Middle Hierarchical Level
The middle level of the somatomotor system hierarchy comprises the primary motor cortex, the basal ganglia, the cerebellum, and motor nuclei of the brainstem. These structures are shown in gray in the block diagram of Figure 12.2, which depicts the main interconnections between the middle- level structures and between these and the rest of the nervous system.
Broadly speaking, the function of the middle level is to receive inputs from the higher centers and generate the required motor program, that is, the spatiotemporal pattern of action potentials that would activate the neurons of the lower levels so as to produce the desired movements. In generating the motor program, the middle-level structures utilize sensory inputs they receive from muscle, joints, eyes, and the vestibular apparatus.
Block diagram of the main connections of the middle-level structures of the somatomotor system.
The distinction between planning and execution of movement is strikingly illustrated by the results of brain-imaging studies that seek to identify the areas of the brain that are involved in the performance of certain motor tasks. The following observations are noteworthy:
- 1. When subjects were requested to perform some finger movements from memory, activity was detected in cortical areas that included the somatosensory cortex, the posterior parietal cortex, parts of the prefrontal cortex, the premotor cortex, the supplementary motor cortex, and the primary motor cortex. But when the subjects were asked to mentally rehearse the movement without actually moving their fingers, no activity was detected in the primary cortex while other areas remained active.
- 2. Voluntary movements can be planned, but their execution can be delayed or cancelled altogether, unlike reflex actions.
- 3. When an activity is repeated often enough and is learned, so that it becomes almost automatic, the motor association areas become less involved, but the primary motor cortex becomes more involved as the activity becomes more precise.
- 4. After one learns to sign one's name with the dominant hand and then attempts to sign with one's toes, for example, the same hand area in the PMA is activated in both cases. This suggests that a learned motor action is stored in the association cortex and can be used to direct movement in a different set of muscles.
- 5. As noted earlier, planning and execution are not sharply divided anatomically. Thus, the motor association areas project directly to the motor nuclei of the brainstem and the spinal cord (Figure 12.2), and the primary motor cortex is involved in some decision-making concerning movement.
Some main features of the middle-level structures and their interconnections are discussed in the following sections.
Primary Motor Cortex
Stimulation of the surface of the primary motor cortex, at a relatively low threshold, elicits movement of different parts of the body. Moving the site of stimulation laterally across the motor cortex reveals a somatotopic organization, the most medial part corresponding to the lowest extremities and the most lateral parts corresponding to the face and tongue (Figure 12.3). However, the distribution of various body parts over the primary motor cortex is not uniform, resulting in a caricature of a human figure referred to as the motor homunculus, or little person. A relatively large cortical area
representation is associated with finer movements, which involve small motor units and hence a larger number of motor neurons that need to be controlled. This is the case, for example, with:
- 1. The hand, including the four fingers and the opposable thumb, which underlies the amazing manual dexterity of humans.
- 2. Facial muscles, conveying a variety of facial expressions, which is important for social interaction.
- 3. The mouth and tongue, involved in vocalization.
It should be noted that the somatotopic organization depicted in Figure 12.3 is a diagrammatic representation of the responses in a frontal plane through the primary motor cortex. It does not imply a one-to-one mapping between cortical neurons and muscles. In fact, at the neuronal level, individual muscles and joints are represented at multiple sites in the primary motor cortex in a complex pattern, and cortical stimulation generally activates several muscles rather than individual muscles.
The primary motor cortex consists of six layers, like the rest of the neocortex. However, the cell-packed granular layer 4 is much less prominent than in the primary sensory areas. Instead, the most distinctive layer of the primary motor cortex is layer 5 that contains the cell bodies of the giant Betz cells, which are pyramidal cells having the largest cell bodies of neurons in humans, reaching 100 gm in diameter and which activate lower motoneurons that control muscles of the distal extremities. Betz cells have more primary dendritic shafts branching out of the soma than typical pyramidal cells. The dendrites of Betz cells project to practically all cortical layers. Their axons, in addition to axons of smaller pyramidal cells of layer 5, project outside the cortex through the pyramidal tract (Section 126.96.36.199). Layers 2 and 3 also contain smaller pyramidal cells that project to other cortical areas.
