Neurodegenerative diseases were one of the first ailments to receive medical attention worldwide [1]. Clinically, these diseases show heterogeneous and overlaying symptoms. Hence, a framework that hones diagnostics will allow better specific treatment and management. Parkinson's disease (PD) is one of the leading neurodegenerative disorders with heightened prevalence [2,3]. It is considered a movement disorder, although it is accepted that the classic motor symptoms are accompanied by a myriad of nonmotor symptoms as well as hyposmia, urinary dysfunction, orthostatic hypotension, memory loss, depression, pain, gastrointestinal dysfunction, and sleep disturbances [4]. Gastrointestinal symptoms include drooling, dysphagia, disabled gastric emptying, constipation, and impaired defecation[5,6].

Several cellular mechanisms, including mitochondrial dysfunction, oxidative stress, neuroinflammation, and deficient protein degradation are implicated in the pathogenesis of PD. Nonetheless, the pathological fingerprint consists of neural inclusions of Lewy bodies (LBs) and Lewy neurites, with cell loss in the substantia nigra and other brain areas. The burgeoning of LBs began from an initial template from alpha-synuclein, which incited the seeding of nearby alpha-synuclein proteins that triggered aggregates, a toxic, insoluble-pleated sheet structure, to form LB. [7]. At the network neurophysiology level, these pathological fingerprints lead to rearrangement in the electrophysiological and neurophysiological activity that is generating the symptoms [8,9,10].

Despite the above, the diagnosis is complicated by the overlap of motor and nonmotor symptoms in addition to the possibility of other neurodegenerative diseases; as a result, disease management remains suboptimal [4]. In recent years, the criteria for diagnosis have been designed and validated and are dependent on the presence of motor symptoms [4,11,12]. In this context, we reviewed insights related to the neurophysiology of brain networks associated with symptoms of PD and included evidence from neurophysiological tests that contribute to better diagnosis and management.

Network implied in motor symptoms.

Motor symptoms dominate the clinical expression of PD. Muscular rigidity, akinesia, bradykinesia, gait instability, and resting tremor form a core of the motor symptoms [9,11,12]. The concept of “parkinsonism” encompasses all motor impairments. For the clinical diagnosis, parkinsonism is defined as bradykinesia accompanied by rest tremor, rigidity, or both [7,11,12]. Dopamine (DA) loss secondary to degeneration of neurons in the substantia nigra pars compacta (SNc) initiates parkinsonism by a leak of modulatory function in the network [7,9,13].

Motor output is modulated by the basal ganglia (BG). The BG is composed of several nuclei: the striatum (Str), the external (GPe) and internal (GPi) segments of the globus pallidus, the subthalamic nucleus (STN), the SNc and the reticulata (SNr). Functionally, the cortex (Cx) sends motor information by excitatory axons to the Str, STN, and thalamus (Th). In this way, the information reaches the circuit through the Str and emerges through the output nuclei, the GPi/SNr, which then sends the information to the thalamus (Th). The circuit is organized according to the projection neurons of the Str that send their axons to the output nuclei in a dual manner. Thus, the connection between the Str and GPi/SNr forms the "direct" pathway. At the same time, the Str establishes a connection before reaching the output nuclei with the GPe and the STN, giving rise to the “indirect” pathway (Str-GPe-STN-GPi/SNr; Fig 1A) [14, 15, 16] In turn, this network is included in larger parallel circuits that involve regions of the frontal Cx and the ventral Th and include two nuclei belonging to the brainstem: the superior colliculus [17] and the pedunculopontine nucleus (PPT) [18]. Based on functions of the cortical area of origin, the BG-Th-Cx network is designated as “motor," “associative/cognitive," “limbic," “motor," “associative/cognitive," and “limbic.”[8,19] In our context, parkinsonism arises from abnormal activity patterns in the motor circuit.

Synthesized in the SNc, DA is a critical modulator in the network. Expressed in both direct and indirect pathways, DA receptors are coupled to different second messenger systems (through Gs or Golf for D1-like receptors and Gi or Go for D2-like receptors). In the direct pathway, DA acts on D1 receptors to inhibit the BG output nucleus. DA acts on D2 receptors in the indirect pathway to suppress activity. Under normal conditions, DA release in the Str reduces the combination of these effects under GPi and SNr activity, reducing the inhibition of thalamocortical neurons that receive the input from the output nucleus (Fig 1A) [20,21]. Thus, the loss of DA by nigrostriatal pathway degeneration induces aberrant transmission of the sensorimotor striatum [21] (more strongly than transmission to the associative and limbic regions); consequently, this pathway degeneration allows GPe-STN activity to go into overdrive, thus raising the inhibition of STN neurons and their projections to output nuclei [22, 23]. This change causes parkinsonism (Fig 1B).

