**2.1 Fundamental mechanisms of gait control**

Activation of different areas in the forebrain evokes different types of goal directed behaviors. On the basis of findings of our studies (Takakusaki et al., 2004b, 2006) as well as those of previous works (Grillner 1981, Mori 1987, Rossignol 1996), current perception of the neuronal pathways involved in locomotor control is illustrated in Fig.1A.

Fig. 1. Fundamental signal flows involved in gait control

A. Schematic illustrations of basic signal flows involved in gait control. (a) - (c) Sensory signals (a) act on the cerebral cortex and the limbic system generate "volitional and cognitive reference" (b) and "emotional reference" (c), respectively. (1) Volitional process requires cortical information processing. (2) Projection from the limbic system to the brainstem is responsible for emotional processes. (3) The brainstem (midbrain, pons and medulla) and spinal cord are involved in automatic processes. The basal ganglia and the cerebellum control volitional and automatic processes via cortico-basal ganglia (C-BG) loop and basal ganglia-brainstem (BG-BS) system, respectively. B. Posture of normal (left) and parkinsonian (right) states.

Motor behaviors require the recruitment of the activities of the entire nervous system (Fig.1A) and musculoskeletal systems (Fig.1B). Sensory signals, derived from both external stimuli and internal visceral information (Fig.1Aa), have the following dual functions (Takakusaki, 2008). One is to generate cognitive information processing that is utilized for working memory to guide future behavior (Fig.1Ab). Another may affect the emotional and arousal states (Fig.1Ac). Accordingly, animals initiate movements depending on either a *"volitional or cognitive reference"* or an *"emotional reference"* (Takakusaki 2008). Goal-directed

automatic aspects. In the second section we demonstrated recent findings obtained in animal experimentation how BG-BS system controlled postural muscle tone and locomotion. Then we propose hypothetical models that can provide rational explanations of motor disturbances in PD. In the third, final, section, we consider the role of the BG-BS systems in non-motor functions with special reference to the regulation of arousal state and awake-

Activation of different areas in the forebrain evokes different types of goal directed behaviors. On the basis of findings of our studies (Takakusaki et al., 2004b, 2006) as well as those of previous works (Grillner 1981, Mori 1987, Rossignol 1996), current perception of the

A. Schematic illustrations of basic signal flows involved in gait control. (a) - (c) Sensory signals (a) act on the cerebral cortex and the limbic system generate "volitional and cognitive reference" (b) and "emotional reference" (c), respectively. (1) Volitional process requires cortical information processing. (2) Projection from the limbic system to the brainstem is responsible for emotional processes. (3) The brainstem (midbrain, pons and medulla) and spinal cord are involved in automatic processes. The basal ganglia and the cerebellum control volitional and automatic processes via cortico-basal ganglia (C-BG) loop and basal ganglia-brainstem (BG-BS) system, respectively. B. Posture of normal (left) and

Motor behaviors require the recruitment of the activities of the entire nervous system (Fig.1A) and musculoskeletal systems (Fig.1B). Sensory signals, derived from both external stimuli and internal visceral information (Fig.1Aa), have the following dual functions (Takakusaki, 2008). One is to generate cognitive information processing that is utilized for working memory to guide future behavior (Fig.1Ab). Another may affect the emotional and arousal states (Fig.1Ac). Accordingly, animals initiate movements depending on either a *"volitional or cognitive reference"* or an *"emotional reference"* (Takakusaki 2008). Goal-directed

neuronal pathways involved in locomotor control is illustrated in Fig.1A.

sleep states.

