Beta-frequency sensory stimulation enhances gait rhythmicity through strengthened coupling between striatal networks and stepping movement

Sridhar, Sudiksha; Lowet, Eric; Gritton, Howard J.; Freire, Jennifer; Zhou, Chengqian; Liang, Florence; Han, Xue

doi:10.1038/s41467-024-52664-0

Download PDF

Article
Open access
Published: 27 September 2024

Beta-frequency sensory stimulation enhances gait rhythmicity through strengthened coupling between striatal networks and stepping movement

Nature Communications volume 15, Article number: 8336 (2024) Cite this article

2647 Accesses
5 Altmetric
Metrics details

Subjects

Abstract

Stepping movement is delta (1–4 Hz) rhythmic and depends on sensory inputs. Stepping-related delta-rhythmic neural activity is coupled to beta (10–30 Hz) frequency dynamics that are also prominent in sensorimotor circuits. We explored how beta-frequency sensory stimulation influences stepping and dorsal striatal regulation of stepping. We delivered audiovisual stimulation at 10 or 145 Hz to mice voluntarily locomoting, while recording locomotion, cellular calcium dynamics and local field potentials (LFPs). We found that 10 Hz, but not 145 Hz stimulation prominently entrained striatal LFPs. Even though stimulation at both frequencies promoted locomotion and desynchronized striatal network, only 10 Hz stimulation enhanced the delta rhythmicity of stepping and strengthened the coupling between stepping and striatal LFP delta and beta oscillations. These results demonstrate that higher frequency sensory stimulation can modulate lower frequency striatal neural dynamics and improve stepping rhythmicity, highlighting the translational potential of non-invasive beta-frequency sensory stimulation for improving gait.

The Ramp protocol: Uncovering individual differences in walking to an auditory beat using TeensyStep

Article Open access 10 October 2024

Neural circuit mechanisms underlying context-specific halting in Drosophila

Article Open access 02 October 2024

Cortical midfrontal theta dynamics following foot strike may index response adaptation during reactive stepping

Article Open access 22 October 2022

Introduction

Stepping movement is highly rhythmic and occurs at delta frequencies (~1–4 Hz) in humans and mice^1,2. While the spinal central pattern generators are critical for pacing movement^3,4, stepping related delta-rhythmic neural activities are broadly detected in many motor-related brain areas, including the motor cortex^5,6,7, dorsal striatum⁸, subthalamic nucleus (STN)^9,10 and cerebellum^7,11,12. Stepping, as with many aspects of voluntary movement, is intricately linked to sensory inputs¹³ and the reliance on sensory cues is often used as a compensation strategy to improve gait in Parkinsonian patients^14,15,16. Indeed, rhythmic auditory cues, such as metronome, music with adapted tempo or beats, or counting, when played around individual subject’s stepping frequencies have been shown to improve patients’ gait and mobility^{17,18,19,20,21,22}.

Beta frequency (8–35 Hz) oscillations are broadly observed across the cortical-basal ganglia circuits and related to sensorimotor processing^{23,24,25,26,27}. During locomotion, transient movement related LFP beta bursts, hundreds of milliseconds long, are prominent across the motor cortex, striatum, globus pallidus externus, and subthalamic nucleus (STN)^9,23. In the striatum, transient LFP beta oscillations have been reported during task initiation²⁶, cue processing²³ and movement completion²⁴. However, persistent, and exaggerated beta oscillations are widely recognized as a functional biomarker of akinesia and bradykinesia in Parkinson’s disease^28,29. Thus, beta oscillations are hypothesized to maintain status quo^25,27, as transient beta also occurs during motor planning without actual movement, and inhibition or desynchronization of beta promotes movement transitions. Consistent with the anti-kinetic effects of persistent beta rhythms, transcranial alternating current stimulation (tACS) at beta frequencies inhibits motor learning³⁰, slows voluntary movement, and reduces force generation in humans^31,32. Similarly, beta-frequency deep brain stimulation (DBS) in the STN leads to further deterioration of bradykinesia and rigidity in Parkinsonian patients^33,34,35.

Dorsal striatum, the largest input structure of the basal ganglia, receives sensory and motor inputs from broad cortical areas^36,37, including somatosensory, visual, auditory and motor cortices, and from many thalamic areas^38,39, such as centromedian, parafascicular, centrolateral, pulvinar and lateral posterior/dorsal nuclei. Striatum plays important roles in various aspects of motor control and motor learning⁴⁰ and is responsive to sensory stimulation⁴¹. Individual striatal neurons exhibit diverse responses during sensorimotor tasks and naturalistic behaviors⁴². At the network level, recent large-scale cellular calcium imaging analyses revealed that functional connectivity between clusters of striatal neurons could increase during specific actions^43,44, even though the overall network correlation strength may decrease⁴⁵, highlighting dynamic recruitment of subpopulations of striatal neurons during behaviors.

The coupling of delta and beta LFP oscillations are broadly observed across the cortico-basal ganglia-thalamic circuits^2,28. Specifically, fluctuation of striatal beta power was found to be coupled to the phase of cortical and thalamic neuronal delta oscillations^2,28,46. Intriguingly, striatal beta was also temporally locked to internally generated delta rhythmic finger tapping behavior⁴⁷, suggesting that both neuronal and behavioral delta rhythms can temporally organize beta rhythmicity^48,49. Recently, using cellular voltage imaging, we discovered that the membrane potentials of many striatal neurons exhibit prominent delta oscillations, which organize beta-rhythmic spike bursting and striatal LFP beta oscillations⁸, highlighting a potential cellular mechanism for the observed circuit level beta and delta coupling.

While strong stimulation of the motor circuits at beta frequencies using tACS or DBS resulted in anti-kinetic effects^{30,31,32,33,34,35}, we hypothesize that sensory stimulation at beta frequencies could boost motor circuit processing of stepping movement via sensory entrainment, without producing exaggerated beta oscillations that are known to be anti-kinetic. To test this hypothesis, we delivered beta rhythmic audiovisual stimulation to mice voluntarily locomoting, since lower frequency stimulations, including sensory stimulation, tACS, DBS, and transcranial magnetic stimulation (TMS) have been found to entrain endogenous neural circuit oscillations^50,51,52,53. The exact beta bands vary widely across individual Parkinsonian patients, animal models, and behavioral states⁵⁴. While beta frequencies generally cover ~12–30 Hz, they span 8–35 Hz^55,56,57,58. In awake, head-fixed mice, our previous studies found that striatal beta band centers around 10 Hz^45,55, and thus we delivered stimulation at 10 Hz to maximize the chance of entraining cortical-striatal circuits. We compared the effect of 10 Hz sensory stimulation to the higher frequency 145 Hz stimulation that would be less effective in entraining neural circuits⁵³. We examined the relationships between cellular calcium dynamics, functional connectivity between simultaneously recorded neurons, population striatal LFPs, and stepping movement.

Results

Population striatal LFP dynamics were better entrained by sensory stimulation at 10 Hz than 145 Hz

To examine how non-invasive sensory stimulation at beta frequencies influences basal ganglia circuitry and locomotion, we performed calcium imaging from individual striatal neurons in head fixed mice freely navigating on a spherical treadmill (Fig. 1a). Briefly, striatal neurons were transduced with the genetically encoded calcium indicator GCaMP7f through intracranial infusion of AAV9-syn-GCaMP7f (n = 5 mice) or AAV9-syn-SomaGCaMP7f (n = 4 mice). GCaMP7f expressing neurons were then recorded using a custom microscope through an imaging window positioned above the dorsal striatum that included the conventionally defined dorsolateral and dorsomedial striatum (Fig. 1a). We also recorded LFPs through a nearby electrode coupled to the imaging window, with the tip being ~200 µm from the bottom of the imaging window (Fig. 1a). During each experiment, mice were presented with five 1-min-long audiovisual stimulation every 5 min, resulting in a total recording of 36 min per session (Fig. 1b). The audiovisual stimulation consisted of concomitant auditory clicks (~77 dB, 50% duty cycle) and visual flashes (50% duty cycle) pulsed at 10 Hz or 145 Hz, in a dimly lit room with ~72 dB constant background noise.

**Fig. 1: 10 Hz sensory stimulation prominently entrained striatal LFPs and suppressed movement onset transitions.**

We first evaluated the sensory entrainment of LFPs by comparing LFP spectral power during stimulation versus the 1-min pre-stimulation period. As expected, 10 Hz stimulation significantly increased 10 Hz LFP power during stimulation (Fig. 1c, d, g, j). Further, the population LFP power increase at 10 Hz lagged the stimulation onset by 5.5 s (Fig. 1g, h) and remained elevated for 8.5 s after stimulation offset (Fig. 1g, i). Finally, 10 Hz stimulation exhibited significant phase locking to the 10 Hz component of the LFP signals (Fig. 1k), during both running and resting (Supplementary Fig. 1a, b), further confirming that the entrainment effect is independent of movement state.

In contrast, 145 Hz stimulation did not alter the 145 Hz LFP spectral power (Fig. 1e, f, l, m). Interestingly, we detected a weak but nonetheless significant phase locking between 145 Hz stimulation and the 145 Hz component of the LFPs, though the phase locking strength was an order of magnitude weaker than that observed with 10 Hz stimulation (Supplementary Fig. 1c, Fig. 1n). Moreover, 10 Hz stimulation did not alter 145 Hz LFP power (Supplementary Fig. 1d), and 145 Hz stimulation did not alter 10 Hz LFP power (Supplementary Fig 1e), suggesting that the observed entrainment effects were specific to the frequencies of audiovisual stimulation. Together, these results demonstrate that audiovisual stimulation at 10 Hz produces prominent entrainment of the striatal network, whereas stimulation at 145 Hz mediates a much weaker effect.

Sensory stimulation at either 10 Hz or 145 Hz increased the time mice spent on running, but only 10 Hz suppressed movement onset transitions

We next examined the influence of sensory stimulation on locomotion. Since each recording session consisted of five stimulation trials at either 10 Hz or 145 Hz, movement features were compared between stimulation versus baseline of the same recording session. Baseline was defined as the period excluding the stimulation periods and the minute after each stimulation to avoid any residual stimulation effects. We first identified running bouts, defined as continuous movement for at least 2 s, and the corresponding onset transitions (Fig. 1o). We found that sensory stimulation at both frequencies increased the fraction of time mice spent on running (Fig. 1p, Supplementary Fig. 2a), without changing the mean velocity during running (Fig. 1q). Interestingly, 10 Hz stimulation, but not 145 Hz stimulation, led to a significant reduction in movement onset transitions and longer running bout durations (Fig. 1r, Supplementary Fig. 2c, d), consistent with the general notion that beta rhythms help maintain the status quo^25,27.

