Behavior- and circuit-specific cortico-striatal decoupling during the early development of Parkinson’s disease-like syndrome

ABSTRACT

Despite that cortico-striatal decoupling has been widely reported in individuals diagnosed with Parkinson’s Disease (PD), its onset, circuit specificity and underlying mechanism remain largely unclear. To investigate these questions, dual fiber photometry is established for the first time to evaluate cortico-striatal coupling during varied motor behaviors, whose cell-type resolution was provided by the usage of Cre transgenic mouse lines. Contralateral turning, digging and licking show distinct coupling patterns, among which digging induces the strongest coupling. Striatal D1R-expressed medium spiny neurons (dMSNs) and D2R-expressed MSNs (iMSNs) similarly contribute to the cortical-striatal coupling during turning and licking but not digging, with much tighter coupling between the dMSNs and the M1 cortex. In PD-like mouse model induced via intra-striatal injection of synthetic mouse wildtype α-synuclein preformed fibril (PFF), digging-associated cortical-striatal decoupling emerges as early as 1-month post induction (Mpi), which becomes significant since 2 Mpi and correlates with later-onset behavioral deficit. Notably, impaired dMSNs but not iMSNs mediate this decoupling, which can be rescued by activation of D1 but not D2 receptor. Mechanistically, we found an inverted U-shape decline in striatal dopamine level along the disease development in PFF-injected mice. Supplement with L-DOPA alleviates the decoupling and motor deficit, suggesting that early dopamine deficiency directly contributes to the cortical-striatal decoupling and the associated motor deficit. These findings provide new insights into the temporal profile and mechanisms underlying the PD-associated cortico-striatal decoupling, which has been implicated as functional biomarker for early diagnosis of PD.

INTRODUCTION

Parkinson’s Disease (PD) is one of the most prevalent neurodegenerative disorders that primarily affects patients’ motor function(1, 2). Various brain regions within the basal ganglia loop collaborate to regulate motor behavior(3, 4), with the functional coupling between the primary motor cortex (M1) and the striatum (STR) playing a crucial role in motor planning, execution, and learning(5–7). Functional magnetic resonance imaging (fMRI) revealed the reduced cortico-striatal connectivity, or cortico-striatal decoupling in PD patients(8–10). Furthermore, this PD-associated decoupling is correlated with both disease duration and behavioral impairment of patients(11, 12). Cortico-striatal decoupling is also associated with motor deficits in other disorders such as dystonia(13) and Huntington’s disease(14). Together, these findings underscore the importance of understanding the mechanism of cortico-striatal coupling and characterizing the cortico-striatal decoupling associated with diseases.

D1 receptor- and D2 receptor-expressed medium spiny neurons (MSN) are the two main types of neurons in the striatum(15, 16). Both of them receive glutamatergic inputs from the cortex and are considered to play distinct yet debated roles in motor control via activation of the direct and indirect pathway, respectively(17–20). How these neurons contribute to the cortico-striatal coupling during motor behavior requires further investigation. In dopamine-depleted PD rodent models, impaired activities of D1R-expressed MSN (dMSN) and/or D2R-expressed MSN (iMSN) have been reported(21). While some studies found similar changes in both dMSN and iMSN firing(22) or increased similarity between their discharge frequency(23) after dopamine deletion, many others suggest the imbalanced activity between them might be the main cause of motor deficit(24–27). Whether and how dMSN and iMSN contribute to the cortico-striatal decoupling associated with parkinsonian remains to be answered.

The intrastriatal injection of α-Syn preformed fibril (PFF) effectively induces aggregation of α-Syn and dopaminergic (DA) neuron cell loss in the substantia nigra (SN), two pathological hallmarks of PD(28) along with motor dysfunction that gradually advances(29–31). This PD-like mouse model allows the monitoring of potential alterations in cortico-striatal coupling preceding the onset of motor deficits.

In the present study, we first established a dual fiber photometry system to examine the cortico-striatal coupling with behavioral and pathway specificity. Using the PFF-induced PD-like mouse model, we then analyzed the characteristic as well as the underlying cellular and circuitry mechanism of cortico-striatal decoupling during disease progression. Our findings validate a new method to study functional coupling with cell type and circuit resolution, and for the first time unveil the temporal profile of M1 cortex-dorsal striatum decoupling with the PFF-induced PD-like mouse model, uncover the distinct contributions of dMSN and iMSN.

RESULTS

Dual fiber photometry measures coupling of cortical and striatal activity during different motor behaviors

Several techniques, including electroencephalography (EEG), local field potential (LFP), and fMRI, have been utilized to assess the coupling between neural activities in various brain subregions(32–34). However, these approaches are limited in their capacity to elucidate the circuitry mechanisms underlying such coupling. Considering that striatal neurons predominantly comprise dMSNs and iMSNs, there is a pressing need for a novel method to explore the cell type-specific cortico-striatal coupling during motor behavior in both healthy and diseased brains.

To address this challenge, we employed fiber photometry, which allows for the concurrent monitoring of activity across different cell types in multiple brain regions in behaving mice(35). Two calcium indicators, jRGECO1a and GCaMp7s, were respectively expressed in neurons in the deeper layers of the primary motor cortex (M1) and dorsal striatum (STR) of the same hemisphere, followed by the implantation of two optical fibers (Fig 1A-B; see Methods). Utilizing this dual fiber photometry approach, we successfully monitored neuronal activities of M1 and the dorsal striatum without discernible crosstalk of fluorescence signals (Fig 1B, Supplementary Figure 1A). Three naturalistic mouse motor behaviors, namely contra-lateral turning, licking, and digging, were chosen to investigate cortico-striatal coupling (Fig 1C, 1G, 1K). The coupling index is determined by calculating the Pearson ’s correlation coefficient between the activities of the two channels within a time window around behavior initiation. Additionally, the phase lag is estimated to assess the temporal alignment of the two responses (see Methods). Contralateral turning has been demonstrated to elicit a striatal response(36). Fibers in the striatum consistently detected robust neuronal activity in response to turning initiation (Fig 1D, right), validating the temporal resolution of fiber photometry for monitoring fast motor behavior-evoked neuronal activity. Interestingly, M1 neuronal activity was suppressed mildly during contralateral turning (Fig 1D, left), resulting in a negative coupling between M1 and the striatum (Fig 1E; see Methods). In contrast, both M1 and the striatum were activated during licking, with the striatal response exhibiting slightly earlier onset and being more phasic compared to M1’s response (Fig 1H-I). The distinct dynamics of these responses consequently resulted in a coupling index close to 0 for licking behavior (Fig 1J). During digging, where mice retrieved food pellets buried underneath the bedding (see Methods), both M1 and the dorsal striatum exhibited robust activation upon initiation (Fig 1L). The digging-induced responses from both channels were temporally aligned (Supplementary Figure 1C) and exhibited similar kinetics (Fig 1M), reflected by a coupling index close to 1 (Fig 1N).

