Cognitive Neuroscience

NeuroReport
NeuroReport 10, 681±685 (1999)

SEVERAL neurophysiological studies have highlighted
the role of the midbrain periaqueductal gray matter
(PAG) in the initiation of vocalization in various
animal species, from frogs to primates. With regard to
humans, only two cases of complete mutism following a
lesion to the PAG have been reported so far. This article
describes a new case of a patient (GM) who, following
an ischemic lesion to the periaqueductal gray region of
the midbrain, presented with complete and irreversible
mutism, though her language comprehension functions
and her non-verbal expression capacity were preserved.
This clinical case provides evidence that in humans the
PAG also acts as a link between different vocalizationeliciting external and internal stimuli (which reach the
PAG from sensory and emotional structures) and the
vocal±motor coordinating mechanisms in the lower
brain stem. NeuroReport 10:681±685 # 1999 Lippincott
Williams & Wilkins.

Key words: Emotion; Midbrain; Mutism; Periaqueductal
gray; Speech; Vocalization

Introduction
Lesions in different structures of the nervous system
may provoke patterns of transitory or persistent
mutism. Lebrun [1] schematically identi®ed ®ve
types of mutism: (i) patients with total (or global)
aphasia who, by de®nition, have lost the ability to
use language both expressively and receptively. They
generally present with hemiplegia or hemiparesis but
usually respond adequately to non-verbal stimulation. (ii) Patients with apallism who look awake but
do not show awareness of their environment. They
do not speak and do not follow verbal instructions
but they may move or shout or moan spontaneously. This condition is believed to be due to a
complete disconnection of the cerebral cortex, which
is responsible for cognitive functions, including
speech, from the underlying nervous structures. (iii)
Patients with `locked-in syndrome', who have spastic quadriplegia, and are mute but alert. Usually
their comprehension of spoken language is preserved
and it is possible to communicate with them by
means of eye movements (i.e. upward gaze for `no'
and downward gaze for `yes'), since they can
voluntarily control a few isolated eye movements
only. (iv) Patients with transitory mutism following
cerebellar damage [2]. (v) Patients with `akinetic
mutism' who do not speak and do not move (or
only weakly and after repeated stimulation) but are
0959-4965 # Lippincott Williams & Wilkins

Complete mutism after
midbrain periaqueductal
gray lesion
Antonio Esposito, Guy Demeurisse,
Barbara Alberti1 and
Franco Fabbro1,2,CA
Service de Revalidation Neurologique, CHU
Brugmann, Universite Libre de Bruxelles, Brussels,
Belgium; 1 Children's Neurological Research
Hospital `E. Medea', Ass. La Nostra Famiglia, Via
Don L. Monza, 20, 23842 Bosisio-Parini (LC);
2
Department of Physiology and Pathology,
University of Trieste, Italy
CA,1

Corresponding Author and Address

aware of their environment and have preserved at
least elementary speech comprehension. The lesion
responsible for akinetic mutism may lie at various
levels between the pons and the frontal lobes. The
most frequently involved structures are the midline
portions of the midbrain and both gyri cinguli [3,4].
A nervous structure which is of fundamental
importance for the initiation of vocalization and
speech is the periaqueductal gray (PAG). Electrical
stimulation and experimental lesion studies highlighted that this structure is responsible for initiating
vocalization in numerous animal species such as
frogs, cats and apes. In these species, a lesion to the
PAG provokes complete and permanent mutism
[5,6]. Experimental neuroanatomical studies in
squirrel monkeys have revealed that numerous motivation-controlling brain structures, such as the mediobasal amygdala, the middle thalamus, different
hypothalamic areas and the anterior cingulate cortex,
have direct projections to the periaqueductal gray
and laterally bordering tegmentum. The PAG has
direct projections to the reticular formation around
the nucleus retroambiguus, a region called periambigual reticular formation. The periambigual reticular
formation projects to the trigeminal motor nucleus
responsible for jaw control, the facial muscle responsible for lip movements, the nucleus ambiguus
controlling the vocal folds, and the thoracic and
upper lobar ventral horn containing the expiratory
Vol 10 No 4 17 March 1999

