For video material relating to movement disorders, please go to Movement Disorders Videos.
Less is known of the function of the cerebellum, thalamus and basal ganglia than of other structures in the brain, but there is an increasing appreciation of their complex role in motor and nonmotor functions of the entire nervous system. These structures exercise functions that far exceed their previously assumed supporting parts as simple ‘relay stations’ between cortex and spinal cord.
The subcortical structures receive massive different inputs from the cerebral cortex and peripheral sense organs and stretch receptors. Through recurrent feedback loops this information is integrated and shaped to provide output which contributes to scaling, sequencing and timing of movement, as well as learning and automatization of motor and nonmotor behaviours.
Functional neuroanatomy—the cerebellum can roughly be divided into (1) vestibulocerebellum—integration of vestibular information, (2) spinocerebellum—integration of sensory information from the body, (3) pontocerebellum—integration of information from the cortex regarding planned or on-going movement.
Function—these are proposed to be as follows: (1) a timing device for movement, (2) facilitation of motor learning, and (3) facilitation and correct scaling and harmonization of muscle activity.
Clinical features of cerebellar lesions—these include impairment of movement with dysmetria (‘past-pointing’), dysdiadochokinesia, truncal and gait ataxia (in midline vermal lesions), dysarthria, and abnormal eye movements (commonly nystagmus).
Functional neuroanatomy—the basal ganglia participate in multiple parallel loops which take information from different (mainly cortical) areas and then feedback (mainly) to those same areas. Input is mainly from the striatum; output comes almost exclusively from either the globus pallidus interna or the substantia nigra pars reticulate, which send inhibitory projections to the thalamus; dopamine is the main neurotransmitter that regulates activity.
Function—four main roles are hypothesized: (1) release of desired movement from inhibitory control, (2) inhibition of undesired movement, (3) facilitation of sequential automatic movements, (4) integration of attentional, reward and emotional information into movement and learning.
Clinical features of basal ganglia lesions—these include rigidity, akinesia, and dystonia.
Functional neuroanatomy—the thalamus receives afferent input from the special senses, basal ganglia, cerebellum, cortex and brainstem reticular formation; efferent output is mainly directed to cortical areas and striatum.
Function—the main thalamic functions are thought to include (1) modulation of sensory information by integration of brainstem (in particular reticular activating complex) and relevant cortical information; and (2) modulation of cortical activity via cortico-thalamocortical loops.
Clinical features of thalamic lesions—these include (1) sensory abnormalities—ranging from loss to deep-seated, severe pain; (2) motor disorders—e.g. hemiplegia; and (3) movement abnormalities—e.g. myoclonus, dystonia—usually in the context of lesions also involving the basal ganglia.
The cerebellum is located in the posterior fossa, bordered above by the tentorium cerebri and below by the foramen magnum. Anteriorly it borders the lower pons and medulla, separated from them by the fourth ventricle. The cerebellum is connected to the pons and medulla by the superior, middle, and inferior cerebellar peduncles. Afferents to the cerebellum enter largely through the inferior and middle peduncles, whereas most of the cerebellar efferents exit through the superior cerebellar peduncle. The cerebellum receives its blood supply from the posterior circulation via (rostrally to caudally) the superior, anteroinferior, and posteroinferior cerebellar arteries.
The anatomical divisions of the cerebellum (as is the case for the other subcortical structures discussed here, particularly the thalamus) are complicated by a number of overlapping classifications. The simplest anatomical division of the cerebellum is into the two cerebellar hemispheres and the midline structure called the vermis. A further division is into the flocculonodular lobe, comprising a nodular structure at the base of the cerebellum and an adjacent area of the hemisphere, the anterior lobe—the part of the cerebellum rostral to the primary fissure—and the posterior lobe—the part of the cerebellum caudal to the primary fissure. This division is in line with the proposed evolutionary development of the cerebellum, something that underlies an alternative classification scheme dividing the cerebellum into archicerebellum (flocculonodular lobe, receiving mainly vestibular input), paleocerebellum (anterior lobe, receiving mainly spinal cord input), and neocerebellum (posterior lobe, receiving mainly cerebral cortical input via the pons). Deep within the cerebellum are the cerebellar nuclei, which both receive input and produce output from the cerebellum. These nuclei, medially to laterally, are called the fastigial, globose, emboliform, and dentate nuclei.