Figure 1. Detection of task-specific cortico-striatal coupling via dual fiber photometry imaging

A. Scheme of virus injection and optic fiber implantation.

B. Anatomical verification of virus expression and optic fiber localization. Scale bar: 200μm.

C-N. Examination of the cortico-striatal couplings during contralateral turning (C-F), licking (G-J), and digging (K-N).

(C, G, K) Behavioral diagrams (top) and example Ca²⁺ activities recorded from the primary motor cortex (M1, red) and the striatum (STR, green) with dotted lines denoted onsets of the specified behaviors. Ca²⁺-independent fluorescence was monitored using 405 nm light pulses for individual channel (purple, bottom).

(D, H, L) Heatmaps showing all compiled trials (Top panels) and the averaged Ca²⁺ signal intensities (dF/F0, bottom panels) aligned (dotted lines) to the initiation of the contralateral turning (D), licking (H), and digging (L). All left panels, M1 signal; all right panels, STR signals.

(E, I, M) Temporally aligned traces showing z-scored average signal intensities of both M1 (red) and STR (green).

(F, J, N) Calculated cortico-striatal coupling indices for contralateral turning (F), licking (J), and digging (N).

Thus, we have developed a dual fiber photometry approach to investigate the activity coupling between M1 and the striatum. Our findings reveal distinct coupling patterns, implying that complex rules govern cortico-striatal communication during various motor behaviors.

dMSNs and iMSNs distinctively contribute to the cortico-striatal coupling in different motor behaviors

To investigate the cell type-specific cortico-striatal coupling, GCaMp7s was selectively expressed in either striatal dMSNs or iMSNs by using D1R-cre or D2R-cre mice (Fig 2A-B, see Methods). The dual fiber photometry system recorded clear calcium activities from M1 neurons and striatal dMSNs or iMSNs concurrently (Supplementary Figure 2A-B). Both striatal dMSNs and iMSNs were similarly activated upon contralateral turning and licking (Fig 2C, E), and their coupling with M1 activity remained consistent across cell types (Fig 2D, F). These findings suggest that both cell types in the striatum participate similarly in these two behaviors, consistent with previous studies (36–38). During digging behavior, striatal dMSNs exhibited digging-evoked activity closely resembling that of total striatal neurons and tightly coupled with M1 activity (Fig 2G-H). Notably, although iMSNs exhibited regular activity during free moving (Supplementary Figure 2B), contralateral turning and licking, they displayed a much weaker response during digging (Fig 2G) and significantly reduced coupling with M1 activity compared to that of dMSN-M1 (Fig 2H). These findings further underscore the importance of investigating cell type-specific M1-striatal coupling in motor behavior.

Figure 2. Task- and cell type-specific cortico-striatal coupling

A. Schematic diagram showing the experimental design for analyzing cell type-specific cortico-striatal coupling.

B. Immunofluorescence images showing the colocalization between DRD1 (red) and DIO-GCaMP (green) in D1R-Cre mice (top panels), and the colocalization between DRD2 (magenta) and DIO-GCaMP (green) in the D2R-Cre mice (bottom panels). Scale bar: 10μm.

C, E, G. Z-scored average traces of M1 (red) and STR dMSN (light green) and iMSN (dark green) Ca²⁺ activities temporally aligned to the initiation of contralateral turning (C), licking (E), and digging (G).

D, F, H. Comparison of cell type-specific cortico-striatal coupling indices for turning (D), licking (F), and digging behavior (H, P = 0.0087). Mann-Whitney test. **P<0.01.

Intrastriatal injection of α-syn PFF induces PD-like pathology in wildtype mice

Cortico-striatal decoupling has been widely reported in patients with mild to severe Parkinson’s disease (PD), correlating positively with motor impairment(11, 39, 40). To investigate the onset and mechanism underlying this PD-associated cortico-striatal decoupling, we adapted an inducible PD-like mouse model established previously(41) by bilaterally inoculating α-synuclein (α-Syn) preformed fibrils (PFF) into dorsal striatum of wildtype mice (Fig 3A). Behavioral and pathological phenotypes were examined at both one- and three-month post PFF inoculation (Fig 3B). One-month post-injection (1Mpi), PFF led to a slight reduction in rotarod learning ability (Fig 3C1) without affecting the rotarod test performance (Fig 3C2) or the general locomotor activity (Fig 3C3). By 3-month post-injection, mice receiving PFF displayed notable motor deficits (Fig 3D1, D2), indicating a progressive PD-like motor impairment. Concurrently, we assessed two pathological hallmarks of PD, α-Syn aggregates, and dopaminergic neuron loss, via immunohistochemistry. PFF inoculation in striatum induced significant increase in the phosphorylated α-Syn immunoreactivity (p-α-Syn), a common marker for synucleiopathy, both 1- and 3-month post injection in both the striatum (Fig 3E-F, L-M) and the substantia nigra (SN) (Fig 3H-I, O-P). The rapid formation and spreading of synucleiopathy coincided with a progressive reduction in striatal dopamine transporter (DAT) signal intensity in the striatum (Fig 3G, N) and a decreased in dopaminergic neuronal cell count in SN (Fig 3J, Q). These findings confirm that PFF can initiate the formation of α-Syn aggregates, which can further spread to the SN and trigger the pathology of dopaminergic neurons in a time-dependent manner.

Figure 3. Pathological characterization of α-synuclein preformed fibril (PFF) induced PD-like mouse model

A. Schematic diagram describing the bilateral inoculation of PFF into striatum of wildtype mice. Right panel, transmitted electron microscope image of PFF. Scale bar: 100 nm.

B. Experimental timeline. IHC: Immunohistochemistry.

C-D. Behavioral analysis of the PFF- and PBS-injected mice 1 (C) and 3 (D) months post injection (Mpi). (C1) Learning curve of rotarod running.

(C2) Latency to fall in the rotarod test.

(C3) Total travel distance in the open field test.

(D1) Latency to fall in the rotarod test 3Mpi (P = 0.0207).

(D2) Total travel distance in the open field test 3Mpi (P = 0.0147).

E-K. Immunofluorescence analyses of pathological hallmarks in PFF- and PBS-injected mice 1Mpi. All data normalized to PBS control.

(E-G) Comparison of the p-α-Syn intensity (F, P = 0.0079) and dopamine transporter (DAT) signal intensities (G) in STR. E, Representative immunofluorescence images. Scale bars: 50 μm.

(H-K) Comparison of the p-α-Syn intensity (I, P = 0.0043), DA neuron number (J, P = 0.0022) and DAT intensity (K) in substantia nigra (SN). H, Representative immunofluorescence images. Scale bars: 250 μm.

L-R. Immunofluorescence analyses of pathological hallmarks in PFF- and PBS-injected mice 3Mpi. All data normalized to PBS control.