**2. General framework of movement control 2.1 Fundamental mechanisms of gait control** 

Fig. 1. Fundamental signal flows involved in gait control

parkinsonian (right) states.

behaviors therefore may require following the three processes; "*volitional process"* (Fig.1A(1))*,* "*emotional process"* (Fig.1A(2)) and "*automatic processes"* (Fig.1A(3))*.* The volitional process is derived from intentionally-elicited motor commands arising from the cerebral cortex based on volitional and cognitive references. This process requires cortical information processing and is executed by the corticoreticular and corticospinal projections. The emotional process is elicited by emotional reference via projections from the limbichypothalamus to the brainstem. This contributes to the emotional motor behaviors including fight or flight reactions. Regardless of whether the locomotion is volitional or emotional, it is accompanied by the automatic processes that are evoked by sequential activation of basic motor programs in the brainstem and spinal cord. The cerebellum regulates volitional and automatic processes by acting on the cerebral cortex and the brainstem, respectively. Sensory feedback via spinocerebellar tract plays an important role in this operation. The basal ganglia control these processes via loops with the cerebral cortex, brainstem and the limbic system. Because output of the basal ganglia is altered in basal ganglia disorders, all these movement processes can be disturbed.

#### **2.2 Mechanisms of integrating posture and locomotion by subcortical structures**

In animal experiments, decerebrate cat preparation has been used to examine subcortical mechanisms of controlling posture and locomotion. When the decerebration was made at the precollicular-postmammillary level (x in Fig.2A), a cat maintained reflex standing posture due to decerebrate rigidity (mesencephalic cat). Repetitive microelectrical stimulation (50 Hz, 30 μA) applied to the cuneiform nucleus (CNF; a blue point in Fig.2B) bilaterally increased the level of extensor (soleus) muscle tone, and then elicited stepping movements which were developed to locomotion by moving a treadmill (an arrowhead in Fig.2Ba). However the same type of stimuli applied to the ventral part of the PPN (red point in Fig.2B) induced muscular atonia, which lasted even after termination of the stimulation (Fig. 2Bc). Stimulation between these two sites (a green point in Fig.2C) evoked stepping movements followed by muscular atonia (Fig.2Bb). Stimulation of the locus coeruleus (LC, an orange point in Fig.2B) bilaterally increased extensor muscle tone (Fig.2Bd). Generally the locomotion evoking sites (blue circles in Fig.2D), i.e. the midbrain locomotor region (MLR), were located in the CNF, while the inhibitory region was located in the PPN (red circles in in Fig.2C). Neurons between these regions may be involved in both locomotion and muscular atonia. As show in Fig.2D, cholinergic neurons were abundantly distributed in the area corresponding to the inhibitory region, indicating that an activation of cholinergic neurons requires muscle tone suppression (Takakusaki et al., 2003a).

Our current perception of neuronal mechanisms of controlling postural muscle tone and locomotion is shown in Fig.3A on the basis of previous studies (Grillner 1981; Mori 1987; Rossignol 1996; Takakusaki et al., 2004b, 2006). Three locomotor regions are identified. They are the MLR, the subthalamic locomotor region (SLR) and the cerebellar locomotor region (CLR). Signals from the MLR may activate "muscle tone excitatory system" and "locomotor system or rhythm generating system". The former is composed of monoaminergic descending pathways such as the coerulospinal and raphespinal tracts, and excitatory reticulospinal tract arising from the ventromedial medullary reticular formation (v-MRF) which approximately corresponds to the nucleus reticularis magnocellularis. The latter is composed of the excitatory reticulospinal tract and central pattern generators (CPG) in the spinal cord. Cortical projections to the MLR have not yet been identified. It is possibly

Possible Contribution of the Basal Ganglia

Fig. 3. Neuronal mechanism of locomotion in cats

Brainstem System to the Pathogenesis of Parkinson's Disease 437

A. Cat locomotor system. Signals from the MLR activate muscle tone excitatory system and locomotor system. Locomotor system is composed of excitatory reticulospinal tract from the ventromedial medullary reticular formation (v-MRF) and central pattern generator (CPG) in spinal cord. The excitatory reticulospinal tract also operates as the muscle tone excitatory system as well as the coerulospinal tract from the LC and raphespinal tract from the raphe nuclei (RN). Signals from the SLR and the cerebellar locomotor region (CLR) act on these systems to evoke locomotion. Cholinergic neurons in the PPN activate muscle tone inhibitory system, which arises from the pontine reticular formation (PRF) neurons and inhibitory reticulospinal tract neurons in the dorsomedial MRF (d-MRF). GABAergic output from the SNr to the MLR/PPN controls locomotion and muscle tone. Output from the basal ganglia to the thalamocortical neurons controls intentional and volitional gait behaviors. Visuospatial information from the visual cortex (Ctx) to motor Ctx via the parietal Ctx requires for programing accurate postural-gait synergy. B. Central and peripheral inputs to

spinal locomotor network. Locomotor rhythm and pattern are generated by spinal

(excitatory reticulospinal system) and muscle tone control systems (raphespinal,

Takakusaki et al., 2008. (c) and (d) in C are modified Mori et al. 1989.

interneurons. Activity of spinal neurons is modified by corticospinal tract, locomotor system

coerulospinal and inhibitory reticulospinal tracts). C. Obstacle clearance by forelimb (a) and hindlimb (b) during locomotion. Locomotor behaviors elicited by stimulating the SLR (c) and MLR (d) in cats with chronically implanted stimulating electrodes. Each picture was depicted at 0.5 sec and 0.1 sec intervals in (c) and (d), respectively. A and B are modified

suppress postural muscle tone and locomotor rhythm (Takakusaki et al., 2003b). This inhibitory system is also thought to induce muscular atonia during the rapid eye movement (REM) sleep (Chase & Morales 1990; Takakusaki et al., 1993). There are serotonergic projections to the PPN (Honda & Semba, 1994) and to the cholinoceptive area of the PRF (Semba, 1993). Serotonergic projections to the PPN likely inhibit cholinergic neurons

mediated by connections via the SLR (Rossignol, 1996). If decerebration was made at precollicular-premammillary level (y in Fig.2A), the cat spontaneously walked without stimulation. Thus the SLR exists between two decerebrate levels, and mostly corresponds to the lateral hypothalamus, and it may contribute to emotional behaviors (Griller et al. 1997). Signals from the SLR activate the locomotor system either directly or indirectly via the MLR (Grillner et al., 1997). The CLR corresponds to the mid-part of cerebellar white matter which contains massive fibers connecting bilateral fastigial nuclei. (Mori et al., 1999).

Fig. 2. Midbrain control of locomotion and muscle tone in decerebrate cat preparation

A. Two decerebrate levels (x and y) in parasagittal plane of the cat brainstem. B. Stimulus sites on coronal plane of the mesopontine tegmentum, which is indicated by dashed line in (A; at P 2.0). (a) - (d) Locomotion (a, b) and muscle tone alteration (b, c, d) induced by repetitive stimuli (30 μA, 50 Hz) applied to each site. Stimulus period is indicated under each record. C. Effective sites where stimulation evoked locomotion (blue circles), muscular atonia (red circles) and a mixture of both (green triangles) on coronal (a) and parasagittal (b) planes of the brainstem. D. Microphotographic presentation of cholinergic neurons identified by choline-acetyltransferase (ChAT) immunohistochemistry with lower (a) and higher (b) magnification of the mesopontine tegmentum. Abbreviations, EMG; electromyograms, IC; inferior colliculus, LDT; laterodorsal tegmental nucleus, LR; left and right, MLR; midbrain locomotor region, Mm; mammillary body, NRPo; nucleus reticularis pontis oralis, PPN; pedunculopontine tegmental nucleus, SC; superior colliculus, SCP; superior cerebellar peduncle, SLR; subthalamic locomotor region, SNr; substantia nigra pars reticulata.

Muscle tone control regions also exist in the brainstem. One is muscle tone inhibitory region in the PPN (Figs.2 and 3A). Cholinergic neurons in the PPN may activate "muscle tone inhibitory system", which is composed of cholinoceptive pontine reticular formation (PRF) neurons (Takakusaki et al., 2003a), reticulospinal neurons arising from the dorsomedial MRF corresponding to the nucleus reticularis gigantocellularis, and spinal inhibitory interneurons in the lamina VII of Rexed (Takakusaki et al., 1994, 2003b). This system then inhibits α- and γ-motoneurons innervating extensor and flexor muscles in parallel to interneurons mediating reflex pathways (Takakusaki et al., 2001). Because CPG is composed of spinal interneuronal circuits, an activation of the inhibitory system can simultaneously