The cellular architecture of the cerebellum is complex but remarkably uniform (Fig. 184.108.40.206). It comprises five cellular types: Purkinje cells, granule cells, basket cells, Golgi cells, and stellate cells. These are arranged in three distinct cortical layers. These are, from the outside in, the molecular layer (layer 1), the Purkinje cell layer (layer 2), and the granule cell layer (layer 3). Afferent input arrives at the cerebellum in the form of mossy fibres and climbing fibres. These are excitatory neurons arising from the input structures to the cerebellum. Only the inferior olivary complex sends mossy fibres to the cerebellum, with the rest of the input structures sending climbing fibres. These fibres may synapse on cerebellar nuclei, or ascend into the cerebellar cortex directly. The only efferents from the cerebellum are the axons of Purkinje cells.
Mossy fibres synapse with granule cells in layer 3, the axons which then ascend to layer 1, there forming parallel fibres that synapse with the dendrites of Purkinje cells directly, or synapse with basket cells and stellate cells in layer 1; these, in turn, form synaptic connections with dendrites of Purkinje cells. Climbing fibres ascend directly to layer 1 where they synapse with the dendrites of Purkinje cells. Axons of Purkinje cells give off collaterals as they descend both to adjacent Purkinje cells and to Golgi cells that lie in the outer part of layer 3.
The above brief description of cerebellar gross and cellular architecture goes some way to showing how the cerebellum is well placed to integrate a large amount of afferent information and to provide output of this integrated information to a large number of cerebral and spinal targets.
A first step to understanding the functional anatomy of the cerebellum is to consider the main input and output pathways. The cerebellum can roughly be divided into three functional areas, which receive particular inputs and produce output to particular areas either directly via the axons of Purkinje cells or via synapses of Purkinje cell axons on to cerebellar nuclei, which then connect to other structures.
The main input is afferent fibres from the ipsilateral vestibular ganglion and vestibular nucleus, and the contralateral inferior olivary complex. This input either goes directly to the flocculonodular lobe or reaches there via the fastigial nucleus of the cerebellum. Output is to the vestibular nuclei either directly or via the fastigial nucleus.
The main inputs are ipsilateral cutaneous and proprioceptive afferents from the body and face via dorsal and ventral spinocerebellar, cuneocerebellar, trigeminocerebellar, and spinoreticular tracts. Further input comes from motor and sensory areas of the cerebral cortex and vestibular nuclei via pontine reticulospinal nuclei and the contralateral red nucleus, and from the contralateral inferior olivary complex. All these inputs either go directly to the anterior lobe of the cerebellum, or reach there via synapses in the globose and emboliform nuclei. Output, either direct or via these same cerebellar nuclei, goes to the pontine reticular nuclei, the contralateral red nucleus, and a major projection to the contralateral posterior division of the ventrolateral nucleus of the thalamus.
This receives input from the contralateral pontine nuclei, which in turn receive massive input from widespread areas of cerebral cortex, in particular the frontal and parietal lobes. Input is also received from the contralateral inferior olivary complex. Input either proceeds directly to the posterior lobes of the cerebellum or reaches there via synapses in the dentate nucleus. Output (either direct or via synapses in the dentate nucleus) goes to the contralateral red nucleus and to the cortex via the contralateral posterior division of the ventrolateral nucleus of the thalamus.
Thus, in simple terms, the cerebellum has three main functional divisions: the vestibulocerebellum, concerned mainly with integrating vestibular information, the spinocerebellum, concerned mainly with integrating sensory information from the body, and the pontocerebellum, concerned mainly with integrating information from the cortex regarding planned or ongoing movement. All areas of the cerebellum also receive input from the contralateral inferior olivary complex. The inputs to the cerebellum are largely excitatory, using glutamate as a neurotransmitter. In contrast, Purkinje cells, the output cells of the cerebellum, are inhibitory, using γ-aminobutyric acid (GABA) as a neurotransmitter.
Recent advances in understanding of cerebellar functional architecture have revealed that the cerebellum appears to be divided into multiple ‘modules’ with similar cell structure, but receiving and giving out highly topographically organized information (Fig. 220.127.116.11). These modules are longitudinally arranged strips of the cerebellar cortex about 1–2 mm across, each 5–6 mm in length. This modular organization has been studied in most detail in relation to receptive fields from the forelimb of the cat. This work has demonstrated that particular sensory receptive fields in the forelimb map to particular areas of the inferior olivary complex, which in turn given off projections to particular areas within cerebellar modules, called cerebellar microzones. Each area of the inferior olive may project to a variety of microzones which may be distributed widely in the cerebellar hemispheres. Crucially, however, these distributed microzones all send output to a specific area of the cerebellar output nuclei. This organizational structure (called multizonal microcomplexes) permits topographically organized input to be fed into a variety of discrete areas in the cerebellum. These discrete microzones might respond to and process a particular aspect of movement control, such as timing, direction, or scaling of movement. These various aspects of movement control could then be integrated by the common output of these areas to a particular part of the cerebellar output nuclei. Parallel fibres may in addition allow integration of information between microzones responsible for movements at a number of different muscles, but which are commonly recruited as a group or ‘synergy’. This would allow for the coordination of complex multi-joint movements such as reaching and grasping.