(L-N) Comparison of the p-α-Syn intensity (M, P = 0.0022) and dopamine transporter (DAT) signal intensities (N, P =0.0411) in STR. L, Representative immunofluorescence images. Scale bars: 100 μm. (O-R) Comparison of the p-α-Syn intensity (P, P =0.0022), DA neuron number (Q, P = 0.0079) and DAT intensity (R) in SN. O, Representative immunofluorescence images. Scale bars: 250 μm.

Mann-Whitney test, unless otherwise noted. Two-way ANOVA with Sidak’s multiple comparisons test (C1). *P<0.05, **P<0.01.

Behavior-specific cortico-striatal decoupling emerges early in the PFF-injected mice

Utilizing the PFF-induced PD-like mouse model, we aimed to investigate the onset timing and mechanisms underlying cortico-striatal decoupling during disease progression by using the dual fiber photometry method. PFF inoculation had no discernible effect on the fiber-based calcium imaging (Supplementary Figure 3A-C). Given its robustness and circuitry dependency, we compared digging-evoked M1-dorsal striatum coupling between PFF and PBS-injected mice. At one month post-injection, cortico-striatal coupling largely persisted (Fig 4B-C), albeit with a moderate increase in trial-by-trial phase lags among PFF mice compared to PBS counterparts (Fig 4B, C3), indicating the emergence of early temporal decoupling. By the second month post-injection, PFF mice exhibited a further increase in phase lag relative to PBS controls (Fig 3D, E2, E3), accompanied by a declining trend in the cortico-striatal coupling index (Fig 4E1) while cortical and striatal response amplitudes remained unchanged (Supplementary Figure 3E). Following three to four months of PFF inoculation, decoupling in PFF mice worsened, contrasting with sustained tight coupling observed in PBS mice between M1 and the striatum during digging (Fig 4F, G1). This decoupling in PFF mice were attributed to both elevated phase lag (Fig 4G2, G3) and a significant reduction in striatal response amplitude (Fig 4H). Temporal profiles of the coupling index (Fig 4I, left), phase lag (Fig 4I, right), and response amplitude (Supplementary Figure 3F-G) collectively suggest an initial impairment in temporal precision of cortico-striatal communication in this PD-like mouse model, with cortico-striatal decoupling exacerbating as pathology progresses. We subsequently investigated whether the cortico-striatal decoupling in digging affect the digging behavior. Digging duration and frequency were quantified and compared between groups (See Methods). It was found that until three months post-injection PFF mice exhibited a shortened digging duration (Fig 5A) and increased frequency (Fig 5B) relative to the PBS controls. Furthermore, the phase lag and coupling index, but not the striatal response amplitude (Supplementary Figure 4C-D), strongly correlated positively (Fig. 5C) and negatively (Fig. 5E) with digging frequency. Conversely, they exhibited opposite correlations with digging duration (Fig 5D, F). These findings suggest that the functional decoupling between M1 and dorsal striatum emerges early, predicting the later-onset behavioral deficit.

Figure 4. Task-specific cortico-striatal decoupling in PD-like mice

A. Experimental timeline.

B-I, Dual fiber photometry imaging of M1 and STR in PFF- and PBS-injected mice at 1 Mpi (B-C), 2 Mpi (D-E), and 3-4 Mpi (F-I) during digging.

(B, D, F) Z-scored average traces of M1 (red) and STR (green) Ca²⁺ activities temporally aligned to the initiation of digging. Left, PBS; Right, PFF.

(C, E, G) Quantitative analyses of the cortico-striatal coupling. (C1, E1, G1) Coupling indices (E1, P = 0.0649. G1, P = 0.0037). (C2, E2, G2) Phase lags (E2, P = 0.0281. G2, P = 0.0082).

(C3, E3, G3) Cumulative distribution of phase lag per trial (C3: PBS = 140 trials, PFF = 124 trials. E3: PBS = 109 trials, PFF = 104 trials; P = 0.0041. G3: PBS = 102 trials, PFF = 118 trials; P = 0.0009. Kolmogorov-Smirnov test). Inserts, Comparison of the trial average phase lag (In C3, P = 0.0188. In E3, P = 0.0072. In G3, P = 0.0110).

(H) Plateau amplitude of M1 (left) and STR (right) Ca²⁺ activities at 3∼4Mpi.

(I) The temporal profiles of coupling index (left) and phase lag (right).

J-K. Dual fiber photometry imaging of M1 and STR at 4 Mpi during licking.

(J) Z-scored average traces of M1 (red) and STR (green) Ca²⁺ activities temporally aligned to the initiation of licking.

(K) Comparison of the coupling indices.

(L) Comparison of the phase lags.

Mann-Whitney test, unless otherwise noted. *P<0.05. **P<0.01. ***P<0.001.

Figure 5. Digging-associated cortico-striatal decoupling correlated with specific behavioral deficits

A. Duration of a single digging event. Left, the temporal profiles of average digging duration. Right, comparison of the duration at 3∼4Mpi (P = 0.0022).

B. Frequency of digging events. Left, temporal profiles of digging frequency. Right, comparison of the frequency at 3∼4Mpi (P = 0.0152).

C, D. Correlation analysis of phase lag between digging frequency (C) and duration (D) from PBS- and PFF-injected mice at 3∼4Mpi. r: Pearson’s correlation coefficient. Each dot represents data from an individual mouse (the same below).

E, F. Correlation analysis of coupling index between digging frequency (E) and duration (F) from PBS- and PFF-injected mice at 3∼4Mpi.

Mann-Whitney test. *P<0.05. **P<0.01.

Of note, although the cortico-striatal coupling during digging was compromised in PFF mice, it remained unaltered during licking even after 4 months of PFF injection (Fig 4K-M, Supplementary Figure 3H), at which point PFF mice displayed normal licking behavior (Supplementary Figure 4E-H). This result suggests that the PD-associated decoupling and motor deficit at early stages is behavior-specific.

Defects in dMSN neuronal activity mediates the digging-associated cortico-striatal decoupling in PFF-injected mice

A functional imbalance between the dMSNs and iMSNs has been proposed to underlie the bradykinesia in PD patients(16). Therefore, we explored whether there is a cell-type effect in the observed early-onset cortico-striatal decoupling in the PFF-induced PD-like mice. To investigate this, we employed a methodology akin to that depicted in Figure 2A (Fig 6A). Additionally, D2R-Cre mice were also utilized to examine the M1-dMSN coupling by expressing CreOFF-GCaMP in the striatum (See Method), which could effectively label dMSNs (Supplementary Figure 5A) and generate reliable dMSNs responses during various behaviors (Supplementary Figure 2C-F, Supplementary Figure 5B). We then tracked cortico-striatal coupling in a pathway-specific manner beginning at 2 months post-injection, when decoupling becomes apparent in PFF mice. The findings revealed that digging-related cortical and striatal dMSN responses were markedly decoupled in PFF mice as early as 2 months post-injection (Fig 6B), as evidenced by a significantly decreased coupling index (Fig 6C1) and an increased phase lag (Fig 6C2). This M1-dMSN decoupling exacerbated further at 3∼4 months post-PFF injection, coinciding with a notably diminished striatal dMSN response compared to PBS mice (Fig 6D-E). In contrast to the M1-dMSN decoupling, coupling between M1 and striatal iMSN during digging remained largely unaffected in the PFF mice throughout the experimental time course (Fig 6F-J), indicating a cell type-selective impairment in dMSNs contributing to the cortico-striatal decoupling in this PD-like mouse model. Moreover, PFF inoculation had no discernible effect on the licking-evoked coupling between M1 and dMSN or iMSN (Supplementary Figure 5G-H), suggesting that cell-type-specific cortico-striatal decoupling in PD is also behavior-specific.