Neurons in the motor association areas involved in a given voluntary movement increase their firing at least a few hundred milliseconds before any movement can be detected. Neurons of the primary motor cortex increase their firing some 50-150 milliseconds before movement occurs and generally during the movement as well. A variety of firing responses have been observed in neurons of the primary motor cortex, mostly related to the magnitude of force and its direction. The increase in the rate of firing of some neurons, particularly corticospinal neurons that terminate directly on spinal motoneurons, is related to the magnitude of the force exerted at a joint, much like motoneurons, whereas in other neurons the increase in the rate of firing is related to the rate of change of force. Some neurons increase their firing rate in relation to the direction of force as well as its magnitude, whereas the activity of other neurons corresponds to kinematic variables, such as position and velocity of movement. Although individual neurons may have a maximum firing rate in a particular direction, the variation of firing rate of these neurons with direction is not highly selective. However, the average increase in the firing rate of a population of these neurons strongly correlates with the actual direction of force. Thus, the direction of the force is encoded not by the rate of firing of single neurons, but rather in the average increase of firing in a population of neurons. In tasks that require precise control of force, some neurons in the primary motor cortex increase their firing when the force decreases rather than increases. These neurons are believed to provide a more closely controlled derecruiting of motor units in the muscles involved. In many cases, neurons of the primary cortex are found to receive strong sensory input from the limb whose muscles they project to.
The primary motor cortex receives inputs from the basal ganglia and the cerebellum, via various thalamic nuclei, as will be elaborated later for outputs from these regions. It has interconnections with the somatosensory cortex, the motor association areas, and the frontal cortex.
In humans, lesions of the primary motor cortex disturb the dexterous execution of movements and cause deficits ranging from muscle weakness and discoordination to paralysis when upper motoneurons are completely destroyed.
SPOTLIGHT ON TECHNIQUES 12A:
RECORDING FROM THE BRAIN
Brain electrical activity is most conveniently recorded by surface electrodes placed on the scalp - a recording known as the electroencephalogram (EEG), referred to in Section 8.1.5. The EEG arises mainly from the synaptic inputs to the large pyramidal, cortical cells. An excitatory synaptic input to the apical dendrites produces an inward current, resulting in a positive charge intracellularly and a negative charge extracellularly (Figure 12.4a); current then flows through the main dendrite and back through the membrane and extracellularly to the apical dendrites, thereby completing the current loop, as required by conservation of charge. This extracellular current flow from lower regions, nearer to the cell soma, toward the apical dendrites, results in a negative voltage being recorded by a surface electrode near the apical dendrites, with respect to a distant reference electrode (Figure 12.4a). On the other hand, excitatory synaptic input close to the soma produces an extracellular current that flows toward the apical dendrites, resulting in a positive voltage being recorded by a surface electrode near the apical dendrites, with respect to a distant reference electrode (Figure 12.4b).
(a) Voltage recorded with excitatory synaptic input applied to apical dendrites of pyramidal cells; (b) voltage recorded with excitatory synaptic input applied close to the soma of pyramidal cells; EEG recorded in awake, relaxed state (c) and in the alert state (d).
Because the pyramidal cells are oriented parallel to one another and perpendicularly to the surface of the cortex, the voltages resulting from the extracellular currents of individual pyramidal cells are additive and are reinforced by the underlying rhythmic cortical activity (Section 8.1.5). The resulting voltages are large enough to be recorded by electrodes placed on the scalp. Nevertheless, the relatively high resistance of the bony skull attenuates the magnitude of the recorded voltages to the range 10-100 pV. To record these small voltages with minimal noise, active electrodes are used that combine a low-noise, Ag/AgCl electrode with an integrated circuit (IC) amplifier. To allow comparative interpretation of the EEG for clinical purposes between different subjects, the placement of EEG recording electrodes is standardized according to the 10-20 electrode system, so called because the spacing is based on intervals of 10 and 20 percent of the distances on the scalp from side to side and from front to back. The original system allowed for 19 electrodes, but has since been extended to 70 electrodes, if necessary. The reference electrode is commonly attached to one earlobe, and the ground electrode is attached to the mastoid, which is the back part of the temporal bone, on the same side of the head. The electrodes are usually embedded in a cap that is worn over the head.
