Driving motor cortex oscillations modulates bradykinesia in Parkinson’s disease

Abstract

In patients with Parkinson’s disease, beta (β) and gamma (γ) oscillations are altered in the basal ganglia, and this abnormality contributes to the pathophysiology of bradykinesia. However, it is unclear whether β and γ rhythms at the primary motor cortex (M1) level influence bradykinesia.

Transcranial alternating current stimulation (tACS) can modulate cortical rhythms by entraining endogenous oscillations. We tested whether β- and γ-tACS on M1 modulate bradykinesia in patients with Parkinson’s disease by analysing the kinematic features of repetitive finger tapping, including movement amplitude, velocity and sequence effect, recorded during β-, γ- and sham tACS. We also verified whether possible tACS-induced bradykinesia changes depended on modifications in specific M1 circuits, as assessed by short-interval intracortical inhibition and short-latency afferent inhibition. Patients were studied OFF and ON dopaminergic therapy. Results were compared to those obtained in a group of healthy subjects.

In patients, movement velocity significantly worsened during β-tACS and movement amplitude improved during γ-tACS, while the sequence effect did not change. In addition, short-latency afferent inhibition decreased (reduced inhibition) during β-tACS and short-interval intracortical inhibition decreased during both γ- and β-tACS in Parkinson’s disease. The effects of tACS were comparable between OFF and ON sessions. In patients OFF therapy, the degree of short-interval intracortical inhibition modulation during β- and γ-tACS correlated with movement velocity and amplitude changes. Moreover, there was a positive correlation between the effect of γ-tACS on movement amplitude and motor symptoms severity. Our results show that cortical β and γ oscillations are relevant in the pathophysiology of bradykinesia in Parkinson’s disease and that changes in inhibitory GABA-A-ergic interneuronal activity may reflect compensatory M1 mechanisms to counteract bradykinesia.

In conclusion, abnormal oscillations at the M1 level of the basal ganglia-thalamo-cortical network play a relevant role in the pathophysiology of bradykinesia in Parkinson’s disease.

Video Abstract

See Oswal (doi:10.1093/brain/awab458) for a scientific commentary on this article.

Introduction

The pathophysiology of bradykinesia, one of the cardinal symptoms of Parkinson’s disease, is still not entirely clear.¹ Evidence suggests a crucial role of altered oscillatory activity within the basal ganglia, particularly in the beta (β, 13–30 Hz) and gamma (γ, 40–80 Hz) frequency bands,^2-4 which are the main natural rhythms of the motor system.^5–8 Local field potential recordings in patients with Parkinson's disease undergoing deep brain stimulation demonstrated pathologically increased β oscillations, which directly correlated with bradykinesia severity.^9–15 Additionally, changes in γ oscillations, including reduced γ power and burst rate, have been described in Parkinson’s disease, and these abnormalities lead to altered movement force, velocity and amplitude.^3,16–18 Notably, β and γ rhythm alterations are restored in the clinical ‘ON’ condition, in parallel to bradykinesia improvement.^4,13,18–21

Besides specific basal ganglia abnormalities in Parkinson’s disease, bradykinesia has recently been interpreted as arising from a wide network dysfunction that includes cortical sensorimotor areas.^1,22 In this context, the primary motor cortex (M1) plays a major role, and its neurophysiological changes relate to specific bradykinesia features in patients.^1,23 Recent EEG and magnetoencephalographic (MEG) studies indicate that increased β power in M1 is associated with bradykinesia severity, while bradykinesia amelioration may relate to increased γ power.^24–26