mediated by connections via the SLR (Rossignol, 1996). If decerebration was made at precollicular-premammillary level (y in Fig.2A), the cat spontaneously walked without stimulation. Thus the SLR exists between two decerebrate levels, and mostly corresponds to the lateral hypothalamus, and it may contribute to emotional behaviors (Griller et al. 1997). Signals from the SLR activate the locomotor system either directly or indirectly via the MLR (Grillner et al., 1997). The CLR corresponds to the mid-part of cerebellar white matter which

contains massive fibers connecting bilateral fastigial nuclei. (Mori et al., 1999).

Fig. 2. Midbrain control of locomotion and muscle tone in decerebrate cat preparation

planes of the brainstem. D. Microphotographic presentation of cholinergic neurons identified by choline-acetyltransferase (ChAT) immunohistochemistry with lower (a) and

electromyograms, IC; inferior colliculus, LDT; laterodorsal tegmental nucleus, LR; left and right, MLR; midbrain locomotor region, Mm; mammillary body, NRPo; nucleus reticularis pontis oralis, PPN; pedunculopontine tegmental nucleus, SC; superior colliculus, SCP; superior cerebellar peduncle, SLR; subthalamic locomotor region, SNr; substantia nigra pars

Muscle tone control regions also exist in the brainstem. One is muscle tone inhibitory region in the PPN (Figs.2 and 3A). Cholinergic neurons in the PPN may activate "muscle tone inhibitory system", which is composed of cholinoceptive pontine reticular formation (PRF) neurons (Takakusaki et al., 2003a), reticulospinal neurons arising from the dorsomedial MRF corresponding to the nucleus reticularis gigantocellularis, and spinal inhibitory interneurons in the lamina VII of Rexed (Takakusaki et al., 1994, 2003b). This system then inhibits α- and γ-motoneurons innervating extensor and flexor muscles in parallel to interneurons mediating reflex pathways (Takakusaki et al., 2001). Because CPG is composed of spinal interneuronal circuits, an activation of the inhibitory system can simultaneously

higher (b) magnification of the mesopontine tegmentum. Abbreviations, EMG;

reticulata.

A. Two decerebrate levels (x and y) in parasagittal plane of the cat brainstem. B. Stimulus sites on coronal plane of the mesopontine tegmentum, which is indicated by dashed line in (A; at P 2.0). (a) - (d) Locomotion (a, b) and muscle tone alteration (b, c, d) induced by repetitive stimuli (30 μA, 50 Hz) applied to each site. Stimulus period is indicated under each record. C. Effective sites where stimulation evoked locomotion (blue circles), muscular atonia (red circles) and a mixture of both (green triangles) on coronal (a) and parasagittal (b)

Fig. 3. Neuronal mechanism of locomotion in cats

A. Cat locomotor system. Signals from the MLR activate muscle tone excitatory system and locomotor system. Locomotor system is composed of excitatory reticulospinal tract from the ventromedial medullary reticular formation (v-MRF) and central pattern generator (CPG) in spinal cord. The excitatory reticulospinal tract also operates as the muscle tone excitatory system as well as the coerulospinal tract from the LC and raphespinal tract from the raphe nuclei (RN). Signals from the SLR and the cerebellar locomotor region (CLR) act on these systems to evoke locomotion. Cholinergic neurons in the PPN activate muscle tone inhibitory system, which arises from the pontine reticular formation (PRF) neurons and inhibitory reticulospinal tract neurons in the dorsomedial MRF (d-MRF). GABAergic output from the SNr to the MLR/PPN controls locomotion and muscle tone. Output from the basal ganglia to the thalamocortical neurons controls intentional and volitional gait behaviors. Visuospatial information from the visual cortex (Ctx) to motor Ctx via the parietal Ctx requires for programing accurate postural-gait synergy. B. Central and peripheral inputs to spinal locomotor network. Locomotor rhythm and pattern are generated by spinal interneurons. Activity of spinal neurons is modified by corticospinal tract, locomotor system (excitatory reticulospinal system) and muscle tone control systems (raphespinal, coerulospinal and inhibitory reticulospinal tracts). C. Obstacle clearance by forelimb (a) and hindlimb (b) during locomotion. Locomotor behaviors elicited by stimulating the SLR (c) and MLR (d) in cats with chronically implanted stimulating electrodes. Each picture was depicted at 0.5 sec and 0.1 sec intervals in (c) and (d), respectively. A and B are modified Takakusaki et al., 2008. (c) and (d) in C are modified Mori et al. 1989.