Function and dysfunction
The main functions of the cerebellum (which are still the subject of much debate) are proposed as (1) a timing device for movement, (2) a structure that facilitates (motor) learning, and (3) a structure that allows integration of information (including information about planning of a movement and sensory feedback information on the progress of a movement) in order to facilitate correct scaling and harmonization of muscle activity.
In patients with cerebellar lesions a number of abnormalities can be seen in simple reaching tasks. Normal movements are usually accompanied by precisely timed agonist–antagonist–agonist bursts that allow the limb to arrive exactly at the desired target. In patients with cerebellar lesions, there is first of all a delay in movement initiation, followed by a delay in the antagonist burst so that the patient frequently overshoots the target. This is the basis for dysmetria or ‘past-pointing’, examined at the bedside during finger–nose or heel–shin testing. This overshoot may be a partial cause of intention tremor (worsening tremor towards the end of movement) as well as an additional effect of cerebellar lesions on the timing of activity in muscle-stretch reflex loops (via cerebellar projections to γ-motoneurons). Clinically, the timing and scaling role of the cerebellum can be assessed by looking for dysdiadochokinesia: a breakdown in force, rate, and rhythm of movement. This is often tested by asking patients to tap gently, regularly, and rapidly on a table or the examiner’s hand with their fingers. This breakdown of smooth repetitive movements can even be detected by feel or by sound (‘listening to the cerebellum’).
Midline vermal lesions usually cause truncal and gait ataxia, often in the absence of limb ataxia. The gait is wide based and particularly precarious on turning or on heel–toe walking. Unilateral cerebellar hemispherical lesions cause deviation or falling to the ipsilateral side. Unlike a sensory ataxia, cerebellar ataxia is not made worse by shutting the eyes.
Cerebellar dysarthria may often simply manifest as slurred speech, as if intoxicated. However, in addition some patients may have either scanning or explosive speech, due to an inability to modulate its rate, rhythm, and force appropriately. Dysarthria is usually present with lesions of the vermis, whole cerebellum, or its connections, but may be absent if one lateral hemisphere alone is involved. Eye movements are frequently abnormal in disease of the cerebellum or its connections. This may relate in part to the extensive connections from vestibular areas to the cerebellum. The following eye-movement abnormalities may be seen: gaze-evoked, rebound, downbeat, or positional nystagmus, dysmetric voluntary saccades and jerky pursuit, square-wave jerks (macrosaccadic oscillations), impaired vestibulo-ocular reflex suppression, and skew deviation.
There is no complete consensus on what structures make up the basal ganglia, but most would include the caudate nucleus, putamen, globus pallidus, subthalamic nucleus, and substantia nigra (Fig. 18.104.22.168). The globus pallidus is subdivided into the globus pallidus externa and interna (GPe/GPi), and the substantia nigra is subdivided into the pars reticulata and pars compacta (SNr/SNc). As with the cerebellum and thalamus, the basal ganglia have additional nomenclature systems that are still in use. The most important term is the word ‘striatum’ (or sometimes neostriatum) which is used to describe the caudate nucleus and the putamen together. The globus pallidus may be called the pallidum or paleostriatum, and the globus pallidus and putamen together may be called the lentiform nucleus. The phrase ‘corpus striatum’ is used to refer to the caudate, putamen, and globus pallidus together.
The basal ganglia occupy a position near the base of the cerebral hemispheres. The putamen lies lateral to the thalamus, separated from it (and from most of the caudate nucleus, except anteriorally) by the internal capsule. The caudate nucleus, with its head lying anterodorsomedial to the putamen, arcs back, following, and progressively tapering with, the lateral ventricles, its tail swinging forward until its anteriorly pointing tip terminates in the amygdaloid nucleus. The pallidum lies medial to the putamen but still lateral to the internal capsule. The substantia nigra lies in the midbrain, transversely above the cerebral peduncles. Its pars reticulata, the termination of the striatonigral pathway, is below the internal segment of the globus pallidus, and its pars compacta contains the dopaminergic neurons that form the nigrostriatal pathway. Below the thalamus, medial to the internal capsule and rostral to the midbrain, is the subthalamic nucleus. Most of the caudate, putamen, and globus pallidus derive their arterial supply from the anterior circulation via the lateral lenticulostriate arteries and branches of the anterior choroidal and middle cerebral arteries. The thalamus, subthalamic region, and substantia nigra are supplied by the posterior circulation.