The EEG has proved valuable for both clinical and research purposes as well as for practical applications, as in brain-computer interfaces (Spotlight on Techniques 13A). The EEG shows distinctive patterns for different, normal states and for some brain abnormalities. Figure 12.4c shows an EEG that is typical for an awake, relaxed adult with eyes closed to avoid distractions. The alpha rhythm (Section 8.1.5) is prominent under these conditions. If the subject is in an attentive state, or is thinking hard about something, the alpha rhythm is replaced by smaller amplitude, faster oscillations, as in Figure 12.4d. The EEG pattern changes markedly in sleep. As a person becomes drowsy, the alpha rhythm gradually disappears and is replaced by slower waves of larger amplitude as sleep deepens. But during the rapid-eye-movement (REM) phases of sleep, the EEG pattern is similar to that of the attentive state (Figure 12.4d). Generally speaking, the amplitude of the EEG varies inversely with the frequencies that are present, these being in the range 1-150 Hz, and EEG records from the frontal and parietal regions tend to show higher frequencies than records from the occipital regions.
The EEG is extensively used in the diagnosis and differentiation of different types of epilepsy, a condition in which groups of neurons in the brain become hyperexcitable and fire in near-synchrony producing distinctive wave patterns that can spread over a large area of the cortex and produce motor, sensory, or behavioral manifestations. The EEG is also useful for diagnosing superficial brain tumors, brain inflammation (encephalitis), stroke, some brain disorders (encephalopathies), and sleep disorders. The absence of an EEG is considered evidence of biological death.
Where a more accurate localization of the source of an abnormal brain activity is required, an invasive procedure is sometimes used to record brain electrical activity directly from the surface of the cerebral cortex, avoiding the skull. Such a recording is an electrocorticogram (ECoG, Spotlight on Techniques 13A). An interesting minimally invasive procedure for brain electrical recording and stimulation uses a stentrode, which consists of a small metallic mesh - the stent - of a few millimeters in diameter, having small metal electrodes on its surface. The stentrode is inserted through blood vessels of the brain and maneuvered into the desired position under X-ray guidance, for example, as in the insertion of a stent through cardiac catheterization.
The currents that give rise to the EEG also produce magnetic fields, which are recorded in magnetoencephalography (MEG). The magnetic fields are exceedingly small, of the order of tens of femtoteslas (fT, where femto denotes 10'15). By comparison, the strength of the earth's magnetic field is of the order of tens of microteslas. Special devices, referred to as SQUIDs (superconducting quantum interference devices) are used for measuring these tiny magnetic fields and are set in a large, fixed "helmet" that covers a large area of the head. Recently, there have been attempts to reduce the "helmet" to a normal size that would allow unrestricted head movements.
There are important differences in the nature of EEG and MEG recordings. First, EEG recordings arise from both the radial components of current, which are normal to the cortical surface, as well as the tangential components of current, which are parallel to the cortical surface. Because of the difference in the spatial orientation of the magnetic fields associated with these current components, MEG detects only the tangential components. MEG is therefore most sensitive to activity in the walls of the cortical sulci, and is not sensitive to activity in the top of the gyri. Second, MEG primarily detects intracellular currents associated with synaptic activity because the magnetic fields associated with extracellular volume currents tend to cancel out. Third, magnetic fields decay with distance more rapidly than electric fields, so that MEG is more sensitive to superficial cortical activity. A distinct advantage of MEG is a more precise localization of the source of activity, such as an epileptic focus, because of a better spatial resolution, of 2-3 mm, compared to a spatial resolution of 7-10 mm for EEG. This is because magnetic fields are less distorted than electric fields by the skull, scalp, and CSF. Both EEG and MEG have millisecond temporal resolution and are therefore able to detect events that last for 10 ms or less. The fast time resolution and precise spatial localization of MEG have made it useful for functional brain imaging (Spotlight on Techniques 12B).