suppress postural muscle tone and locomotor rhythm (Takakusaki et al., 2003b). This inhibitory system is also thought to induce muscular atonia during the rapid eye movement (REM) sleep (Chase & Morales 1990; Takakusaki et al., 1993). There are serotonergic projections to the PPN (Honda & Semba, 1994) and to the cholinoceptive area of the PRF (Semba, 1993). Serotonergic projections to the PPN likely inhibit cholinergic neurons

Possible Contribution of the Basal Ganglia

(Matsuyama & Drew 1997).

**2.3.2 Emotional locomotor behaviors** 

system contributes to appetitive behaviors.

**ganglia dysfunction** 

Brainstem System to the Pathogenesis of Parkinson's Disease 439

environment– play an important role in shaping the basic pattern of locomotion. Lajoie & Drew (2007) observed, after unilateral lesion of area 5 of the posterior parietal cortex, that cats frequently hit the obstacle as they stepped over it. They also frequently hit the obstacle with their hindlimbs even when the forelimbs negotiated the obstacle successfully. These findings suggest an important role for the posterior parietal cortex in the coordination of the forelimbs and hindlimbs and in the planning and programming of visually-guided gait modification (Fig.3C). Neuroanatomical studies indicate that the posterior parietal cortex sends selected projections to the motor cortical areas from layer III, while those to the lateral cerebellum via the pontine nuclei arise from layer V (Andujar & Drew, 2007). Neurons in the primary motor cortex and those in the premotor/supplementary motor areas (PM/SMA) mainly project to the spinal cord and the reticular formation via corticospinal and corticoreticular projections, respectively

The MLR was initially established as a functional region involved in the initiation of locomotion on the basis of its connections with limbic structures and the basal ganglia (Armstrong, 1986; Megensen et al., 1991). Regardless of the nature of emotional stimuli, they usually elicit alert responses that produce stereotyped movements such as increased postural muscle tone and/or locomotion that accompanies autonomic sympathetic responses. The limbic-hypothalamic systems play crucial roles in these processes. Sinnamon (1993) proposed the following three types of locomotor systems that function in different behavioral or motivational contexts; an appetitive system, a primary defensive system, and an exploratory system. In cats with chronically implanted electrodes, stimulation of the SLR elicited alerting responses followed by exploratory (searching) or defensive behaviors (Fig.3Cc; Mori et al., 1989). Signals from the SLR are mediated by dense fibers in the medial forebrain bundle projecting to the midbrain (Rossignol, 1996). On the other hand, stimulation of the MLR abruptly elicited machine-like explosive locomotion (Fig.3Cd). Neural circuits connecting the nucleus accumbens (the oldest part of the striatum), the hippocampus, and the amygdala, are involved in emotional memory, and projections from the nucleus accumbens to the MLR may contribute to the expression of exploratory behaviors (Mogenson, 1991). In addition, projections from the lateral and the medial hypothalamic areas to the MLR are thought to operate as defensive and appetitive systems, respectively (Grillner et al., 1997; Jordan, 1998). The orexin-containing neurons located in the prefornical lateral hypothalamic area are considered to control appetite, energy balance, and vigilance states via projections to various areas in the nervous system (Peyron et al., 1998; Sakurai, 2002; Siegel, 2004). The orexinergic projections to the MLR facilitated the activity of the locomotor system (Takakusaki et al., 2005), indicating that the hypothalamic orexinergic