The basal ganglia receive a huge variety of input from the cerebral cortex, limbic system, and cerebellum. Although the role of the basal ganglia in motor control has been heavily emphasized, it is clear that the basal ganglia have a key role in many other aspects of behaviour, reflected in the diversity of input to and output from many ‘nonmotor’ areas of the brain.
A key concept of basal ganglia functional anatomy is their participation in several parallel loops, which take information from different (mainly cortical) areas, and then feed back, mainly to those same areas. Although the basal ganglia would seem well set up to integrate information from these various loops, in fact they seem not to do so, and information is kept remarkably separate.
Five main loops are recognized: motor, oculomotor, dorsolateral prefrontal, orbitofrontal, and anterior cingulate. The motor loop has received by far the most attention, given its presumed role in movement disorders such as Parkinson’s disease. The main interest has focused on how activity in this loop is modulated by the basal ganglia, and in particular how dopamine plays a role in this.
Basic basal ganglia pathways
An important first step to aid understanding of the functional organization of the basal ganglia is to consider that input to the basal ganglia, whatever its source, arrives almost exclusively at the putamen or caudate (i.e. the striatum). Therefore, the striatum is the main input structure of the basal ganglia. Output from the basal ganglia comes almost exclusively from either the GPi or the SNr. Therefore the GPi/SNr forms the main output from the basal ganglia. Crucially, these output structures send inhibitory projections to the thalamus (Fig. 22.214.171.124).
Ninety-eight per cent of the neurons in the striatum are medium spiny neurons, which mainly receive excitatory input from glutamatergic neurons of the cerebral cortex. The rest of the striatal neuronal population is made up of large nonspiny cholinergic interneurons and GABA-ergic interneurons. The medium spiny neurons are inhibitory and use GABA as their neurotransmitter. They form two main groups (bundled together in structures called ‘Wilson’s pencils’) which have different routes to get to the output structures (GPi/SNr). One group projects directly to the GPi/SNr—the direct pathway—there colocalizing substance P and dynorphin as neurotransmitters. Activity in this pathway therefore inhibits basal ganglia output. The other group has an indirect route to the GPi/SNr, projecting first to the GPe, there colocalizing enkephalin and neurotensin as neurotransmitters. The pathway continues via an inhibitory projection from GPe to subthalamic nucleus (STN), and finally via an excitatory glutamatergic projection from STN to GPi/SNr. The net effect of activation of this indirect pathway (combining two inhibitory and one excitatory synapse) is to excite GPi/SNr. Activity in the indirect pathway therefore facilitates basal ganglia output (see Fig. 126.96.36.199).
Output from the basal ganglia is via GABA-ergic projections from GPi and SNr to the thalamus. The projections from GPi travel in two fibre bundles through the internal capsule; the one from the outer part of GPi is called the ansa lenticularis, and the other from the inner part of GPi is called the lenticular fasciculus. After traversing the internal capsule they meet with neurons from the SNr and join to form the thalamic fasciculus, which terminates in various nuclei of the thalamus (for motor fibres this is mainly the ventrolateral medial nucleus). Additional output from GPi and SNr goes to the pedunculopontine nucleus and the superior colliculus. A crucial point is that the basal ganglia inhibits the structures to which it projects—this is vital to understand the models of basal ganglia function described below.
The most important neurotransmitter that regulates the activity of the basal ganglia is dopamine. Dopamine has different effects on the direct and indirect pathways. Dopaminergic neurons from the SNc ascend and synapse on striatal neurons (this is called the nigrostriatal pathway). Striatal neurons that will form the direct pathway express mainly dopamine D1-receptors at the nigrostriatal synapses: these are stimulated by dopamine. In contrast, striatal neurons that will form the first projection of the indirect pathway express mainly dopamine D2-receptors, which are inhibited by dopamine. Therefore the net effect of dopamine on the striatum is to increase direct pathway activity, decrease indirect pathway activity, and therefore reduce the inhibitory output of GPi/SNr (Fig. 188.8.131.52).
How is dopamine release controlled in the basal ganglia? The answer may lie in recent discoveries regarding the exact make-up of the striatum. It appears that, as well as the direct and indirect pathways, the striatum also sends out projections direct to the SNc which stimulate activity in dopaminergic neurons projecting to the striatal neurons that form the direct and indirect pathways. So, the striatum is not formed by a homogeneous population of medium spiny neurons, but in fact is divided into two distinct subpopulations. One population forms ‘striosomes’, which are more densely packed groups of medium spiny neurons that have less cholinergic input and are rich in opiate receptors. These are the neurons that send projections to the SNc (the striatonigral projection). The other population forms the ‘matrix’, which is made up of less densely packed medium spiny neurons with no output to SNc. These are the neurons that form the direct and indirect pathways. The putamen is almost all matrix, whereas the caudate has many striosomes. The striosomes receive input mainly from limbic structures, whereas the matrix receives input from a variety of cortical areas, but not limbic structures. The connections of these different striatal neurons has led to the idea of two striatal systems: the ventral striatum (striosome), which, via its connections with the limbic system, feeds emotional, reward and attentional information into the basal ganglia, and via its ability to modulate dopaminergic output from the SNr it can influence activity in the dorsal striatum (matrix).
