In this study, it was characterized how different types of neurons in the M1 and M2 areas represent remodeling of lateralized movement representation in the parkinsonian state.
Specifically, the effect of partial dopamine depletion on the forelimb selectivity of the motor cortex was investigated; recorded the neuronal activity from the output layer of M1 and M2, differentiating between RS and FS neurons; and identified the IT and PT neurons using the Multi-Linc method. The neuronal activity in rats performing the right–left pedal task was performed under three different conditions: healthy rats, non-lesioned hemisphere, and 6-OHDA lesioned hemisphere in hemiparkinsonian rats. The main finding of this study is that the partial dopamine depletion decreased the characteristic contralateral lateralized activity of the motor cortex. This effect was observed in the RS neurons of M1 and M2. clearly observed in this population. The non-lesioned hemisphere had the opposite effect, increasing the contralateral preference, particularly in the M1-FS neurons, indicating a possible pathological or compensatory change in the hemisphere contralateral to the lesion. In addition, a contrasting effect of the 6-OHDA lesion effect was observed in the PT and IT neuron activity of M1 and M2. In M1, only the PT neurons decreased their contralateral preference, whereas in M2, the affected population was the IT.
Partial dopaminergic depletion and task performance
In this study a partial hemiparkinsonian rat model was produced injecting 6-OHDA into the right DLS. The rats exhibited rotational behavior after apomorphine injection, with at least 150 contralateral turns per hour. This rotational behavior has been used as a reliable method for demonstrating motor imbalance in PD rat models (Bankiewicz, 2004). It has been detected rotational behavior in partial lesioned animals with cell loss as low as 40% (Truong 2006). Also,
the loss of TH staining in motor cortex and striatum was evident around three weeks after the 6-OHDA injection, where the TH staining was practically completely lost around the injection side, with a smaller, but significant decrease in M1. The loss of TH+ cells in the ipsilateral SNc was around 60%, similar to previous studies using partial lesions. Also, fast-scan cyclic voltammetry demonstrated that the localized striatum 6-OHDA lesions reduced evoked dopamine release by about 75% relative to the contralateral striatum (Amalric, 1987; Lemaire, 2012). However, even with this important reduction in the TH+ cells, the rats did not exhibit overt parkinsonian symptoms. Compensatory mechanisms in the SNc (such as acceleration of DA turnover or receptor hypersensitivity), or broader changes in the cortico-BG network, may explain why clinical signs of parkinsonism emerge only after a massive degeneration of DA neurons (Melamed, 1982). In PD patients and PD monkey models, the loss of dopamine is predominantly observed in the posterior putamen, a region of the BG associated with the control of habitual behavior (Hornykiewicz, 1998; Kish, 1988; Brooks, 1990; Redgrave, 2010).
Thus, partial dorsal striatum lesions may resemble this dopamine loss pattern. However, it is important to mention that this kind of lesion also affects the direct dopaminergic input to the motor cortex (Guo, 2015), which was evident in our study with the reduction of TH staining in M1. Therefore, the changes observed in motor cortex activity could be produced by a direct effect of the loss of dopamine in the cortex or by greater dysregulation of the cortico-BG circuit.
In any case, we observed an increase in the beta-frequency power in the cortical LFP, and increased coherence between cortical regions, which is characteristic of the parkinsonian state.
Somatosensory striatum 6-OHDA lesion similar to the used in this study produced amplified oscillations across beta and low gamma frequencies in the dopamine depleted regions during a T-maze task, which was normalized by L-DOPA treatment (Silberstein, 2005; Mallet, 2008;
Stoffers, 2008; Lemaire, 2012). Lesions in the dorsal striatum may resemble presymptomatic
or mild stages of PD, in which early neuronal changes have already occurred, but the overt symptomatology may not still be present.
The study of PD in animal models is usually performed under a complete dopamine depletion in one or both hemispheres (Emborg, 2007; Magill, 2001; Mallet, 2008). This limits the assessment of movement-related activity due to the reluctance of the animals to move (Leblois, 2006). Likewise, under operant conditions, a more extensive striatal activity disruption may affect the velocity and number of movements (Panigrahi, 2015), or the motivational and motor processing (Hori, 2019). However, the partial DLS lesion used in this work permitted the rat to perform skillful forelimb movements after training, and allowed the study of brain activity during task performance under similar forelimb movement conditions.
