Z.\u003cem\u003e et al.\u003c/em\u003e Comparative analysis of functional network dynamics in high and low alcohol preference mice. \u003cem\u003eExp Neurol\u003c/em\u003e \u003cstrong\u003e389\u003c/strong\u003e, 115238 (2025). https://doi.org/10.1016/j.expneurol.2025.115238\u003c/li\u003e\n\u003cli\u003eGuenthner, C. J., Miyamichi, K., Yang, H. H., Heller, H. C. \u0026amp; Luo, L. Permanent genetic access to transiently active neurons via TRAP: targeted recombination in active populations. \u003cem\u003eNeuron\u003c/em\u003e \u003cstrong\u003e78\u003c/strong\u003e, 773-784 (2013). https://doi.org/10.1016/j.neuron.2013.03.025\u003c/li\u003e\n\u003cli\u003eCorbit, V. L.\u003cem\u003e et al.\u003c/em\u003e Dissociable roles of central striatum and anterior lateral motor area in initiating and sustaining naturalistic behavior. \u003cem\u003eCell Rep\u003c/em\u003e \u003cstrong\u003e44\u003c/strong\u003e, 115181 (2025). https://doi.org/10.1016/j.celrep.2024.115181\u003c/li\u003e\n\u003cli\u003eGe, J.\u003cem\u003e et al.\u003c/em\u003e Diverse Neuronal Activity in Barrel Cortex During Self-Grooming and Whisking in Mice. \u003cem\u003eNeurosci Bull\u003c/em\u003e (2026). https://doi.org/10.1007/s12264-025-01583-1\u003c/li\u003e\n\u003cli\u003eHuang, K.\u003cem\u003e et al.\u003c/em\u003e A hierarchical 3D-motion learning framework for animal spontaneous behavior mapping. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 2784 (2021). https://doi.org/10.1038/s41467-021-22970-y\u003c/li\u003e\n\u003cli\u003eLi, G.\u003cem\u003e et al.\u003c/em\u003e Neural substrates for regulating self-grooming behavior in rodents. \u003cem\u003eJ Zhejiang Univ Sci B\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 841-856 (2024). https://doi.org/10.1631/jzus.B2300562\u003c/li\u003e\n\u003cli\u003eKalueff, A. V.\u003cem\u003e et al.\u003c/em\u003e Neurobiology of rodent self-grooming and its value for translational neuroscience. \u003cem\u003eNat Rev Neurosci\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 45-59 (2016). https://doi.org/10.1038/nrn.2015.8\u003c/li\u003e\n\u003cli\u003eSong, C., Berridge, K. C. \u0026amp; Kalueff, A. V. \u0026apos;Stressing\u0026apos; rodent self-grooming for neuroscience research. \u003cem\u003eNat Rev Neurosci\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 591 (2016). https://doi.org/10.1038/nrn.2016.103\u003c/li\u003e\n\u003cli\u003eKalueff, A. V., Lou, Y. R., Laaksi, I. \u0026amp; Tuohimaa, P. Abnormal behavioral organization of grooming in mice lacking the vitamin D receptor gene. \u003cem\u003eJ Neurogenet\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, 1-24 (2005). https://doi.org/10.1080/01677060590949683\u003c/li\u003e\n\u003cli\u003eKalueff, A. V. \u0026amp; Tuohimaa, P. Contrasting grooming phenotypes in three mouse strains markedly different in anxiety and activity (129S1, BALB/c and NMRI). \u003cem\u003eBehav Brain Res\u003c/em\u003e \u003cstrong\u003e160\u003c/strong\u003e, 1-10 (2005). https://doi.org/10.1016/j.bbr.2004.11.010\u003c/li\u003e\n\u003cli\u003eKomorowska, J. \u0026amp; Pellis, S. M. Regulatory mechanisms underlying novelty-induced grooming in the laboratory rat. \u003cem\u003eBehav Processes\u003c/em\u003e \u003cstrong\u003e67\u003c/strong\u003e, 287-293 (2004). https://doi.org/10.1016/j.beproc.2004.05.001\u003c/li\u003e\n\u003cli\u003ePires, G. N., Tufik, S. \u0026amp; Andersen, M. L. Grooming analysis algorithm: use in the relationship between sleep deprivation and anxiety-like behavior. \u003cem\u003eProg Neuropsychopharmacol Biol Psychiatry\u003c/em\u003e \u003cstrong\u003e41\u003c/strong\u003e, 6-10 (2013). https://doi.org/10.1016/j.pnpbp.2012.11.006\u003c/li\u003e\n\u003cli\u003eBerridge, K. C., Aldridge, J. W., Houchard, K. R. \u0026amp; Zhuang, X. Sequential super-stereotypy of an instinctive fixed action pattern in hyper-dopaminergic mutant mice: a model of obsessive compulsive disorder and Tourette\u0026apos;s. \u003cem\u003eBMC Biol\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, 4 (2005). https://doi.org/10.1186/1741-7007-3-4\u003c/li\u003e\n\u003cli\u003eWang, C. C. \u0026amp; Shyu, B. C. Differential projections from the mediodorsal and centrolateral thalamic nuclei to the frontal cortex in rats. \u003cem\u003eBrain Res\u003c/em\u003e \u003cstrong\u003e995\u003c/strong\u003e, 226-235 (2004). https://doi.org/10.1016/j.brainres.2003.10.006\u003c/li\u003e\n\u003cli\u003eKita, H. \u0026amp; Kitai, S. T. Amygdaloid projections to the frontal cortex and the striatum in the rat. \u003cem\u003eJ Comp Neurol\u003c/em\u003e \u003cstrong\u003e298\u003c/strong\u003e, 40-49 (1990). https://doi.org/10.1002/cne.902980104\u003c/li\u003e\n\u003cli\u003eInagaki, H. K.\u003cem\u003e et al.\u003c/em\u003e A midbrain-thalamus-cortex circuit reorganizes cortical dynamics to initiate movement. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e185\u003c/strong\u003e, 1065-1081 e1023 (2022). https://doi.org/10.1016/j.cell.2022.02.006\u003c/li\u003e\n\u003cli\u003eCassidy, R. M. \u0026amp; Gale, K. Mediodorsal thalamus plays a critical role in the development of limbic motor seizures. \u003cem\u003eJ Neurosci\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e, 9002-9009 (1998). https://doi.org/10.1523/JNEUROSCI.18-21-09002.1998\u003c/li\u003e\n\u003cli\u003eDevienne, G., Vantomme, G. \u0026amp; Huguenard, J. R. Disrupted Development of the mPFC-Thalamic Circuit in Shank3(-/-) mice, an autism-associated model. \u003cem\u003eMol Psychiatry\u003c/em\u003e (2025). https://doi.org/10.1038/s41380-025-03390-6\u003c/li\u003e\n\u003cli\u003eO\u0026apos;Neill, P. K., Gore, F. \u0026amp; Salzman, C. D. Basolateral amygdala circuitry in positive and negative valence. \u003cem\u003eCurr Opin Neurobiol\u003c/em\u003e \u003cstrong\u003e49\u003c/strong\u003e, 175-183 (2018). https://doi.org/10.1016/j.conb.2018.02.012\u003c/li\u003e\n\u003cli\u003eKim, J., Pignatelli, M., Xu, S., Itohara, S. \u0026amp; Tonegawa, S. Antagonistic negative and positive neurons of the basolateral amygdala. \u003cem\u003eNat Neurosci\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, 1636-1646 (2016). https://doi.org/10.1038/nn.4414\u003c/li\u003e\n\u003cli\u003eFerguson, E. F., Spackman, E., Cai, R. Y., Hardan, A. Y. \u0026amp; Uljarevic, M. Characterizing associations between emotion dysregulation, anxiety, and repetitive behaviors in autistic youth with intellectual disability. \u003cem\u003eAutism Res\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 1810-1817 (2024). https://doi.org/10.1002/aur.3207\u003c/li\u003e\n\u003cli\u003eCao, X.\u003cem\u003e et al.\u003c/em\u003e Cadherin 11 regulates presynaptic vesicle trafficking and neuronal activity in autism-related brain circuit. \u003cem\u003eMol Psychiatry\u003c/em\u003e \u003cstrong\u003e31\u003c/strong\u003e, 192-208 (2026). https://doi.org/10.1038/s41380-025-03141-7\u003c/li\u003e\n\u003cli\u003eMark, V. W. Acute versus chronic functional aspects of unilateral spatial neglect. \u003cem\u003eFront Biosci\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, e172-189 (2003). https://doi.org/10.2741/973\u003c/li\u003e\n\u003cli\u003eGrandi, L. C. \u0026amp; Gerbella, M. Single Neurons in the Insular Cortex of a Macaque Monkey Respond to Skin Brushing: Preliminary Data of the Possible Representation of Pleasant Touch. \u003cem\u003eFront Behav Neurosci\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 90 (2016). https://doi.org/10.3389/fnbeh.2016.00090\u003c/li\u003e\n\u003cli\u003eBoyd, B. A., McBee, M., Holtzclaw, T., Baranek, G. T. \u0026amp; Bodfish, J. W. Relationships among Repetitive Behaviors, Sensory Features, and Executive Functions in High Functioning Autism. \u003cem\u003eRes Autism Spectr Disord\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, 959-966 (2009). https://doi.org/10.1016/j.rasd.2009.05.003\u003c/li\u003e\n\u003cli\u003eRamirez-Armenta, K. I.\u003cem\u003e et al.