Basic pathways and the rate model of basal ganglia function
The discussion above about the connections of the various nuclei of the basal ganglia sets the scene for the most influential model of basal ganglia function proposed by DeLong and colleagues in 1990 (against the background of work by many others). The key to understanding this model is to appreciate that:
◆ output from the GPi and SNr is inhibitory to the thalamus (and therefore to the cortex), and therefore, from a motor circuit point of view, an increase in rate of GPI/SNr firing is hypothesized to inhibit movement
◆ as the direct and indirect pathways have opposite effects on basal ganglia output, the rate model hypothesizes that they will have opposite effects on movement
◆ dopamine has opposing effects on the direct and indirect pathways, tending to increase direct pathway activity via D1-receptors and decrease indirect pathway activity via D2-receptors. The net result of dopaminergic stimulation is, therefore, to decrease GPi/SNr rate of firing, promoting movement.
This model, sometimes described as the ‘rate model’, is successful in explaining a number of aspects of motor dysfunction related to the basal ganglia, e.g. the pathology of Parkinson’s disease leads to a dopamine-depleted state which would be predicted to decrease direct pathway activity and increase indirect pathway activity. This would tend to cause an increase in GPi/SNr (inhibitory) output to the thalamus, therefore inhibiting movement (Fig. 184.108.40.206). In contrast, hemiballism (flinging movements of the arm and leg) is known to occur frequently with damage to, or close to, the subthalamic nucleus. The rate model would predict the effect of a subthalamic nucleus lesion to be a drop in indirect pathway activity, leading to a reduction in GPi/SNr activity and a consequent increase in thalamic activity, promoting movement (Fig. 220.127.116.11). There is experimental support for this model, e.g. the finding from functional imaging studies that in Parkinson’s disease there is hypermetabolism of the GPi which reverses with the administration of levodopa.
However, problems occur when considering other clinical aspects of movement disorders, e.g. a lesion of the GPi in experimental animals tends not to cause excessive movement, as the model would predict. In fact, a lesion of the GPi can be a very successful treatment for both levodopa-induced dyskinesia in Parkinson’s disease and primary dystonia. In addition, the movement disorders characterized by excessive movement (the hyperkinetic movement disorders—dystonia, tremor, tics, myoclonus, chorea) are variable in their clinical features, something that is difficult to explain via a model based solely on rate of GPi/SNr firing.
Beyond basic basal ganglia connections: the hyperdirect pathway and basal ganglia oscillations
The connections between the basal ganglia are considerably more complex than the rate model permits, e.g. there is a ‘hyperdirect’ pathway—a glutamatergic pathway that directly links the supplementary motor area (SMA) and the subthalamic nucleus. In addition, there are numerous basal ganglia–basal ganglia pathways, e.g. a direct excitatory connection from the subthalamic nucleus to GPe, and a direct inhibitory connection from GPe to GPi and SNr. These additional connections suggest the presence of two networks within the basal ganglia: an ‘extrastriatal network’ where the subthalamic nucleus is the main player, with links to GPe, GPi, and SNr, and a ‘striatal network’ that connects directly to GPi/SNr.
The development of animal models of parkinsonism, and more recent developments in human deep brain stimulation surgery for Parkinson’s disease and dystonia, have permitted direct recording from the basal ganglia. These recordings have demonstrated that alterations in pattern and synchrony of basal ganglia firing may be more important than changes in rate alone. So, for example, in patients with Parkinson’s disease, direct recordings from the basal ganglia show an increase in bursting activity in the subthalamic nucleus that oscillates at a low frequency (in the beta band: 10–20 Hz) and is synchronized across the basal ganglia and motor cortex. Successful treatment with levodopa is associated with a shift to higher-frequency oscillations (into the gamma band >60 Hz).