Figure 6. Cell type-specific contribution of digging-related cortico-striatal decoupling in PFF-injected mice

A-E. Dual fiber photometry imaging of M1 and striatal dMSN in PFF- and PBS-injected mice at 2 Mpi (B-C), and 3-4 Mpi (D-E) during digging.

(A) Scheme of virus injection and optic fiber implantation. Note the CreON-GCaMP virus was injected into the STR of D1R-Cre mice, while CreOFF-GCaMP virus was injected into the STR of D2R-Cre mice.

Right panel, experiment timeline.

(B, D) Z-scored average traces of M1 (red) and STR dMSN (aqua) Ca²⁺ activities temporally aligned to the initiation of digging. Left, PBS; Right, PFF.

(C, E) Quantitative analyses of the M1-dMSN coupling and plateau amplitude of dMSN Ca²⁺ activities at 2 Mpi (C) and 3-4 Mpi (E). The hollow signal represents D1R-Cre mice; the cross signal represents D2-Cre mice.

(C1, E1) Coupling indices (C1, P = 0.0043. E1, P = 0.0062). (C2, E2) Phase lags (E2, P = 0.0043. G2, P = 0.0062).

(C3, E3) Plateau amplitude of dMSN Ca²⁺ activities (E3, P = 0.0451).

F-J. Dual fiber photometry imaging of M1 and striatal iMSN in PFF- and PBS-injected mice at 2 Mpi (G-H), and 3-4 Mpi (I-J) during digging. Right panel, experiment timeline.

(F) Scheme of virus injection and optic fiber implantation.

(G, I) Z-scored average traces of M1 (red) and striatal iMSN (dark green) Ca²⁺ activities temporally aligned to the initiation of digging. Left, PBS; Right, PFF.

(H, J) Quantitative analyses of the M1-iMSN coupling and plateau amplitude of iMSN Ca²⁺ activities at 2 Mpi (H) and 3-4 Mpi (J).

(H1, J1) Comparison of the coupling indices. (H2, J2) Comparison of the phase lags.

(H3, J3) Comparison of the plateau amplitude of iMSN Ca²⁺ activities. Mann-Whitney test. *P<0.05. **P<0.01.

Altered locomotion-associated striatal dopamine level in the PFF-injected mice

Insufficient striatal dopamine released from SN dopaminergic terminals is the primary cause of the disturbed basal ganglia circuit that mediates the motor deficits in PD(16). Recent studies have indicated a correlation between dopamine depletion and abnormal functional connectivity in both rodent models and PD patients(42–45). Consequently, we hypothesized that abnormal dopamine level might underlie the early-onset, cell-type-specific cortico-striatal decoupling observed in PFF-induced PD mice. To monitor extracellular dopamine level at various time points, we expressed the dopamine sensor DA4.4 in striatal neurons (Fig 7A-B). In PBS mice, the fluorescence signal remained stable throughout the 3-month recording period (Supplementary Figure 6A). The extracellular dopamine level was first examined when PFF and PBS mice freely explored new cages (Fig 7C). PFF mice exhibited normal general activity (Supplementary Figure 6B) and unaltered dopamine signals (Fig 7D-E), measured as the summation of the fluorescent intensity during the exploration (see Methods). Since exploration of new environments primarily triggers novelty-seeking behavior driven by dopamine release from the ventral tegmental area (VTA)(46, 47), these findings suggest minimal impact of PFF inoculation on VTA-mediated dopamine release up to 3 months post-injection. Dopaminergic neurons in the substantia nigra compacta (SNc) are pivotal for voluntary movement(40, 48). To assess whether dopamine release by SNc dopaminergic neurons is affected, mice were trained to run at a constant speed on a treadmill (Fig 7F). Both PBS and PFF mice maintained their performance up to 3-month post-injection (Supplementary Figure 6C). Interestingly, dopamine signal during constant speed running exhibited a inverted U-shaped pattern of changes in PFF mice compared to PBS mice, with significant reductions at 3-4 weeks and 10-14 weeks post-PFF injection, but a transient recovery at 6-7 weeks (Fig 7G-H). These findings suggest impaired SNc-mediated dopamine release in PFF mice.

Figure 7 Cortico-striatal decoupling in PFF-injected mice accompanied by deficiency of striatal extracellular dopamine

A. Schematic diagram showing the experimental design for assessing striatal extracellular dopamine (DA) level. DA4.4 is a florescence-based DA sensor. Right panel, anatomical verification of virus expression and optic fiber localization. Scale bar: 200μm.

B. Experimental timeline.

C. Schematic diagram for free-exploring paradigm.

D. Representative magnitude heatmap of normalized (dF/F0) of DA sensor activity 3∼4 weeks post-injection (wpi), 6∼7wpi, 10wpi, and 13∼14wpi from one PBS-injected mouse and one PFF-injected mouse during 2-minute free-exploring.

E. Normalized sum activity of DA sensor during free-exploring from PBS- and PFF-injected mice 3∼4wpi, 6∼7wpi, 10wpi, 13∼14wpi. Data batch-normalized to the PBS group (the same below).

F. Schematic diagram for treadmill running paradigm.

G. Representative magnitude heatmap of normalized (dF/F0) of DA sensor activity 3∼4wpi, 6∼7wpi, 10wpi, and 13∼14wpi from same two mice as in (D) during 2-minute treadmill running.

H. Normalized sum activity of DA sensor during free-exploring from PBS- and PFF-injected mice at 3∼4wpi (P = 0.0007), 6∼7wpi, 10wpi, 13∼14wpi (P = 0.0043).

Mann-Whitney test. **P<0.01. ***P<0.001.

