1287_ch07.pdf

(566 KB) Pobierz
Motor Cortex in Voluntary Movements
7
Wheels of Motion:
Oscillatory Potentials
in the Motor Cortex
William A. MacKay
CONTENTS
ABSTRACT
From their earliest recognition, the oscillatory electroencephalogram (EEG) signals
in the sensorimotor cortex have been associated with stasis: a lack of movement,
static postures, and possibly physiological tremor. It is now established that 10-, 20-,
and 40-Hz motor cortical oscillations are associated with constant, sustained muscle
contractions, again a static condition. Sigma band oscillations of about 14 Hz may
be indicative of maintained active suppression of a motor response. The dynamic
phase at the onset of an intended movement is preceded by a marked drop in
oscillatory power, but not all frequencies are suppressed. Fast gamma oscillations
coincide with movement onset. Moreover, there is increasing evidence that oscilla-
tory potentials, even of low frequencies (4–12 Hz), may be linked to dynamic
episodes of movement. Although their overall power is reduced, these oscillations
0-8493-1287-6/05/$0.00+$1.50
© 2005 by CRC Press LLC
Copyright © 2005 CRC Press LLC
 
appear to exert motor executive or preparatory functions. Most surprisingly, the 8-Hz
cortical oscillation — the neurogenic component of physiological tremor — is
emerging as a major factor in shaping the pulsatile dynamic microstructure of
movement, and possibly in coordinating diverse actions performed together.
7.1 INTRODUCTION
Within 10 years of Hans Berger’s discovery of the EEG and brain rhythms, many
features of oscillatory potentials in the sensorimotor cortex were established. For exam-
ple, what might be called the Kornmüller doctrine appeared in 1932. Based on a
thorough study of rabbit cortical activity, Kornmüller postulated that alpha rhythms
dominate in granular cortex, and beta rhythms in agranular cortex.
1
In the rolandic area,
On the other hand, the func-
tional significance of brain rhythms in motor control remains a topic of much debate.
Extracellular field potentials are generated by neuronal dipoles created within
elongated dendritic fields, aligned in parallel arrays. Cortical pyramidal cells with
their long apical dendrites are the classic example of dipole generators. The current
sink is the site of net depolarization, and the source is the site of normal membrane
polarity or of hyperpolarization. Oscillatory potentials are generated by a combination
of mechanisms. Many cortical neurons have pacemaker-like membrane properties
such that they can produce oscillatory potentials at a variety of frequencies; generally,
the higher the depolarization, the faster the frequency.
2,3
For the oscillations to be
stabilized and sustained, however, a resonant circuit needs to be recruited. Inevitably
such circuits involve inhibitory interneurons to reinforce the excitation–inhibition
alternation.
4
Furthermore, the circuits entrain components both within the cerebral
cortex and the associated parts of the thalamus to create a thalamocortical network.
5,6
5
That is, there is no phase delay between them. Therefore, oscillations offer a
useful strategy for circumventing, or at least offsetting, conduction delays within
the nervous system, but the price to pay is a preparatory, oscillatory lead-up time.
Brain rhythms can be monitored with noninvasive EEG or magnetoencephalo-
graphic (MEG) recording, electrocorticograms (ECoG) recorded with subdural grids
of electrodes placed on the surface of the brain, or local field potentials (LFPs)
recorded with microwires or microelectrodes within the brain tissue. They can also
be seen in the firing patterns of single neurons, but not as clearly and reliably. In
sensorimotor cortex, mu and beta rhythms, which may be prominent prior to a
movement, are commonly suppressed at the onset of a movement. These changes
in oscillatory potentials are time-locked to the event, but unlike event-related poten-
tials, they are not usually phase-locked to the event.
Therefore they cannot be
extracted by averaging, but require frequency analysis for detection. Such phenom-
ena are manifested as frequency-specific changes of the ongoing electromagnetic
8
Copyright © 2005 CRC Press LLC
this concept has been repeatedly supported to this day.