One way to try to unify these disparate aspects of basal ganglia physiology into a functional whole is to first consider the basal ganglia as having a strong inhibitory bias. Therefore, although STN neurons fire quite consistently in response to cortical activity, fed to them via the hyperdirect pathway, this is not translated on the whole into changes in firing from basal ganglia output nuclei (GPi/SNr), due to strong inhibitory control from the striatum, and therefore the tonic inhibitory discharge of the basal ganglia output continues. However, in the presence of dopamine, this situation is reversed, and the net effect of dopamine on the direct and indirect pathways causes a shift in basal ganglia output firing, allowing the information carried in the subthalamic nucleus firing patterns to be fed through to the thalamus. This occurs in a strictly segregated way, and the topography of input is preserved.
In disease, there is a shift towards more synchronous firing within the basal ganglia with, in the case of Parkinson’s disease, a shift towards low-frequency oscillations even when movement is attempted, reflecting a loss of the normal modulation of firing patterns during movement. In dystonia, a hyperkinetic disorder, the GPi shows lower firing rates compared with Parkinson’s disease (as would be predicted by the rate model), but in addition there are more frequent and irregular bursts seen with long pauses of absent activity. This might link to the clinical picture of dystonia with excessive muscle activation that stops and starts with shifting coactivation of agonists and antagonists, leading to abnormal posture, writing movements, and often a jerky tremor. Synchronization of firing across the basal ganglia undermines its ability to focus and concentrate activation in a topographically discrete manner.
Function and dysfunction
The above discussion is complex, but reflects the evolving understanding of the functional role of the basal ganglia. The basal ganglia are hypothesised to have four main roles, all of which have most often been related to the motor function of the basal ganglia:
1 To release a desired movement from inhibitory control, e.g. before a desired eye movement the tonic discharge of the basal ganglia output nuclei drops, and this allows the movement to occur.
2 To inhibit undesired movement: in the motor system this would be reflected in the highly topographically organized nature of basal ganglia input and output. Therefore, as well as releasing the desired movement, the basal ganglia appear to play a key role in inhibiting other movements. This focusing role is also known as centre-surround inhibition, where the desired movement (centre) is surrounded by an area of undesired movement that is actively inhibited.
3 To facilitate sequential automatic movements: in motor learning experiments, basal ganglia activity tends to increase as learning occurs. This is thought to reflect a role for the basal ganglia in coding sequences of movements that become automated. This may explain the particular difficulty showed by patients with Parkinson’s disease in performing multi-stage automatic movements such as turning over in bed.
4 To integrate attentional, reward and emotional information into movement and learning: via the connections of the limbic system with the ventral striatum, the basal ganglia form an important location for the integration of motivational and emotional information with motor behaviour. This is particularly the case for reward-based learning. It has been suggested that the basal ganglia can be seen as integrating two aspects of reward-based learning: the ‘critic’—the ventral striatum system that holds information on how motivated the organism is towards a particular goal—and the ‘actor’—the dorsal striatum that holds information on the motor behaviour needed to achieve that goal.
These various functions are certainly biased towards the motor system, but it is clear, from both the discussion of basal ganglia connections above, and the symptoms displayed by patients with disorders of the basal ganglia, that nonmotor aspects of behaviour are strongly linked to the function of the basal ganglia. It may be particularly the case for motivation and reward-based learning, e.g. lesions of the caudate nucleus have been associated with the psychiatric syndrome of abulia—a syndrome of apathy and lack of motivation that is thought to reflect failure of normal reward-based motivational mechanisms.
The movement disorders are hypothesized to reflect dysfunction within the basal ganglia, although, surprisingly, it is difficult to mimic some of these disorders simply by lesions to the basal ganglia alone. Thus tics and myoclonus rarely occur in humans as a consequence solely of basal ganglia lesions. Likewise, chorea rarely occurs from lesions to the caudate nucleus alone, as one might expect given the degeneration of this nucleus in Huntington’s disease. Parkinsonism, combining akinesia (slowness—bradykinesia) and progressive fatiguing of repetitive movement), rigidity (stiffness of muscles in flexion and extension), rest tremor of 5–6 Hz, and postural instability, can be seen in response to discrete lesions of the SNc. In terms of the various functions of the basal ganglia outlined above, both rigidity and akinesia could be seen as reflecting an inability to release the desired movement (akinesia) and a failure to inhibit undesired movement (rigidity). In Parkinson’s disease, clear deficits in reward-related learning and performance of integrated automatic movements are seen, together with emotional and motivational problems. Dystonia can also be produced by discrete basal ganglia lesions (usually to the putamen) and, in terms of the basal ganglia functions outlined above, dystonia could reflect an inability to inhibit unwanted movement, leading to the typical clinical picture of overflow of activity into adjacent muscles and co-contraction of agonists and antagonists. The huge variety of clinical presentation of movement disorders no doubt reflects the interaction of basal ganglia dysfunction with dysfunction caused by neurological disease elsewhere in subcortical and cortical areas.