Despite preliminary observations, whether and through which mechanisms changes in β and γ oscillatory activity of M1 significantly contribute to the pathophysiology of bradykinesia in Parkinson’s disease is still unknown. A possible approach to better clarify this issue is to assess whether driving β and γ cortical rhythms through non-invasive brain stimulation modulates bradykinesia in patients. We used transcranial alternating current stimulation (tACS), a neurophysiological technique that entrains the activity of resonant neurons at the stimulating frequency and enhances the power of entrained cortical oscillations, as also recently demonstrated in vivo.^27–31 In a double-blind, randomized and sham-controlled study, we applied tACS at β and γ frequencies over M1 in patients with Parkinson's disease during the execution of repetitive finger tapping and objectively assessed movement kinematics. To clarify the possible mechanisms underlying changes in kinematic features of bradykinesia during tACS, we also tested M1 intracortical excitability by applying transcranial magnetic stimulation (TMS) techniques.³² We specifically studied short-interval intracortical inhibition (SICI), a measure of GABA-A-ergic activity,^33,34 and short-latency afferent inhibition (SAI), reflecting cholinergic neurotransmission,^33,35 since previous research in healthy subjects indicated that γ-tACS modulates SICI, while β-tACS modulates SAI.^36–39 To evaluate whether dopaminergic treatment modifies possible β- and γ-tACS effects, all patients were assessed both OFF and ON therapy. Data were also compared with those collected in a cohort of healthy subjects.

Materials and methods

Participants

Eighteen patients with Parkinson’s disease and 16 age- and gender-matched healthy subjects were enrolled in the study (Table 1). All participants were right-handed. Patients were recruited from the Department of Human Neurosciences, Sapienza University of Rome. Parkinson’s disease diagnosis was based on clinical criteria.^40,41 The clinical motor assessment included the Hoehn and Yahr scale and the motor section of the Movement Disorder Society-sponsored Unified Parkinson's Disease Rating Scale (MDS-UPDRS-III).⁴² The Beck Depression Inventory (BDI-II)⁴³ was used to exclude patients with moderate or severe depressive symptoms, while the Montreal Cognitive Assessment (MoCA)⁴⁴ and Frontal Assessment Battery (FAB)⁴⁵ were adopted to exclude dementia. No subject had additional neuropsychiatric comorbidities or was taking drugs known to influence M1 excitability.³³ The experimental procedures, which conformed to the Declaration of Helsinki and adhered to international safety guidelines,^46,47 were approved by the local institutional review board. All participants gave their written informed consent to the study.

Kinematic recording and analysis

The methodology for repetitive finger tapping recordings and analysis was adopted from our previous studies.^23,48–51 Participants were asked to repetitively open and close the index finger and thumb (finger tapping) as wide and fast as possible for 15 s. A 3D optoelectronic system was used (SMART motion system, BTS Engineering) to follow the displacement of reflective markers taped to the participant’s index finger, thumb and hand (more affected side in patients and dominant side in heathy subjects). Movement analysis was performed using a dedicated software (SMART Analyzer, BTS Engineering). We determined the total number of performed movements and the movement rhythm [coefficient of variation (CV) of the inter-tap intervals, with higher values representing a lower regularity of repetitive movements]. We used linear regression techniques to determine movement amplitude and velocity (regression line intercept), as well as amplitude and velocity decrement across the 15-s trials (regression line slope) (Supplementary Fig. 1).^23,48–51 During kinematic assessment, a video recording of finger tapping was also acquired in patients with Parkinson's disease (OFF therapy). Three blinded neurologists with experience in movement disorders subsequently rated the videos according to the MDS-UPDRS-III item 4 score.

TACS

We delivered tACS using a BrainSTIM (EMS) connected to two electrodes (5 × 5 cm) enclosed in sponges soaked with saline solution. One electrode was centred over the first dorsal interosseus hotspot and the other over Pz (according to the 10–20 international EEG system). Impedance was kept at <10 kΩ. We delivered tACS with no direct current offset, a peak-to-peak amplitude of 1 mA and 3-s ramp-up (i.e. progressive increase in the stimulation amplitude) and ramp-down (i.e. progressive decrease in the stimulation amplitude) periods. Stimulation frequency was set at 20 Hz for β-tACS and 70 Hz for γ-tACS. Sham tACS consisted of a 7 s stimulation delivered at 20 Hz, which included 3-s ramp-up, 1 s stimulation at 1 mA amplitude and 3-s ramp-down. The stimulation ended automatically after the ramp-down period. We specifically used the same tACS montage and stimulation parameters adopted in previous studies because these settings were demonstrated to ensure that sufficient current reached M1 and produced relevant neurophysiological and behavioural motor effects.^{36–39,51–58} No participant reported visual or cutaneous sensation during tACS. Accordingly, subjects were unable to distinguish between the different stimulation conditions.