**3. Basal ganglia control of movements and motor disturbances by the basal** 

It is established that the C-BG loop is required for volitional movement (Delong & Wichmann, 2007; Middleton & Strick, 2000). Neural circuits between the prefrontal cortex and the caudate nucleus (cognitive loop) are involved in the regulation of complex, visuallyguided limb movements and the planning and programming those movements. Neural

(Leonald & Llinás, 1994), and those to the PRF may reduce activity of the inhibitory system (Takakusaki et al., 1994). In contrast, the inhibitory system suppresses the activity of the coerulospinal tract (Mileykovskiy et al., 2000). Accordingly muscle tone can be regulated by a counterbalance between the inhibitory and the excitatory systems (Takakusaki et al., 2006). It was reported that a patient with a lesion in the dorsolateral mesopontine tegmentum did not lose muscle tone during REM sleep ("REM without atonia") (Boeve et al., 2007; Culebras & Moore 1989). Also, a patient with a lesion in the dorsal part of mesopontine tegmentum could not stand and walk (Masdeu et al., 1994). These clinical case reports suggest that both a muscle tone inhibitory region and a MLR are realities in the mesopontine tegmentum of the human.

Spinal mechanisms of locomotor control are schematically illustrated in Fig.3B. Signals from the cerebral cortex and the brainstem, and those from peripheral sensory afferents are integrated at spinal cord to achieve appropriate locomotor control. Various combinations of spinal reflexes operate during locomotion. Those mediating flexion reflex and crossed extension reflex undertake major roles in the generation of locomotor rhythm (Rossignol 1996; Rossignol et al., 2006; Takakusaki et al., 2001, 2003b; McCrea & Rybak, 2008). Spinal interneurons that constitute CPG generate detailed locomotor rhythm. The locomotor rhythm is then translated to next order interneuronal groups which shape "locomotor pattern". Finally signals are sent to last-order interneurons, including reciprocal Ia interneurons, Ib interneurons and Renshaw cells. They are located in lamina IV-VII of Rexed and project to target motoneurons. Lamina VIII interneurons project to the contralateral side of spinal cord and may control alternating limb movements (Matsuyama & Takakusaki, 2008). Signals generated by spinal locomotor network are then transmitted back to the cerebral cortex, the brainstem and the cerebellum so that they monitor events in the spinal cord (Fig.3B).

#### **2.3 Initiation of movements by the forebrain structures 2.3.1 Cortical control of locomotor behaviors**

Drew et al. (1996) demonstrated, in cats with chronically implanted electrodes in the cerebral cortex, that a majority of motor cortical neurons exhibited simple rhythmic firing in relation to step cycles during steady-state locomotion. However their discharge rates considerably increased when the cats initiated to walk and had to accurately step over obstacles. Thus, commitment of cortical processing seems unnecessary during the automatic locomotor movements. On the other hand, stepping movements that accompany accurate foot placement resemble to the forelimb reaching of higher primates (Drew et al., 2004; Georgopoulos & Grillner, 1989). Such an accurate movement requires visuomotor cognitive processes (Fig.3A), which are controlled by neural circuits involving the cerebral cortex, basal ganglia, and cerebellum (Middleton & Strick, 2000). Subjects are aware of the locations of obstacles around them, and they are able to alter their stepping patterns even without available visual information of the location of the obstacles relative to the body (Fig.3C). McVea & Pearson (2007) reported that perturbing walking cats in a consistent manner evoked lasting changes to the walking pattern that were expressed only in the context in which walking was disturbed. Moreover, cats that had stepped over an obstacle by forelimb (Fig.3Ca) remembered the location of the obstacle and could use working memory to guide stepping for the hindlimb (Fig.3Cb). Therefore, sensory inputs that signal context –the surrounding visual and auditory