The coupling of oscillators sharing a common frequency may be found not just
between the cortex and thalamus, but among a set of cortical areas, and between
the cortex and the spinal cord. (The cerebral cortex and cerebellum and the cortex
and striatum may also have coupled oscillations.) Essentially these are pulse-coupled
oscillators, linked by a quasiperiodic train (or rhythmic burst) of action potentials.
Linked oscillators, with close to the same frequency, invariably synchronize over
time.
7
activity of the brain. A decrease in power in a given frequency band is thought to
be due to a decrease in synchrony of the underlying neuronal population, and
therefore is commonly termed an event-related desynchronization (ERD). Con-
versely an increase in power in a frequency band is termed an event-related syn-
chronization (ERS) of the neuronal population.
8
However, it is not sufficient to equate an oscillation with inhibition or any
other single process, when it is, in fact, cyclic. Even if the inhibitory duty cycle is
more than 50%, there is a regular period of depolarization following the inhibition.
Neglecting this dichotomy can severely limit the usefulness of many ideas about the
function of rhythms in sensorimotor cortex.
The appearance of a strong LFP oscillation in a cortical area does have an
influence on trains of action potentials. As Figure 7.1 shows, action potentials tend
to occur near the negative peaks of oscillatory field potentials, corresponding to the
A
LFP
Spikes
100 ms
B
Spike-triggered
average of LFP
0
100 ms
Relationship of cortical LFP oscillation to single-unit discharge pattern. (A) The
extracellularly recorded spike train shows more regular discharge during the LFP oscillation.
Both the spikes and LFP were recorded simultaneously with the same electrode. (B) For a
similar unit to that shown in A, and only using spikes occurring during an oscillatory episode,
spike-triggered averaging of the LFP shows that spiking occurs preferentially at the time of
the negative peak in the LFP oscillation. (Adapted from Reference 10, with permission.)
Copyright © 2005 CRC Press LLC
The major frequency bands of cortical oscillation considered here are theta
(4–8 Hz), mu (or sensorimotor alpha, 8–12 Hz), sigma (12–15 Hz), beta (15–30 Hz),
and gamma (>30 Hz). Because of the well-established ERD of both mu and beta
rhythms at the time of movement onset, and their reappearance when movement
stops, they are commonly equated with “inhibition” or “deactivation” of the motor
cortex.
8,9
FIGURE 7.1
356521774.001.png
In effect, the oscillation tends to
regularize the spike train into a quasiperiodic pattern,
10–12
which would be useful for
some purposes but not others. Most importantly, pyramidal tract (PT) neurons are
entrained to the oscillatory LFPs in the motor cortex,
10
12,13
thereby conveying the
oscillation to spinal motoneurons.
As a result of these findings in the past 10 years, consensus seems to be settling
on the hypothesis that motor cortical rhythms accompany intervals of stationary
sensorimotor processing. The duration of an oscillatory potential would correspond
to an episode of relatively stable activity in the neuronal territory exhibiting it. Shifts
in power among frequency bands, or in correlation strength among oscillatory
potentials in different regions, indicate functional transitions within the motor system
from one state to another. During stationary (postural) states, cortical oscillations
provide an economic way of driving motor units. Partial synchronization of the
discharge of corticomotoneuronal cells would allow them to recruit motor units
while maintaining as low a firing rate as possible.
12,14,15
As will be seen below,
this is not universally true. It is possible that motor cortical oscillations may provide
an economical means of driving motor units or spinal interneurons during dynamic
as well as stationary phases of motor control.
8,9,16
7.2 THETA BAND
working in neocortical slices, showed that it is possible
to elicit realistic theta rhythm (4–8 Hz) by applying carbachol, a cholinergic agonist,
and bicuculline, a gamma-amino butyric acid (GABA
17
) antagonist. The circuitry for
generating theta oscillations was localized to superficial cortical layers. A prominent
current sink was found in layers 2 to 3, with the source in layer 5. Rhythmic activity
required intact glutamatergic transmission as well as inhibition. By analogy with
the hippocampus, the inhibitory interneurons in the network probably have a slow
spiking frequency, and terminate on distal dendrites of pyramidal neurons.