The two thalami sit at the head of the brainstem, their medial borders largely separated by the third ventricle, but often partially fused as the massa intermedia. They constitute the largest nuclear mass in the diencephalon (the others being the hypothalamus and subthalamus). On the lateral surface of the thalamus is the external medullary lamina, containing thalamocortical and corticothalamic fibres either entering or exiting the internal capsule. The external medullary lamina and the internal capsule are separated by a thalamic nucleus called the reticular nucleus. The internal structure of the thalamus, already complex, is further confused by the existence of different nomenclatures (the one used here being that of Wessler). Inside the thalamus the internal medullary lamina (consisting of fibres leaving or entering the various thalamic nuclei) roughly divides the thalamus into three groups of nuclei—lateral, medial, and anterior—with each subdivided into ventral and dorsal areas. There are further nuclei that are not defined by this ventral/dorsal system such as those that lie within the internal medullary lamina (the intralaminar nuclei), and others such as the lateral and medial geniculate and the pulvinar. The blood supply to the thalamus derives from the posterior circulation via the posterior cerebral arteries and perforators from the terminal part of the basilar artery.
Before discussing the functional anatomy of the thalamus, we briefly summarize its cellular structure. The main output cells of the thalamus are called relay cells. These form excitatory glutamatergic projections to the cortex. These cells receive multiple inputs including GABA-ergic inputs from interneurons within the thalamus, cholinergic input from the brainstem reticular formation, as well as glutamatergic input from particular cortical areas (usually those areas to which the relay cells then project back, forming corticothalamocortical loops). Relay cells have two modes of firing—a burst mode and a tonic mode—which may have different functions (see below). Relay cells are mainly contained in the dorsal thalamic nuclei (the relay nuclei), whereas nuclei in the ventral thalamus (particularly the intralaminar nuclei) project mainly to the basal ganglia via glutamatergic projections.
The thalamus is in an ideal position to modulate information flow to and from the cortex. Although previously this role had been thought of as a mainly passive relay station, it is clear that the thalamus has a much greater role in moulding the information that passes through it than previously realized.
Thalamic afferents arrive from five main sources.
1 Afferents from special senses (except olfaction): touch (from the body—ventral posterolateral nucleus; face—ventral posteromedial nucleus), taste (ventral posteromedial nucleus), vision (lateral geniculate nucleus), and hearing (medial geniculate nucleus)
2 Afferents from the output nuclei of the basal ganglia: GPi (centromedian nucleus, ventral anterior nucleus, ventral lateral nucleus oralis and medialis) and SNr (mediodorsal nucleus and ventral anterior nucleus magnocellularis)
3 Afferents from the cerebellum: ventral lateral nucleus caudalis, to the ventral posterolateral nucleus oralis
4 Cortical afferents from many cortical areas: mainly synapse on dorsal thalamic nuclei
5 Afferents from the brainstem reticular formation.
Efferents from the thalamus from three main groups:
1 Efferents from thalamic nuclei to representative areas of the cortex determined by the input to the nucleus (e.g. afferents from the retina project to the lateral geniculate nucleus, which then projects to the visual cortex)
2 Efferents to cortical areas that project directly to the thalamus (corticothalamocortical loops)
3 Efferents to the striatum (mainly from the intralaminar nuclei).
The functional anatomy of thalamic circuits has been most closely studied for the visual system, and this can serve as a model for other thalamic circuits (Fig. 18.104.22.168). In the visual system, the main input to be relayed to the appropriate area of cortex comes from the retina, but, interestingly, this forms only about 5% of the input to the relevant thalamic nucleus: the lateral geniculate body. The rest of the input comes from a variety of sources including inhibitory input from thalamic interneurons and the thalamic reticular nucleus, excitatory input from the brainstem reticular formation, and layer 6 of the visual cortex. Output from the lateral geniculate is then primarily to layers 4 and 6 of the visual cortex. This system therefore has a primary function: transfer of visual information from the retina to the visual cortex (sometimes called the driver function or first-order relay), but this is subject to a huge amount of modulation from other areas, both cortical and brainstem.
A secondary system, often called the higher-order relay, is distinguished from this first-order system. This system takes cortical information down to the thalamus (typically the dorsal nuclei), and then back again to the same area (corticothalamocortical loops). As for the first-order system this circuit is subject to multiple modulatory inputs at the thalamic level. Of course, the cortical areas projecting as higher-order relays may have themselves been influenced by first-order relays, leading to a complex series of loops integrating and modulating information flow to and from the cortex.