Consistent with our results, partial dopamine depletion does not impede learning of other tasks (Molina-Luna, 2009; Lemaire, 2012). In the right-left pedal task, under healthy conditions, the rats exhibited a preference for either right or left movements in the early training period, which disappeared along with the training progress. In the lesioned rats, the preference for movement of the right forelimb (contralateral to the non-lesioned hemisphere) was consistent across all rats. However, the performance of the lesioned rats was similar in both forelimbs after 2 weeks of training, and similar to that of the healthy rats. This correction may be a consequence of the training itself, or instead the result of compensatory changes from the non-lesioned hemisphere (Remple, 2001). This partially lesioned hemiparkinsonian model under task performance may be useful in future studies of movement-related activity in mild parkinsonism. However, previous studies show that skilled reach training was not necessary for the display of a normal motor map, as has been reported after other form of injuries in animals (Metz, 2004; Friel, 2000).
Decreased limb specificity in M1 and M2 in parkinsonian state
Selectivity during ipsilateral or contralateral movements suggests diversity in the degree of lateralization between the motor cortices, in which regions representing higher-order information exhibit less lateralization (PM, SMA in monkeys; M2 in rats) and have a greater role in bilateral coordination; by contrast, M1 encodes concrete motor information, and therefore requires higher specificity for unilateral forelimb movements (Kurata, 2007;
Nakayama, 2015; Soma, 2017). In this study, the contralateral preference, mainly in M1, was decreased by dopamine depletion. These results are in agreement with fMRI studies showing higher ipsilateral cortical activation during voluntary movements in PD patients (Wu, 2015).
The disruption of cortical interhemispheric inhibition is likely to be the primary mechanism responsible for promoting the decrease in motor cortex lateralization and the generation of mirror movements (Johnson, 1998; Almeida, 2002; Li, 2007; Wu, 2015). However, these changes are not restricted to the motor cortex. In the parkinsonian state, a large fraction of BG neurons exhibit related to movements of more than one body part, including activity modulations related to movements of both the arm and the leg, movements of multiple joints, and movements of the ipsilateral limbs (Filion, 1988; Levy, 2001; Williams, 2005; Baker, 2010;
Erez, 2011).
According to the classic rate model of PD, dopamine depletion produces an imbalance between the direct and indirect pathways of the cortico-BG network, decreasing activity in the former and increasing activity in the latter (although we did not find an overall decreased activity). In humans and primates, initial studies using a metabolic estimate of afferent terminal activity suggested increased activation of iMSNs, the STN, and the GPi (Crossman, 1985; Mitchell, 1986) and decreased activation of the GPe and thalamus (Schwartzman and
Alexander, 1985). Later, extracellular single-unit recordings in parkinsonian primates and humans revealed increased average firing of GPi neurons, suggesting that parkinsonian motor deficits may be due to excessive basal ganglia output (Filion and Tremblay, 1991; Boraud, 1996;
Heimer et al., 2002), though similar changes were not observed within the SNr (Wichmann et al., 1999). With this assumption, these findings support classical model predictions that STN hyperactivity contributes to PD motor symptoms (Bergman, 1990; Wichmann, 1994). This imbalance explains reduced thalamic output to the motor cortex, culminating in the characteristic hypokinetic manifestation of PD (Albin, 1989; DeLong, 1990; Galvan, 2008;
Pasquereau, 2011). Consistent with these predictions, firing rates in primary motor cortex were reduced in parkinsonian primates (Pasquereau, 2011; Pasquereau, 2016). While electrical stimulation may induce complex effects on activity in target structures, high-frequency DBS was initially hypothesized to decrease output of target structures, functionally acting as a reversible lesion (see Chiken and Nambu, 2016 for review).
Whereas studies in parkinsonian nonhuman primates are conflicting regarding changes in striatal activity, rodent studies demonstrate bidirectional changes in iMSN and dMSN activity consistent with classical models. In anesthetized parkinsonian rats, antidromically identified dMSNs showed decreased firing as compared to healthy animals, while presumed iMSNs showed elevated firing (Mallet, 2006; Kita, 2011). Optical activation of cortico-STN terminals, which are likely activated during electrical stimulation, robustly reversed parkinsonian deficits (Gradinaru, 2009). These findings have led to the hypothesis that STN DBS may alleviate motor deficits by altering cortical activity through antidromic activation rather than by altering basal ganglia output.
Additionally, the reduction in cortical motor output is associated with reduced control
specificity, in which the signal to the contralateral part of the body is reduced relative to the ipsilateral signal. In the case of monkey or rodent hemiparkinsonian models or early-stage PD in humans, in which one hemisphere is more affected than the other (Kish, 1988), the activity in the less affected hemisphere is associated with increasing ipsilateral control (Viaro, 2011;
Bronfeld, 2011; Wu, 2015). This reorganization of the cortical map may act as a compensatory mechanism to control the side of the body contralateral to the lesioned hemisphere (Figure 10B) (Netz, 1997). Nonetheless, these potentially compensatory changes in the non-lesioned side may produce other motor problems in the attempt to preserve the necessary output to both sides of the body, e.g., mirror movements (Vidal, 2003; Espay, 2005; Li, 2007).
In disagreement with the classic model of PD and the possible compensatory modifications, opposite changes were found in the lesioned and non-lesioned hemispheres regarding preferred movement–related activity during contralateral and ipsilateral forelimb movements. In M1-RS and M2-RS neurons we observed a decrease in the selectivity for contralateral movements in the lesioned hemisphere and an increased proportion of ipsilateral-preferring neurons (Figure 6 and 7). On the other hand, we found a higher proportion of contralateral neurons and a stronger bias for contralateral movement representation on the non-lesioned side, relative to the healthy and lesioned conditions (Figure 7B). Previous studies in patients with hemiparkinsonian symptoms revealed that during the movement of the forelimb ipsilateral to the less affected hemisphere, there is a stronger inhibitory influence from the most affected hemisphere to the opposite side (Wu, 2015). These results are in agreement with our study, which revealed reduced activity during ipsilateral movements in the non-lesioned hemisphere, resulting in a greater difference in activity between contralateral and ipsilateral movements. One mechanism that could produce
these changes in the most affected hemisphere is reduced activation of the striatum, which would decrease its influence, primarily in the cortical motor areas on the same side.
Meanwhile, the increased activation of the striatum in the less affected hemisphere may promote higher selectivity for contralateral movements (Playford, 1992; Holden, 2006;
Prodoehl, 2010). The increased lateralized activity of FS neurons in the non-lesioned hemisphere may contribute to the potentially compensatory changes, not only modulating the activity on the same side, but also modifying the activity of IT neurons projecting to the contralateral hemisphere.
The activity of the lesioned hemisphere during contralateral movement remained similar to the activity in a healthy animal, implying that the decrease in contralateral selectivity was due to increased activity during ipsilateral movements. We speculate that in order to maintain an adequate performance, this abnormally increased drive to the ipsilateral forelimb from the lesioned hemisphere requires additional control from the non-lesioned side. Studies in monkeys indicate that a large number of nonprimary motor cortex neurons exhibit changes in activity that were not entirely related to muscle activity, but were associated with higher-order motor control (Tanji, 1987). Therefore, the preferred activation of the non-lesioned hemisphere during contralateral movements that we observed may not be entirely related to muscle activation but may act as a control of the abnormal ipsilateral activity from the lesioned-hemisphere (Figure 10C).
Differential effect of dopaminergic depletion on IT and PT neurons
The main finding regarding the effect of the parkinsonian state on the PT and IT identified neurons was the differential change of the forelimb selectivity depending of the cortical region, showing a specific reduction in the contralateral preference only in PT neurons
in M1, and only in IT neurons in M2. Considering the variation in anatomical traits and inputs, different cortical projection neurons may play different roles in the pathology and compensatory mechanisms in PD. Several studies support the idea of a differential effect of dopamine depletion depending on cortical region and neuron type. In resting MPTP-treated parkinsonian monkeys, only M1-PT neurons exhibit a reduction in firing rate and elevation in bursting activity, in agreement with the classic rate models, suggesting that these cells are particularly sensitive to the loss of dopamine (Pasquereau, 2011). Additionally, in subsequent studies, MPTP produced a reduction in kinematic encoding, principally in the PT neurons (Pasquereau, 2016). This PT sensitivity could be explained by direct input from the thalamic nuclei, which themselves receive direct inputs from the BG, and are affected by the propagation of abnormal BG activity (Strick, 1974; Rathelot, 2009). However, other features such as different laminar distributions, morphologies, or firing patterns may contribute to the responses of these neurons to pathological conditions (Stewart, 2000; Morishima, 2006; Groh, 2010).
In this study, we observed a differential modification in the lateralized activity depending of the neurons subpopulations depending on the motor area. In M1, PT neurons and putative interneurons on the lesioned side exhibited a reduction in contralateral preference, which can coexist with a decrease in the action selectivity (Figure 10D, left). The non-lesioned hemisphere exhibited an increased contralateral preference in FS neurons, acting as a possible compensatory mechanism. PT neurons are positioned to play a relatively direct role in the expression of PD motor signs due to their direct projections to the spinal cord (Magill, 2001).
In previous studies PT neurons can be classified into fast and slow. However, it is possible that antidromic stimulation in vivo failed to activate smaller sized axons because of their higher
threshold. If so, most of the large pyramidal cells studied in rats in vivo would correspond to the fast PT neurons in cats and primates; smaller cells with finer axons, might be the slow PT neurons (Tseng 1993). In our experiments, in order to evoke antidromic spikes in specific axonal projections from the IT or PT neurons of M1 or M2, a blue LED light pulse (intensity, 5–
10 mW; duration, 0.5–2 ms, typically 1 ms) was applied through each of the two optical fibers.
The intensity of the light pulse was adjusted by changing the duration (0.1–2.0 ms, typically 1.0 ms), with the goal of making the antidromic-like spikes visible most reliably on the oscilloscope. Also, we checked whether they were elicited constantly by repetitive stimuli at 4–10 ms intervals (frequency-following test). Once we considered a neuron as a candidate, its spontaneously occurring spikes, detected using a window discriminator (2, Dagan; or WD-2010, custom-made by O'hara & Co., Ltd), were used as a source to trigger antidromic stimulation with a delay (1.0–2.0 ms) for the collision test (Saiki, 2018). However, due to the reliability of these method, we did not attempt to increase the stimulation intensity which may have activated the smaller (putative slow) PT neurons. In the case of rodents, the most reliable way to differentiate these subpopulations is throughout morphological features (Oswald, 2013) or distinct gene expression (Economo, 2018).
By contrast, in M2, the IT neurons were the affected population (Figure 10D, right). IT neurons form the basis of interhemispheric interactions and provide important glutamatergic input to motor regions of the striatum, influencing the disorder physiology of the dopamine-depleted striatum (Mallet, 2006). The importance of dopamine in interhemispheric connectivity is strongly suggested by studies in hemiparkinsonian rats, which exhibit increased functional coupling between the left and right motor cortices, as well as between the motor cortex on the non-lesioned side and the STN on the lesioned side (Degos, 2009; Jávor-Duray,
2015). Therefore, the possible compensatory increased activity from the dopamine-dominant hemisphere, which is aimed at promoting movement in the lesioned hemisphere (through IT neurons), comes at the cost of reduced specificity of PT neuron activation. Mirror movements are the extreme manifestation of this dysregulation, which is commonly expressed as impaired bimanual coordination. The dopaminergic system has considerable and wide-spread modulatory cortical influences (Steiner, 2001). Thus, the relationship between cortico-cortical interaction and clinical symptoms may be a direct effect of the impairment of dopaminergic tone in the cortical function, such as a reduction in the intracortical inhibition (Ridding, 1995;
Cantello, 2002). Decrements in cortico-cortical abnormal coupling with STN stimulation or levodopa have shown to produce a considerable clinical improvement (Silberstein, 2005). The effect of dopaminergic therapy may relate, in part, to a complex effect of dopamine on the patterning of striatal output, or extra-striatal effects, such as direct dopaminergic effects on the cerebral cortex, acting to suppress the abnormal synchronization between cortical regions.
PD subjects during off-medication also exhibited higher ipsilateral area recruitment of the cerebellum and primary motor cortex,being partially normalized by levodopa (Palmer, 2009).
Altogether, the levodopa replacement therapy has shown to have positive effects on the impaired striatum activity, decreased transcallosal inhibition, and compensatory efforts from cortical motor regions, or other areas such as the cerebellum, which are possible reasons contributing to the decreased motor lateralization in PD. Therefore, it is possible that this kind of therapeutic strategies may reverse the changes observed in the lateralized activity of M1 and M2.
Considering the selectivity during ipsi- or contralateral movements, it can be suggested a diversity of the motor cortex regarding the degree of lateralization, where regions
representing higher-order information exhibit less laterality (PMC, SMA in monkeys; M2 in rats), having a greater role in bilateral coordination; while M1 encodes concrete motor information, requiring a higher specificity for forelimb movements (Kurata 2010, Nakayama 2015, Soma 2017). The decrease of the lateralized activity observed in the lesioned hemisphere, may be a consequence of an abnormal hierarchical organization. Studies in 6-OHDA hemiparkinsonian rats show that dopaminergic cell loss altered cortical hierarchical organization, resulting in an increased drive from ‘top-down’ frontal areas onto primary motor and somatosensory regions. Direct influences from the more rostral non-primary frontal areas in the non-lesioned hemisphere towards areas in the lesioned hemisphere tended to increase, suggesting a flattening of cortical network hierarchy (Jávor-Duray 2017). These studies are in agreement with our results where the limb selectivity was similar in M1 and M2 in the lesioned hemisphere. These changes may lead to behavioral deficits and make patients more reliant on top-down prediction, in other words, this might lead to the higher cortical areas attempting to control the areas lower in the cortical hierarchy (Jávor-Duray 2017).
Repetitive, sequential and bimanual movements are especially difficult for Parkinson’s disease patients, particularly when volitional (Benecke, 1986; Fleminger, 1992; Georgiou, 1994). Parkinson’s disease patients may compensate for some of their motor deficits by the use of sensory guidance (Dietz, 1990; Georgiou, 1994). Not only bimanual but also sequential movements are more difficult than unimanual ballistic movements. They may have problems integrating two or more motor programs. Supplementary motor area is crucially involved with internal generation of movements, particularly when sequential patterns involved. PD patients show significantly reduced supplementary motor area activation but overactivity of the lateral premotor and inferior parietal cortices. These results imply
sequential rather than simple ballistic movements may require a switch to lateral premotor-parietal cortex circuits when the striatal-frontal circuits dysfunction in PD, providing sensory guidance (Samuel, 1997). Healthy people tend to use automatic control instead of attentional control to perform daily behaviors. Deterioration of motor automaticity is a general feature in PD patients. The sensorimotor striatum is critical for support the automatization of motor skills.
Significant dopamine depletion in the posterior putamen induces a failure to shift automated motor skills to the sensorimotor striatum, which in turn likely contributes to the difficulty of acquiring/executing automatic program/actions in PD (Wu, 2014). As the ability to perform automatic movements is impaired, the motor skills revert to the early-learning stages, PD patients have to rely on attentional control. PD patients may be forced into a progressive dependence on the goal-directed mode of action control that is mediated by comparatively preserved processing in the rostromedial striatum (Redgrave, 2010).
Conclusion. In this study, we evaluated the effect of a dopaminergic system lesion during voluntary movement, and found that the changes in neuronal activity characteristic of the parkinsonian state may be present without clear behavioral changes. Also, dopamine depletion had different effects depending on the motor area. Moreover, the reduction in dopamine in one hemisphere caused changes (either pathological or compensatory) in the non-lesioned hemisphere. Further studies are needed to determine whether changes in the non-lesioned hemisphere act as a compensatory mechanism to maintain similar performance in both forelimbs, as well as how the training process modifies the normal course of the effect of reduced dopamine on the motor cortex activity.