Activation of D1R but not D2R rescues the cortico-striatal decoupling and behavioral defects of the PFF-injected mice

To investigate whether the impaired dopamine release is responsible for the dMSN-mediated cortico-striatal decoupling and related behavioral deficit, PFF mice at 4-month post-injection received intraperitoneal (i.p.) administration of L-dopa, SKF-81297 (SKF, a D1R receptor agonist), quinpirole (Quin, a D2R receptor agonist), or saline (SAL) (Fig 8A). Consistent with our previous findings, both L-dopa and SKF, but not Quin, rescued the digging-associated decoupling phenotypes, reflected by decreasing phase lag (Fig 8C) and a trend of increasing coupling index in PFF mice (Fig 8D). Besides, SKF but not Quin enhanced the striatal response amplitude in them (Fig 8E). These results further support the notion that at this stage of PD development, deficient striatal dopamine leads to cortico-striatal decoupling via impairing dMSN neuronal activation during digging.

Figure 8 Selective activation to D1R rescues both digging-associated cortico-striatal decoupling and digging behavioral defects

A. Schematic diagram for pharmacological experiments. i.p., intraperitoneal administration.

B-E. Dual fiber photometry imaging of M1 and STR in PFF- and PBS-injected mice at 4Mpi during digging, after ip. SAL (saline), L-dopa, SKF (SKF-81297, a D1R selective agonist), or Quin (quinpirole, a D2R selective agonist).

(B) Z-scored average traces of M1 (red) and STR (green) temporally aligned, digging-associated Ca²⁺ activities after administration of drugs. Left, PBS; Right, PFF.

(C, D, E) Quantitative analyses of the digging-associated cortico-striatal coupling (C,D) and amplitude of striatal Ca²⁺ activities (E) after administration of drugs.

(C1) Comparison of the phase lags (SAL-PBS vs. SAL-PFF: Mann-Whitney test, P = 0.0152. PFF groups: SAL-PFF as control, F(1.588, 9.529) = 9.270, SAL-PFF vs. Ldopa-PFF, P =0.0156; SAL-PFF vs. SKF-PFF, P = 0.0328).

(C2) Comparison of the phase lags, after saline or quinpirole administration.

(D1) Comparison of the coupling indices (SAL-PBS vs. SAL-PFF: Mann-Whitney test, P = 0.0043). (D2) Comparison of the coupling indices, after saline or quinpirole administration.

(E1) Comparison of the plateau amplitude of striatal activity (SAL-PBS vs. SAL-PFF: Mann-Whitney test, P = 0.0424. PFF groups: SAL-PFF as control, F(1.453, 8.719) = 6.152, SAL-PFF vs. SKF-PFF, P = 0.0459).

(E2) Comparison of the plateau amplitude of striatal activity, after saline or quinpirole administration. F-G. Effect of dopaminergic drugs on digging behavior.

(F) Comparison of digging duration (PFF groups: SAL-PFF as control, F(1.463, 7.317) = 19.97, SAL-PFF vs. Ldopa-PFF, P = 0.0018; SAL-PFF vs. SKF-PFF, P = 0.0388).

(G) Comparison of digging frequency (PFF groups: SAL-PFF as control, F(2.120, 10.60) = 11.78, SAL-PFF vs. SKF-PFF, P = 0.0143; SAL-PFF vs. Quin-PFF, P = 0.0192).

One-way ANOVA followed by Dunnett’s multiple comparisons test, unless otherwise noted. Wilcoxon matched-pairs test (C2, D2, E2). *P<0.05. **P<0.01.

Subsequently, the behavioral effects of these drugs were tested. Both L-dopa and SKF improved the digging performance of the PFF mice (Fig 8F-G) without changing their general locomotion (Supplementary Figure 7C). Moreover, SKF significantly increased the success rate, further supporting that targeting dMSNs alone is more beneficial for the digging behavior (FigureS7D). In contrast, Quin did not rescue the digging performance (Fig 8F, Supplementary Figure 7D), consistent with the absence of digging-associated M1-iMSN decoupling in PFF mice.

DISCUSSION

Many clinical studies have revealed abnormal cortico-striatal functional coupling in patients diagnosed with Parkinson’s disease(8–11), among which the decoupling between M1 and striatum has been found to tightly correlate with parkinsonian(49–51) even in the early stages of the disease(51). Therefore, abnormal brain connectivity has been proposed as a functional biomarker for early diagnosis of neurodegenerative diseases including PD(52, 53) and Alzheimer’s disease(54, 55). The phenotypic characterization and the underlying mechanistic investigation of such decoupling in PD animal model may provide insight to facilitate not only the early diagnosis but also intervention.

Despite the development of various methods to detect inter-region brain coupling, such as LFP, EEG, and fMRI(56), none of these techniques offer cell type-specific precision. Our current study establishes a dual fiber photometry imaging method to investigate M1-striatal coupling with circuitry resolution by using cell type-specific cre mouse lines. Although photometry imaging has lower temporal resolution compared to LFP and EEG(57, 58), our method demonstrates its capability of differentiating distinct M1-striatal coupling patterns associated with different motor behaviors (Fig 1). Among the three behaviors, spontaneous contralateral turning as a reflexive behavior(59) evokes striatal but no obvious M1 response, resulting in the weakest coupling reflected by the negative coupling index. Both tone-cued licking and food pellet-driven digging evoke strong M1 and striatal activities, which are tightly coupled with each other in particularly during digging. Interestingly, we found that only coupling associated with digging, but not the other two motor behaviors with relatively weak coupling, displays early abnormality in the PD-like mice. These findings suggests that motor tasks requiring effective cortico-striatal co-activation might facilitate the early detection of PD-like pathological decoupling, which is supported by neuroimaging studies showing task-related fMRI such as finger tapping(50, 60, 61) detects abnormal brain activity and/or connectivity in early-stage PD(62).

Our current method further allows us to determine the contribution of dMSNs and iMSNs to the cortico-striatal coupling. Consistent with previous studies(37, 63), we found that both types of MSNs were activated during contralateral turning and licking, validating the accuracy of our method. Such co-activation results in comparable contributions of dMSNs and iMSNs to the cortico-striatal coupling during these behaviors. In contrast, our results reveal that M1 couples much more strongly with dMSNs than iMSNs during digging behavior, mainly due to the greater and more timely activation of dMSNs. Our findings are in line with previous studies using methods such as in vivo two-photon imaging (64) or cell type specific optogenetic manipulation(19) that show distinct activation patterns of the dMSNs and iMSNs during motor behaviors such as lever pressing(19, 64). Therefore, by simultaneously monitoring M1 and striatal dMSNs or iMSNs activities, we provide evidences for the first time that the M1-striatal coupling is regulated in a behavior- and circuit-specific manner.

Among all three motor behaviors we tested, only cortico-striatal coupling associated with digging is impaired even before the behavioral deficit manifested. Digging is also the only behavior evoking unequal activity of dMSNs and iMSNs. which may render it more susceptible to PD pathology. Striatal MSNs form functional units, whose activity is responsible for driving specific behavior and is tightly regulated by the complex lateral inhibitory network(65). Indeed, we found the significantly impaired digging-evoked dMSN but not iMSN activity correlates with the behavioral deficit. It is noteworthy that the phase lag, a parameter describing the relative timing of the M1 and striatal responses, shows the strongest correlation with the digging performance (Fig 5C-F). This suggest that the effective connectivity between the M1 and the striatum also, if not more, potently regulates specific behaviors.