\u003c/em\u003e Optogenetic inhibition of indirect pathway neurons in the dorsomedial striatum reduces excessive grooming in Sapap3-knockout mice. \u003cem\u003eNeuropsychopharmacology\u003c/em\u003e \u003cstrong\u003e47\u003c/strong\u003e, 477-487 (2022). https://doi.org/10.1038/s41386-021-01161-9\u003c/li\u003e\n\u003cli\u003eGivon, L., Edut, S. \u0026amp; Klavir, O. The role of fear and dopamine-striatal pathways in grooming. \u003cem\u003eNeuropharmacology\u003c/em\u003e \u003cstrong\u003e269\u003c/strong\u003e, 110323 (2025). https://doi.org/10.1016/j.neuropharm.2025.110323\u003c/li\u003e\n\u003cli\u003eMondragon-Gonzalez, S. L., Schreiweis, C. \u0026amp; Burguiere, E. Closed-loop recruitment of striatal interneurons prevents compulsive-like grooming behaviors. \u003cem\u003eNat Neurosci\u003c/em\u003e \u003cstrong\u003e27\u003c/strong\u003e, 1148-1156 (2024). https://doi.org/10.1038/s41593-024-01633-3\u003c/li\u003e\n\u003cli\u003eMerrick, C. \u0026amp; Filingeri, D. The evolution of wetness perception: A comparison of arachnid, insect and human models. \u003cem\u003eJ Therm Biol\u003c/em\u003e \u003cstrong\u003e85\u003c/strong\u003e, 102412 (2019). https://doi.org/10.1016/j.jtherbio.2019.102412\u003c/li\u003e\n\u003cli\u003eOblong, J. E. Comparison of the impact of environmental stress on male and female skin. \u003cem\u003eBr J Dermatol\u003c/em\u003e \u003cstrong\u003e166 Suppl 2\u003c/strong\u003e, 41-44 (2012). https://doi.org/10.1111/j.1365-2133.2012.10928.x\u003c/li\u003e\n\u003cli\u003eWiersielis, K. R.\u003cem\u003e et al.\u003c/em\u003e Sex differences in corticotropin releasing factor-evoked behavior and activated networks. \u003cem\u003ePsychoneuroendocrinology\u003c/em\u003e \u003cstrong\u003e73\u003c/strong\u003e, 204-216 (2016). https://doi.org/10.1016/j.psyneuen.2016.07.007\u003c/li\u003e\n\u003cli\u003eJung, H.\u003cem\u003e et al.\u003c/em\u003e Sexually dimorphic behavior, neuronal activity, and gene expression in Chd8-mutant mice. \u003cem\u003eNat Neurosci\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 1218-1228 (2018). https://doi.org/10.1038/s41593-018-0208-z\u003c/li\u003e\n\u003cli\u003eWilson, C.\u003cem\u003e et al.\u003c/em\u003e Experience-dependent grooming microstructure alterations and gastrointestinal dysfunction in the SAPAP3 knockout mouse model of compulsive behaviour. \u003cem\u003eJ Affect Disord\u003c/em\u003e \u003cstrong\u003e363\u003c/strong\u003e, 520-531 (2024). https://doi.org/10.1016/j.jad.2024.07.143\u003c/li\u003e\n\u003cli\u003eGammie, S. C. \u0026amp; Stevenson, S. A. Effects of daily and acute restraint stress during lactation on maternal aggression and behavior in mice. \u003cem\u003eStress\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 171-180 (2006). https://doi.org/10.1080/10253890600969106\u003c/li\u003e\n\u003cli\u003eKabra, M., Robie, A. A., Rivera-Alba, M., Branson, S. \u0026amp; Branson, K. JAABA: interactive machine learning for automatic annotation of animal behavior. \u003cem\u003eNat Methods\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 64-67 (2013). https://doi.org/10.1038/nmeth.2281\u003c/li\u003e\n\u003cli\u003eOhayon, S., Avni, O., Taylor, A. L., Perona, P. \u0026amp; Roian Egnor, S. E. Automated multi-day tracking of marked mice for the analysis of social behaviour. \u003cem\u003eJ Neurosci Methods\u003c/em\u003e \u003cstrong\u003e219\u003c/strong\u003e, 10-19 (2013). https://doi.org/10.1016/j.jneumeth.2013.05.013\u003c/li\u003e\n\u003cli\u003eLiu, H.\u003cem\u003e et al.\u003c/em\u003e Comprehensive Brain-wide Mapping of Afferent and Efferent Nuclei Associated with the Heart in the Mouse. \u003cem\u003eNeurosci Bull\u003c/em\u003e \u003cstrong\u003e41\u003c/strong\u003e, 1743-1760 (2025). https://doi.org/10.1007/s12264-025-01384-6\u003c/li\u003e\n\u003cli\u003eYue, Y.\u003cem\u003e et al.\u003c/em\u003e Mapping Brain-Wide Neural Activity of Murine Attentional Processing in the Five-Choice Serial Reaction Time Task. \u003cem\u003eNeurosci Bull\u003c/em\u003e \u003cstrong\u003e41\u003c/strong\u003e, 741-758 (2025). https://doi.org/10.1007/s12264-025-01380-w\u003c/li\u003e\n\u003cli\u003eHu, Y.\u003cem\u003e et al.