Transcranial magnetic stimulation

Single- and paired-pulse TMS was performed using a MAGSTIM 200 connected to a figure-of-eight coil (Magstim Company), with the handle positioned at a ∼45° angle from the midline and pointing backward. The hotspot of the first dorsal interosseus muscle (which was identified both on the scalp and over the tACS electrode), and resting and active motor thresholds were identified following international guidelines.⁵⁹ The motor threshold intensity that elicited motor evoked potentials (MEPs) of ∼1 mV amplitude was then determined (MT_1mV). SICI and SAI were studied using standard techniques.^34,35,60 For SICI, a conditioning stimulus at 80% active motor threshold was followed by a test stimulus at MT_1mV, with a 2-ms interstimulus interval. For SAI, median nerve stimulation was performed at the wrist using a 0.1-ms electrical rectangular pulse (Digitimer DS7A) through a bipolar electrode and an intensity that induced a painless thumb twitch. TMS intensity was set at MT_1mV and the interstimulus interval between peripheral and cortical stimulation was 23 ms. EMG activity was recorded from the first dorsal interosseus of the more affected side in patients and of the dominant side in heathy subjects using surface electrodes. EMG signals were amplified (Digitimer D360, Digitimer), digitized at 5 kHz (CED 1401; Cambridge Electronic Design), and stored on a computer for offline analyses (Signal software, Cambridge Electronic Design, UK). Peak-to-peak MEP amplitude was measured and averaged for each condition. SICI and SAI were expressed as the ratio between the amplitude of conditioned and single-pulse MEP.

Experimental design

The experimental protocol included a kinematic assessment of repetitive finger tapping and TMS recordings of M1 excitability measures (MEPs evoked by single TMS pulses, SICI and SAI). Both kinematic and TMS recordings were performed during sham, β-, and γ-tACS delivered in random order with predefined time intervals between conditions (see Fig. 1 legend for details). Importantly, both participants and the researchers who collected the various neurophysiological measures were blinded to stimulation conditions. Only the additional researcher who set up the stimulation frequencies and applied tACS was unblinded. All patients with Parkinson's disease underwent two sessions, which were randomly conducted at least 1 week apart, in the clinical OFF and ON state. In the OFF session, patients were studied after withdrawal (>12 h) of dopaminergic therapy, while in the ON session they were tested on their usual therapeutic regimen.

Statistical analysis

The Mann–Whitney U-test was used to analyse differences in age and clinical scores between patients with Parkinson's disease and heathy subjects, while the Wilcoxon test was used to compare MDS-UPDRS-III scores between patients in the OFF and ON state. Fisher’s exact test was adopted to evaluate possible differences in categorical variables between groups. Two-tailed t-tests were applied to verify motor threshold changes between patients and heathy subjects (unpaired t-test) and between patients in the OFF and ON state (paired t-test).

The effects of tACS on kinematic parameters and TMS measures were evaluated using separate repeated-measures (rm)ANOVAs, with the within-group factor ‘frequency’ (three levels: sham, β, γ) and the between-group factor ‘group’ (two levels: Parkinson’s disease OFF, heathy subjects). In case of significant interactions, an rmANOVA with ‘frequency’ as a factor was then used for each group. To assess kinematic and TMS variables OFF and ON therapy, we used an rmANOVA with ‘frequency’ and ‘state’ (two levels: Parkinson’s disease OFF, Parkinson’s disease ON) as within-group factors. Greenhouse–Geisser corrections were applied when a violation of sphericity was detected. Post hoc analyses in the rmANOVAs were performed using t-tests, with the Tukey HSD test applied to correct for multiple comparisons.

The effects of tACS on MDS-UPDRS-III item 4 score were assessed by the Friedman test with ‘frequency’ as a factor (median values across blocks and raters for each stimulation condition). Inter-rater agreement was tested by K statistic (index of Fleiss),⁶¹ and Landis–Koch classification was used to define the agreement level (<0 poor; 0.01–0.20 slight; 0.21–0.40 fair; 0.41–0.60 moderate; 0.61–0.80 substantial; and 0.81–1 almost perfect).

Pearson’s correlation test was used to evaluate possible relationships between neurophysiological variables. For this purpose, tACS effects were quantified by calculating the ratio between values recorded during β- or γ-tACS and values recorded during sham tACS (e.g. ratio velocity β-tACS/sham tACS). Spearman’s rank-correlation test was used to assess possible clinical-neurophysiological relationships. Values are presented as mean values ± 1 standard error of the mean (SEM), unless otherwise stated. The level of significance was set at P < 0.05. Statistical analyses were performed using Statistica (TIBCO software, USA).

Sample size was computed using G*Power software.⁶² We set a desired power of 0.80 and an alpha error of 0.05 and based the effect size values on our previously published data regarding tACS effects on finger tapping,⁵¹ SICI⁵² and SAI.³⁶ The minimum required sample size to detect a difference in the effect of the factor ‘frequency’ in patients (main aim of the study) was 16 for kinematic and 15 for TMS measures.

Data availability

The datasets analysed during the current study are available from the corresponding author on reasonable request.

Results

No differences were found in age (P = 0.37), gender distribution (P = 0.70), BDI-II (P = 0.11), MoCA (P = 0.68) or FAB scores (P = 0.16) between patients (OFF state) and heathy subjects. As expected, MDS-UPDRS-III score was higher in patients in the OFF than ON state (P < 0.001; Table 1).

Parkinson's disease patients OFF versus heathy subjects

Kinematic parameters

Movement velocity and amplitude were lower in patients with Parkinson's disease than heathy subjects, as demonstrated by the significant effect of the factor ‘group’ in rmANOVAs. Also, patients showed amplitude decrement (sequence effect) and altered movement rhythm (higher CV values) (Table 2). Importantly, the analysis demonstrated a significant Frequency × Group interaction for velocity and amplitude, suggesting different tACS effects in the two groups. In patients, tACS modified movement velocity [F(2,34) = 8.23, P = 0.001], with lower values (i.e. decreased velocity) observed during β-tACS than γ-tACS (P < 0.01) and sham tACS (P = 0.02) (percentage change in velocity during β-tACS compared to sham tACS, mean ± 1 SD −5.3 ± 8.5%, range −24.5 to +8%; Fig. 2). In contrast, movement velocity was comparable in the three stimulation conditions in heathy subjects [F(2,30) = 0.79, P = 0.46] (Supplementary Fig. 2). Furthermore, movement amplitude changed during tACS in patients [F(2,34) = 8.85, P < 0.001], with higher values (i.e. increased amplitude) during γ-tACS than β-tACS (P < 0.001) and sham tACS (P = 0.02) (percentage change in amplitude during γ-tACS compared to sham tACS, mean ± 1 SD: +6.5 ± 11.3%, range: −5.3 to +31.7%; Fig. 2). Again, this effect did not occur in heathy subjects [F(2,30) = 1.11, P = 0.34]. The number of movements and velocity decrement were similar between groups and conditions (Table 2).

Correlation analysis showed a positive relationship between the effects produced by tACS on movement amplitude and velocity (β-tACS: r = 0.70, P = 0.001; γ-tACS: r = 0.87; P < 0.001). Fisher’s test demonstrated that changes in velocity and amplitude during tACS were concordant in most cases (β- and γ-tACS both P < 0.001; Fig. 2).

Finger tapping performance ratings on video recordings

MDS-UPDRS-III item 4 scores were comparable between sham, β- and γ-tACS, as shown by the non-significant factor ‘frequency’ in the Friedman test (mean rank of sham tACS: 1.92, β-tACS: 2.22, γ-tACS: 1.86; P = 0.14). Inter-rater agreement yielded a moderate concordance level for the ratings of β-tACS (K = 0.48) and γ-tACS trials (K = 0.46), and a substantial concordance level for sham tACS trials (K = 0.63).

Transcranial magnetic stimulation measures

Active (Parkinson's disease OFF versus heathy subjects: 50.1 ± 2.4% versus 50.0 ± 1.7%; P = 0.97) and resting motor thresholds (Parkinson's disease OFF versus heathy subjects: 61.1 ± 2.9% versus 63.1 ± 2.1%; P = 0.58) were comparable between patients and heathy subjects.

The amplitude of MEPs evoked by single TMS pulses was similar between groups and stimulation conditions (Table 3). SICI was overall less effective (i.e. reduced inhibition) in patients than heathy subjects, as indicated by the significant factor ‘group’. Importantly, the effect exerted on SICI by tACS differed in the two groups, as suggested by the Group × Frequency interaction (Table 3). In particular, although tACS modified SICI in both groups [heathy subjects: F(2,30) = 7.23, P < 0.01; Parkinson’s disease OFF: F(2,34) = 10.56, P < 0.001], post hoc analyses demonstrated decreased SICI (i.e. reduced inhibition) only during γ-tACS in heathy subjects (γ-tACS versus β-tACS: P = 0.01; γ-tACS versus sham tACS: P < 0.01), while SICI was less effective during both γ-tACS (γ-tACS versus sham tACS: P < 0.001) and β-tACS (β-tACS versus sham tACS: P = 0.01; β-tACS versus γ-tACS: P = 0.31) in patients (Fig. 3 and Supplementary Fig. 2). SAI was comparable between patients and heathy subjects, and it was modulated by tACS in both groups, as suggested by the significant effect of the factor 'frequency', lack of effect of the factor 'group', and lack of a Group × Frequency interaction (Table 3). Post hoc analysis showed higher SAI values (i.e. reduced inhibition) during β-tACS than γ-tACS (P < 0.01) and sham tACS (P < 0.001) (Fig. 3 and Supplementary Fig. 2).

Parkinson's disease patients OFF versus ON

Kinematic parameters

As expected, the number of performed movements, movement velocity and amplitude were higher, and movement rhythm improved (lower CV values) in patients in the ON than OFF state, as indicated by the significant factor ‘state’ in the various rmANOVAs; conversely, amplitude and velocity decrement did not change between sessions (Table 2). Importantly, the analysis demonstrated a comparable tACS effect on movement velocity and amplitude between patients in the ON and OFF state, as shown by the significant effect of the factor ‘frequency’ and the lack of a Frequency × State interaction (Table 2).

Transcranial magnetic stimulation measures

Active (Parkinson's didease ON versus OFF: 50.3 ± 2.5 versus 50.1 ± 2.4%; P = 0.88) and resting motor thresholds (Parkinson's didease ON versus OFF: 60.4 ± 2.5 versus 61.1 ± 2.9%; P = 0.71) were comparable between patients in the ON and OFF state. Dopaminergic therapy did not modify the amplitude of MEPs evoked by single TMS pulses, SAI or SICI, as indicated by the non-significant effect of the factor ‘state’ in rmANOVAs (Table 3). Similar to kinematic parameters, the analyses demonstrated that tACS effects on SAI and SICI were comparable between patients in the ON and OFF state, as suggested by the significant effect of the factor ‘frequency’ and the lack of a Frequency × State interaction (Table 3).

Clinical and neurophysiological correlations

In patients in the OFF state, there was a relationship between β-tACS-induced changes in SICI and movement velocity (r = 0.65, P < 0.01) and amplitude (r = 0.52, P = 0.02). That is, the greater the modulation of SICI (higher ratio SICI β-tACS/sham tACS), the lower the detrimental effect of β-tACS on movement parameters (higher ratio velocity β-tACS/sham tACS) (Figs 3 and 4). Also, there was a positive correlation between γ-tACS-induced changes of SICI and movement velocity (r = 0.72, P < 0.001) and amplitude (r = 0.65, P < 0.01), i.e. the greater the modulation of SICI (higher ratio SICI γ-tACS/sham tACS), the higher the beneficial effect of γ-tACS on movement (higher ratio amplitude γ-tACS/sham tACS) (Figs 3 and 4). Moreover, we found a positive correlation between motor symptom severity (total MDS-UPDRS-III score) and the effect of γ-tACS on movement amplitude (r = 0.57, P = 0.01). The higher the MDS-UPDRS-III score, the greater the beneficial effect of γ-tACS (higher ratio amplitude γ-tACS/sham tACS). No relationship emerged between MDS-UPDRS-III scores and the effect of β-tACS on movement velocity (r = −0.23, P = 0.35) or the effect of β- (r = 0.02, P = 0.94) or γ-tACS on SICI (r = 0.36, P = 0.14). Finally, no significant correlations between tACS effects on kinematic and TMS parameters (r ranging from 0.07 to 0.37; P ranging from 0.13 to 0.78) or clinical-neurophysiological correlations were detected in patients in the ON state (r ranging from −0.40 to 0.26; P ranging from 0.09 to 0.53).

Discussion

This study demonstrated two novel findings. First, we found that β-tACS reduced movement velocity and γ-tACS increased movement amplitude in patients, while the sequence effect did not change during β- and γ-tACS. Velocity and amplitude changes during tACS correlated, and most patients showed concordant modifications of both parameters during tACS. Second, γ-tACS decreased SICI and β-tACS decreased SAI in both patients with Parkinson's disease and heathy subjects, while β-tACS decreased SICI only in patients. Importantly, tACS effects on M1 excitability correlated with kinematic changes. SICI modulation was more pronounced in patients showing less movement velocity and amplitude worsening during β-tACS and greater improvements in these movement parameters during γ-tACS. Finally, tACS effects on bradykinesia and M1 excitability were comparable between patients in the ON and OFF state. These data overall demonstrate that driving β and γ oscillations at the M1 level modulates bradykinesia in Parkinson’s disease and provide insight into the mechanisms responsible for this phenomenon.

TACS-induced modulation of bradykinesia

In line with previous studies, we found that patients with Parkinson's disease showed movement amplitude and velocity reduction, progressive amplitude decrement during finger tapping (sequence effect) and altered rhythm as compared with heathy subjects.^{23,48,63–65} The novel finding was that β-tACS further decreased movement velocity in patients (i.e. slowness worsening) compared with γ- and sham tACS. The tACS mechanism of action is entraining the firing of neuronal populations and boosting entrained cortical oscillations.^27–31 We thus hypothesize that the bradykinesia worsening we observed during β-tACS was due to β rhythm enhancement in M1. Beta activity in the human M1 is antikinetic,^66–69 and increased β synchronization in the basal ganglia-thalamo-cortical loop is a well-known pathophysiological marker in Parkinson’s disease^2,4,70 that correlates with bradykinesia severity.^9–15 Specifically, bradykinesia is thought to result from the increased firing rate of indirect pathway neurons, which determines an abnormal β frequency in output basal ganglia nuclei and, in turn, excessive β oscillations at the cortical level.^71,72 Previous studies have suggested that cortical β can drive subcortical activity in Parkinson’s disease.^24,73,74 In line with these data, the slowness worsening we observed during β-tACS may suggest that entraining β resonant neurons in M1 enhances pathological β activity in the entire basal ganglia-thalamo-cortical network. Accordingly, increasing β oscillations in M1 promotes bradykinesia in patients (‘pro-bradykinetic’ effect).

We also found that γ-tACS modified movement amplitude in patients, which increased during γ-tACS as compared with β- and sham tACS. Gamma activity has a prokinetic role in humans since it synchronizes during voluntary movement execution.^6,75 In Parkinson’s disease, γ power reduction is associated with motor impairment, and the degree of local γ synchronization in the subthalamic nucleus (STN) correlates with motor performance.^11,16 STN γ rhythm is coherent with M1 activity,⁷⁶ and cortical γ oscillations and STN-cortical γ coherence increase in parallel to motor symptoms improvement.^{18,25,76–78} Interestingly, it has been hypothesized that increased γ activity in Parkinson’s disease may exert a compensatory role in response to excessive antikinetic β activity.^17,25 Our results showing improved movement amplitude during γ-tACS are consistent with the idea that entraining γ rhythm counteracts bradykinesia in patients with Parkinson's disease (‘anti-bradykinetic’ effect). Notably, we also found that the effect of γ-tACS on movement amplitude was positively related to motor symptoms severity (total MDS-UPDRS-III score). This relationship suggests that boosting γ oscillations determines greater anti-bradykinetic effects in more affected patients. In line with these data, a recent study demonstrated greater γ activity impairment in patients with more severe symptoms.¹⁸

Despite the selective effect of β- and γ-tACS on a single movement parameter, correlation analyses demonstrated that the amplitude-velocity relationship is preserved in patients during stimulation,⁷⁹ and proportional amplitude and velocity modifications occurred during tACS. In addition, most patients showed concordant kinematic changes during tACS. Overall, these findings indicate that amplitude and velocity are symmetrically shifted upward or downward in patients with Parkinson's disease by γ- and β-tACS, respectively.

Another study finding concerns the lack of the sequence effect modulation by tACS, a typical component of Parkinson’s disease bradykinesia.^1,41,80 Specific mechanisms within M1 and abnormalities in other brain areas are known to concur in the pathophysiology of the sequence effect in Parkinson’s disease.^1,12,81,82 Our results may suggest that the sequence effect is not significantly influenced by changes in β and γ rhythms of M1, and the role of abnormal oscillations in generating this bradykinesia feature may prevail at the basal ganglia level.¹² Moreover, the neurophysiological substrates of sequence effect in Parkinson’s disease include M1 plasticity impairment.²³ Since previous studies demonstrated that β- and γ-tACS delivered alone on M1 does not induce cortical plasticity changes,^{37,38,52,58,83} we speculate that mechanisms underlying impaired M1 plasticity in Parkinson’s disease do not ameliorate during tACS, thus leaving the sequence effect unmodified.

Previous observations have shown that tACS modifies motor behaviour in young heathy subjects,^51,53,55,56 while we found no tACS effects on movement kinematics in our sample of elderly heathy subjects. Since we recently demonstrated that tACS effects on M1 decrease with physiological ageing,⁸⁴ it is possible that age-related mechanisms could have contributed to the lack of behavioural changes during tACS in our heathy subjects group. This is not necessarily the case in patients with Parkinson's disease, where behavioural tACS effects are still present despite ageing. These findings possibly indicate that pathological brain rhythms are more susceptible to be modulated by tACS than physiological rhythms.

TACS-induced modulation of intracortical excitability and relationship with movement changes

In line with previous evidence,^36,37,39 we found that γ-tACS selectively decreased SICI (reduced inhibition) while β-tACS decreased SAI in heathy subjects, two TMS parameters reflecting the activity of different inhibitory intracortical circuits within M1. Animal and human studies revealed that different subpopulations of cortical interneurons are resonant to the γ^85–87 and β rhythm.^88–92 Accordingly, we suggest that γ- and β-resonant inhibitory interneurons in M1 are activated by γ- and β-tACS, respectively.^36–39,58 Importantly, in patients with Parkinson's disease, both γ-tACS and β-tACS modified SICI, a measure of GABA-A-ergic interneuronal activity. Since SICI was impaired at baseline in our cohort of patients, we speculate that these γ- and β-resonant interneurons are still functionally active in Parkinson’s disease.^52,93 Furthermore, since SICI changed during β-tACS in patients and not in heathy subjects, we hypothesize that the susceptibility of GABA-A-ergic interneurons to β oscillations is enhanced in Parkinson’s disease. Importantly, SICI changes during β-tACS were negatively related to the effect of stimulation on movement velocity and amplitude, i.e. the greater the SICI modulation (reduced inhibition), the less the detrimental effect on movement kinematics. This may indicate that reduced GABA-A-ergic activity within M1 represents a possible compensatory mechanism against the pro-bradykinetic effect of β oscillations in Parkinson’s disease. In line with our hypothesis, a recent MEG study demonstrated that GABA-ergic modulation alleviates bradykinesia driven by increased cortical β activity during movement.²⁴ Moreover, facilitation of corticospinal output through reduced inhibitory processes has been identified as a compensatory mechanism to counteract motor symptoms in Parkinson’s disease.^1,94,95 Finally, we also found that SICI modulation during γ-tACS was positively related to movement velocity and amplitude changes, i.e. the greater the SICI reduction, the wider and faster the movement. This relationship further emphasizes the role of reduced GABA-A-ergic inhibition within M1 as a possible neurophysiological substrate of compensatory anti-bradykinetic effects in Parkinson’s disease. Overall, our findings point to the modulation of β- and γ-resonant interneurons as the underlying neurophysiological mechanism responsible for tACS-induced bradykinesia changes in Parkinson’s disease.

