| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Vol. 83, Issue 5, 1438-1447, 1997
1 Department of Respiratory
Medicine, ABSTRACT Murphy, K., D. R. Corfield, A. Guz, G. R. Fink, R. J. S. Wise, J. Harrison, and L. Adams. Cerebral areas associated with
motor control of speech in humans. J. Appl.
Physiol. 83(5): 1438-1447, 1997.
articulation; control of breathing; vocalization
INTRODUCTION THE ACT OF SPEAKING demands simultaneous control of
respiratory, laryngeal, and articulatory systems. The underlying
mechanisms have been studied since the late 17th century, and the
difference between language formation and the control of speaking has
been commented on by various authors [see McFarland and Lund (15) and Whitcombe (30) for reviews].
The principles of speech mechanisms were outlined by Gutzmann (9), but
the mechanisms of neural control of breathing for speech remained
obscure until Draper et al. (3) described how, during a continuous
utterance (counting), coordinated activity of expiratory and
inspiratory muscles resulted in the production of a constant subglottic
pressure while the thorax and lung were becoming smaller. This analysis
has been broadly confirmed in studies of singing (22). All these
studies, together with reviews of both respiratory function in speech
(11) and phonatory mechanisms (1), testify to the complexity of the
required neural control mechanisms.
It is well established that partial or complete inability to speak
properly as part of motor linguistic impairment results from lesions
centered on the posterior third of the inferior frontal gyrus in the
dominant hemisphere (2). More recently, a view has developed in the
clinical literature that classical Broca's aphasia only occurs when
damage extends to areas in the frontal and parietal lobes and also to
deeper structures (16). However, beyond this, it is difficult to assess
the importance of Broca's area in the central nervous organization of
"speaking" without language processing. The significance of the
controversies surrounding the function of Broca's area and the
associated disorder of motor linguistics was that it made it necessary
to design an experimental paradigm that would permit the determination
of the cerebral areas concerned with speaking while at the same time
minimizing the higher processes involved in language formation and
recall. We chose to use positron emission tomography (PET) in an
experimental design where the combination of different speech-related
conditions would allow us to 1)
highlight breathing for speech and vocalization while eliminating the
effects of articulation and 2)
highlight articulation independently of breathing for speech and
vocalization. We have utilized repetition of a simple phrase that
requires only labial movements (12). Results of these studies have been
presented in preliminary form (17).
METHODS Subjects, Measurements, and Protocol
Experimental Paradigm
Training
We have defined
areas in the brain activated during speaking, utilizing positron
emission tomography. Six normal subjects continuously repeated the
phrase "Buy Bobby a poppy" (requiring minimal language
processing) in four ways: A) spoken aloud, B) mouthed silently,
C) without articulation, and
D) thought silently. Statistical
comparison of images from conditions A
with C and
B with
D highlighted areas associated with
articulation alone, because control of breathing for speech was
controlled for; we found bilateral activations in sensorimotor cortex
and cerebellum with right-sided activation in the thalamus/caudate nucleus. Contrasting images from conditions
A with B and
C with D highlighted areas associated with
the control of breathing for speech, vocalization, and hearing, because
articulation was controlled for; we found bilateral activations in
sensorimotor and motor cortex, close to but distinct from the
activations in the preceding contrast, together with activations in
thalamus, cerebellum, and supplementary motor area. In neither
subtraction was there activation in Broca's area. These results
emphasize the bilaterality of the cerebral control of "speaking"
without language processing.
Table 1.
Elements present in 4 speech-related conditions
Breathing for Speech and Vocalization
No
Breathing for Speech and No Vocalization
Element
Articulation
Condition A: Speech
Condition B: Mouthing speech
+
+
Language
+
0
Breathing for speech
+
0
Vocalization
+
+
Articulation
+
0
Hearing
No articulation
Condition C: Unarticulated speech
Condition
D: Internal speech
+
+
Language
+
0
Breathing for speech
+
0
Vocalization
0
0
Articulation
+
0
Hearing
The 4 speech-related conditions studied have been placed into
appropriate rows [depending on whether articulation was present (+) or
absent (0)] and appropriate columns (depending on whether or not there
was breathing for speech and vocalization). Under each condition
presence or absence of elements within that condition is indicated.
Column on right lists these elements. Subtraction of
condition B from A and of condition D from
C leaves breathing for speech, vocalization, and hearing.
Subtraction of condition C from A and of condition
D from B leaves articulation.
Fig. 1.
Breathing patterns expected during each of 4 speech-related conditions
(A through D; see Table 2), based on
observations during pilot studies. Conditions
A and C are
characterized by rapid inspiration followed by a controlled expiration
while the phrase "Buy bobby a poppy" is being "spoken"
repetitively. During conditions B and
D there is little disturbance of
breathing pattern.
[View Larger Version of this Image (20K GIF file)]
Table 2.
Exact instructions given to subjects both during training phase and
before each scan during experimental runs
Condition A: Speech
I want you to say the phrase
"Buy Bobby a poppy" repeatedly in a normal speaking voice and at
a normal speaking rate until I tell you to stop.
Condition B:
Mouthed speech
I want you to mouth the phrase "Buy Bobby a
poppy" repeatedly without making any sound and at a normal speaking
rate until I tell you to stop.
Condition C: Unarticulated
speech
I want you to say to yourself the phrase "Buy Bobby
a poppy" repeatedly and at a normal speaking rate until I tell you
to stop. In time with this, I want you to make an "aah" sound;
you should do this with your mouth slightly open but without moving
your tongue, lips, or jaw. Please ensure that you keep repeating the
phrase to yourself in your head in time with the sound.
Condition D: Internal speech
I want you to say to
yourself the phrase "Buy Bobby a poppy" repeatedly and at a
normal speaking rate until I tell you to stop. There should be no sound
and no movement. Please ensure that you keep repeating the phrase to
yourself in your head.
PET Scanning
PET scanning was performed, measuring relative regional cerebral blood flow (rCBF) by recording the regional distribution of cerebral radioactivity after the iv slow bolus injection of radiolabeled water (H215O). 15O is a positron emitter with a half-life of 2.1 min. Scans were performed in random order during the four conditions described above (Table 1; Fig. 1); each was repeated three times. Before each scan, the appropriate instruction was given, as used during the familiarization visit. Each scan was separated by a period of 10 min to allow for sufficient decay of radioactivity.Scanning Protocol and Data Acquisition
Subjects lay supine, with eyes closed, in a dimmed room with low ambient noise. They were positioned on an adjustable scanner table attached to the PET camera (model 953B, Siemens CTI). The head was positioned to minimize movement during the tasks and was supported in a stable position by use of an individually fitted head mold. This ensured that, during execution of the various tasks, the head movement would be <2 mm. Emission data were acquired in three-dimensional mode with interdetector collimating septa removed (28) and corrected for effects of tissue attenuation by use of measurements from a transmission scan (68Ge/68Ga ring sources) carried out before the first emission scan.For each measurement of relative rCBF, 10 mCi of H215O were given via an indwelling cannula placed in an antecubital vein. Emission data were collected sequentially over 150 s, comprising a 30-s background scan (scan A), a 30-s delay, and a 90-s scan (during which the different speech conditions were performed) after tracer arrival in the brain (scan B). H215O was given as a slow bolus 15-s before scan B (25). At the same time, the subject was told to start the appropriate speech condition. The integrated counts during scan B were corrected for background activity (from scan A). This gave an estimate of rCBF. Because no arterial blood samples were taken (14), no calibration was possible, and the term "relative" rCBF implies that there was no absolute quantification of rCBF. The corrected emission scan data were then reconstructed as 31 axial planes (plane thickness, 3.4 mm) with the use of filtered backprojection and a Hanning filter (cut-off frequency, 0.5 cycles/pixel). The resolution of the resulting raw images (determined by the physical characteristics of PET and the camera used) was 8.5 (x) × 8.5 (y) × 4.3 (z) mm at full-width half-maximum (FWHM) filtration (26). For further processing, image planes were displayed in a 128 × 128 pixel format, with a pixel size of 2 × 2 mm.
Magnetic Resonance Imaging
On a separate occasion, a magnetic resonance (MR) image of the subject's brain was obtained using a 2-Tesla Siemens Vision scanner (2T), utilizing relative T1 weighting to give good gray-white matter contrast.Image Processing and Data Analysis
Image manipulations and calculations were performed on SPARC computers (Sun Computers, London, UK) with the use of ANALYZE (BRU, Mayo Foundation, Rochester, MN) and PROMATLAB (Mathworks) software. Statistical parametric modeling software (Wellcome Dept. of Cognitive Neurology) was used to create statistical maps of significant relative rCBF changes (7, 8). First, all scans of each individual were realigned to the first scan of the series, thus ensuring spatial congruency. All PET scans and the MR image were transformed to standard stereotactic space (27) to account for individual differences in brain size and shape (7, 8). PET images were filtered in three dimensions by use of a 10-mm FWHM isotropic Gaussian kernel filter for the following reasons: 1) to improve signal-to-noise ratio, 2) to reduce intersubject variability caused by differences in local anatomy, and 3) to condition the data so that they conform more closely to the Gaussian field theory (8) that forms the basis for the model from which the statistical inferences are drawn when correcting for multiple comparisons. Subsequently, a pixel-based analysis of covariance, treating "global" activity (reflecting global blood flow) as the covariate, controlled for state-dependent differences in global blood flow associated with the different conditions (7, 8). A simple general linear model (8) was used to model relative rCBF as a linear combination of condition, subject, and global effects. In this analysis, each scan was treated as an independent observation. Thereafter, contrasts of condition effects were assessed by using t-statistics (at each voxel), and a transformation into normal-distribution Z statistics was performed. The resulting sets of spatially distributed Z values constitute statistical parametric maps showing regions of significant condition-associated relative rCBF changes (7, 8). These regions were displayed with a statistical threshold at either P < 0.001 and Z
3.10 or
P < 0.01 and
Z 2.33 (not corrected for multiple
comparisons). Within each region, the local maxima of significant
relative rCBF changes were then derived in terms of
x, y,
and z coordinates (see definitions in
Fig. 4 legend). The anatomic locations of local maxima were defined
both with reference to the stereotactic atlas of Talairach and Tournoux
(27) and also by superimposition on the group mean MR image that had
been transformed into standard stereotactic space (as described above).
C) and
(BD) are shown as
through projections onto representations of stereotactic space (27).
Areas reaching statistical significance
(P < 0.001;
Z > 3.10) are imaged. Sagittal
image views brain from side; coronal image is from back; transverse image is from top. R, right; VPC,
vertical plane through posterior commissure; VAC, vertical plane
through the anterior commissure; nos. are coordinates in stereotatic
space. x, Distance (in mm) to right
(+) or left () of midsagittal (interhemispheric) line; y, distance anterior (+) or posterior
() to vertical plane through anterior commissure (VAC);
z, distance above (+) or below
() intercommissural line through anterior and posterior
commissures. Areas of activation of increasing significance are shown
as arbitrary gray scale ranging from light gray to black. smc1,
Sensorimotor cortex; t/c, thalamus/caudate nucleus; c, cerebellum.
Coordinates of peak activation in smc1
(left and
right) are tabulated beneath coronal
image, together with Z statistic
(far right). Lower
right, adjusted rCBF for right side smc1 for all
subjects by condition. Bars show increased activity in
conditions A and
B where articulation is present. Solid
marks represent Fisher's least significant difference between any mean
value at P < 0.05. Other significant
activations are seen in coronal and transverse images within c and t/c
areas on right.
The orthogonal design (Table 1) of the experimental paradigm outlined
above allowed two principal contrasts to be defined, focusing on areas
associated with control of breathing for speech and vocalization
(AB and
CD), on the one hand, and
articulation (AC and
BD), on the other hand. In
each case, the two contrasts could be combined to increase the power of
the analysis. The elements of
(AB) + (CD) double the contribution
of breathing for speech and vocalization (together with hearing) but
eliminate the contribution from articulation. Similarly, the elements
of (AC) + (BD) double the contribution
from articulation but eliminate the contributions from breathing for
speech, vocalization, and hearing. Areas activated during breathing for
speech and vocalization cannot be distinguished from each other in the
experimental paradigm used in the present study.
RESULTS
Performance of the Task
All subjects reported that they were able to perform the tasks as instructed. During condition A (speaking aloud), normal conversational loudness and a normal speaking rate were easily maintained. During condition B (mouthing silently), the throat microphone confirmed the absence of sound production, and the video camera, together with the EMG recordings, confirmed that the subjects were mouthing appropriately. During condition C (vocalized without articulation, ah-ahah-ah-ahah), the cadence was similar to condition A, and the video confirmed that the mouth was not moving. During condition D (internal speech), there was no evidence of sound production, facial movement, or EMG activity; the subjects reported that they were saying the phrase to themselves as requested.Respiratory Data
An example of the breathing patterns from one individual throughout one scan for each of the four conditions is shown in Fig. 2, and the average values for the individual respiratory variables during these scans and also before scanning in the absence of any speech (Rest) are shown for each condition in Fig. 3. Speech, whether articulated or unarticulated, showed a controlled, near constant, expiratory airflow lasting 8-10 s, interrupted by rapid inspirations of ~0.75-s duration. The similarity between the breathing variables in these two conditions (A and C) was confirmed for the group as a whole (P > 0.05 by ANOVA; Fig. 3) for all the variables shown. Breathing variables in the two conditions without vocalization (B and D) were similar (P > 0.05 by ANOVA; Fig. 3) to each other and to the breathing in the "Rest" condition. However, the breathing traces showed a variable degree of "control" of expiratory airflow, as judged by periods of constancy of slope of the expired volume trace (together with minor slope disturbances), in condition B in four of the six subjects. This is not represented in Fig. 2, nor were these disturbances enough to affect the mean respiratory variables (Fig. 3).
I) is not different across conditions.
Expected breathing pattern (see Fig. 1) can be seen with a shorter
inspiratory duration (TI), larger expiratory duration
(TE), and larger total volume (VT) in those
conditions that have "breathing for speech and vocalization" in
them. Those conditions predicted to have little modulation of breathing
(Rest, Mouthing, and Internal speech) are similar. Change in functional
residual capacity (
FRC) was most variable during speech, but this
was not significantly different from other conditions. TT,
total breath time.
MR Imaging
MR structural images of the brain were obtained in five of the six subjects.Relative rCBF Measurements
rCBF associated with control of articulation. The mean results of the main contrast (AC) + (BD)
are presented in Fig. 4. This showed a
significant bilateral activation in the sensorimotor cortex (smc1),
together with activation in the cerebellum bilaterally
(Z = 3.5) and on the right side in the
thalamus/caudate nucleus (Z > 3.6);
there was no activation in the supplementary motor area (SMA). The bar
graph shows the relative rCBF (arbitrarily adjusted to a global value
of 50 ml · min
1 · dl1)
at the pixel of maximum significance for each condition for the
right-sided smc1 activation alone. There were significant differences
in the adjusted rCBF values between conditions
A or B with
C or
D, with
D being significantly lower than
C. Thus adjusted rCBF was greater in
those conditions in which articulation was present.
The pattern of activation of smc1 for one representative subject,
co-registered on his own strucural MR image, is shown in Fig.
5A. The
coronal, sagittal, and transverse planes are displayed at the group
mean coordinates of the right-sided focus of activation applied onto
that subject's MR images. This clearly shows the cortical sites of
these activations.
In view of the evidence that the breathing in condition B may have been "contaminated" by some "breathing for speech" (see above), we examined the component (A
C) of the main contrast
(Table 1; Fig.
6A). The
foci of activation, smc1, were still present and had the same
coordinates. The only difference was that the areas of activation were
more circumscribed, with a highly significant Z score in spite of the reduced
statistical power of the contrast.
C) alone. Comparison of
result of this simple subtraction with that from complex subtraction in
Fig. 4, which includes data from condition
B, gives similar result with better definition but with
a lower, but highly significant, Z
value. B: group results of relative
rCBF increases associated with control of breathing for speech,
vocalization, and hearing oneself, using the contrast
(CD) alone. This result from
simple subtraction should be compared with that from complex
subtraction in Fig. 7, which includes data from
condition B. This simple subtraction gives a similar but less-well-defined result.
rCBF associated with control of breathing for speech and vocalization and hearing. The mean results of the main contrast, (A
B) + (CD), are presented in Fig.
7, which highlights bilateral areas within
the sensorimotor cortex (smc2) together with a unilateral activation in
a strip on the surface of the precentral gyrus (mc), most probably in
the motor cortex (Brodman area 4/6 junction). Reduction of the level of significance to P < 0.01 (Z > 2.33) demonstrates a
similiar area of activation on the other side (mc, shown for a single
subject in Fig. 5B).
The coordinates of mc shown in Fig. 7 are the site of the focus of
activation, but the area of activation extends above and below this
level by at least 4 mm in both directions. The coordinates of the
region with no activation ("hole") lying between mc and smc2 are
similar to that of smc1 in the other main contrast (described above;
Fig. 4). Activations were also present in the SMA in the midline
(Z = 3.8), in the thalamus bilaterally (Z = 4.1), and in the cerebellum in
the midline (Z = 3.94). The bilateral
activations seen in the superior part of the temporal lobes, presumably
associated with "hearing" oneself speaking, were expected because
"hearing" could not be removed by this contrast (Table 1;
METHODS).
B) and
(CD) are shown as through
projections onto representations of stereotactic space. Areas reaching
statistical significance (P < 0.001;
Z > 3.1) are shown. st, Superior
temporal; s, supplementary motor area; t, thalamus. Coordinates of peak
activation in smc2 (left and
right) and peak activation in strip
mc (right) are tabulated beneath coronal image, together with Z
statistics. Lower left: adjusted rCBF
for pixel of maximum significance of right smc2 is shown for all
subjects, by condition. This shows increased activity in those conditions, A and
C, in which control of breathing for speech and vocalization are present. Lower
right: Right mc activation is shown for all subjects,
by condition, showing similar activations in
A, B,
and C, all significantly greater than
D (see
DISCUSSION). Other details are same
as in Fig. 4.
The bar graphs in Fig. 7 show the relative rCBF (arbitrarily adjusted to a global value of 50 ml · min
1 · dl1) at the pixel of
maximum significance, for each condition for the right-sided smc2 area
and for the right-sided mc area. For the smc2 activation, there were
significant differences in the adjusted rCBF values between
conditions A and
B; between A,
B, or C with
D; but not between
B and
C or
A and
C. Thus adjusted rCBF in smc2 tended
to be greater in this area when there was breathing for speech together
with vocalization and to be lowest in the conditions with least
modulation of breathing. For the mc activation, the bar graph shows
that there were no significant differences between the adjusted rCBF
values in conditions A, B, and
C but they were all significantly
greater than those in condition D.
The pattern of activation for one representative subject, co-registered
on his own structural MR image, is shown in Fig.
5B. The coronal, sagittal, and
transverse planes are at the group mean coordinates of the right-sided
focus of activation (smc2) applied to the individual subject's images
and show the cortical sites of the main activations, including that of
mc (bilaterally in this individual).
The evidence of some degree of control of expiratory airflow in
condition B compared with
condition D in some of the subjects led to a suspicion that the breathing pattern seen might have been
contaminated by some breathing for speech as in
conditions A and
C. If this were true, then any
contrast involving scans taken during condition
B might be similarly contaminated. We therefore examined the contrast (CD)
alone (Table 1; Fig. 6B). In spite of the reduced statistical power of this contrast, a similar pattern of
activation was found as in the main contrast
(AB) + (CD), except that areas smc2,
and especially mc, were not so clearly defined.
Other Areas Activated
Neither of the above contrasts highlighted any unilateral activations classically associated with speaking a language. In particular, Broca's area centered on the triangular part of the left inferior frontal cortex (approximate Talariach coordinates: x,40 to 60;
y, +10 to +12; and
z, +8 to +12) showed no foci of
activation.
DISCUSSION
We have demonstrated bilateral areas in the sensorimotor cortex that are activated in the control of breathing for speech and vocalization. In addition, we have identified separate bilateral areas in the sensorimotor cortex that are activated in the control of articulation. There were no unilateral activations in the area described by Broca (2).
Experimental Paradigm
The phrase. The cadence and repetitive nature of the task resembled that utilized by Draper and co-workers (3) in counting, in which evidence of sophisticated neural control of the respiratory musculature had already been found. A major advantage of the phrase (12) was that it consisted predominantly of labial consonants.Thus all the mouthing activity was visible to the investigator and could be monitored. The importance of this was that the control conditions C and D, in which an absence of mouth movement was required, could be shown to have been carried out correctly. In addition, inspection of mouth movements could be made qualitatively between conditions A and B to assess their similarity. Language content. The task used a phrase which was overlearned, repetitive, and required no particular attention. There was therefore minimal language content, with little probability of finding activation in the classic Broca's area. Those aspects of speaking and language that are associated with activation in the left dorsolateral prefrontal cortex have been recently studied by Warburton et al. (29) using PET; noun and verb generation were the optimal stimulus to such activation. Herholz et al. (10), using MR imaging-guided PET, showed that the most common anatomic areas of activation with verb generation were in the left inferior frontal cortex centered around the classic Broca's site in the pars triangularis. The subtraction paradigm used in the present study (Table 1) was designed to remove activation associated with any residual language content likely to be constant across all four conditions. The finding that Broca's area was not activated was therefore consistent with this analysis and adds evidence that this area is not involved with the control of breathing for speech, vocalization, or articulation. Breathing pattern. The simple phrase chosen for repetition in this study clearly brings out the control of airflow so that the expected pattern (Fig. 1A) was realized (Fig. 2A). This was also true for unarticulated speech (Fig. 1C; compare Fig. 2C). The expected breathing pattern during the mouthing of speech, i.e., no obvious change from a normal resting level (Fig. 1B), was not completely realized because of the evidence that a degree of respiratory modulation that resembled control of breathing for speech was present in some subjects. The breathing pattern during internal speech (Fig. 2D) was as expected (Fig. 1D). The normal mean resting level of ventilation (Fig. 3) maintained throughout the various tasks testifies that the subjects, when vocalizing, were doing so at a normal conversational level and when not vocalizing were not hyperventilating. This suggests that they were relaxed and had had sufficient training.Statistical Inference
Because we had a priori hypotheses for the regions of interest that we planned to test with our experimental paradigm, i.e., sensorimotor cortex, SMA, cerebellum, and Broca's area, we made no correction in our statistical analysis for multiple comparisons (8). Of the other brain regions for which we had no strong a priori hypotheses, only the activation in the superior temporal gyrus survived correction for multiple comparisons (P < 0.001).Validity of Using Elements of Speech as Independent Variables
The validity of manipulating the elements of a task that has both cognitive and sensorimotor components as if they were simple arithmetical symbols entirely depends on the absence of an interaction between these elements. This is clearly highlighted in the present studies (Table 1) in which, with the objective of increasing the power of the analysis, the results from simple contrasts between two conditions have been combined, utilizing an orthogonal design. This makes the subtraction complex. The problems that this may cause have been extensively reviewed by Friston (6). Formal investigation by Sidtis et al. (24) of these problems during speaking with minimum language content concluded that often, but not always, statistical maps produced by compound subtractions (as in the present study) diverge more from the results of nonimaging studies of the functional anatomy of speech than do those produced by simple subtractions. Examination of a linguistic task combined with finger movements (13) demonstrated that cognitive and behavioral processes do not combine in a simple additive manner. Nevertheless, in the present study, if images produced by complex subtractions are compared with those from simple subtractions (Fig. 4 compared with Fig. 6A; Fig. 7 compared with Fig. 6B), they are found to be surprisingly alike. Indeed, the foci of activation for control of breathing for speech and vocalization with articulation present (AB) were similar to
the foci when articulation was absent
(CD). There is further
support for the use of complex subtractions in the fact that, in the
simple subtraction (AC; Table
1), the superior temporal lobe activation associated with hearing
disappears (even though the nature of the sound being produced is very
different) as would be predicted. Addition of the second simple
subtraction (BD), which does
not contain hearing, does not interfere with the disappearance of
hearing in the complex contrast (Fig. 4).
General Motor Control
The cerebellum was activated in both main contrasts defining areas concerned with articulation (Fig. 4) and breathing for speech with vocalization (Fig. 7). By contrast, the SMA was activated only in the contrast defining control of breathing for speech and vocalization (Fig. 7). These activations are entirely typical of those found with any volitional action (18), presumably reflecting planning and control mechanisms. The apparent lack of SMA activation in the contrast defining areas concerned with the control of articulation is likely to have resulted from an effective subtraction of SMA activations in the components of the contrast (Table 1). It is of interest that these general motor control areas are activated in speaking, even when speech is overlearned, repetitive, and requires no attention.Articulation
The main contrast designed to show areas in the cerebrum activated with articulation uncontaminated by other activities has shown clearly defined areas in the sensorimotor cortex (smc1) with very high statistical probability (Fig. 4). The areas are symmetrical and large. They are entirely compatible with the map of Penfield and Rasmussen (see Fig. 22 in Ref. 20), with the homunculus showing the lip representation on the motor homunculus dorsal to the jaw representation and ventral to the face representation but occupying an area larger than both. The tongue representation on the map is more ventral and close to the Sylvian fissure. In the present study, the tongue should not have been tensed or moved, but there is no objective evidence of this. Some tongue movement may be associated with the ventral extensions of the activations, especially seen on the left side in the main contrast (Fig. 4) but not seen in the simple subtraction (Fig. 6A) in which condition C is contrasted with condition A alone.We found no definite evidence for the articulatory premotor area,
concerned with the planning of articulatory movements and recently
described from studies of infarcts in patients with apraxia of speech,
to be in the precentral gyrus of the insula on the left side (4). This
small area with left-sided coordinates (using the same Talairach system
as in the present study) of x = 41, y = 2,
z = +10 may possibly have been
activated during articulation in the present study (see Fig. 4).