\u003c/em\u003e Comparative brain-wide mapping of ketamine- and isoflurane-activated nuclei and functional networks in the mouse brain. \u003cem\u003eElife\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e (2024). https://doi.org/10.7554/eLife.88420\u003c/li\u003e\n\u003cli\u003eShannon, P.\u003cem\u003e et al.\u003c/em\u003e Cytoscape: a software environment for integrated models of biomolecular interaction networks. \u003cem\u003eGenome Res\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 2498-2504 (2003). https://doi.org/10.1101/gr.1239303\u003c/li\u003e\n\u003cli\u003eKraeuter, A. K., Guest, P. C. \u0026amp; Sarnyai, Z. The Open Field Test for Measuring Locomotor Activity and Anxiety-Like Behavior. \u003cem\u003eMethods Mol Biol\u003c/em\u003e \u003cstrong\u003e1916\u003c/strong\u003e, 99-103 (2019). https://doi.org/10.1007/978-1-4939-8994-2_9\u003c/li\u003e\n\u003cli\u003eDanduga, R. \u0026amp; Kola, P. K. Elevated Plus Maze for Assessment of Anxiety and Memory in Rodents. \u003cem\u003eMethods Mol Biol\u003c/em\u003e \u003cstrong\u003e2761\u003c/strong\u003e, 93-96 (2024). https://doi.org/10.1007/978-1-0716-3662-6_8\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"grooming microstructure, dorsolateral striatum, mediodorsal thalamic nucleus, basolateral amygdala, anxiety, locomotion, autism","lastPublishedDoi":"10.21203/rs.3.rs-9128688/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9128688/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Self-grooming is an evolutionarily conserved behavior characterized by a stereotyped sequence that is often disrupted in neuropsychiatric disease models, such as obsessive‒compulsive disorder (OCD) and autism spectrum disorder (ASD). While the dorsolateral striatum (DLS) has been implicated in modulating the grooming microstructure, the underlying circuit mechanisms remain unclear. By combining automated behavioral analysis, whole-brain activity mapping, and circuit-specific manipulations, we identified two opposing DLS circuits governing this microstructure. The glutamatergic mediodorsal thalamus (MD)→DLS pathway facilitates the core motor program by specifically amplifying mid-sequence grooming phases. Conversely, the GABAergic basolateral amygdala (BLA)→DLS pathway couples grooming with anxiety states by modulating sequence initiation and termination. Relevant disease models show distinct alterations and treatment methods: in stressed mice, fragmented grooming with disrupted termination is rescued by inhibiting the BLA→DLS pathway, whereas in Shank3B knockout (KO) mice, hyperpersistent mid-sequence grooming is normalized by either the alleviation of MD→DLS hyperactivity or the activation of the BLA→DLS pathway. These findings reveal phase-specific circuits that coordinate stereotyped behavior, providing new insights into the pathophysiology of related disorders.","manuscriptTitle":"Opposing thalamic and amygdalar projections to the dorsolateral striatum sculpt the grooming microstructure in stress and autism models","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-26 09:22:15","doi":"10.21203/rs.3.rs-9128688/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"63e7e6b4-9305-4e11-9a99-3c1d47271bca","owner":[],"postedDate":"March 26th, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"revise","date":"2026-05-08T19:06:03+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"This content is not available.","date":"2026-05-04T17:33:05+00:00","index":1,"fulltext":"This content is not available."}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":65065939,"name":"Biological sciences/Neuroscience/Neural circuits"},{"id":65065940,"name":"Health sciences/Diseases/Psychiatric disorders/Autism spectrum disorders"}],"tags":[],"updatedAt":"2026-05-08T19:10:33+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-26 09:22:15","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9128688","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9128688","identity":"rs-9128688","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}