Clinical effects of tACS and influence of dopaminergic therapy

The video rating of motor performance during tACS demonstrated similar MDS-UPDRS-III item 4 scores between stimulation conditions in patients. This result either indicates that kinematic effects of tACS are subclinical or reflects the lack of the sequence effect modulation, which is instead incorporated into MDS-UPDRS-III bradykinesia items. Our data also show that β- and γ-tACS neurophysiological effects were comparable between patients in the OFF and ON medication condition. This observation suggests that tACS effects on movement execution in Parkinson’s disease are not influenced by the dopaminergic status of patients, i.e. amplitude and velocity are proportionally modified during stimulation, independent of the basal level of motor performance. Accordingly, tACS effects on the activity of both β- and γ-resonant neurons in M1 were similar, regardless of the level of central dopaminergic activity. The lack of a dopaminergic effect on tACS-related modulation of movement kinematics may also indicate that β- and γ-tACS effects do not depend on baseline cortical β and γ activity, i.e. they are similar despite possible β and γ oscillation changes following L-DOPA intake.^13,18,20,96 In addition, previous evidence has shown that L-DOPA administration variably modifies neurophysiological measures in Parkinson’s disease, and this variability may have contributed to our results.^23,97

Confounding factors and limitations

The experimental design of the study allowed us to exclude several factors that could have influenced our data. First, we ensured several minutes of rest between kinematic trials and blocks and 10 min between the various TMS recordings to exclude fatigue or carryover effects of tACS. Furthermore, the three stimulation conditions were always applied in a randomized order. Since we used sham tACS as a control condition and no visual or cutaneous sensations were perceived during β-tACS or γ-tACS, we could exclude that unspecific attentive or placebo effects influenced our results. Again, baseline level of cortical excitability was similar between groups and between patients in the OFF and ON state. However, although our findings can be interpreted as the result of resonance phenomena, we did not provide direct evidence of cortical rhythms entrainment in our patients. In this regard, β- and γ-tACS effects on M1 are known to occur only during stimulation,^{30,37,39,52,58,83} and to date it is difficult to reliably record EEG during tACS in humans due to the stimulation artefact, which is significantly greater than cortical activity. In addition, methods to avoid artefact-related contamination on EEG recordings are not standardized and are still a matter of debate.^98–100 Direct evidence of effective tACS entrainment in humans is still lacking, and thus future studies are needed to resolve this relevant issue. Moreover, following the methodology of previous studies that demonstrated significant β- and γ-tACS effects on M1 functions,^{39,51,53,55,56,58} we applied the stimulation at fixed frequencies of 20 and 70 Hz. Thus, we cannot exclude that using tACS at individual β and γ frequencies would have produced even greater effects. Finally, although some TMS measures are known to be modulated according to tACS phase,³⁶ we applied TMS independent of phase in this study.

Conclusions

This study provides new insights on the role of β and γ oscillations at the M1 level and further understanding of the pathophysiological mechanisms underlying bradykinesia in Parkinson’s disease. Our data suggest a relevant role of β and γ rhythms in the motor cortex in modulating movement performance in Parkinson’s disease. Increased cortical β oscillations favour bradykinesia (pro-bradykinetic effect), whereas enhanced cortical γ oscillations may have an anti-bradykinetic effect. Interestingly, the effects of cortical β and γ oscillations on bradykinesia are linked to changes in the activity of inhibitory GABA-A-ergic interneurons within M1. Cortical disinhibition may act as a compensatory mechanism to bradykinesia in Parkinson’s disease.^94,95 Our data also support the hypothesis that changes in cortical β and γ rhythms in Parkinson’s disease reflect the propagation of altered oscillations within the basal ganglia-thalamo-cortical network,^73,74,76 which has a relevant role in generating bradykinesia.^1,4 Moreover, we provide the first proof-of-principle that targeting the cortical node of the basal ganglia-thalamo-cortical oscillatory network may determine significant changes in movement performance in Parkinson’s disease. In this regard, the ameliorative effects produced by γ-tACS may support the development of novel non-invasive brain stimulation approaches with therapeutic purposes in Parkinson’s disease.

Acknowledgements

The authors wish to thank all patients and healthy subjects for their participation in this research.

Funding

No funding was received towards this work.

Competing interests

The authors report no competing interests.

Supplementary material

Supplementary material is available at Brain online.

References

Abbreviations

 
• M1

  primary motor cortex

 
• MDS-UPDRS

  Movement Disorder Society-sponsored revision of the Unified Parkinson's Disease Rating Scale

 
• MEP

  motor evoked potential

 
• SAI

  short-latency afferent inhibition

 
• SICI

  short-interval intracortical inhibition

 
• tACS

  transcranial alternating current stimulation

 
• TMS

  transcranial magnetic stimulation