Neuronal activity patterns play an essential role in determining the integrative functions of the BG. In neural ensembles, information is transmitted through temporal patterns of action potentials [24]. Therefore, it is accepted that the information is encoded in the firing rate of individual neurons [25]. In this context, changes in the firing rate of individual neurons in some specific nuclei of the BG induced by DA loss explain the pathophysiology of parkinsonism. In the network, loss of DA reduces the direct pathway's tonic excitation and the indirect pathway's tonic inhibition [13,21]. Both changes increase the mean firing rates of output nuclei. Consequently, the BG overinhibits their thalamic and brainstem targets [26]. This causes decreased activity in Th and Cx, resulting in akinesia.

Other nuclei that have shown changes in the firing rate secondary to the loss of DA are the Str and the Th. In the Str, the neurons projecting to the direct pathway decrease their spontaneous activity, while those of the indirect pathway increase it [8]. In Th, neurons slow their firing rate after the loss of DA [27], and their firing is modulated during reaching movements [8]. The GPe decreases its firing rate after DA depletion [28], and at the same time, the local levels of GABA are increased [29,30]. Reports of changes in the motor Cx in parkinsonism are scarce. Recently, the decreased firing of neurons projecting to the pyramidal tract was shown but did not affect those projecting to the Str [31]. These results suggest that transmission from cortical neurons to pyramidal tract neurons might be involved in motor symptoms associated with parkinsonism [8]. It has been proposed that cortical firing activity could be secondary to dysfunction in the BG and Th and possible changes in the Cx secondary to the loss of DA and other neurotransmitters [32]. However, other studies show findings related to the alterations associated with parkinsonism, both in firing frequency and patterns and network synchrony.

Firing pattern implication in the neurophysiology of the BG network.

Physiologically, the action potential is the canonical form of information transmission in the brain. Nevertheless, it is known that specific neurons, based on their intrinsic electrical properties, can show increased mean firing frequency over a short time; this type of activity is described as the burst pattern [13,33]. Thus, the previously described firing frequency changes have been added to support action potential bursts.

In the BG normal network, the firing pattern of the GPi, GPe, and STN neurons is random, although action potentials do not occur in bursts. In this condition, the GPi fires action potentials continuously at high frequency. In the same way, the GPe fires action potentials at high frequency, although with pauses, and the STN also fires action potentials continuously but in a medium range of frequencies [13] . These characteristics change considerably secondary to DA depletion. Extensive evidence allows us to accept that DA depletion modifies the intrinsic properties of the neurons of different nuclei of the circuit [34,35,36]. Similarly, the GPe increases the range in which burst firing occurs, adding to the diminished firing rate[38,39]. Similarly, the STN modifies its firing pattern so that it is similar to bursting and increases its firing frequency[40]. Notably, burst activity in the STN has been resolutely correlated with clinically severe parkinsonism in patients with PD [41].

The reciprocal connection between the GPe and STN inside the network is physiologically transcendental. Both cores are considered the pacemaker of the network [42], and this notion is given particular importance in the development of bursts [23]. Higher activity from the indirect pathway onto the GPe that guides rhythmicity is demonstrated after DA loss by increased density in the synaptic link between both cores; consequently, the GPe increases inhibition that causes a hyperpolarization-induced higher burst pattern in STN neurons [23,42]. In addition, during parkinsonism, neurons of the GPe and GPi increase synchronization, and the GPi tends to fire in a burst pattern [43].

Thalamic neurons have intrinsic properties that allow them to burst under normal conditions or exhibit tonic bursting depending on the physiological state [44]; the burst pattern trends for the BG nuclei are similar to those in the Th [45], especially in the motor region [40,45]. Notably, burst activity results from a convergence of axons from the cerebellum in the Th motor regions [. In this sense, a connection between the cerebellum and BG was described recently and suggested implications for parkinsonism symptoms [8,46]. Therefore, the cerebellum-Th-Cx network contributes to parkinsonism [47].

Network oscillation.

Burst firing is fundamental in network physiology. This pattern augments the reliability of communication between neurons and contributes to integrating local and distal network information [48]. When bursting occurs rhythmically in an ensemble, it results in oscillatory activity [13]. At the same time, the oscillatory activity that is caused by the temporal interaction of neural activity in the network causes synchronization [49]. As previously described, pathophysiological changes induced by DA loss modify the firing pattern of all nuclei integrated into the BG-Th-Cx network; as a result, abnormal oscillatory activity and dysfunctional synchronization originate [49]. A wide range of current evidence holds that motor output and parkinsonism occur in the beta frequency band [50,51,52,53,54].