A
18
Tha-
There-
fore, a thalamocortical feedback loop is likely involved in sustaining theta oscillations.
Although the circuit to generate it may be present, theta rhythm is not conspic-
uous in motor cortical recordings. Relationships of theta oscillations to movement
typically involve neighboring regions. For example, in cats, a widespread 5-Hz LFP
oscillation with foci in somatosensory (S1) and in the visual cortex correlates to
general disinterest in the environment and drowsiness.
5
However, theta oscillations
can occur simultaneously with faster rhythms, in the context of a movement. Popi-
vanov et al.
19
recorded scalp EEG in subjects who moved a manipulandum with the
right hand to direct a light beam at a target. The authors developed an autoregressive
model of the linear dynamics of EEG, and found that abrupt short transients of
model coefficients occured during the movement preparatory period, related to
dynamic changes. These transients marked temporal nonstationarities in the alpha,
beta, and gamma bands that seemed to reflect boundaries separating successive
phases of motor preparation. When the EEG was averaged with respect to the first
Copyright © 2005 CRC Press LLC
depolarization phase of local pyramidal cells.
During intervals of overt
movement and dynamic fluxes in joint torque, it is generally observed that cortical
oscillations below 50 Hz are blocked or desynchronized.
Lukatch and colleagues,
lamic neurons also commonly have oscillatory membrane potentials at 6 Hz.
20
:
µ V
ECoG (mean)
A
75
0
300
3 single trials
B
0
300
EMG
(f. digitorum communis)
0
−
3 s
−
2
−
1
0
Preparatory 7-Hz rhythm recorded in human inferior parietal cortex. (A) Mean
ECoG of 30 trials of hand grip aligned on EMG onset. (B) Subset of three successive trials
showing superimposed ECoG waveforms and surface EMG (rectified) of the flexor digitorum
communis muscle. (Adapted from Reference 21, with permission.)
transient, it coincided with increased theta activity (3–7 Hz) and increased gamma
activity (35–40 Hz), which continued in sporadic bursts during the movement. These
theta plus gamma oscillations were observed mainly over the supplementary motor
area (SMA), premotor and parietal cortex. The authors interpreted the oscillatory bursts
as markers of the succession of dynamic stages in the production of movement.
20
found a remarkable 6–9 Hz
oscillation in the inferior parietal and superior temporal area immediately preceding
movement onset. The movement was the grasping of a joystick, self-paced about
every 10 sec. As shown in Figure 7.2, the rhythm was clearly preparatory to move-
ment because it was phase-locked to the onset of the electromyogram (EMG) onset,
but rapidly disappeared when movement started. Some leads showing this activity
were in the rolandic area, or over the cingulate motor area. The patients were on
reduced antiepileptic medication, and had no motor deficits.
21
Nonetheless, this
rhythm may have been influenced by medication. The theta rhythm was obviously
not driving movement, but it could have been influencing a premotor network
involved in setting up the motor performance.
Because the theta rhythm was phase-locked to EMG onset, there is a possibility
that it was used as a temporal reference for preparatory activity, analogous to the
use of theta rhythm in hippocampus for coding spatial position.
21
22
Again with ECoG
have found extensive theta oscillations
both within rolandic cortex and far beyond it, as subjects performed a virtual “taxi
23
Copyright © 2005 CRC Press LLC
FIGURE 7.2
Recording with subdural electrodes directly from the cortex of epileptic patients
undergoing surgical intervention, Turak and colleagues
recording in humans, Caplan and coworkers
356521774.002.png

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika:

Zgłoś jeśli naruszono regulamin