Striatal LFPs are known to be modulated by locomotion⁵⁹. During baseline without stimulation, we found that LFP theta (6–8 Hz) and gamma (40–100 Hz) power increased during movement (Supplementary Fig. 1f, g). The theta power increase was particularly prominent with a peak at ~7 Hz and a harmonic at ~14 Hz. However, the narrow band 10 Hz beta power was suppressed during running in the absence of stimulation (Supplementary Fig. 1g). Due to the strong theta harmonic, we focused on higher frequency beta (15–30 Hz) band, which was also decreased during running (Supplementary Fig. 1f, g), consistent with the general suppression of beta power during movement²⁵ and our previous findings in mice under similar behavioral conditions⁴⁵. During rest, sensory stimulation at either 10 Hz or 145 Hz did not change the LFP power at theta, beta, and gamma bands (Supplementary Fig. 1h, i). During running, beta power was further suppressed by 10 Hz stimulation, compared to the baseline without stimulation, whereas theta power was increased compared to the baseline (Supplementary Fig. 1j, l, m). In contrast, 145 Hz stimulation did not alter LFP spectrum power across theta, beta and gamma bands in running periods (Supplementary Fig. 1k). Together, these results demonstrate that even though sensory stimulation at both frequencies promoted movement, only 10 Hz stimulation, but not 145 Hz, suppressed movement onset transitions. These behavioral results along with the differential effects of the two stimulation frequencies on striatal LFPs highlight that beta-frequency sensory stimulation is effective at reducing movement flexibility, likely through entrainment of locomotor circuits.

10 Hz stimulation, but not 145 Hz, enhanced stepping rhythmicity and improved the coupling between stepping and LFP beta

Since locomotion in mice and humans occurs in the delta frequency range¹, we calculated the spectral power of the treadmill speed, as treadmill speed fluctuations closely track animals stepping cycles during running^8,60. Consistent with our previous observations under similar locomotion conditions⁸, mice stepping occurred at delta frequencies with a peak around 3-4 Hz (Fig. 2a, b). Upon 10 Hz stimulation, animal’s movement speed during running became more rhythmic around delta frequencies (3-4 Hz), showing a significant increase in delta power (Fig. 2c, e). In contrast, during 145 Hz stimulation, the power spectral density of movement speed exhibited a much wider range, and 145 Hz stimulation did not alter movement speed delta power (Fig. 2d, e). After detecting a change in stepping rhythmicity during 10 Hz stimulation, we next examined how sensory stimulation influences the temporal relationship between striatal population LFP dynamics and stepping by computing the phase-locking value (PLV) between LFPs and movement speed. During running, in the absence of sensory stimulation, there was a significant phase locking between LFPs and movement speed at delta frequencies (Fig. 2f–h), consistent with our previous findings⁸. Interestingly, 10 Hz stimulation, but not 145 Hz stimulation, further promoted LFP-movement speed phase locking at delta frequencies during running (Fig. 2f–h).

**Fig. 2: 10 Hz, but not 145 Hz, enhanced gait rhythmicity and elevated coupling of stepping with striatal LFP delta and beta oscillations.**

Our previous study also showed that striatal LFP beta and gamma powers were nested within the delta-rhythmic stepping cycles, peaking at a similar phase as spikes in delta-rhythmic striatal neurons⁸. To understand how sensory stimulation influences the relationship between higher frequency components of the LFP power and stepping movement, we filtered the movement speed trace in the broader delta frequency range (2-4 Hz) and computed the mean LFP spectrogram around all movement delta-peaks. Consistent with our previous observations⁸, in the absence of stimulation, LFP beta and broader gamma (40–100 Hz) powers were modulated by the movement delta (Fig. 2i–k).

During stimulation at both 10 Hz and 145 Hz, the nesting of LFP beta and gamma within the delta-rhythmic movement speed trace remained evident (Fig. 2j, k). Thus, delta-rhythmic movement organizes the timing of LFP beta and gamma regardless of sensory stimulation. Interestingly, 10 Hz stimulation, but not 145 Hz, selectively increased the coupling between movement delta phase and LFP beta power, but not gamma power (Fig. 2j, k). Together, these results demonstrated that while audiovisual stimulation at both 10 Hz and 145 Hz promoted movement (Fig. 1p), only 10 Hz stimulation enhanced the delta rhythmicity of stepping (Fig. 2c, e), suggesting distinct neuronal mechanisms supporting movement vigor versus rhythmicity. 10 Hz stimulation strengthened delta-frequency phase locking between LFP and movement and increased the coupling between LFP beta power and movement delta phase (Fig. 2j), highlighting that beta-frequency sensory modulation of stepping rhythmicity is mediated by improved coupling between striatal neural activity and stepping movement on a moment-to-moment basis.

Cellular calcium activity preceded locomotion

To examine how sensory stimulation modulates striatal encoding of locomotion, we performed single cell calcium imaging from hundreds of striatal neurons simultaneously using GCaMP7f that is better at capturing spiking than the older generations of GCaMPs⁴⁴ (Fig. 3a, b). While intracellular calcium concentration fluctuates due to various biochemical and biophysical activities, the rising phase of the transiently occurring calcium events is generally associated with elevated spiking probabilities, whereas the falling phase is influenced by the slow kinetics of calcium ion dissociation from GCaMP7f and calcium clearance from the cytosol. Thus, we identified calcium events in the GCaMP7f fluorescence traces and binarized the event rising phase as ones and everywhere else as zeros to capture periods of heightened neuronal activity (Supplementary Fig. 3a). During baseline without stimulation, we detected 2.45 ± 1.58 calcium events/min (mean ± standard deviation, n = 6096 neurons from 24 sessions), consistent with previous calcium imaging studies of striatal neurons using GCaMP7f⁴⁴, but higher than that using GCaMP6f^43,45. Using the rising phase of each calcium event, we obtained an activity rate of 0.24 ± 0.15 Hz (mean ± standard deviation, n = 6096 neurons from 24 sessions). As with our previous study⁴⁴, we detected no difference between GCaMP7f and Soma-GCaMP7f in terms of calcium event rate, activity rate, or pairwise Pearson correlation coefficients (Supplementary Fig. 3b–d). Thus, all subsequent analyses were performed by combining the two mice groups.

**Fig. 3: Striatal calcium activity preceded locomotion.**

As many striatal neurons are known to increase their activity during movement^43,45,61, we first determined the fraction of neurons that were modulated by locomotion under our experimental conditions. Briefly, we computed the difference in activity rate between running and resting and then compared it to a shuffled distribution computed from two randomly chosen baseline periods (details see Methods). Since calcium activity rate is generally low, we identified movement-responsive neurons as those with a significant increase in activity rate (>97.5th percentile in shuffled distribution) during running compared to resting. During the baseline period without sensory stimulation, we found 70.4% neurons were activated by movement (Fig. 3c, 4289/6096 neurons from 24 sessions), with a mean activity rate of 0.44 ± 0.24 Hz during movement across movement-responsive neurons.

We also noticed that the activity of movement-responsive neurons preceded the rise in speed (Fig. 3d, e). To quantify this temporal relationship, we computed the cross correlation between the population activity of movement-responsive neurons and locomotion speed with different temporal lags and identified the time lag when the peak correlation occurred. Across baseline sessions without stimulation, the peak correlation was 0.68 ± 0.02 (mean ± standard error, n = 16 sessions, Supplementary Fig. 3e), with neuronal responses preceding locomotion by 156 ± 42.13 ms (mean ± standard error, n = 16 sessions; range: 50–525 ms; median: 75 ms). This result is in general agreement with prior electrophysiology findings showing striatal neuron activation before movement^62,63. Further, audiovisual stimulation did not alter the time lags (Fig. 3f) or the peak correlation coefficient (Fig. 3g). Thus, even though sensory stimulation heightened locomotion bouts (Fig. 1p), it did not disrupt the temporal relationship of striatal activity and locomotion.

Cellular neuronal activity tracked delta-rhythmic movement and sensory stimulation did not alter this temporal relationship

After observing a temporal relationship between population LFP delta and movement stepping cycles, we further evaluated the relationship between individual neurons and stepping. As our imaging frame rate was 20 Hz, though sufficient to capture intrinsically slow cytosolic calcium dynamics, it limited our analysis to frequencies below 10 Hz. Surprisingly, we found that calcium event onsets exhibited significant phase locking to the broader delta (2-4 Hz) frequency component of the movement speed (Fig. 4a–g). Thus, cellular calcium dynamics tracks gait cycles, similar to previous studies that recorded striatal neuron spiking using extracellular electrodes^{8,61,64,65,66}. However, sensory stimulation at either frequency did not change the phase locking strength (Fig. 4l, Supplementary Fig. 4a, c), or the preferred phase (Supplementary Fig. 4b, d). Furthermore, event onsets were also phase locked to the delta component of the LFP trace (Fig. 4a–c, 4h–k), and stimulation did not alter the phase locking strength or the preferred phase (Fig. 4m, Supplementary Fig. 4e–h). The prominent phase locking of calcium events to stepping movement and LFP delta oscillations confirms that individual striatal neurons are coordinated with step-by-step movement, as demonstrated in previous studies^8,64,65. The fact that sensory stimulation did not alter the temporal relationship between striatal cellular dynamics and stepping movement further supports the notion that physiological audiovisual stimulation boosts intrinsic striatal circuit processing of locomotion.

**Fig. 4: Striatal cellular activity was phase coupled to stepping movement and LFP delta oscillations and this temporal relationship was unaffected by sensory stimulation.**

Sensory stimulation at 10 Hz and 145 Hz modulated a greater and yet balanced fraction of activated versus suppressed neurons during movement

Aligning the response of individual neurons to audiovisual stimulation onset, we found that stimulation evoked responses were most prominent at the stimulation onset, with some neurons exhibiting increased calcium activity rates and others decreased rates (Fig. 5a, d). As we did not detect any trend across the five stimulation trials within each recording session (Supplementary Fig. 5a-f), evoked responses were combined across trials. As a population, evoked responses showed a sharp increase at stimulation onset, peaking around 0.3 s after onset, and then settling into a sustained increase after about 1 s, for both stimulation frequencies (Fig. 5b, e). While the transient population increase within one second of stimulation onset was significant for both frequencies (Fig. 5g), the sustained increase was only significant for 145 Hz stimulation, but not for 10 Hz (Fig. 5h). At stimulation offset, we also detected a significant rebound, peaking around 0.5 s after the offset (Fig. 5c, f, i).

**Fig. 5: Sensory stimulation at 10 Hz and 145 Hz evoked a transient increase in activity rate at both onset and offset, and about equal fractions of neurons were activated and suppressed.**

Next, we determined whether sensory stimulation modulated each neuron, either positively or negatively, by comparing the activity rate during resting or running separately, using the same shuffling procedure applied for determining movement-responsive neurons (details in Methods). During resting, about 44.0% neurons were modulated by sensory stimulation (Fig. 5j, Supplementary Table 1), and during running, a significantly larger fraction (~75.5%) were modulated (Fig. 5j, Supplementary Table 1-2, Fisher’s test, 10 Hz: p = 1.17e^-142; 145 Hz: p = 5.24 e^-89). Intriguingly, the proportions of positively and negatively modulated neurons were similar during both resting and running for both stimulation frequencies (Fig. 5k, Supplementary Table 1).

Sensory stimulation inhibited movement-responsive neurons during running

Further evaluation of movement-responsive neurons revealed that ~ 84% were also responsive to sensory stimulation (Fig. 6a, Supplementary Table 3), suggesting that many striatal neurons integrate sensory and motor information. We next characterized how sensory stimulation modulated movement-responsive neurons and found that more neurons were inhibited by sensory stimulation during running than during resting (Fig. 6b, Supplementary Fig. 5g, Supplementary Table 4, Fisher’s test, 10 Hz: p = 3.4e^-64; 145 Hz: p = 5.6e^-78). Similarly, at the population level, sensory stimulation led to a greater suppression of movement-responsive neurons during running but not resting (Fig. 6c, d, Supplementary Fig. 5h). Thus, sensory stimulation led to heterogenous changes of single neuron activities in a locomotion-state dependent manner. While the overall fraction of activated neurons was balanced by the fraction of inhibited neurons regardless of locomotion state, sensory stimulation robustly inhibited movement-related population dynamics during running.

**Fig. 6: Sensory stimulation robustly suppressed movement-responsive neurons during running.**

Both locomotion and sensory stimulation desynchronized movement-responsive neuronal networks

To further understand how sensory-evoked single neuron responses relate to network dynamics, we calculated pairwise Pearson correlations between movement-responsive neurons, as these neurons are most relevant to the locomotion behavior tested. We found that the overall correlation coefficients between these neurons across the entire recording sessions showed a negative linear relationship with the fraction of time mice spent in running (Fig. 7a, R² = 0.66, p = 7.6 e^-7, n = 25 sessions) and the average speed of the session (Supplementary Fig. 6a), suggesting that locomotion generally desynchronizes movement-responsive neural network, consistent with our previous study⁴⁵.

**Fig. 7: Locomotion and sensory stimulation both desynchronized the network of movement-responsive neurons.**

To further evaluate the impact of sensory stimulation and locomotion on functional connectivity, we identified neuron pairs that exhibited significant correlations given their corresponding activity rates. Briefly, we compared the correlation coefficient of a neuron pair to a shuffled distribution formed by assigning random lags between the neuron pair. If the observed correlation value was >97.5th percentile of the shuffled distribution (1000 shuffles), the neuron pair was deemed a ‘correlated pair’, otherwise a ‘random pair’. We found that the fraction of correlated pairs during running (run-relevant pairs) was significantly higher than during resting (rest-relevant pairs) (Fig. 7b). Since the activity rates of these movement-responsive neurons increased during running (Supplementary Fig. 6b), we computed the normalized correlation coefficient of each neuron pair by subtracting the median correlation of the corresponding shuffles. Interestingly, the normalized correlation coefficients among run-relevant pairs were significantly weaker than the rest-relevant pairs (Fig. 7c), which cannot be attributed to the activity rate differences between the populations (Supplementary Fig. 6b).

During sensory stimulation, the fraction of correlated pairs was significantly lower than baseline during both resting and running (Fig. 7d, e). Furthermore, stimulation enhanced the correlation strength of the correlated pairs compared to baseline, regardless of the movement state (Fig. 7f, g), which again cannot be explained by the variation in activity rates (Supplementary Fig. 6c-f). Together, these results demonstrate that both locomotion and sensory stimulation desynchronized the overall movement-responsive neural networks. However, locomotion was accompanied by an increased fraction of correlated pairs during movement, whereas sensory stimulation reduced the fraction, highlighting the distinct circuit effects of sensory versus motor processing in the striatum.

Discussion

To understand the impact of beta-rhythmic sensory stimulation on stepping movement and basal ganglia circuits, we delivered audiovisual stimulation at either 10 Hz or 145 Hz to awake head-fixed mice voluntarily locomoting, while recording locomotion, dorsal striatal LFPs and cellular calcium dynamics. Consistent with the general ability for low frequency stimulation to entrain brain circuits, we found that sensory stimulation at 10 Hz effectively paced striatal population dynamics. While sensory stimulation at both frequencies promoted locomotion and desynchronized striatal network, only 10 Hz stimulation suppressed movement onset transitions, enhanced stepping rhythmicity at delta frequencies, and strengthened the coupling of LFP delta and beta power to the phase of delta-rhythmic stepping. Thus, beta-frequency sensory stimulation enhances gait rhythmicity by strengthening the interactions between striatal neural dynamics and stepping movement, distinct from the generally anti-kinetic effects observed with strong and non-physiological beta-frequency tACS or DBS electrical stimulation of cortico-basal ganglia circuits^{30,31,32,33,34,35}. These results highlight the potential of beta-rhythmic sensory stimulation for improving gait via strengthening the coupling of striatal networks and stepping movement.

We found that 10 Hz stimulation resulted in a significant increase in 10 Hz LFP spectral power, which lagged the stimulation onset for ~ 5.5 s and remained elevated for ~ 8.5 s after stimulation offset (Fig. 1g, h). The long delay in LFP entrainment suggests a network mechanism that likely involves broad sensorimotor regions interconnected with the striatum^36,67. Indeed, large-scale recordings revealed that visual stimuli evoked neural activity changes across many brain circuits in rodents^68,69. Similarly, human EEG and fMRI studies demonstrated that sensory stimuli engaged broad brain regions^70,71. As LFPs are dominated by synaptic inputs, it is highly plausible that inputs from various cortical and subcortical areas contribute to the LFP entrainment effects observed in the striatum. While each brain region could support 10 Hz rhythms with distinct local circuit mechanisms, the coupling across cortico-basal ganglia-thalamic circuits most likely underlies the gradual strengthening of the 10 Hz striatal rhythm, resulting in the seconds’ long delay.

Stepping movement exhibits characteristic delta frequencies in both mice and humans^1,8. Under our experimental conditions, mice movement rhythmicity peaked around 3-4 Hz. This delta component of movement speed was enhanced by 10 Hz sensory stimulation, but not 145 Hz stimulation (Fig. 2c, e), suggesting that 10 Hz stimulation leads to more rhythmic and consistent stepping patterns. These findings are supportive of the idea that 10 Hz sensory stimulation can engage delta-rhythmic neuronal activities via cross-frequency coupling that occur during natural physiological processes. This effect is distinct from the strong and non-physiological electrical or magnetic stimulation of the motor circuits that may evoke exaggerated oscillations which do not occur during physiological conditions^33,34,35. Beta oscillations are conventionally thought to maintain status quo and once exaggerated, as in Parkinson’s disease, leads to akinesia²⁸. We also observed that 10 Hz sensory stimulation lengthened running bouts (Fig. 1p, Supplementary Fig. 2c) and reduced movement onset transitions (Fig. 1r), consistent with the general notion that beta oscillations promote status quo and help maintain ongoing movement^25,27.

We found delta-rhythmic stepping is synchronized with striatal LFP delta oscillations, which temporally organizes LFP beta and gamma power regardless of sensory stimulation (Fig. 2i–k). This observation is consistent with our previous findings⁸ and further confirms that dorsal striatum plays an important role in orchestrating stepping movement. Even though audiovisual stimulation at both 10 Hz and 145 Hz promoted movement, only 10 Hz stimulation improved stepping rhythmicity and strengthened the coupling of striatal LFP delta rhythms and stepping (Fig. 2f, h). The general increase in movement vigor during sensory stimulation at both frequencies likely arises from increased excitatory inputs to the striatum. In contrast, the selective enhancement of the delta-rhythmicity of stepping movement by 10 Hz stimulation may be attributed to beta-frequency entrainment of thalamic and cortical sensory inputs to the striatum and subsequent beta-delta cross-frequency coupling. However, proprioceptive inputs due to delta rhythmic movement could also contribute to the strengthened coupling between LFP delta and stepping².

In addition to stepping, striatal neurons also respond to rhythmic whisking^62,72. Rodent whisking occurs around 8-16 Hz during active exploration^60,73,74, which was found to be phase coupled to stepping in head-fixed mice⁶⁰. We observed significant 10 Hz LFP entrainment during both running and resting, suggesting that stepping-coupled whisking has limited contribution to the LFP entrainment effect detected. Nonetheless, it is possible that 10 Hz sensory stimulation also entrains neural circuits related to whisking, which further strengthens the observed LFP entrainment. Future analysis of the effect of sensory stimulation on whisking will help delineate whether sensory entrainment of rhythmic movement is a broad phenomenon, beyond the specific stepping movement examined here. We also observed prominent movement-dependent theta oscillations, which however showed little coupling to stepping rhythmicity (Fig. 2f, g, i), suggesting a potentially external origin, e.g. the hippocampus as previously demonstrated^75,76.

It is intriguing that the faster beta oscillations augmented by sensory stimulation could modulate the slower delta oscillations to promote stepping automaticity. Neuronal delta rhythms and beta rhythms are temporally linked via cross-frequency coupling across motor circuits^2,28,77,78, but generally with the slower delta rhythms setting up the temporal window to modulate the power of the faster beta rhythms. Our previous cellular voltage study showed that many striatal neurons exhibit delta rhythmic subthreshold membrane voltage fluctuations, with spiking occurring at beta frequency around the peak of membrane voltage delta oscillations⁸. It is highly probable that the membrane biophysical mechanisms supporting the generation of delta and beta rhythms are linked, and thus selective enhancement of cellular beta rhythms could impact the slower delta rhythms. For example, beta rhythmic sensory inputs to a neuron could engage intrinsic ion channels involved in supporting cellular delta oscillations. Future voltage imaging studies that directly measure the cellular voltage effects of sensory stimulation could reveal insights on how higher beta frequencies impact the lower stepping-related delta frequencies at the individual neuron level.

We used GCaMP7f calcium imaging to probe the effect of sensory stimulation on individual striatal neurons and population striatal networks. To estimate neuronal spiking probability from cytosolic calcium that is directly measured by GCaMP7f, we identified calcium events as those with sharp rising phases and performed calcium event analysis using the onset or the entire rising phase. We were pleasantly surprised that calcium event onsets were phase-locked to the delta component of both stepping movement and LFP (Fig. 4a–k), despite the slow kinetics of calcium imaging, highlighting the ability of calcium imaging to track neuronal dynamics. While striatal LFPs are dominated by presynaptic inputs, the observation that cellular calcium was phase coupled to LFP delta provides strong evidence that striatal LFP delta rhythmicity organizes striatal outputs. The temporal relationship of calcium events with movement and LFP delta rhythms were largely unaffected by sensory stimulation, in agreement with the general notion that calcium events reflect cellular spiking output probabilities. This result highlights that sensory stimulation boosts locomotion without interfering with striatal locomotion encoding ability (Fig. 4l, m).

Interestingly, we found that ~70% of neurons exhibited increase in cellular calcium during movement, similar to prior electrophysiological studies⁶¹, but greater than previous calcium imaging studies using GCaMP6f^43,45, in line with the improved sensitivity of GCaMP7f. Furthermore, consistent with previous electrophysiology studies, striatal cellular calcium rise preceded locomotion by approximately a hundred milliseconds⁶², supporting the role of striatum in regulating movement (Fig. 3d–g). We observed a sharp and transient increase of calcium activity at stimulation onset, and a transient rebound at stimulation offset (Fig. 5a–i). Upon sensory stimulation, individual striatal neurons exhibited diverse responses, with some increasing their activity rate and others decreasing. However, we found an equal proportion of activated and inhibited fractions across the population regardless of locomotion states (Fig. 5k), suggesting potential network regulation to maintain balanced excitability across behavioral states as reported in CA1 previously⁷⁹.

When focusing on movement-responsive neurons that are relevant for the locomotion task condition, we found that sensory stimulation has an inhibitory effect when mice were running, but not resting (Fig. 6b–d). It is possible that it is easier to detect inhibition when these movement-responsive neurons have higher activity rate during running in contrast to resting. Nonetheless, our results are consistent with previous studies showing sensory stimulation generates smaller responses in medium spiny projecting neurons (SPNs) during up-state than down-state in anesthetized mice⁸⁰, and sensory suppression of spontaneous whisking-evoked striatal neuron activity in awake mice⁶². Previous studies also reported differential sensory responses between D1-SPNs and D2-SPNs^62,81, though both SPN types tracked gait cycles⁶⁴. As most of the neuron recorded here are SPNs, it will be interesting to examine whether some of the observed heterogeneity arises from systemic differences between D1- versus D2-SPNs. In addition to SPNs, striatal cholinergic interneurons, known to be critical in regulating striatal dynamics may be robustly recruited by sensory stimulation. In particular, cholinergic drugs have consistently been found to improve gait stability in Parkinsonian patients, whereas dopamine replacement has mixed effects on gait and sometimes worsens gait^82,83. However, it remains to be determined how much of the gait improvement effect of cholinergic drugs is mediated by striatal cholinergic interneurons versus a general enhancement of cognitive ability and sensorimotor integration^82,84. Similarly, parvalbumin-positive fast spiking interneurons and other GABAergic interneurons have also been shown to be critical in sensorimotor behaviors⁴³. Future cell-type specific analysis will offer insights on how sensory stimulation engages different striatal neuron subtypes and their involvement in gait regulation.

At the network level, we found that sensory stimulation and locomotion both desynchronized movement-responsive neuron populations, though through distinct effects (Fig. 7). During locomotion, we observed an increase in the proportion of correlated neuron pairs, similar to previous findings of increased functional connectivity within neuron clusters during movement⁸⁵ (Fig. 7b). However, the connectivity strength between these correlated neuron pairs was lower during running compared to resting, pointing to a desynchronization effect (Fig. 7c). Desynchronization increases network information coding, which would be consistent with elevated processing of the environment as animals move⁸⁶, even though mice were head fixed under our experimental condition. During stepping, although striatal neurons’ calcium activity rates increased, we detected a reduction in pairwise correlation coefficients that may stem from greater asynchrony between calcium event timing across neurons. Such asynchrony could arise from many mechanisms, including increased membrane conductance of neurons that render them less sensitive to synaptic inputs, recruitment of inhibitory cells, non-overlapping or desynchronized inputs to striatal neurons, lateral inhibition between striatal neurons, or temporally jittered spiking outputs to membrane depolarizations^87,88,89. During sensory stimulation, the heightened sensory inputs to striatum could engage similar cellular and network mechanisms as during locomotion to enhance network information coding by reducing neuronal synchrony. While sensory stimulation reduced the fraction of correlated cell pairs in the movement-responsive striatal network, the correlation strength within these neurons increased (Fig. 7d–g), suggesting that sensory inputs have the potential to synchronize specific neuron populations receiving joint inputs. Overall, the sensory stimulation evoked desynchronization supports a potential therapeutic benefit in boosting movement by making the striatal network more dynamic and flexible.

There is a natural decline in gait automaticity with aging, and Parkinson’s disease is in particular characterized by various gait impairments, including heightened temporal variability and asymmetry, diminished rhythmicity, and shortened step length^19,83. The malfunction of basal ganglia circuits in Parkinson’s disease is thought to disrupt habitual or automatic processes⁹⁰, resulting in heightened cognitive effort during stepping and a loss of gait automaticity. To compensate for the diminished gait automaticity, Parkinsonian patients often rely on sensory cues during walking^14,15,16, as sensory cue may help regularize input rhythmicity to provide a temporal framework during stepping^91,92. Further, it has been suggested that abnormal neuronal synchrony in Parkinson’s disease may result from a loss of active decorrelation within the basal ganglia^45,93, and our previous study linked abnormal increase in striatal synchrony to motor deficits in low-dopamine Parkinsonian mouse models⁴⁵. The observation that sensory stimulation desynchronizes striatal network also supports a desynchronization mechanism underlying heightened locomotion during sensory stimulation. Given the non-invasive nature of sensory stimulation, our study highlights an exciting translational potential of using sensory stimulation to influence various aspects of movement, including gait rhythmicity, propensity to move, and movement transitions in patients with movement and gait disorders.

Methods

Animal surgery

All animal procedures were approved by the Boston University Institutional Animal Care and Use Committee. The study included both female and male C57BL/6 mice (n = 9, 7 females and 2 males) 8–12 weeks at the start of the experiments, obtained from Charles River Laboratories. Data from male and female mice were pooled. Animal facilities were maintained around 70 °F and 50% humidity and were kept on a 12 h light/dark cycle. Under general anesthesia with isoflurane (~1–3%), mice were surgically implanted with a sterilized custom imaging window over the dorsal striatum (anteroposterior 0.5 mm, mediolateral 1.8 mm, dorsoventral -1.6 mm from the brain surface), a small ground pin over the cerebellum as the LFP recording reference, and a headplate. The imaging window consisted of a stainless-steel cannula (outer diameter 3.17 mm, inner diameter 2.36 mm, height 2 mm) adhered to a circular coverslip (size 0, outer diameter 3 mm) with ultraviolet curable optical adhesive (Norland Products), coupled to an infusion cannula (26 gauge; No.C135G-4; Plastics One) and an insulated stainless-steel wire (Diameter: 130 µm, Plastics One Inc., 005SW-30S, 7N003736501F) for LFP recordings. The stainless-steel wire was glued to the infusion cannula and protruded from the bottom of the infusion cannula by about 200μm. To image the dorsal striatum, the overlying cortical striatal tissue was gently aspirated to gain optical access^43,45,94. While the tissue removal was necessary and not ideal, this animal preparation did not result in any noticeable behavioral deficits, including rotational behaviors as reported previously⁴⁵. Mice were treated with buprenex (0.1 mg/kg) for 48 h following surgery or with sustained release (SR) buprenorphine (3.25 mg/kg) at the beginning of surgery and single-housed to prevent any damage to the head bar or window implant.

Following recovery from surgery (14–21 days post-surgery), one cohort of mice (n = 4) were injected with a 1 μl AAV9-Syn-GCaMP7f.WPRE.SV40 virus (titer 6.6 × 1012 GC/mL) through the attached guide cannula. The other cohort (n = 5) was injected with a 1 μl AAV9-Syn-SomaGCaMP7f.WPRE.SV40 virus (titer 6.6 × 1012 GC/mL). Mice were awake and head-fixed throughout the injection period. After viral infusion (21–28 days), animals were habituated to voluntary locomotion on a spherical treadmill for 2–3 days per week for 2 weeks. Each habituation session lasted for ~30 min per day, in a dimly lit room under similar conditions as during recording sessions.

Locomotion, imaging and LFP data acquisition

During each recording session, mice were awake, head fixed above a spherical treadmill that permitted voluntary locomotion. The spherical treadmill consisted of a Styrofoam ball supported by pressurized air ( ~ 50 psi) in a custom 3D printed plastic housing as in our previous studies^43,45,95. Movement was monitored with two computer universal serial bus sensors fixed to the plastic housing at an angle of 75 degrees at the equator. The x- and y- displacement acquired by each sensor was collected at 20 Hz using a custom Teensy system and recorded via a custom MATLAB script as described in Romano et al.⁹⁵.

Image acquisition was carried out using a custom microscope equipped with a scientific complementary metal oxide semiconductor (sCMOS) camera (ORCA-Flash4.0 LT Digital CMOS camera, No. C11440-42U, Hamamatsu), a Leica N Plan 10 × 0.25 PH1 microscope objective lens, an excitation filter (No. FF01-468/553-25), a dichroic mirror (No. FF493/574-Di01-25×36), an emission filter (No. FF01-512/630-25; Semrock) and a tube lens (Nikon Zoom-NIKKOR 80-200 mm f/4 AI-s). GCaMP7f fluorescence was excited by a 5 W light-emitting diode (LZ1-00B200, 460 nm; LedEngin). The imaging field of view was 1.343 × 1.343 mm², with each pixel corresponding to 1.312 × 1.312 μm². Image acquisition was though HC Image Live (Hamamatsu) at 20 Hz. Image data were stored in dcimg format and converted into tiff format for offline processing.

LFPs were recorded with OmniPlex (PLEXON) at 1 kHz. To synchronize movement data, calcium imaging data, LFPs and sensory stimulation, OmniPlex also recorded the TTL pulses used for each sCMOS image frame acquisition, Teensy motion sensor reading, and audiovisual sensory stimulation triggers.

Sensory stimulation

Visual stimulation was delivered by a white LED placed ~ 5 cm away from the eye. Auditory stimulation was through a speaker within earshot at ~ 77 decibel. Both light and sound pulses were delivered at 50 % duty cycle via a function generator (33220 A, Keysight). Each recording session consisted of five trials of stimulation at either 10 Hz or 145 Hz. Only one stimulation frequency was used per day. Each trial contains a 1-min-long sensory stimulation, followed by 5 min of baseline without stimulation (Fig. 1b). There was also a 6-min period before the start of the first trial, resulting in a total of 36 min per recording session. One or two sessions were recorded for each mouse for each stimulation frequency.

Movement speed trace extraction

Animal’s movement speed was computed as the linear velocity of the treadmill movement using the following formula:

$${\widehat{y}}_{L}=\frac{{y}_{L}-{y}_{R} \cos (\alpha )}{\sin (\alpha )} \quad v \,=\, \sqrt{{\widehat{y}}_{L}^{2}+{y}_{R}^{2}}$$

(1)

where α is the angle between the two sensors (75 degrees), v is the instantaneous velocity of the mouse, and y_L, y_R are the vertical readings from the left and right sensors as in Romano et al.⁹⁵. We then identified sensor artifacts as points >100 cm/s and removed these artifacts by replacing them using linear interpolation.

Identification of rest period and running bouts

We first downsampled the movement speed trace using a moving average with a window size of 20 frames (1 s) (movmean in MATLAB R2023b). Since the variance of speed during resting is low, we computed the variance of the down-sampled speed trace using a 2 s window (40 frames) and identified “putative stationary periods” as periods with variance <0.1, a mean speed <5 cm/s, and a duration >2 s. We then computed the low-speed threshold as the mean + 2 standard deviations of the speed during the putative stationary periods. For two sessions in which mice had extensive movement, the putative rest periods could not be reliably estimated, and thus we set the low-speed threshold to 2.5 cm/s based on manual inspection of the speed traces.

Using the low-speed threshold, we then identified rest periods utilizing a sigmoid function. Specifically, we first assigned each movement speed data point (v) a value using the function F (Eq. 2), which scales all points higher than the threshold towards zero.

$$F(v,\, a,\, c)=1-\frac{1}{1+\exp \left(-a\left(v-c\right)\right)}$$

(2)

where v is the movement speed trace, c is the low-speed threshold, a is set to 0.8 (sigmoid slope). We then computed the scaled speed trace F using a moving window of 2 s. Data points of the scaled speed trace F >0.45 for at least 2 s were deemed to be the rest periods. Note, each rest period was at least 2 s long.

Similarly, running bouts were identified using the sigmoid function defined in Eq. 3. The scaled speed trace G was computed using a moving window of 2 s. Data points of the scaled speed trace G >0.55 for at least 2 s were deemed to be the running bouts. Note, each running bout was at least 2 s long.

$$G(v,\, a,\, c)=\frac{1}{1+\exp \left(-a\left(v-c\right)\right)}$$

(3)

Movement onset transition identification

Movement onset transition was defined as the onset of running bouts preceded by resting periods. To determine the onset point of each running bout, we first computed a high-speed threshold for each session as the 50^th percentile of the speed during all periods other than resting periods. A “putative onset point” is then defined as the first time point within each running bout where movement speed exceeded the value of the low-speed threshold + 10% of the difference between the high-speed threshold and the low-speed threshold. We then determined the exact onset point using a search-box to find the time point within 2 s before and after the putative onset point, where the acceleration exceeded the threshold defined as (high speed threshold-low speed threshold)/3, and stayed above the threshold for at least 250 ms. Visual inspection was performed to confirm the accuracy of the onset detection.

Movement speed rhythmicity analysis

Movement speed rhythmicity analysis was restricted to running bouts. Stepping rhythmicity was estimated by calculating the delta frequency power of the movement speed trace. Specifically, movement speed trace was first band pass filtered using a Butterworth filter in the delta frequency range (3–4 Hz), and then Morelet wavelets power spectrograms were calculated with the FieldTrip toolbox⁹⁶ (https://www.fieldtriptoolbox.org/) for Matlab.

Calcium imaging pre-processing, ROI segmentation and calcium event detection

Each calcium imaging video was first motion corrected as detailed in Tseng et al.⁹⁷. We then calculated a maximum fluorescence minus minimum fluorescence projection image of the motion-correct image stack, and manually identified regions of interests (ROIs) corresponding to individual cells as circles with a radius of 6 pixels (7.8 µm). The ROI intensity trace was computed as the mean intensity across all pixels within each ROI. A local background ‘donut’ trace was calculated as mean intensity of all pixels within 50 pixels of the centroid of the ROI excluding pixels within any identified ROIs. This resulting background trace was further subtracted from the corresponding ROI’s trace to minimize tissue fluorescence scattering. Next, the fluorescence traces (Δf/f) for each neuron were computed as ROI intensity trace minus the local background ‘donut” trace, normalized by the mean across the entire session. GCaMP7f fluorescence traces were then interpolated to exactly 20 Hz and linearly detrended for every neuron. All traces were visually inspected before further analysis. The last minute (35–36 min) of all recording session was excluded for subsequent analyses as some sessions showed evident photobleaching.

Calcium event detection was based on Friedrich et al.⁹⁸, which uses a generalized pool adjacent violators algorithm to infer the most likely event train. Specifically, we used a threshold that is 2.5 times of the standard deviation of the detrended fluorescence trace on the discretized deconvolved output. We performed a manual inspection of detected events in about 100 traces from multiple mice to ensure the accuracy of event detection. To generate calcium activity traces, the rising phase of each calcium event was binarized as ones, and the rest as zeros. Calcium activity rate across all recorded neurons was 18.39 ± 3.69 frames/min at 20 Hz frame rate (mean ± std dev, n = 6096 neurons from 9 mice).

Identification of movement responsive neurons

A neuron was categorized as movement responsive if its calcium activity rate was significantly higher during running than resting. Specifically, we concatenated all the rest or run periods in baseline without stimulation and computed the difference between the calcium activity rate during running versus resting (D).

$$D=\,\left(\frac{{\sum }_{{run}}y}{{\sum }_{{run}}t}\,-\,\frac{{\sum }_{{rest}}y}{{\sum }_{{rest}}t}\right)$$

(4)

A shuffled distribution was created by randomly selecting equivalent running and resting periods from the entire baseline period 1000 times and calculating the difference in each shuffle. If the observed difference was >97.5th percentile of the shuffled distribution, the neuron was classified as a movement responsive neuron. Two sessions that had less than a minute of rest periods in baseline were excluded from this analysis due to the lack of sufficient baseline for estimating the shuffled distribution.

Cross correlation analysis between population activity rate and movement

We estimated the correlation coefficient between the speed trace and the summed calcium activity rate across movement responsive neurons at different time lags using the Python 3.11.5 package scipy.correlate. Significance in correlation was compared to the shuffled distribution built by temporally shuffling (1000 times) the summed calcium activity rate relative to the speed trace with random lags and calculating the corresponding correlation coefficients. The temporal lag was estimated as the timepoint corresponding to the peak correlation coefficient. Sessions with lags >1 s were excluded from this analysis.

Detection of stimulation responsive cells

We identified stimulation responsive neurons by comparing the mean activity rate during stimulation versus baseline in the same locomotion state (rest or run). The shuffled distribution (1000 shuffles) was built by computing the mean activity rate during randomly chosen baseline period in a particular locomotion state during each shuffle. If the observed value during stimulation period was significantly higher than the 97.5th percentile of the shuffled distribution, the neuron was considered activated by sensory stimulation. Similarly, if the activity rate was lower than 2.5th percentile of the shuffled distribution, the neuron was deemed inhibited. Sessions with <1 min of stimulation period in a given locomotion state or if the duration of baseline period was less than twice the duration of the corresponding stimulation period, were excluded for this analysis.

Neuron pairwise correlation analysis

Correlation coefficient was calculated as the Pearson correlation coefficient of a neuron pair’s calcium activity traces. A shuffling procedure was used to determine whether a neuron pair was significantly correlated. For each shuffle, we shifted the calcium activity traces of the neuron pair by a random delay and computed the corresponding correlation coefficients. This process was repeated 1000 times to build a shuffle distribution for every neuron pair. If the observed value was greater than 97.5th percentile of the distribution, the neuron pairs were deemed ‘correlated pairs’. Due to the sparsity of calcium events, negatively correlated pairs ( <2.5th percentile) were excluded for subsequent analyses.

To evaluate the modulation of correlation during specific periods of interest, we concatenated activity traces in those periods and computed the median correlation coefficient across the correlated neuron pairs. To estimate the impact of event rate on the identification of correlated pairs, we grouped neuron pairs by onset frequencies and computed the threshold (97.5^th percentile) of the shuffled distribution and found the threshold had no correlation with the onset frequency (Supplementary Fig. 7b). Similarly, to estimate the impact of onset frequency on correlation coefficients, we examined the mean and median correlation coefficient for neuron pairs with different onset frequencies (Supplementary Fig. 7a, c, d). The median across shuffles increases with high onset frequency rates (Supplementary Fig. 7c). Thus, we computed the correlation coefficients by subtracting the median correlation coefficient across 1000 shuffles from the observed correlation coefficient for every neuron pair to correct for any bias due to activity rate difference.

LFP power spectrum

Intrinsic to awake recordings, LFP recordings were occasionally affected by motion artifacts. We first excluded potential motion artifacts, points greater than mean ±5 standard deviations. The excluded periods were <1 min for each of the 36 min long recording session, except one recording session that was completely excluded due to long-lasting artifacts. LFP power spectrum was computed using Complex Morlet wavelet transformation with Pywavelets package⁹⁹ (‘cmor14.0-1.5’ with bandwidth 14, center frequency 1.5, and scaled between 5 and 5000). The power spectrogram was further normalized by the total power across all frequencies at each time point to probe the relative distribution of power at a specific frequency band. LFP power for theta (6–8 Hz), beta (15–30 Hz) and gamma (40–100 Hz) bands were computed by averaging the power across all frequencies within these bands. To compare the power between running and resting periods during baseline without stimulation, we combined baseline periods of 10 Hz and 145 Hz recording sessions.

To determine LFP power change around movement delta rhythmicity (Fig. 2i), we first aligned LFP power spectrogram to each movement delta oscillation peak and normalized LFP power to the mean power across the given delta-peak triggered window. We then averaged the normalized LFP power spectrogram across all delta peaks in all recording sessions to generate the movement speed delta oscillation peak triggered LFP spectrogram.

Phase locking analysis

We first applied Hilbert transform to find the instantaneous power at each frequency. The speaker system produced artifacts in some sessions that can be easily identified as a sharp and consistent change in LFP waveforms during each stimulation. We excluded these sessions from phase locking analysis by manually inspecting each individual LFP trace, the average LFP waveform triggered by all stimulation pulses of a session, and the LFP power spectrum.

To compute phase locking value (PLV) between LFP and stimulation pulses, we used the following equation:

$${{{\rm{PLV}}}}({{ \, \, f}})=\left|\frac{1}{N}\right|{\sum}_{N}{e}^{i\varphi ( \, \, f,n)}$$

(5)

where the phase of a given frequency f was obtained from the Hilbert Transform of the stimulation pulse train using a Butterworth filter (filter order = 2) around 10 Hz.

To compute the PLV between LFP and movement, we used the same Eq. (5) where we bandpass filtered movement speed traces and LFP traces at delta frequencies (2–4 Hz) using a Butterworth filter and calculated the instantaneous phase using Hilbert Transform.

To compute the phase-locking of calcium event onsets to movement speed (Fig. 4a–g, l) and LFPs (Fig. 4a–c, h–k, m), we first bandpass filtered movement speed traces or LFP traces at delta frequencies (2–4 Hz) using a Butterworth filter (MATLAB R2023b ‘butter’ built-in function) and calculated the instantaneous phase using Hilbert Transform. Since PLV is biased by low event numbers, we utilized an adjusted phase locking value (PLVu²) that accounts for the number of events and only included neurons with at least 20 events. PLVu² were computed using the following equation as in our previous study⁸:

$${{{{\rm{PLVu}}}}}^{2}({{ \, \, f}})=\frac{1}{N-1}({PLV}{\left(f\right)}^{2}\,x\,N-1)$$

(6)

where N is the number of calcium events. To test for statistical significance, a shuffled distribution for each neuron was generated by computing PLVu² with randomized event times. We then used mixed effect model between the observed value and the shuffled distribution. Phase angle polar histograms were created by plotting the preferred phase (circular mean value) of each neuron relative to the delta-filtered movement speed or LFP signals.

Cross-frequency coupling (CFC) between movement speed and LFP

We first computed the instantaneous phase of movement speed delta oscillation using Hilbert transform from the delta filtered movement speed trace (Butterworth filter, filtered order = 2). Similarly, we computed the instantaneous amplitude of LFPs across beta/gamma frequencies, and then derived the instantaneous phase of beta or gamma amplitude fluctuations. CFC of movement speed delta and LFP beta/gamma were computed as the PLV between these phase vectors.

Statistical analyses

To estimate sample size, power analysis was based on effect size differences found in our previous calcium imaging studies^43,79,100. Power analysis was performed using G*Power 3.1.9.7 (http://www.gpower.hhu.de), applying a two-tailed Wilcoxon signed-rank test utilizing α = 0.05 and a power (1-β) value of 0.80, and this revealed a minimal sample size of 150 neurons. We exceeded this number in nine animals, so further statistical analysis was performed. All comparisons were made with versus without sensory stimulation in the same mice to rule out any potential impact due to variation of animal preparation procedure between mice. To compare LFP and movement changes between stimulation and baseline, we used Wilcoxon signed rank test, with a significance level α = 0.05 across trials or sessions. To compare fractions, we used Fisher’s exact test with error bars computed as the 95% confidence interval defined as⁸⁶:

$${{Confidence\; Interval}}={{{\rm{P}}}}\pm 1.96\sqrt{\frac{P(1-P)}{n}}$$

(7)

where P denotes the percentages, and n denotes the number of samples.

To compare changes involving individual neurons, we used Mixed effect models (MATLAB R2023b function fitglme, Poisson distribution with log link function). The fixed effects in the model were movement speed (rest and run), sensory stimulation (on and off) and an interaction term (speed x stimulation). The random effect was session IDs and neurons.

To determine neuronal calcium activity rate changes during stimulation, we used the following model:

$$Y \sim {{{\rm{Speed}}}}+{{{\rm{Stimulation}}}}+{{{\rm{Speed}}}}\; {{{\rm{x}}}}\; {{{\rm{Stimulation}}}}+(1|{{{\rm{Session}}}}\; {{{\rm{ID}}}})$$

where Y is the calcium activity rate. We computed the R² for the model and tested the coefficients for significance (10 Hz: R² = 0.59, n = 2189; 145 Hz: R² = 0.49, n = 2041). We also compared the fit of the model to an intercept-only model, speed-only model, speed, and stimulation without interaction model using log likelihood and Akaike information criterion (AIC). Higher AIC with lower log likelihood indicates better fit of the model. We confirmed that the chosen model exhibited highest AIC and lowest log likelihood compared to other models tested. Using this model, speed was identified as a significant factor in the absence of sensory stimulation (mixed effect model, 10 Hz: p_approx = 0, 145 Hz: p_approx = 0), confirming that movement responsive neurons showed a robust increase in activity rate during sustained locomotion.

To test for the influence of stimulation on PLVu² between baseline and stimulation during running, we used the following mixed effect model:

$$Y \sim {{{\rm{Stimulation}}}}+(1{{{\rm{|Session\; ID}}}})$$

where Y is the PLVu². As the preferred phase is sensitive to both dispersion and direction, we assessed statistical significance using the Kuiper’s test (circ_cmtest), which is a circular analogue of the Kruskal-Wallis test.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

Small datasets of example recordings are available along with the software/code at the Github repository: https://github.com/HanLabBU/Sridhar-Sensory-Modulation-Striatum (https://doi.org/10.5281/zenodo.13567696). Additional experimental data are available from the lead contact upon request due to large file sizes. The source data for all relevant statistics are provided in the Source Data file. Source data are provided with this paper.

Code availability

The code used for data analysis is available at the Github repository https://github.com/HanLabBU/Sridhar-Sensory-Modulation-Striatum