Weakening of dMSNs while strengthening of iMSNs is observed in PD rodent models mimicking late stages of PD, which lead to imbalance activity of the direct and indirect pathways and are considered to mediate the motor dysfunction in PD(16, 66). Here we found that the activity of dMSNs but not iMSNs is impaired during digging early in the PFF-induced PD-like mice. A previous study also found a selective disconnection of dMSNs but not iMSN with the M1 in a 6-OHDA-induced partial dopamine lesion mouse model(67). These findings together suggest that dMSNs might be more vulnerable at the early stages of PD development. Interestingly, recent research has demonstrated that gene therapy selectively activating striatal dMSNs can effectively reverse PD-like symptoms in both rodent and primate models of PD(68). Here we also showed that administration of D1R agonist but not D2R agonist, rescued both the digging-associated decoupling and behavioral deficits. Therefore, dMSNs might be better targets for early PD intervention.

The lack of dopamine (DA) from SN dopaminergic terminals in the striatum can cause dysfunction in the basal ganglia circuit and consequently lead to motor defects in PD patients(16, 69–71). Research has revealed abnormal coordination between cortical and striatal ensembles in acutely dopamine-depleted mice(72). Moreover, a clinical study reported an increase in M1-putamen connectivity after L-DOPA intake in some PD patients(73). These findings collectively link the cortico-striatal coupling with DA level in the PD state. In fibril-induced PD-like mice, a reduction in dopaminergic neurons in SNpc begins as early as 4 weeks post inoculation, which is consistent with a significant reduction in the striatal DA level during constant speed running. Interestingly, DA release transiently recovers 6-7 weeks post inoculation, potentially due to a compensatory increase in spiking of the remaining dopaminergic neurons in the SNpc(74). Therefore, the impaired DA level during motor behavior may mediate the decoupling and behavioral deficit observed in the PD-like mice. Indeed, supplement of L-dopa almost fully rescues these defects. Accumulating evidence suggests that abnormal axonal dopamine release could occur before dopaminergic cell lost(71, 75), further supporting the notion that dopamine-dependent, motor task-specific cortical-striatal decoupling might serve as an functional biomarker for early PD.

Another mechanism regulating the cortico-striatal coupling might involve striatal fast-spiking interneurons (FSIs), who is well known to dictate the temporal precision of neural information processing(76–78). Within the striatal functional units that drive specific motor behavior, potent lateral inhibitions are provided by the soma-targeting FSIs in addition to the MSN collaterals(65). In addition, these FSIs are found to be activated faster with lower thresholds compared to MSNs(79, 80), making them ideal for regulating the timing and dynamic of the striatal response. Moreover, FSI activity is sensitive to DA(81) and its alteration has been shown to exacerbate striatal imbalance(24). It remains to be determined whether the impaired DA level in the early stage of PD affects cortico-striatal coupling by altering FSI activity.

CONCLUSION

Our study establishes a methodology utilizing fiber photometry to detect motor task-driven inter-regional brain coupling with cell-type resolution. By employing this approach, we reveal the behavior-specific M1-striatum decoupling preceding the behavioral deficit in the α-Syn PFF-induced PD-like mice, supporting the use of motor task-specific cortico-striatal decoupling as a functional biomarker of early PD. Additionally, we found that the decoupling associated with digging is selectively mediated by impaired dMSNs in a dopamine-dependent manner, unveiling for the first time the circuitry mechanism mediated task-specific decoupling in the PD state.

Author contributions

Conceptualization, K.-W.H.; Primary Investigation, X.-R.Y.; Assistant Investigation, Y.L., W.-T.Z.; Data Analysis, X.-R.Y., Y.L.; Writing, X.-R.Y., Y.L. and K.-W.H.; Funding Acquisition, K.-W.H; Supervision, K.-W.H.

List of abbreviations

PD

Parkinson’s disease

M1

Primary motor cortex

Str

Striatum

SN

Substantia nigra

DA

Dopamine

DAT

Dopamine transporter

D1R

Type 1 dopamine receptor

D2R

Type 2 dopamine receptor

MSN

Medium spiny neuron

dMSN

direct pathway MSN

iMSN

indirect pathway MSN

α-Syn

α-Synuclein

PFF

Preformed fibril

Mpi

Month(s) post injection

METHODS

Animals

All procedures were approved by the Institutional Animal Care and Use Committees at the Interdisciplinary Research Center on Biology and Chemistry (IRCBC), Chinese Academy of Science. C57BL/6J mice were purchased from Shanghai Lingchang Biotechnology Co., Ltd.. The transgenic mice (D1R-Cre: MMRRC_034258-UCD; D2R-Cre: MMRRC_034258-UCD) were bred in-house. Mice were group housed with a 12 h:12 h light:dark cycle and ad libitum access to food and water until experiments. Male mice of 2-8 months of age were used with exception stated specifically. All experiments were carried out at ZT (Zeitgeber time) 15∼19.

Preparation and characterization of mouse wildtype α-Syn PFF

The procedures followed the same protocol as described previously(31). In brief, pET22 vector plasmid containing α-syn gene was transfected into BL21 (DE3) cells. α-syn expression was induced by IPTG (isopropyl-1-thio-D-galactopyranoside). Cells were then harvested and sonicated followed by boiling, streptomycin (20 mg/mL) treatment, pH adjustment, and dialysis overnight in turn. The α-syn monomer was then incubated in solution containing 100 μM in 50 mM Tris, pH 7.5, 150 mM KCl, and 0.05% NaN3 buffer at 37 °C with constant agitation (900 rpm) in ThermoMixer (Eppendorf) for one week to form fibril structure. The fibril was quantified by subtracting the residual soluble α-syn from the total amount of α-syn monomer and the fibril pellet was suspended with PBS to 2 μg/μL. The obtained fibril was sonicated (20% power 15 times (1 s on, 1 s off) on ice, JY92-IIN sonicator) and the structure was confirmed by TEM (FEI Company, USA).

Stereotaxic injection

Craniotomy was performed as described previously(31, 76). Briefly, mice were anesthetized with isoflurane vapor (1∼2% in air) and head-fixed in a stereotaxic frame (RWD Life Science, China) with heating pad (∼37 °C) underneath to maintain body temperature. The scalp was removed, and virus/fibril/PBS was delivered at a rate of 80 nl/min to either M1 (bregma/midline/depth: 1.77/1.5/-0.67 mm) or dorsal lateral striatum (dlSTR. For GCaMP7s: -0.11/2.5/2.3 mm; for PFF/PBS, DA sensor, 0.2/2.0/2.6 mm) by using a glass pipette connected to a micro syringe pump (Stoelting, USA). For photometry calcium imaging, 150 nl of AAV-hSyn-NES-jGRECO1a was injected into M1 layer V; 350 nl of AAV-hSyn-GCaMP7s (or 400nL AAV-hSyn-DIO-GCaMP7s for Cre mice) was injected into dlSTR. To measure the striatal extra-cellular dopamine level, 400nl of AAV-hSyn-GRabeen-DA4.4 was injected into dlSTR. All viruses were diluted to a titer of 10¹² vg./ml. For establishing the α-Syn PFF model, PFF or PBS (0.1 μl/g body weight, 2.5 μl at most) was injected at a speed of 0.5 μl/min for the first 0.2 μL and 0.2 μL/min for the rest.

Fiber photometry surgery and imaging

Optical fibers (NA: 0.37; diameter: 200μm; Inper, China) were implanted using the same M1 and dlSTR coordinates except 0.05 mm less in depth right after the virus injection. Fiber was secured to the skull using dental cement (C&M Super Bond, Japan). Mice were then housed individually and allowed to recover for at least 2 weeks before imaging.

Photometry recording during behavioral paradigms was conducted during the dark cycle utilizing a commercial multichannel fiber photometry system (Inper, China). Excitation light at wavelengths of 470 nm, 561 nm, and 410 nm was used to excite GCaMP7s/DA-4.4, RGECO1a, and Ca²⁺-independent internal channel control, respectively. Fluorescence signal was collected at 40 Hz using InperStudio (MultiColorEVAL15) software, which automatically synchronized the calcium signal with video footage from a camera positioned above the arena. Throughout the imaging, the patch cord attached to the optical fiber was suspended by a helium-filled balloon to minimize its weight impact on the mouse behavior. All behavioral trials lasted for less than 30 min.

Immunofluorescence (IF) staining

IF was performed as previously describe(31). Briefly, mice were anesthetized with isoflurane vapor and perfused with PBS followed by about 40 ml of 4% paraformaldehyde (PFA). Subsequently, whole brain was removed and fixed overnight in PFA at 4°C. The fixed brains were then saturated in PBS containing 20% sucrose for overnight to ensure the mice brains sinking bottom, and 30% sucrose for 2 d. Afterward, the brains were embedded in tissue freezing medium (OCT, Leica) and rapidly frozen and stored at -80°C. Next, 20 μm coronal slices were sectioned using the cryostat (Leica, CM3050s-1-1-1). These sections were incubated in blocking solution (PBS containing 5% normal goat serum (Sigma) and 0.1% Triton X-100) for 2 hours at room temperature, followed by overnight incubation at 4°C with the primary antibody diluted in the blocking solution. After incubation, the primary antibody solution was washed out three times with PBST (0.1% Tween-20 in PBS) and once with PBS, followed by incubation with the fluorophore-conjugated secondary antibodies for 2 hours at room temperature. For MSN type-specific labeling, DAPI (diluted at 1:500) was then used to stain the nuclei for 10 minutes at room temperature. Information of all antibodies used were summarized in Supplementary Table 1.

To validate the expression and location of the calcium and dopamine sensor, PFA-fixed brains were sectioned into 60 μm coronal slices using a vibratome (Leica VT 1000S). Slices with optical fiber track were selected for subsequent fluorescence imaging. Slices were further stained for either D1R or D2R for cell-type specific colocalization.

Confocal Imaging

Fluorescence images were captured using a spinning disk confocal microscope (Andor, UK). Different lasers with wavelengths of 630 nm, 561 nm, 488 nm, and 405 nm were sequentially utilized to excite the Fluor 647, Fluor 568 or RGECO1a, Fluor 488 or GCaMP7s or DA4.4, and DAPI respectively. Imaging was performed using a 20× air (NA = 0.75), or 40× water (NA = 1.25) objective, with consistent settings maintained for each set of experiments. The image resolution is 0.6 μm:0.6 μm/pixel (x:y) for 20× and 0.15 μm:0.15 μm/pixel (x:y) for 40× objective. Z-stack images were acquired for all brain regions and then projected in the z-direction to obtain the maximum intensity for subsequent analysis in ImageJ. Step size (μm) and thickness (μm) of each brain region were as followed: SN: 0.5, 1; Striatum: 1, 5 (for characterization of the PFF-induced pathology); Others: 2, 10. The whole view of STR was analyzed, while the SN region was manually selected based on the mouse brain atlas. The average intensity of phosphorylated α-synuclein (P-α-syn) and dopamine transporter (DAT) was measured using the same background subtraction and contrast transformation applied to each dataset, and then normalized to the control (PBS group). DAT-positive dopamine neurons in the Substantia Nigra pars compacta (SNpc) were automatically selected and counted using the Spot module of Imaris (Oxford Instruments), with settings consistent with previous descriptions(41).

Behavioral task or test

a. Digging test

Mice were food deprived for 24h before behavioral digging test. During the deprivation, two food pallets (25mg, sucrose and milk powder in a 1:1 ratio. Xietong Bio-engineering Co., Ltd., China) were placed in its home cage to introduce novel food. The test environment consisted of a lidless, semitransparent box (FENGSHI group, China), identical in size and texture to the mouse’s home cage (l:w:h: 29 cm×18 cm×13 cm), serving as the arena for the digging test. The bottom of the arena was covered with approximately 3cm-thick clean bedding, and twelve food pallets were evenly buried in the bedding. Additionally, a solitary food pallet was strategically positioned at the center of the arena’s surface on top of the bedding, easily accessible for the mice upon entering the arena, thereby enticing them to dig for buried pallets. The mice’s performance in the arena was recorded using two cameras—one providing a top view and the other a side view—for subsequent post-hoc analysis. After completing the test, the bedding was discarded, and the arena was cleaned with 75% ethanol.

b. Sound-cued licking task

Mice were habituated and trained for three days prior the final experiment.

Day0. Completely water deprivation

Day1. Mice were required to lick water from the water outlet connected to a sensor. Once the mice lick the outlet, a single dose of 5μl water was pumped out from the outlet. Each mouse repeated this process 100 times.

Day2. A sound cue (3000Hz, 200ms) with random interval (10-20 s) was introduced, and mice had to associate it with water. 5μl water was pumped out if mice licked the outlet within 3s after the cue; otherwise, they received nothing. A 5-second interval allowed mice to consume the water before moving to the next random time interval. Each mouse was allowed to complete 80 water trials

Day3. The random time interval from Day 2 was replaced with a 5-second non-licking interval. Mice had to refrain from licking the outlet for 5s to trigger the sound cue. If they licked during this interval, they had to wait another 5s for the sound cue. The time window for mice to respond to the cue and to consume water remained the same as Day 2. Each mouse was allowed sufficient time to complete 120 watered trials.