However, the nearest area of significant
(Z = 2.65) activation in the present
study is at coordinates x = 46,
y = 6,
z = +16. There also appears to be an
area significantly (Z = 2.5) activated
at similar coordinates on the opposite side. These are close to the
coordinates given by Dronkers (4).
Breathing for Speech and Vocalization
Penfield and Bouldrey (19) made the seminal observation that vocalization, usually without lip or tongue movements, may result from electrical stimulation over a defined area of the surface of the exposed lateral motor cortex of man during neurosurgery over a range of sites in either hemisphere. The continuous involuntary sounds produced were terminated at maximum expiratory volume and resumed after an inspiration. The areas we have described as being associated with the control of breathing for speech and vocalization, namely smc2 and mc, are in similar regions of the cortex. The significance of the activations seen in the mc area was weaker than in the smc2 area. The level of significance had to be reduced to P < 0.01 to show that these areas were bilaterally activated. The likelihood that mc is indeed an area concerned with the control of breathing for speech and vocalization is increased by its preservation as an activated area when data from condition B (mouthing speech) was removed in the simple subtraction shown in Fig. 6B. The bar graph for mc in Fig. 7 showing that the mean adjusted rCBF in condition B is at the same level as in conditions A and C (although significantly higher than for condition D) also suggests that breathing for speech (but not vocalization) is contaminating condition B. The higher levels of significance of the activation in area smc2, both in the complex and simple subtractions, give greater confidence that these areas are concerned with breathing for speech and vocalization. The bar graph in Fig. 7 for smc2 also shows a reduction in mean adjusted rCBF for condition B that is below that for conditions A and C but not to the level of condition D. We argue that the effect of contamination of condition B with breathing for speech is of less importance when activations are highly significant (as in smc2) than when such activations are less significant (as in mc).Ramsay and colleagues (21) reported on activations in the motor cortex bilaterally that occur with a controlled willful expiration without vocalization (see Fig. 3C in Ref. 21). The y and z coordinates of the pixels of maximum significance are similar to those found for the mc area in the present study, but the x coordinate is 8 mm more medial. When the coordinates in the study of Ramsay et al. are compared with those of smc2 in the present study, they are 14-15 mm more anterior, but this is still within the overall areas of activation of both studies. We cannot therefore be certain that areas of the sensorimotor cortex activated by willful expiration are similar to those areas activated by behavioral expiration (as in speech), but there is certainly an overlap. This question cannot be resolved with the present data because of the difficulty of geometrically defining the limits of activations which have length and depth.
Conclusion
These studies have used the simplest possible paradigm to focus on speech of a very limited and stereotyped form so that we can separate the elements of its complex motor control. The challenge now is to extend these studies to more natural speech so that the mechanisms underlying the interaction between language and speaking can be revealed.ACKNOWLEDGEMENTS
We thank Professor R. S. J. Frackowiak, Wellcome Department of Cognitive Neurology, Institute of Neurology, London, and Professor D. Brooks, Medical Research Council Cyclotron Unit, Hammersmith Hospital, London, UK, for helpful preliminary discussions. We also thank Dr. K. Friston, Wellcome Department of Cognitive Neurology, Institute of Neurology, London, for statistical advice during the planning of these experiments. We are grateful to Elizabeth Whitcombe (Spinal Injuries Unit, Stoke Mandeville Hospital) for persistence in urging that this type of study be done.
FOOTNOTES
Address for reprint requests: K. Murphy, Dept. of Respiratory Medicine, Imperial College of Medicine, Charing Cross Campus, St. Dunstan's Rd., London W6 8RP, UK (E-mail: K.Murphy{at}cxwms.ac.uk).
Received 5 March 1997; accepted in final form 8 July 1997.
REFERENCES
| 1. | Broad, D. J. Phonation. In: Normal Aspects of Speech, Hearing, and Language, edited by F. D. Minifie, T. J. Hixon, and F. Williams. Englewood Cliffs, NJ: Prentice-Hall, 1973, p. 127-168. |
| 2. | Broca, P. Perte de la parole ramellissment chronique et destruction partielle du lobe anterieur gauche du cerveau. Bull. Soc. Antrop. Paris 2: 235-238, 1861. |
| 3. | Draper, M. H., P. Ladefoged, and D. Whitteridge. Respiratory muscles in speech. J. Speech Hearing Res. 2: 16-27, 1959. |
| 4. | Dronkers, N. F. A new brain region for coordinating speech articulation. Nature 384: 159-161, 1996[Medline]. |
| 5. | Fisher, R. A. The Design of Experiments. Edinburgh: Oliver and Boyd, 1935. |
| 6. | Friston, K. J. Statistical parametric mapping and other analysis of functional imaging data. In: Brain Mapping. The Methods, edited by A. W. Toga, and J. C. Mazziotta. Boca Raton, FL: Academic, 1996, p. 363-375. |
| 7. | Friston, K. J., J. Ashburner, C. D. Frith, J. B. Poline, J. D. Heather, and R. S. J. Frackowiak. Spatial registration and normalization of images. Human Brain Mapping 3: 165-189, 1995. |
| 8. | Friston, K. J., A. P. Holmes, K. J. Worsley, J.-B. Poline, J. D. Heather, and R. S. J. Frackowiak. Statistical parametric maps in functional imaging: a general linear approach. Human Brain Mapping 2: 1-25, 1995. |
| 9. | Gutzmann, H. Zur physiologie und pathologie der atemsbewegungen (Pneumographie). Berlin: Klin-Wochenschrift, 1906. |
| 10. | Herholz, K., A. Thiel, K. Wienhard, U. Pietrzyk, H. M. Vonstockhausen, H. Karbe, J. Kessler, T. Bruckbauer, M. Halber, and W. D. Heiss. Individual functional anatomy of verb generation. Neuroimage 3: 185-194, 1996. [Medline] |
| 11. | Hixon, T. J. Respiratory function in speech. In: Normal Aspects of Speech, Hearing, and Language, edited by F. D. Minifie, T. J. Hixon, and F. Williams. Englewood Cliffs, NJ: Prentice-Hall, 1973, p. 73-126. |
| 12. |
Hunker, C. J.,
J. H. Abbs,
and
S. M. Barlow.
The relationship between Parkinsonian rigidity and hypokinesia in the orofacial system: a quantitative analysis.
Neurology
32:
749-754,
1982 |
| 13. | Jennings, J. M., A. R. McIntosh, S. Kapur, E. Tulving, and S. Houle. Image subtractions do not always add up: the interaction between semantic processing and response modes. Neuroimage 3: S226, 1996. |
| 14. | Mazziotta, J. C., S. C. Huang, M. E. Phelps, R. E. Carson, N. S. MacDonald, and K. Mahoney. A noninvasive positron computed tomography technique using oxygen-15-labeled water for the evaluation of neurobehavioral task batteries. J. Cereb. Blood Flow Metab. 5: 70-78, 1985[Medline]. |
| 15. | McFarland, D. H., and P. Lund. The control of speech. In: Neural Control of Skilled Human Movement, edited by F. W. J. Cody. London: Portland, 1995, p. 61-75. |
| 16. |
Mohr, J. P.,
M. S. Pessin,
S. Finkelstein,
H. H. Funkenstein,
G. W. Duncan,
and
K. R. Davis.
Broca aphasia: pathologic and clinical.
Neurology
28:
311-324,
1978 |
| 17. | Murphy, K., D. R. Corfield, A. Guz, G. Fink, J. Harrison, R. J. S. Wise, and L. Adams. Cerebral areas associated with the motor control of speech in man. J. Physiol. (Lond.) 497P: 22P, 1996. |
| 18. | Passingham, R. E. The Frontal Lobes and Voluntary Action. Oxford, UK: Oxford Univ. Press, 1993. |
| 19. |
Penfield, W.,
and
E. Bouldrey.
Somatic motor and sensory representation in the cerebral cortex of man as studied by electrical stimulation.
Brain
60:
389-443,
1937.
|
| 20. | Penfield, W., and T. Rasmussen. Vocalization and speech arrest. In: The Cerebral Cortex of Man. A Clinical Study of Localization of Function, edited by W. Penfield, and T. Rasmussen. New York: Macmillan, 1950, p. 88-108. |
| 21. |
Ramsay, S. C.,
L. Adams,
K. Murphy,
D. R. Corfield,
S. Grootoonk,
D. L. Bailey,
R. S. Frackowiak,
and
A. Guz.
Regional cerebral blood flow during volitional expiration in man: a comparison with volitional inspiration.