In the resting state, the primary motor cortex (CxM) exhibits widespread oscillatory activity in the beta band [55]. Currently, the beta band is described in two subbbands: low (13-19 Hz) and high (20-30 Hz). Cortical oscillations induced oscillations in the BG; this functional link leads to the spread of oscillatory behavior inside the Cx-BG-Th network [49,54]. The BG structures associated with oscillations in the beta band at rest are the Str [56], the GPe-STN microcircuit [23], and the STN [57]. Beta band oscillations were also observed in the premotor area and cerebellum. Importantly, the main oscillatory activity at rest occurs at a very slow time scale [58].

The physiological effect of network synchronization in the motor system was identified during movement kinetics. The parameter analyzed was beta power. In this way, during motor behavior, the beta band displays minimal power in movement stages and high power amid postural maintenance, such as during stable object holding [59,60,61,62]. Based on these observations, beta oscillations are postulated that preserving the current motor state.(i.e., “status quo”) [63]. The motor task is a shared resource for studying beta bands during motor performance. Thus, during movement execution and changes in isometric muscle contraction, beta power is lowest. The drop in beta power during movement is observed bilaterally, sometimes with a contralateral preponderance. A decrease in beta power also occurs when no active muscle contraction is needed, such as during action observation or passive movement [62].

A relative increase in beta-band power is present during static postural maintenance, such as to keep an object stable [62,64]. The increment onset is approximately 300 ms after the beginning of the stable grasp [65]. In the course of static holding, the tonically contracting muscles display significant coherence and phase synchronization in the beta band. Likewise, the firing of cortical motor neurons (including pyramidal tract neurons) is phase-locked to the beta band [62]. Notably, the thalamocortical network displays coherence in the beta band during movement preparation, isometric contraction, and at rest [49,65]. Throughout movement kinetics, beta power shows a progressive decline approximately 1–2 s before movement starts. Importantly, this power decrease seems to be specific to limb movement coordination. Conversely, 300 to 1000 ms after movement is completed, beta power exhibits a prominent increase known as postmovement beta rebound [55,62, 67].

Network oscillation in parkinsonism.

In parkinsonism, the structures that show aberrant oscillation in the beta-band frequency are the Cx, Str, GPe, STN, GPe-STN loop, and output nuclei GPi/SNr. Additionally, cortical coupling occurs at 10-35 Hz, which correlates with parkinsonism-related acerbity [54]. This process diminishes after both DAergic treatment and STN stimulation [54]. The 11–30 Hz frequency range is mainly anti-kinetic [68]. In this sense, synchronized oscillations between the STN and GPi were shown in PD patients in this frequency range [69,74], which has implications in akinesia or bradykinesia. Importantly, coherence between the CxM and STN-Gpi is displayed in PD patients without pharmacological therapy, which is reduced after they begin pharmacology therapy and in response to voluntary movement [72].

In contrast, tremors exhibit a low range of oscillatory activity: 3-10 Hz. The nuclei implicated are the GPi, STN, and Th [49]. Notwithstanding, the network implicated in genesis is not limited to the BG-Th; it also includes the cerebellum-Th-Cx network and the interaction among these regions [47,73]. In this way, oscillatory activity in the CxM is consistent with the oscillations in the Th, BG, and cerebellum; simultaneously, it propagates to the STN and STN–GP networks. Therefore, the primary motor cortex takes part in the network generating tremor.

During choice-reaction tasks, parkinsonism exhibits cortical and subcortical regions that display event-related desynchronized (ERD) activity before and during performance. In addition, during task performance, the moment of early motor preparation presents high beta-band desynchronization, and the high beta band increases in power during the ‘stop’ phase of movement [54,74,75]. Relevantly, in rest and isometric muscular contraction, the sensorimotor Cx displays higher power in the beta band in the early stage of PD[76] .

Similarly, the gait in parkinsonism was modified by beta band activity. In stepping or gait, beta activity shows alternating suppression in PD patients with respect to the resting state. Based on results in the stepping and walking tasks, an alternating suppression of beta activity during stepping or gait in PD patients, equivalent to that in the resting state, was reported. Additionally, the movement between cued upper and lower extremities induced greater desynchronization of high beta oscillations (24–31 Hz) modulated by movement. PD patients with akinetic rigidity also depict beta desynchronization when walking. In regular walking, the suppression of beta oscillations is related to a pattern of a left-right alternation[77,78,79]. Overall, the previous results show that network dynamics disorders in PD patients are centered around increased beta-band frequency.