We have shown an increase in subthalamic beta-gamma PAC between standing and walking in PD. Patients with poor PAC modulation and low beta-gamma PAC during walking spent more time in the weight-bearing (stability) phase of the gait cycle. We also showed that striatal dopamine promotes gait-related cross-frequency coupling in the STN of parkinsonian patients.
Human walking is a complex motor behavior that requires timed coordination of several cortical and subcortical brain areas [1–3]. The STN is a key node of the supraspinal locomotor network that is directly connected to the supplementary motor and parietal areas and projects to the mesencephalic locomotor region [52, 53]. In recent years, technological advances [19] have enabled important information to be obtained about the composite subthalamic dynamics of human locomotion [6, 10]. Some studies reporting basal ganglia field potential recorded from implanted DBS leads in patients with PD showed modulation of beta oscillations during stepping and actual gait [23–26]. Still, the analysis of power spectral modulation might not fully capture the complex large-scale network dynamics of gait control [27, 28], which eventually encompass multiple frequencies [16] to direct the information flow across distant brain areas [32, 54–57]. Indeed, there is growing evidence that frequency modulation conveys information about movement execution in a patient-specific and frequency-related manner [15–17, 31, 56], and that dopamine deficiency results in impaired encoding of this information [15, 17].
As a special case of cross-frequency coupling, PAC is modulation of the amplitude of high-frequency oscillations by the phase of low-frequency ones and represents information processing and transmission across brain areas that are involved in multiple activities, such as cognition, perception, and movement [33, 42, 47, 58, 59]. The entrainment of oscillatory activity in one frequency band according to the phase of another frequency band has been proposed to be a gateway mechanism to selectively allow task-relevant inputs to be processed [54, 60–65]. Low-frequency oscillations would act as a carrier [16] that coordinates neural activity of local and remote brain region for long-range communication through frequency and amplitude modulation [17, 66]. This has been shown in the hippocampus, for example, to organize the readout from long-term memory of the discrete sequence of upcoming places, as cued by current position [67]. Similarly, we envision a key role for the STN, where beta phase-coding is a mechanism to selectively allow task-relevant inputs to be processed and passed through to basal ganglia output regions for an action release (gamma rhythm) [68, 69]. Interestingly, low beta frequencies may be involved in the keying of a (motor) behavior, while high beta frequencies would carry information about its execution [17]. This interesting perspective needs further specific investigation.
In PD, finely-tuned gamma oscillations (60–90 Hz) are a prokinetic network phenomenon [66, 70] that increase during voluntary movements [68, 71] and correlate positively with movement velocity [72]. Previous studies in PD identified excessive beta-phase coupling to broadband high-gamma amplitude in M1 [73] and STN phase-M1 amplitude [61, 74]. This was associated with the parkinsonian motor state and was modulated by dopaminergic medication and DBS therapy [47, 73, 75]. As for beta power modulation, however, these results are read on the understanding of a direct link between beta oscillations and akinetic-rigid symptoms, without allowing for the physiological or compensatory contribution of these signals [17, 18]. Similar reasoning can be applied to freezing of gait [76], where it is necessary to distinguish the actual episode of gait freezing from the component of standing while freezing, and attempting to overcome the freezing episode, as well as the physiological neural activity related to gait modulation [35]. This aspect has not yet been studied in a precise and standardized method [77].
Our data favor the hypothesis of a physiological contribution of subthalamic beta-gamma modulation to human gait or a compensatory activity based on the residual dopaminergic availability to promote locomotion in parkinsonian patients. This also partly derives from the fact that all subjects enrolled showed substantially normal gait kinematic measures (see Table 2 in [28]) and from the fact that poor PAC modulation results in a longer time in the stance (weight-bearing) phase of the gait cycle. In addition, striatal dopaminergic innervation promotes subthalamic phase-amplitude modulation.
A major limitation of our study is that we have not described a causal or exclusive relationship between increased subthalamic beta-gamma PAC and gait. We were also unable to define the origin of beta-gamma PAC in the STN of parkinsonian patients. de Hemptinne and coll. have proposed that this activity results from the organization of STN spikes into synchronized bursts, with a short interval within the bursts [74]; however, comparison of PAC and spike-phase locking values have shown no correlations [78]. It would also have been very interesting to evaluate beta-gamma PAC between the STN and cortical areas; however, in our study we aimed to further explore the contribution of the STN to parkinsonian gait. Reliable subthalamic input signals (which may include PAC [79]) that code for gait are still an unmet need when using novel stimulators in adaptive mode while preventing patients from additional implants. In this study, cortical EEG were used (see Methods) to obtain trustworthy subthalamic LFPs given the numerous artifacts in the recordings, which cannot be simply visually discharged [24, 34, 36]. Further studies will investigate the role of different cortical areas and the interesting cortico-STN interactions during human walking. The Activa PC + S device gave us the possibility of performing recordings in chronically-implanted patients, reducing the impact of the stun effect. However, this added an additional variable to consider, in that chronic treatment with DBS (in patients also stimulated for years with dopaminergic drugs) may result in long-term effects. The “exhaustion” of these effects was partly monitored in our studies by a return to symptom severity after turning the stimulator off, comparable to the pre-implantation assessment (see Table 1 in [28]).
Being able to understand the contribution of specific brain areas such as the STN in the context of complex motor behaviors remains one of the frontiers of neuroscience. The difficulty of coding the physiological, compensatory, and pathological aspects of human gait can be perceived from the fact that we have no effective pharmacological or neuromodulation treatment for gait disorders. We are hopeful that advances in technology will allow us to collect more and better data, even in an ecological context, to learn more about bipedal walking – a simple yet complex motor act that defines our species [21].