J. Physiol. (Lond.)
461:
85-101,
1993 |
| 22. | Sears, T. A., and J. Newsom-Davis. The control of respiratory muscles during voluntary breathing. Ann. NY Acad. Sci. 155: 183-190, 1968. |
| 23. | Shea, S. A., and A. Guz. Personnalité ventilatoire-an overview. Respir. Physiol. 87: 275-291, 1992[Medline]. |
| 24. | Sidtis, J. J., J. R. Anderson, K. Rehm, S. C. Strother, K. A. Schaper, and D. A. Rottenberg. Are brain functions really additive? Neuroimage 3: S93, 1996. |
| 25. | Silbersweig, D. A., E. Stern, C. D. Frith, C. Cahill, L. Schnorr, S. Grootoonk, T. Spinks, J. Clark, R. Frackowiak, and T. Jones. Detection of thirty-second cognitive activations in single subjects with positron emission tomography: a new low-dose H215O regional cerebral blood flow three-dimensional imaging technique. J. Cereb. Blood Flow Metab. 13: 617-629, 1993[Medline]. |
| 26. | Spinks, T. J., T. Jones, D. L. Bailey, D. W. Townsend, S. Grootoonk, P. M. Bloomfield, M. C. Gilardi, M. E. Casey, B. Sipe, and J. Reed. Physical performance of a positron tomograph for brain imaging with retractable septa. Phys. Med. Biol. 37: 1637-1655, 1992[Medline]. |
| 27. | Talairach, J., and P. Tournoux. Co-Planar Stereotaxic Atlas of the Human Brain. Stuttgart, Germany: Thieme, 1988. |
| 28. | Townsend, D. W., A. Geissbuhler, M. Defrise, E. J. Hoffman, T. J. Spinks, D. L. Bailey, M. C. Gilardi, and T. Jones. Fully 3-dimensional reconstruction for a PET camera with retractable septa. IEEE Trans. Med. Imaging 10: 505-512, 1991. |
| 29. |
Warburton, E.,
R. J. Wise,
C. J. Price,
C. Weiller,
U. Hadar,
S. Ramsay,
and
R. S. Frackowiak.
Noun and verb retrieval by normal subjects. Studies with PET.
Brain
119:
159-179,
1996 |
| 30. | Whitcombe, E. The Physical Basis of Speech and Its Disorders. Oxford, UK: Oxford Univ. Press., 1997. |
This article has been cited by other articles:
|
|
 |
|
  N. S. Dhanjal, L. Handunnetthi, M. C. Patel, and R. J. S. Wise Perceptual Systems Controlling Speech Production J. Neurosci., October 1, 2008; 28(40): 9969 - 9975. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Padberg, J. G. Franca, D. F. Cooke, J. G. M. Soares, M. G. P. Rosa, M. Fiorani Jr, R. Gattass, and L. Krubitzer Parallel Evolution of Cortical Areas Involved in Skilled Hand Use J. Neurosci., September 19, 2007; 27(38): 10106 - 10115. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. E. Warren, D. A. Sauter, F. Eisner, J. Wiland, M. A. Dresner, R. J. S. Wise, S. Rosen, and S. K. Scott Positive Emotions Preferentially Engage an Auditory-Motor "Mirror" System J. Neurosci., December 13, 2006; 26(50): 13067 - 13075. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. N. Caviness, J. M. Liss, C. Adler, and V. Evidente Analysis of high-frequency electroencephalographic-electromyographic coherence elicited by speech and oral nonspeech tasks in Parkinson's disease. J Speech Lang Hear Res, April 1, 2006; 49(2): 424 - 438. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Saarinen, H. Laaksonen, T. Parviainen, and R. Salmelin Motor Cortex Dynamics in Visuomotor Production of Speech and Non-speech Mouth Movements Cereb Cortex, February 1, 2006; 16(2): 212 - 222. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G.M. Schulz, M. Varga, K. Jeffires, C.L. Ludlow, and A.R. Braun Functional Neuroanatomy of Human Vocalization: An H215O PET Study Cereb Cortex, December 1, 2005; 15(12): 1835 - 1847. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Haslinger, P. Erhard, C. Dresel, F. Castrop, M. Roettinger, and A. O. Ceballos-Baumann "Silent event-related" fMRI reveals reduced sensorimotor activation in laryngeal dystonia Neurology, November 22, 2005; 65(10): 1562 - 1569. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. J. McCrory, A. Mechelli, U. Frith, and C. J. Price More than words: a common neural basis for reading and naming deficits in developmental dyslexia? Brain, February 1, 2005; 128(2): 261 - 267. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Pinto, S. Thobois, N. Costes, D. Le Bars, A.-L. Benabid, E. Broussolle, P. Pollak, and M. Gentil Subthalamic nucleus stimulation and dysarthria in Parkinson's disease: a PET study Brain, March 1, 2004; 127(3): 602 - 615. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. C. McKay, K. C. Evans, R. S. J. Frackowiak, and D. R. Corfield Neural correlates of voluntary breathing in humans J Appl Physiol, September 1, 2003; 95(3): 1170 - 1178. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. C. Blank, S. K. Scott, K. Murphy, E. Warburton, and R. J. S. Wise Speech production: Wernicke, Broca and beyond Brain, August 1, 2002; 125(8): 1829 - 1838. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. E. Watkins, F. Vargha-Khadem, J. Ashburner, R. E. Passingham, A. Connelly, K. J. Friston, R. S. J. Frackowiak, M. Mishkin, and D. G. Gadian MRI analysis of an inherited speech and language disorder: structural brain abnormalities Brain, March 1, 2002; 125(3): 465 - 478. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. J. S. Wise, S. K. Scott, S. C. Blank, C. J. Mummery, K. Murphy, and E. A. Warburton Separate neural subsystems within `Wernicke's area' Brain, January 1, 2001; 124(1): 83 - 95. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. R. Corfield, C. A. Roberts, A. Guz, K. Murphy, and L. Adams Modulation of the corticospinal control of ventilation by changes in reflex respiratory drive J Appl Physiol, November 1, 1999; 87(5): 1923 - 1930. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||
ABSTRACT