One of the most important modulating forces at work in the thalamus arises from the brainstem reticular activating complex. This is demonstrated by the massive decease in thalamic activity seen during sleep, and the potential of certain thalamic lesions to cause coma. The influence of the reticular activating complex may occur via its ability to cause the ‘burst’ pattern of firing in thalamic relay cells. It is hypothesized that this is a ‘wake-up’ signal to the cortex, causing diversion of attention to the particular input in question, following which relay cells switch to their normal regular tonic discharge.
Function and dysfunction
The above discussion clearly demonstrates the role of the thalamus as more than a neuronal rest stop on the way to and from the cortex. The main functions of the thalamus are thought to include:
◆ modulation of sensory information by integration of brainstem (in particular the reticular activating complex) and relevant cortical information
◆ modulation of cortical activity via corticothalamocortical loops.
A diverse range of clinical consequences of thalamic lesions has been described, as one would expect from a region where so many different information flows coalesce, e.g. sensory abnormalities are reported with thalamic lesions, from pure hemisensory loss to deep-seated, severe pain. Mild hemiplegia may be seen with thalamic lesions, sometimes in combination with hemisensory loss, dysaesthesia, hemiataxia, astereognosis, and hemichorea as in the thalamic syndrome of Déjèrine and Roussy. Other lesions, often spreading outside the thalamus to involve the basal ganglia, have been associated with myoclonus, dystonia, or a slow 3–4 Hz tremor of the limbs on one side of the body. Lesions of the ventral lateral nucleus caudalis (also known as the ventral intermediate nucleus) have been used as a treatment for parkinsonian and essential tremor.
These three subcortical structures, the cerebellum, basal ganglia, and thalamus, provide the bridge over which information passes to and from the periphery and the cerebral cortex. Through their intricate structure and interconnections they play a major role in modulating and integrating this information. The recent discovery of a hitherto unknown direct connection between the cerebellum and the basal ganglia again underlines the importance of considering these structures as part of a coordinated system rather than in isolation. The question ‘What does the cerebellum/basal ganglia/thalamus do?’ therefore becomes slightly nonsensical, because in fact they do nothing in isolation, and function only as part of a system. This system can certainly be affected in particular ways by dysfunction of one of its parts, but the results of discrete lesions are often hard to predict and may have wide-ranging consequences for motor and nonmotor behaviour.
Apps R, Garwicz M (2005). Anatomical and physiological foundations of cerebellar information processing. Nat Rev Neurosci, 6, 297–311. [Information about cerebellar microzones.]Find this resource:
Brown P (2003). Oscillatory nature of human basal ganglia activity: relationship to the pathophysiology of Parkinson’s disease. Mov Disorders, 18, 357–63. [Basal ganglia oscillations and disease.]Find this resource:
Hammond C, Bergman H, Brown P (2007). Pathological synchronization in Parkinson’s disease: networks, models and treatments. Trends Neurosci, 30, 357–64.Find this resource:
Hoshi E, et al. (2005). The cerebellum communicates with the basal ganglia. Nature Neurosci, 8: 1491–3. [Yet more complexity: this paper shows that there is a direct connection between cerebellum and basal ganglia.]Find this resource:
Lehericy S, et al. (2001). Clinical characteristics and topography of lesions in movement disorders due to thalamic lesions. Neurology, 57, 1055–66.Find this resource:
Lera G, et al. (2000). A combined pattern of movement disorders resulting from posterolateral thalamic lesions of a vascular nature: a syndrome with clinico-radiologic correlation. Mov Disorders, 15, 120–6.Find this resource:
Middleton FA, Strick PL (1994). Anatomical evidence for cerebellar and basal ganglia involvement in higher cognitive function. Science, 266, 458–61. [Evidence for the importance of subcortical structures in ‘higher’ functions.]Find this resource:
Middleton FA, Strick PL (2000). Basal ganglia and cerebellar loops: motor and cognitive circuits. Brain Res Brain Res Rev, 31, 236–50. [Review of cerebellar and basal ganglia anatomy and physiology.]Find this resource:
Nambu A (2004). A new dynamic model of the cortico-basal ganglia loop. Prog Brain Res, 143, 461–6. [Model of basal ganglia function focusing on the hyperdirect pathway.]Find this resource:
Obeso JA, Rodriguez MC, DeLong MR (1997). Basal ganglia pathophysiology. A critical review. Adv Neurol, 74, 3–18. [Difficulties with the rate model of basal ganglia function explored.]Find this resource:
Obeso JA, et al. (2000). Pathophysiologic basis of surgery for Parkinson’s disease. Neurology, 55, (12 Suppl 6), S7–12.Find this resource:
Rothwell JC (1994). Control of human voluntary movement, 2nd edition. Croom Helm, London. [An excellent short textbook on the physiology of human motor disorders.]Find this resource: