Parkinson’s disease (PD) is a progressive neurodegenerative disorder characterized by bradykinesia/akinesia, rigidity, resting tremor, and gait instability (Marsden, 1994). The pathological hallmark of the disease is the degeneration of dopamine neurons in the substantia nigra pars compacta (Parent and Parent, 2010). The motor signs commonly begin on one side of the body and extend to the other side with disease progression. The onset side generally remains more severely affected over the course of the disease (Toth et al., 2004). These phenomena are accompanied by greater neuronal loss in the substantia nigra and more impaired putaminal uptake of ¹⁸F-fluorodopa contralateral to the side of initial onset (Kempster et al., 1989) or more severely affected side (Leenders et al., 1990). In particular, the clinical decline parallels the course of neuronal degeneration and an inverse linear relation between the UPDRS score and the number of nigral dopaminergic cell was observed (Lee et al., 1994; Tatsch et al., 1997). Thus, it is most likely that the asymmetry in clinical manifestations between the two sides of the body is associated with the asymmetry in the pathological progression in the two hemispheres. However, the mechanisms underlying the motor symptoms severity and its progress are still poorly understood.

The subthalamic nucleus (STN) is a favorite target for deep brain stimulation (DBS) for PD because of its fundamental role in modulating the activity of the basal ganglia (DeLong and Wichmann, 2015). It is believed that the nucleus is a main relay station of the indirect pathway of the basal ganglia, is under the control of the external pallidal segment (GPe), and directly links the cortex and cerebellum. Its action, together with signals from the direct inhibitory GABAergic pathway, modulates the activity of the globus pallidus internus (GPi) and the substantia nigra pars reticularis (SNr). Successful DBS of the STN for PD motor symptoms further implies its central role in the control of movement (DeLong and Wichmann, 2015). The implantation of DBS electrode into the STN allows microelectrode recording from this structure in both hemispheres.

Based on the classical model of pathophysiology of PD, alterations in neuronal activity in the basal ganglia are predicted to underlie parkinsonian motor symptoms (DeLong, 1990; Galvan et al., 2015). These altered neuronal activities include changes in firing rate and patterns, pathologic oscillation and synchronization in the STN and GPi. In support of the prediction, electrophysiology studies in 1-methyl-4-phenyl-1,2,3,6- tetrahydropyridine (MPTP)-treated non-human primates and in patients with PD have found an increase in firing rate, bursting activity and oscillatory activity as well as synchronization in the subthalamic neurons (Bergman et al., 1994; Hassani et al., 1996; Wichmann and Soares, 2006; Breit et al., 2007). Both microelectrode and local field potential recordings show that there are tremor frequency (4–6 Hz) oscillatory activity and β frequency (10–30 Hz) oscillatory activity observed in the STN. Close relationships between the β oscillations and rigidity–bradykinesia (Weinberger et al., 2006, 2009) and between the tremor frequency oscillations and limb tremor in the STN have been reported (Hutchison et al., 1997; Zaidel et al., 2010; Guo et al., 2013). In accordance with the model, intervention by lesioning or by high-frequency stimulation in the STN in animal model and patients with PD results in the improvement of parkinsonian symptoms (DeLong and Wichmann, 2015).

A recent study (Remple et al., 2011) demonstrated the recordings of neuronal activity from the STN of patients with early-stage PD compared with that from age- and sex-matched patients with advanced PD. They found that STN neurons have a significantly higher firing rate in advanced versus early-stage PD and the overall activity of the STN is significantly lower in early versus late PD. The data indicate that neuronal firing in the STN increases with PD progression. However, there are no significant differences between groups in the bursting or oscillatory activity in the STN.

Furthermore, some other studies found that a phase coherence in the β frequency oscillatory activity is associated with the severity of rigidity–bradykinesia (Pogosyan et al., 2010) and the spectral power at β frequency and percentage of single units oscillating at β frequency correlate with rigidity scores (Sharott et al., 2014). In particular, subthalamic neuronal oscillations and phase amplitude coupling are greater in the more affected hemisphere than that in the less affected hemisphere (Shreve et al., 2017). Although the clinical importance of the STN in PD is clear, whether the subthalamic oscillatory neurons, particular β frequency oscillatory neurons, play a critical role in PD severity and progression is still unclear.

Here, we recorded neuronal activity in the STN during targeting for the placement of DBS electrodes in PD patients. The STNs of the hemispheres contralateral to the body side of initial symptoms and contralateral to the body side of extended symptoms within individual PD patients were compared. We expect to provide important data to understand the pathophysiology underlying PD symptoms severity and its progression.

Five hundred seventy-one neurons were identified in 60 STNs of 30 patients, 30 STNs associated with the BSIS and 30 STNs associated with the BSES. The mean duration of the neuronal recordings (17.6 s ± 6.3 s vs 16.5 s ± 4.6 s) and the length of the STN (5.1 ± 1.0 mm vs 5.0. ± 1.0 mm, all p > 0.05, paired t-test) showed no significant difference between the BSIS and BSES. The main analysis focused on oscillatory neurons.

Here, we compared for the first time the subthalamic oscillatory activity in the hemispheres between BSIS and BSES in individual PD patients. There were three findings. First, the MSFR of neurons in the STN of the BSIS was significantly higher than that of the BSES. Second, the proportion of β frequency oscillatory neurons was higher in the BSIS than that in the BSES. Overall, the correlation of β frequency oscillatory neurons with UPDRS sub-scores of rigidity and bradykinesia was greater in the BSIS than that in the BSES. Third, the tremor frequency oscillatory neurons were fewer in the BSIS than in the BSES. The results indicated that increased firing rate and ß frequency oscillatory neurons in the STN were more frequent on the hemisphere opposite the more severe bradykinesia. These findings support the view that increased neuronal firing and ß frequency oscillatory neurons in the STN are associated with severity of PD and its progression. Conversely, tremor frequency oscillatory neurons are less frequent in the STN of the BSIS suggesting that ß oscillatory activity dominates and tremor frequency oscillatory activity reciprocally declines.

Our finding of a17.7% increased STN firing rate in the BSIS versus the BSES (41.3 ± 11.0 Hz vs. 35.1 ± 10.0 Hz) of the individual patients is consistent with the recent report (Remple et al., 2011) which found a 26.4% elevation in the STN firing rate of advanced stage versus early stage groups (36.3 vs. 28.7 Hz). These data support the prediction that STN neuronal firing rate increase with PD severity and its progression. According to pathophysiology “rate model” of parkinsonism, the deficits of dopaminergic input to the striatum result in disinhibition of the STN, causing overdrive the output nuclei of GPi and SNr, leading to increase inhibition of the motor thalamus, resulting in the cardinal symptoms of PD, such as akinesia and bradykinesia (DeLong, 1990). Our findings with the results of Remple et al. (2011) are consistent with the report from studies of animal model that show increased discharge rates in the STN following the application of toxic compounds to the substantia nigra (Bergman et al., 1994; Hassani et al., 1996; Wichmann and Soares, 2006; Breit et al., 2007). Examples show that monkeys exhibit a 36.8% increase in mean firing rate (19 vs. 26 Hz) when given a dose of MPTP sufficient to create severe parkinsonian symptom(Bergman et al., 1994); an increase in the average discharge rate of STN neurons from a control value of 25.7 Hz to 36.1 Hz has been reported in parkinsonian monkeys (DeLong, 1990). Accordantly, the firing rate of subthalamic neurons seems to be increased at an early stage of the disease, when patients have minor or mild symptoms, and becomes more elevated as the disease progresses and symptoms worsen. In our case, a finding was that patients had significant difference in the duration of symptoms between the BSIS and BSES groups. In accordance with the symptom duration, there was a significant difference in the lateralized UPDRS III scores between the two groups (20.8 ± 4.3 vs 15.2 ± 4.3, p < 0.01) during the “Off” state (Table 2). Thus, the increased MSFR in the STN of the BSIS group was most likely associated with more severe symptoms. Our findings, along with those of others (Remple et al., 2011) strongly support the rate model of PD pathophysiology (Albin et al., 1989; DeLong, 1990).

In this study, we demonstrated that the proportion of β frequency oscillatory neurons was higher in the STN contralateral to the BSIS than that in the STN contralateral to the BSES. Furthermore, correlations of β frequency oscillatory neurons with the UPDRS sub-scores for rigidity and bradykinesia in the BSIS were greater than those in the BSES. The results suggest that ß frequency oscillatory neurons in the STN are more frequent on the hemisphere opposite the more severe bradykinesia and rigidity. The findings are also consistent with a recent study which demonstrated that subthalamic neurons with β frequency oscillations and phase amplitude coupling are greater in the more affected hemisphere in PD (Shreve et al., 2017).

A large number of physiological studies (Brown et al., 2001; Cassidy et al., 2002; Levy et al., 2002; Priori et al., 2002; Kuhn et al., 2004, 2008, 2009; Doyle et al., 2005) have suggested that basal ganglia β frequency oscillatory neurons in the 8–35 Hz range are involved in parkinsonian rigidity–bradykinesia (Brown, 2003). Single units and local field potentials recorded in the STN from patients with PD have demonstrated that there is a relationship between β frequency oscillatory activity and rigidity–bradykinesia (Weinberger et al., 2006, 2009). Decreased β oscillation is correlated with the degree of improvement in rigidity–bradykinesia after dopamine medication (Zaidel et al., 2010). A significant correlation between L-dopa-induced β-frequency power suppression and rigidity–bradykinesia has been reported (Pogosyan et al., 2010). The β activity in the STN is also suppressed by DBS in parallel with improved motor performance in patients with PD (Ray et al., 2008; Pogosyan et al., 2010; Zaidel et al., 2010). Furthermore, a recent study found synchrony of the β oscillation in the STN and the severity of parkinsonian rigidity–bradykinesia (Weinberger et al., 2009). Moreover, a recent study (Feng et al., 2016) showed that STN β frequency oscillatory neurons are coherent with limb muscular activity in PD patients with the akinetic/rigid type. The proportion of β frequency oscillatory neurons is higher in the STN of the akinetic/rigid type than in the mixed type. All of the above evidence indicates a close relationship between subthalamic β oscillatory neurons and rigidity/bradykinesia.

In contrast, we found that the proportion of tremor-frequency neurons in the STNs of the BSES was higher than that in the BSIS (34.9 vs 21.4%). A significant correlation of tremor-frequency neurons with the tremor score in the BSES was also found. The results suggest that tremor-frequency neurons are less pronounced in the hemisphere contralateral to initiate symptom with progression. Similar findings (Toth et al., 2004) have also been reported in a clinical study. In that study, 33 patients followed for 10 years who had parkinsonian signs of greatest severity on one side but subsequently over ∼ 5.4 years, the signs gradually became more prominent on the opposite side. Forty-five percent of these patients demonstrated evolution to a rigid form of parkinsonism with disappearance of rest tremor over an average of 7.1 years after presentation. The conclusion is that tremor becomes less pronounced with progression of the disease in some patients.

The pathophysiology of PD tremor certainly differs from that underlying the akinesia/bradykinesia and rigidity (Zaidel et al., 2009). The clinical manifestations are separate, and the response of tremor to L-dopa agents is lack certain than that of bradykinesia (Hallett, 2012, 2014). Tremor often occurs early in the disease. It can occur on the body side contralateral to the otherwise most affected side where bradykinesia and rigidity are most prominent. Its severity is not always correlated with that of akinesiai/bradykinesia or rigidity and does not worsen at the same rate as akinesia/bradykinesia (Jankovic and Kapadia, 2001; Zaidel et al., 2009). Moreover, unlike rigidity and bradykinesia, the tremor severity is not often associated with striatal dopamine deficits. This suggests that non-L-Dopa mechanisms likely outside the basal ganglia play an additional important role (Helmich et al., 2012). Our findings further support the idea that the pathophysiology of PD tremor differs from that of akinesia/bradykinesia and rigidity.

The clinical feature of PD is that the motor signs usually start on one body side in an asymmetrical fashion and spreads to the other side with disease progression. The onset side generally remains more severe along the course of the disease (Leenders et al., 1990). In the current study, 30 patients present a clear unilateral motor symptom onset that gradually spread to other side after a mean time of 5.4 ± 3.4 years. The significantly different lateralized UPDRS III sub-scores for bradykinesia, rigidity, and tremor between BSIS and BSES indicated that the side with initial symptoms remains more severe with disease progression. The finding was supported by postmortem pathological studies which have shown greater neuronal loss in the substantia nigra contralateral to the initially affected side (Kempster et al., 1989). Further positron emission tomography studies demonstrated that putaminal uptake of ¹⁸F-flurodopa is more impaired contralateral to the initially more affected side (Leenders et al., 1990). Our findings indicated that increase in firing rate and β frequency oscillatory neurons are associated with the clinically more severe body side in PD patients.

In the current study, the clinical and physiological information was collected from each side of the body within individuals, so the age, gender, and L-DOPA equivalent daily dose did not influence our results. The major limitation was the small sample size of 30 patients that might decrease statistical power. Additionally, considering the issue of STN nucleus sub-region, there may be differences in neuronal cells of different regions (motor, limbic, associative). The comparison of mean values used in this study was not corrected by neuroimaging, limiting the ability to define electrophysiological indicators based on sub-region. Another limitation was that we only collected pre-surgery UPDRS III data but without data of post- surgery, so that the clinical outcome of patients was not evaluated. In addition, previous studies have demonstrated that right-handed patients tend to have initial motor symptoms on the right side (Kempster et al., 1989; Toth et al., 2004) suggesting left hemisphere predominance of nigrostriatal dysfunction in PD. However, in the current study, 83.3% (25/30) of the patients were right-handed and 36.0% (9/25) of them had initial motor symptoms on right side. Thus, the hypothesis of right-handedness and the left hemisphere dominance of nigrostriatal dysfunction in PD cannot fully explain this phenomenon, so other factors might be involved. No doubt, more strict inclusion criteria in future studies with larger patient populations are needed to further confirm the findings.

The findings suggest that the increased neuronal firing rate and proportion of ß frequency oscillatory neurons are associated with severity of PD, such as akinesia/bradykinesia, and its progression. These findings support the pathophysiology model of PD that the altered neural activities underlie parkinsonian motor symptoms. The reduced tremor frequency oscillatory neurons in the STN of the BSIS suggests that ß frequency oscillatory activity takes over the basal ganglia making it difficult for tremor neuronal activity to sustain itself. Our data also support the clinical observations that the onset of motor symptoms in PD is usually asymmetric and the side on which they are initiated is often more severely affected over the course of the disease (Toth et al., 2004).

The original contributions presented in this study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

The research was conducted based on the Declaration of Helsinki and was approved by the Xuanwu Hospital Ethics Committee, Capital Medical University. All patients signed a written informed consent for the surgery that involved microelectrode recording.

XMZ and PZ contributed by collecting the data, analysis and interpretation of the data, and drafting and revising the manuscript. MH was involved in critically revising the manuscript and provided the intellectual guidance. YQZ, JYL, YW, JPL, YPW, YSH, and YJL contributed to the surgery, data collection, and conceptualization of the study. All authors contributed to the article and approved the submitted version.