(Leonald & Llinás, 1994), and those to the PRF may reduce activity of the inhibitory system (Takakusaki et al., 1994). In contrast, the inhibitory system suppresses the activity of the coerulospinal tract (Mileykovskiy et al., 2000). Accordingly muscle tone can be regulated by a counterbalance between the inhibitory and the excitatory systems (Takakusaki et al., 2006). It was reported that a patient with a lesion in the dorsolateral mesopontine tegmentum did not lose muscle tone during REM sleep ("REM without atonia") (Boeve et al., 2007; Culebras & Moore 1989). Also, a patient with a lesion in the dorsal part of mesopontine tegmentum could not stand and walk (Masdeu et al., 1994). These clinical case reports suggest that both a muscle tone inhibitory region and a MLR are realities in the mesopontine tegmentum of

Spinal mechanisms of locomotor control are schematically illustrated in Fig.3B. Signals from the cerebral cortex and the brainstem, and those from peripheral sensory afferents are integrated at spinal cord to achieve appropriate locomotor control. Various combinations of spinal reflexes operate during locomotion. Those mediating flexion reflex and crossed extension reflex undertake major roles in the generation of locomotor rhythm (Rossignol 1996; Rossignol et al., 2006; Takakusaki et al., 2001, 2003b; McCrea & Rybak, 2008). Spinal interneurons that constitute CPG generate detailed locomotor rhythm. The locomotor rhythm is then translated to next order interneuronal groups which shape "locomotor pattern". Finally signals are sent to last-order interneurons, including reciprocal Ia interneurons, Ib interneurons and Renshaw cells. They are located in lamina IV-VII of Rexed and project to target motoneurons. Lamina VIII interneurons project to the contralateral side of spinal cord and may control alternating limb movements (Matsuyama & Takakusaki, 2008). Signals generated by spinal locomotor network are then transmitted back to the cerebral cortex, the brainstem and the cerebellum so that they monitor events in the spinal

Drew et al. (1996) demonstrated, in cats with chronically implanted electrodes in the cerebral cortex, that a majority of motor cortical neurons exhibited simple rhythmic firing in relation to step cycles during steady-state locomotion. However their discharge rates considerably increased when the cats initiated to walk and had to accurately step over obstacles. Thus, commitment of cortical processing seems unnecessary during the automatic locomotor movements. On the other hand, stepping movements that accompany accurate foot placement resemble to the forelimb reaching of higher primates (Drew et al., 2004; Georgopoulos & Grillner, 1989). Such an accurate movement requires visuomotor cognitive processes (Fig.3A), which are controlled by neural circuits involving the cerebral cortex, basal ganglia, and cerebellum (Middleton & Strick, 2000). Subjects are aware of the locations of obstacles around them, and they are able to alter their stepping patterns even without available visual information of the location of the obstacles relative to the body (Fig.3C). McVea & Pearson (2007) reported that perturbing walking cats in a consistent manner evoked lasting changes to the walking pattern that were expressed only in the context in which walking was disturbed. Moreover, cats that had stepped over an obstacle by forelimb (Fig.3Ca) remembered the location of the obstacle and could use working memory to guide stepping for the hindlimb (Fig.3Cb). Therefore, sensory inputs that signal context –the surrounding visual and auditory

the human.

cord (Fig.3B).

**2.3 Initiation of movements by the forebrain structures** 

**2.3.1 Cortical control of locomotor behaviors** 

environment– play an important role in shaping the basic pattern of locomotion. Lajoie & Drew (2007) observed, after unilateral lesion of area 5 of the posterior parietal cortex, that cats frequently hit the obstacle as they stepped over it. They also frequently hit the obstacle with their hindlimbs even when the forelimbs negotiated the obstacle successfully. These findings suggest an important role for the posterior parietal cortex in the coordination of the forelimbs and hindlimbs and in the planning and programming of visually-guided gait modification (Fig.3C). Neuroanatomical studies indicate that the posterior parietal cortex sends selected projections to the motor cortical areas from layer III, while those to the lateral cerebellum via the pontine nuclei arise from layer V (Andujar & Drew, 2007). Neurons in the primary motor cortex and those in the premotor/supplementary motor areas (PM/SMA) mainly project to the spinal cord and the reticular formation via corticospinal and corticoreticular projections, respectively (Matsuyama & Drew 1997).