Day4. Fiber photometry recording was conducted while mice performed the task, following the paradigm of Day3. The timestamp of licks within each watered trial (see Supplementary Figure 4) was recorded by a sensor controlled by customized Arduino code. The initiation of licking was defined as the first lick detected by the sensor after the sound cue.

c. Rotarod

The procedures followed the protocol described in a previous report(31). Briefly, mice underwent three days of training, consisting of two adaptive trials at a low rolling speed of 5 rpm/min, followed by a third trial with accelerating speed ranging from 0 to 40 rpm/min. After a rest day on the fourth day, mice were tested twice on the rotarod on the fifth day. Each testing session lasted for a total of 5 minutes. During the first 1.5 minutes, the speed of the rotarod accelerated from 0 to 40 rpm/min, which was maintained for the remaining time. Mice were tested twice, and the maximum duration they stayed on the rotarod was recorded for final analysis. Between each trial, the rotarod was thoroughly cleaned with 75% alcohol.

d. Open field test

After a period of habituation, mice were introduced into the center of the open field chamber (40 × 40 cm) and given 10 minutes to freely explore its surroundings. After the test, the arena was cleaned with 75% ethanol. Their activities were observed and recorded using a camera positioned on the ceiling, and the data were subsequently analyzed using EthoVision XT 11.5 software (Noldus, Netherlands).

e. Free exploring in a new cage

The procedures and settings for the open field test were replicated with the exception of utilizing a lidless and empty cage resembling to the mouse’s home cage (l:w:h: 29 cm×18 cm×13 cm) as the arena.

f. Free moving

The procedures and settings were exactly same with free exploring but the arena covered with clean bedding.

g. Treadmill running

Mouse was placed on the treadmill (Zhenghua Biologic apparatus, China) 10mins before testing for habitation. During the test, the rolling velocity of conveyor was constantly 10m/min and the testing session would last for 10mins. To prevent interaction between mice, each mouse was isolated by baffles while running on the treadmill. After the test, conveyor and baffles were cleaned with 75% ethanol. Both male and female mice were included in this experiment but they were separated by gender during the test. Any mice that failed to keep up with rolling conveyor during the test was excluded from the analysis.

Pharmacology

The experimental design followed a previous study that investigated the effects of dopaminergic drugs on striatal MSNs using calcium imaging in parkinsonian mice(26). The drugs (all compounds from Merck), including D1R agonist (SKF81297; 1mg/kg), D2R agonist (quinpirole; 0.5mg/kg), L-DOPA (1mg/kg) together with benserazide (15mg/kg), were dissolved in 0.9% saline and filtered through a PES membrane filter unit (0.22μm, Millex). These solutions were stored at 4℃ and used within a week. The drugs were administered via intraperitoneal (i.p.) injection, with a volume determined according to body weight (0.1 ml/10 g). Both behavioral and photometry experiments commenced 15 minutes after i.p. except quinpirole (60 minutes post i.p.). Information of all chemicals used were summarized in Supplementary Table 1.

Quantification and statistics

Behavioral analysis

Digging behavior. Trials was manually picked from the video by the investigator. The initiation of digging behavior was determined when the bedding beneath the paws of the mice were visibly displaced. The termination of a digging trial was defined as either no bedding displacement for 1 second or the mice engaging in any other behaviors such as rearing, grooming, or eating pallets. The average duration of all digging trials within the first ten minutes after the mice acquired the pallet on the top of the bedding was considered as the digging duration. Digging frequency was defined as the number of digging per minute during this period. The success rate was calculated as the percentage of trials in which the mice successfully obtained the pallet out of the total number of trials.

Licking behavior. It was recorded using a touch sensor controlled by an Arduino chip, with each lick automatically captured and timestamped. The data were then written into a text file. Customized R code was used to analyze the data from the text file and calculate various metrics to characterize licking behavior. A series of successive licks with intervals less than 1 s was considered as a bout. The duration of a bout was defined as the bout length.

Qualitative analysis of photometry signal

a. Photometry signal processing

The photometry data collected was pre-processed for bleaching and motion correction using InperDataProcess v0.7.2 software. The onset of digging and contra-turning behavioral events was manually marked based on the synchronized videos. For digging behavior, only trials having ≥ 10 s of dig-less baseline and ≥ 3 s of digging were selected for final analysis. For turning behavior, only trials with ≥ 4 s intertrial interval and ≥ 90 degree turning were chosen for final analysis.

All fiber signal was presented as dF/ F0 using the equations: F represents the fluorescence value of each timepoint. F0 is the average of F of baseline.

For digging and licking behavior, the time window was -5 to 5 seconds (relative to behavior initiation), during which the -4 to -2 s was used to calculate the F0. For contra-turning, the time window was -1 to 1 second, and the baseline was -1 to 0 s.

The data were further transformed into z-score using the formula to better compare the dynamic of M1 and striatal responses: Where is the population mean, and σ is the population standard deviation.

b. Coupling index & Phase lag

The coupling index represented the Pearson’s correlation coefficient of time-aligned averaged M1 and STR signals. It was calculated using the R cor function. Additionally, the ccf function in R was utilized to determine the lag at which the maximum cross-correlation between the signals occurred. The lag.max parameter was set to be twice the sampling rate.

c. Plateau & peak amplitude

The plateau amplitude was defined as the mean of the normalized (dF/F0) signal within the plateau stage (1∼2 s, relative to initiation) for digging behavior. For licking behavior, the peak amplitude was that within 0∼0.5s.

d. Sum dFF activity

The data of each mouse collected during 10 minutes of treadmill running or free-moving was separated into bins of equal length, typically 8 seconds each. Each bin of data was normalized to dF/F0, where F0 represented the 20th percentile of the data within that bin. In the meanwhile, we calculated the Pearson’s correlation coefficient between the sequence of dF/F0 values and the raw F sequence. If this correlation coefficient lower than 0.8, the mouse was excluded from the further analysis. The sum dFF activity was then determined by summing the resultant dF/F0 values across all bins.

Statistical analysis

Sample sizes are represented in all figures as the number of mice included in the analysis. Statistical analysis was conducted using Prism V9.0 software (GraphPad Software, Inc.). For unpaired data, the non-parametric Mann-Whitney (M-W) test was employed, while one-way balanced ANOVA followed by Dunnet’s multiple-comparison test was used for comparisons among multiple groups. Cumulative distributions were compared using the Kolmogorov-Smirnov (KS) test. The significance level was set at P < 0.05, denoted as *P < 0.05, **P < 0.01, and ***P < 0.001. Additionally, when the P-value fell between

0.05 and 0.1, it was labeled accordingly. Error bars in all figures represent the standard error of the mean (s.e.m).

Acknowledgement

We thank Dr. Cong Liu and his students for their help in preparing the PFF. We also thank the staff members of the animal facility at the National Facility for Protein Science in Shanghai (NFPS), Shanghai Advanced Research Institute, Chinese Academy of Sciences, China for excellent support. This work is supported by NSFC (Grant No. 32271004, 32070963), and Shanghai Municipal Science and Technology Major Project (Grant No. 2019SHZDZX02) to K.-W.H.