References

Broom, L. et al. A translational approach to capture gait signatures of neurological disorders in mice and humans. Sci. Rep. 7, 3225 (2017).
Article ADS PubMed PubMed Central Google Scholar
Magnusson, J. L. & Leventhal, D. K. Revisiting the “paradox of stereotaxic surgery”: insights into basal ganglia-thalamic interactions. Front Syst. Neurosci. 15, 725876 (2021).
Article PubMed PubMed Central Google Scholar
Grillner, S. & El Manira, A. Current principles of motor control, with special reference to vertebrate locomotion. Physiol. Rev. 100, 271–320 (2020).
Article PubMed Google Scholar
MacKay-Lyons, M. Central pattern generation of locomotion: a review of the evidence. Phys. Ther. 82, 69–83 (2002).
Article PubMed Google Scholar
Viana Di Prisco, G., Marlinski, V. & Beloozerova, I. N. Activity of cat premotor cortex neurons during visually guided stepping. J. Neurophysiol. 130, 838–860 (2023).
Article PubMed Google Scholar
Zelenin, P. V. et al. Activity of motor cortex neurons during backward locomotion. J. Neurophysiol. 105, 2698–2714 (2011).
Article CAS PubMed PubMed Central Google Scholar
Armstrong, D. M. & Drew, T. Discharges of pyramidal tract and other motor cortical neurones during locomotion in the cat. J. Physiol. 346, 471–495 (1984).
Article CAS PubMed PubMed Central Google Scholar
Shroff, S. N. et al. Striatal cholinergic interneuron membrane voltage tracks locomotor rhythms in mice. Nat. Commun. 14, 3802 (2023).
Article ADS CAS PubMed PubMed Central Google Scholar
Petra, F. et al. Alternating modulation of subthalamic nucleus beta oscillations during stepping. J. Neurosci. 38, 5111 (2018).
Article Google Scholar
Hell, F., Plate, A., Mehrkens, J. H. & Bötzel, K. Subthalamic oscillatory activity and connectivity during gait in Parkinson’s disease. NeuroImag. Clin. 19, 396–405 (2018).
Article Google Scholar
Marple-Horvat, D. E. & Criado, J. M. Rhythmic neuronal activity in the lateral cerebellum of the cat during visually guided stepping. J. Physiol. 518, 595–603 (1999).
Article CAS PubMed PubMed Central Google Scholar
Armstrong, D. M. Supraspinal contributions to the initiation and control of locomotion in the cat. Prog. Neurobiol. 26, 273–361 (1986).
Article CAS PubMed Google Scholar
Soto-Faraco, S., Kingstone, A. & Spence, C. Multisensory contributions to the perception of motion. Neuropsychologia 41, 1847–1862 (2003).
Article PubMed Google Scholar
Morris, M. E., Iansek, R., Matyas, T. A. & Summers, J. J. Stride length regulation in Parkinson’s disease: normalization strategies and underlying mechanisms. Brain 119, 551–568 (1996).
Article PubMed Google Scholar
Cooke, J. D., Brown, J. D. & Brooks, V. B. Increased dependence on visual information for movement control in patients with Parkinson’s disease. Can. J. Neurol. Sci. 5, 413–415 (1978).
Article CAS PubMed Google Scholar
Hallett, M. The intrinsic and extrinsic aspects of freezing of gait. Mov. Dis. 23, S439–S443 (2008).
Article Google Scholar
Arias, P. & Cudeiro, J. Effects of rhythmic sensory stimulation (auditory, visual) on gait in Parkinson’s disease patients. Exp. Brain Res. 186, 589–601 (2008).
Article PubMed Google Scholar
Thaut, M. H. & Abiru, M. Rhythmic auditory stimulation in rehabilitation of movement disorders: a review of current research. Music Perception 27, 263–269 (2010).
Article Google Scholar
del Olmo, M. F. & Cudeiro, J. Temporal variability of gait in Parkinson disease: effects of a rehabilitation programme based on rhythmic sound cues. Parkinsonism Relat. Disord. 11, 25–33 (2005).
Article PubMed Google Scholar
Forte, R., Tocci, N. & De Vito, G. The impact of exercise intervention with rhythmic auditory stimulation to improve gait and mobility in Parkinson disease: an umbrella review. Brain Sci. 11, 685 (2021)
Cochen De Cock, V. et al. BeatWalk: Personalized music-based gait rehabilitation in Parkinson’s disease. Front Psychol. 12, 655121 (2021).
Article PubMed PubMed Central Google Scholar
Rocha, P. A., Porfírio, G. M., Ferraz, H. B. & Trevisani, V. F. M. Effects of external cues on gait parameters of Parkinson’s disease patients: a systematic review. Clin. Neurol. Neurosurg. 124, 127–134 (2014).
Article PubMed Google Scholar
Leventhal, D. K. et al. Basal ganglia beta oscillations accompany cue utilization. Neuron 73, 523–536 (2012).
Article CAS PubMed PubMed Central Google Scholar
Feingold, J., Gibson, D. J., DePasquale, B. & Graybiel, A. M. Bursts of beta oscillation differentiate postperformance activity in the striatum and motor cortex of monkeys performing movement tasks. Proc. Natl Acad. Sci. USA 112, 13687–13692 (2015).
Article ADS CAS PubMed PubMed Central Google Scholar
Engel, A. K. & Fries, P. Beta-band oscillations—signalling the status quo? Curr. Opin. Neurobiol. 20, 156–165 (2010).
Article CAS PubMed Google Scholar
Bartolo, R. & Merchant, H. beta oscillations are linked to the initiation of sensory-cued movement sequences and the internal guidance of regular tapping in the monkey. J. Neurosci. 35, 4635–4640 (2015).
Article CAS PubMed PubMed Central Google Scholar
Thomas, G. et al. Existing motor state is favored at the expense of new movement during 13-35 Hz oscillatory synchrony in the human corticospinal system. J. Neurosci. 25, 7771 (2005).
Article Google Scholar
Boraud, T., Brown, P., Goldberg, J. A., Graybiel, A. M. & Magill, P. J. In The Basal Ganglia Vol. VIII (eds. Bolam, J. P., Ingham, C. A., & Magill, P. J.) 630 (Springer, 2005).
Ron, L., William, D. H., Andres, M. L. & Jonathan, O. D. Synchronized neuronal discharge in the basal ganglia of Parkinsonian patients is limited to oscillatory activity. J. Neurosci. 22, 2855 (2002).
Article Google Scholar
Bologna, M. et al. Transcranial alternating current stimulation has frequency-dependent effects on motor learning in healthy humans. Neuroscience 411, 130–139 (2019).
Article CAS PubMed Google Scholar
Pogosyan, A., Gaynor, L. D., Eusebio, A. & Brown, P. Boosting cortical activity at beta-band frequencies slows movement in humans. Curr. Biol. 19, 1637–1641 (2009).
Article CAS PubMed PubMed Central Google Scholar
Joundi, RaedA., Jenkinson, N., Brittain, J.-S., Aziz, TipuZ. & Brown, P. Driving oscillatory activity in the human cortex enhances motor performance. Curr. Biol. 22, 403–407 (2012).
Article CAS PubMed PubMed Central Google Scholar
Chen, C. C. et al. Stimulation of the subthalamic region at 20 Hz slows the development of grip force in Parkinson’s disease. Exp. Neurol. 231, 91–96 (2011).
Article PubMed Google Scholar
Timmermann, L. & Florin, E. Parkinson’s disease and pathological oscillatory activity: is the beta band the bad guy? - new lessons learned from low-frequency deep brain stimulation. Exp. Neurol. 233, 123–125 (2012).
Article CAS PubMed Google Scholar
Eusebio, A. et al. Effects of low-frequency stimulation of the subthalamic nucleus on movement in Parkinson’s disease. Exp. Neurol. 209, 125–130 (2008).
Article PubMed PubMed Central Google Scholar
Hunnicutt, B. J. et al. A comprehensive excitatory input map of the striatum reveals novel functional organization. eLife 5, e19103 (2016).
Article PubMed PubMed Central Google Scholar
Johansson, Y. & Silberberg, G. The functional organization of cortical and thalamic inputs onto five types of striatal neurons is determined by source and target cell identities. Cell Rep. 30, 1178–1194.e1173 (2020).
Article CAS PubMed PubMed Central Google Scholar
Smith, Y. et al. The thalamostriatal systems: anatomical and functional organization in normal and parkinsonian states. Brain Res. Bull. 78, 60–68 (2009).
Article CAS PubMed Google Scholar
Mandelbaum, G. et al. Distinct cortical-thalamic-striatal circuits through the parafascicular nucleus. Neuron 102, 636–652.e637 (2019).
Article CAS PubMed PubMed Central Google Scholar
Graybiel, A. M. & Grafton, S. T. The striatum: where skills and habits meet. Cold Spring Harb. Perspect Biol. 7, a021691 (2015).
Article PubMed PubMed Central Google Scholar
Alloway, K. D., Smith, J. B., Mowery, T. M. & Watson, G. D. R. Sensory processing in the dorsolateral striatum: the contribution of thalamostriatal pathways. Front. Syst. Neurosci. 11, 53 (2017).
Article PubMed PubMed Central Google Scholar
Markowitz, J. E. et al. The striatum organizes 3D behavior via moment-to-moment action selection. Cell 174, 44–58.e17 (2018).
Article CAS PubMed PubMed Central Google Scholar
Gritton, H. J. et al. Unique contributions of parvalbumin and cholinergic interneurons in organizing striatal networks during movement. Nat. Neurosci. 22, 586–597 (2019).
Article CAS PubMed PubMed Central Google Scholar
Shemesh, O. A. et al. Precision calcium imaging of dense neural populations via a cell-body-targeted calcium indicator. Neuron 107, 470–486.e411 (2020).
Article CAS PubMed PubMed Central Google Scholar
Zemel, D. et al. Dopamine depletion selectively disrupts interactions between striatal neuron subtypes and LFP oscillations. Cell Rep. 38, 110265 (2022).
Article CAS PubMed PubMed Central Google Scholar
López-Azcárate, J. et al. Delta-mediated cross-frequency coupling organizes oscillatory activity across the rat cortico-basal ganglia network. Front. Neural Circuits 7, 155 (2013).
Article PubMed PubMed Central Google Scholar
Bartolo, R., Prado, L. & Merchant, H. Information processing in the primate basal ganglia during sensory-guided and internally driven rhythmic tapping. J. Neurosci. 34, 3910–3923 (2014).
Article CAS PubMed PubMed Central Google Scholar
Arnal, L. H., Doelling, K. B. & Poeppel, D. Delta–beta coupled oscillations underlie temporal prediction accuracy. Cereb. Cortex. 25, 3077–3085 (2015).
Article PubMed Google Scholar
Morillon, B., Arnal, L. H., Schroeder, C. E. & Keitel, A. Prominence of delta oscillatory rhythms in the motor cortex and their relevance for auditory and speech perception. Neurosci. Biobehav. Rev. 107, 136–142 (2019).
Article PubMed Google Scholar
Krause, M. R., Vieira, P. G., Csorba, B. A., Pilly, P. K. & Pack, C. C. Transcranial alternating current stimulation entrains single-neuron activity in the primate brain. Proc. Natl Acad. Sci. USA 116, 5747–5755 (2019).
Article ADS CAS PubMed PubMed Central Google Scholar
Thut, G. et al. Rhythmic TMS causes local entrainment of natural oscillatory signatures. Curr. Biol. 21, 1176–1185 (2011).
Article CAS PubMed PubMed Central Google Scholar
Notbohm, A., Kurths, J. & Herrmann, C. S. Modification of brain oscillations via rhythmic light stimulation provides evidence for entrainment but not for superposition of event-related responses. Front Hum. Neurosci. 10, 10 (2016).
Article PubMed PubMed Central Google Scholar
Lowet, E. et al. Deep brain stimulation creates informational lesion through membrane depolarization in mouse hippocampus. Nat. Commun. 13, 7709 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Neuville, R. S. et al. Differential effects of pathological beta burst dynamics between Parkinson’s disease phenotypes across different movements. Front. Neurosci. 15, 733203 (2021).
Article PubMed PubMed Central Google Scholar
McCarthy, M. M. et al. Striatal origin of the pathologic beta oscillations in Parkinson’s disease. Proc. Natl Acad. Sci. USA 108, 11620–11625 (2011).
Article ADS CAS PubMed PubMed Central Google Scholar
Brown, P. & Williams, D. Basal ganglia local field potential activity: character and functional significance in the human. Clin. Neurophysiol. 116, 2510–2519 (2005).
Article PubMed Google Scholar
Brown, P. Abnormal oscillatory synchronisation in the motor system leads to impaired movement. Curr. Opin. Neurobiol. 17, 656–664 (2007).
Article CAS PubMed Google Scholar
Gatev, P., Darbin, O. & Wichmann, T. Oscillations in the basal ganglia under normal conditions and in movement disorders. Mov. Dis. 21, 1566–1577 (2006).
Article Google Scholar
Lemaire, N. et al. Effects of dopamine depletion on LFP oscillations in striatum are task- and learning-dependent and selectively reversed by l-DOPA. Proc. Natl Acad. Sci. USA 109, 18126–18131 (2012).
Article ADS CAS PubMed PubMed Central Google Scholar
Sofroniew, N. J., Cohen, J. D., Lee, A. K. & Svoboda, K. Natural whisker-guided behavior by head-fixed mice in tactile virtual reality. J. Neurosci. 34, 9537 (2014).
Article PubMed PubMed Central Google Scholar
Shi, L. H., Luo, F., Woodward, D. J. & Chang, J. Y. Neural responses in multiple basal ganglia regions during spontaneous and treadmill locomotion tasks in rats. Exp. Brain Res. 157, 303–314 (2004).
Article CAS PubMed Google Scholar
de la Torre-Martinez, R., Ketzef, M. & Silberberg, G. Ongoing movement controls sensory integration in the dorsolateral striatum. Nat. Commun. 14, 1004 (2023).
Article ADS PubMed PubMed Central Google Scholar
Schultz, W. & Romo, R. Role of primate basal ganglia and frontal cortex in the internal generation of movements. Exp. Brain Res. 91, 363–384 (1992).
Article CAS PubMed Google Scholar
Yang, L., Singla, D., Wu, A. K., Cross, K. A. & Masmanidis, S. C. Dopamine lesions alter the striatal encoding of single-limb gait. eLife 12, RP92821 (2024).
Article PubMed PubMed Central Google Scholar
West, M. O. et al. A region in the dorsolateral striatum of the rat exhibiting single-unit correlations with specific locomotor limb movements. J. Neurophysiol. 64, 1233–1246 (1990).
Article CAS PubMed Google Scholar
Rueda-Orozco, P. E. & Robbe, D. The striatum multiplexes contextual and kinematic information to constrain motor habits execution. Nat. Neurosci. 18, 453–460 (2015).
Article CAS PubMed PubMed Central Google Scholar
Charpier, S., Pidoux, M. & Mahon, S. Converging sensory and motor cortical inputs onto the same striatal neurons: An in vivo intracellular investigation. PLoS ONE 15, e0228260 (2020).
Article CAS PubMed PubMed Central Google Scholar
Scott, B. B. et al. Imaging cortical dynamics in GCaMP transgenic rats with a head-mounted widefield macroscope. Neuron 100, 1045–1058.e1045 (2018).
Article CAS PubMed PubMed Central Google Scholar
Das, A. et al. Large-scale recording of neuronal activity in freely-moving mice at cellular resolution. Nat. Commun. 14, 6399 (2023).
Article ADS CAS PubMed PubMed Central Google Scholar
Daly, I. et al. Electroencephalography reflects the activity of sub-cortical brain regions during approach-withdrawal behaviour while listening to music. Sci. Rep. 9, 9415 (2019).
Article ADS PubMed PubMed Central Google Scholar
Toader, C. et al. Cognitive crescendo: how music shapes the brain’s structure and function. Brain Sci. 13, 1390 (2023)
Carelli, R. M. & West, M. O. Representation of the body by single neurons in the dorsolateral striatum of the awake, unrestrained rat. J. Comparative Neurol. 309, 231–249 (1991).
Article CAS Google Scholar
Arakawa, H. & Erzurumlu, R. S. Role of whiskers in sensorimotor development of C57BL/6 mice. Behav. Brain Res. 287, 146–155 (2015).
Article PubMed PubMed Central Google Scholar
Chakrabarti, S., Nambiar, J. & Schwarz, C. Adaptive whisking in mice. Front. Syst. Neurosci. 15, 813311 (2022)
Lalla, L., Rueda Orozco, P. E., Jurado-Parras, M. T., Brovelli, A. & Robbe, D. Local or not local: investigating the nature of striatal theta oscillations in behaving rats. eNeuro 4. ENEURO.0128-17 (2017)
DeCoteau, W. E. et al. Learning-related coordination of striatal and hippocampal theta rhythms during acquisition of a procedural maze task. Proc. Natl Acad. Sci. 104, 5644–5649 (2007).
Article ADS CAS PubMed PubMed Central Google Scholar
McGregor, M. M. & Nelson, A. B. Circuit mechanisms of Parkinson’s disease. Neuron 101, 1042–1056 (2019).
Article CAS PubMed Google Scholar
Sherman, M. A. et al. Neural mechanisms of transient neocortical beta rhythms: converging evidence from humans, computational modeling, monkeys, and mice. Proc. Natl Acad. Sci. USA 113, E4885–E4894 (2016).
Article CAS PubMed PubMed Central Google Scholar
Mount, R. A. et al. Distinct neuronal populations contribute to trace conditioning and extinction learning in the hippocampal CA1. eLife 10, e56491 (2021).
Article CAS PubMed PubMed Central Google Scholar
Reig, R. & Silberberg, G. Multisensory integration in the mouse striatum. Neuron 83, 1200–1212 (2014).
Article CAS PubMed PubMed Central Google Scholar
Sippy, T., Lapray, D., Crochet, S. & Petersen, CarlC. H. Cell-type-specific sensorimotor processing in striatal projection neurons during goal-directed behavior. Neuron 88, 298–305 (2015).
Article CAS PubMed PubMed Central Google Scholar
Albin, R. L. et al. Cholinergic systems, attentional-motor integration, and cognitive control in Parkinson’s disease. Prog. Brain Res. 269, 345–371 (2022).
Article PubMed PubMed Central Google Scholar
Bohnen, N. I. et al. Cholinergic system changes of falls and freezing of gait in Parkinson’s disease. Annal. Neurol. 85, 538–549 (2019).
Article CAS PubMed Google Scholar
Sarter, M., Avila, C., Kucinski, A. & Donovan, E. Make a left turn: cortico-striatal circuitry mediating the attentional control of complex movements. Mov. Dis. 36, 535–546 (2021).
Article Google Scholar
Barbera, G. et al. Spatially compact neural clusters in the dorsal striatum encode locomotion relevant information. Neuron 92, 202–213 (2016).
Article CAS PubMed PubMed Central Google Scholar
Mount, R. A. et al. The autism spectrum disorder risk gene NEXMIF over-synchronizes hippocampal CA1 network and alters neuronal coding. Front. Neurosci. 17, 1277501 (2023).
Article PubMed PubMed Central Google Scholar
Tiesinga, P. H. & Sejnowski, T. J. Rapid temporal modulation of synchrony by competition in cortical interneuron networks. Neural. Comput. 16, 251–275 (2004).
Article CAS PubMed PubMed Central Google Scholar
Kincaid, A. E., Zheng, T. & Wilson, C. J. Connectivity and convergence of single corticostriatal axons. J. Neurosci. 18, 4722 (1998).
Article CAS PubMed PubMed Central Google Scholar
Stern, E. A., Jaeger, D. & Wilson, C. J. Membrane potential synchrony of simultaneously recorded striatal spiny neurons in vivo. Nature 394, 475–478 (1998).
Article ADS CAS PubMed Google Scholar
Redgrave, P. et al. Goal-directed and habitual control in the basal ganglia: implications for Parkinson’s disease. Nat. Rev. Neurosci. 11, 760–772 (2010).
Article CAS PubMed PubMed Central Google Scholar
Hidalgo-Balbuena, A. E., Luma, A. Y., Pimentel-Farfan, A. K., Peña-Rangel, T. & Rueda-Orozco, P. E. Sensory representations in the striatum provide a temporal reference for learning and executing motor habits. Nat. Commun. 10, 4074 (2019).
Article ADS PubMed PubMed Central Google Scholar
Nombela, C., Hughes, L. E., Owen, A. M. & Grahn, J. A. Into the Groove: can rhythm influence Parkinson’s disease? Neurosci. Biobehav. Rev. 37, 2564–2570 (2013).
Article PubMed Google Scholar
Wilson, C. J. Active decorrelation in the basal ganglia. Neuroscience 250, 467–482 (2013).
Article CAS PubMed Google Scholar
Klaus, A. et al. The spatiotemporal organization of the striatum encodes action space. Neuron 95, 1171–1180.e1177 (2017).
Article CAS PubMed PubMed Central Google Scholar
Romano, M. et al. A teensy microcontroller-based interface for optical imaging camera control during behavioral experiments. J. Neurosci. Methods 320, 107–115 (2019).
Article PubMed PubMed Central Google Scholar
Oostenveld, R., Fries, P., Maris, E. & Schoffelen, J.-M. FieldTrip: Open source software for advanced analysis of MEG, EEG, and invasive electrophysiological data. Comput. Intell. Neurosci. 2011, 156869 (2011).
Article PubMed Google Scholar
Tseng, H.-A. et al. Region-specific effects of ultrasound on individual neurons in the awake mammalian brain. iScience 24, 102955 (2021).
Article ADS PubMed PubMed Central Google Scholar
Friedrich, J., Zhou, P. & Paninski, L. Fast online deconvolution of calcium imaging data. PLOS Comput. Biol. 13, e1005423 (2017).
Article ADS PubMed PubMed Central Google Scholar
Lee, G. R., Gommers, R., Wasilewski, F., Wohlfahrt, K. & O’Leary, A. PyWavelets: A python package for wavelet analysis. J. Open Source Softw. 4, 1237 (2019).
Article ADS Google Scholar
Mohammed, A. I. et al. An integrative approach for analyzing hundreds of neurons in task performing mice using wide-field calcium imaging. Sci. Rep. 6, 20986 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We thank members of Han Lab for providing invaluable help throughout the study, and Dr. Mark Kramer for guidance on statistical tests. X.H. acknowledges funding NIH 1R01NS115797, NIH 1RF1NS129520, NSF CBET-1848029 and NSF CIF-1955981.

Author information

Authors and Affiliations

- Department of Biomedical Engineering, Boston University, Boston, MA, USA
Sudiksha Sridhar, Eric Lowet, Howard J. Gritton, Jennifer Freire, Chengqian Zhou, Florence Liang & Xue Han
- Department of Neuroscience, Erasmus MC, Rotterdam, the Netherlands
Eric Lowet
- Department of Comparative Biosciences, University of Illinois at Urbana-Champaign, Urbana, IL, USA
Howard J. Gritton
- Department of Pharmacology, Boston University, Boston, MA, USA
Jennifer Freire

Authors

Sudiksha Sridhar
View author publications
You can also search for this author inPubMed Google Scholar
Eric Lowet
View author publications
You can also search for this author inPubMed Google Scholar
Howard J. Gritton
View author publications
You can also search for this author inPubMed Google Scholar
Jennifer Freire
View author publications
You can also search for this author inPubMed Google Scholar
Chengqian Zhou
View author publications
You can also search for this author inPubMed Google Scholar
Florence Liang
View author publications
You can also search for this author inPubMed Google Scholar
Xue Han
View author publications
You can also search for this author inPubMed Google Scholar

Contributions

S.S. and X.H. designed the study. S.S conducted the experiments, collected and analyzed the data. E.L. assisted with data analysis. H.J.G. and J.F performed mouse surgeries and assisted with data collection. F.L. helped with data collection. C.Z assisted with data analysis. S.S. and X.H. prepared the manuscript, and all authors edited the manuscript. X.H. oversaw all aspects of the project and supervised the study.

Corresponding author

Correspondence to Xue Han.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Pavel Rueda-Orozco, and the other, anonymous, reviewers for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Sridhar, S., Lowet, E., Gritton, H.J. et al. Beta-frequency sensory stimulation enhances gait rhythmicity through strengthened coupling between striatal networks and stepping movement. Nat Commun 15, 8336 (2024). https://doi.org/10.1038/s41467-024-52664-0

Download citation

Received: 17 January 2024
Accepted: 18 September 2024
Published: 27 September 2024
DOI: https://doi.org/10.1038/s41467-024-52664-0