681

NeuroReport
motoneurons [7]. These neuroanatomical and neurophysiological studies highlighted that the PAG is
the phylogenetically oldest structure for the generation of sounds. It seems that the periaqueductal
region does not participate in the motor coordination of vocalization itself, but it would couple
speci®c motivations (depending on different internal
and external stimuli) with the corresponding vocal
expression [7,8]. Numerous neurobiological research
studies have shown that the PAG not only is
responsible for the initiation of vocalization, but it is
also involved in at least four other interactive major
functions, namely pain processing, autonomic regulations, fear and anxiety, and sexual behavior components [7].
In humans, only a few cases of mutism following
lesions to the periaqueductal gray matter have been
described in detail [9]. Steriade et al. [10] reported
the case of an 11-year-old boy with hemorrhage
secondary to an aneurysm of the basilar artery. The
clinical ®ndings favoured the existence of a lesion in
the higher part of the brainstem. After the coma
period, the patient displayed dissociation between
complete mutism and the ability to perform movements spontaneously or upon request. He chewed
and swallowed normally. The second patient was
described by Botez [11]. This 59-year-old patient
had suffered two cerebral vascular strokes which left
him with right hemiparesis, left abducens oculomotor paresis and mental disturbances. The patient had
been completely mute for 5 months. He had not
uttered even the slightest mumbling sound (not even
in reaction to nociceptive stimuli). Language comprehension was suf®ciently preserved because he
was capable of performing simple commands correctly. In contrast to this picture of mutism, kinesia
was normal. The autopsy revealed a macroscopic
lesion in the periaqueductal gray matter. The two
cases described by Botez et al. [10,11] suggest that
in humans the periaqueductal gray matter is also
essential for the starting mechanism of speech.
The present article presents another case of complete and permanent mutism following a lesion to
the periaqueductal gray matter. So far very few cases
describing this mutism pattern have been reported.
The more accurate neuro-imaging techniques used
in this study allow for a better understanding of the
neurofunctional mechanisms of the brainstem involved in the initiation of speech.

Patient and Methods
Case report: G.M. is a 74-year-old right-handed
(native French), retired of®ce worker. In December
1994 G.M. suffered a stroke which provoked a
right-sided hemisyndrome with consequent Broca's
682

Vol 10 No 4 17 March 1999

A. Esposito et al.
aphasia. After the ®rst stroke the patient showed a
good functional recovery, both at motor and linguistic level. She could live on her own since she was
completely self-suf®cient. After 10 months (September 1995) G.M. underwent a neurological and neuropsychological examination which revealed a mild
paresis of her right hand (1/5) and a slight de®cit of
the seventh right cranial nerve. When walking she
slightly dragged her right leg. Deep re¯exes were
more pronounced on the right side. No ocular
motility or visual discrimination disorders were
detected. Cutaneous sensitivity, vibration discrimination and pain sensitivity were preserved. Her
neuropsychological recovery was good. The patient
showed a good orientation as to time and space.
Verbal comprehension and repetition were preserved. Verbal expression was good. A slightly
reduced verbal ¯uency and phonemic paraphasias
were the only detectable disorders.
On June 30, 1997 G.M. was found at her home in a
comatose state. An investigation revealed that the
second stroke had occurred ,2 days before the
patient was found. G.M. was admitted to the Brugmann Hospital, Universite Libre de Bruxelles. Upon
admission she did not open her eyes, nor did she
produce any verbal utterance, but she reacted to
painful stimuli by trying to avert them with a hand
movement. An emergency CT scan showed a hypodense (ischemic) area located in the brain stem (Fig.
1), a hypodense image in the right thalamic region,
and an old ischemic area related to the ®rst stroke
(December 1994) and affecting the left lenticular
nucleus, a part of the head of the caudate nucleus and
the left internal capsule. After a few days the patient
was transferred from the intensive care department to
the neurological rehabilitation department where she
stayed until April 1998, when she was transferred to
a nursing home. Neurological examination (April
1998) revealed a paralysis of the third bilateral cranial
nerve (ptosis, large ®xed pupils and divergent strabismus), hypersomnia, mutism and right-sided hemiparesis. The presence of ophthalmoplegia, mutism and
hypersomnia indicated a rostral brain stem lesion.
The signs of dysfunction of the pupils and of the
third nerve clearly indicated a lesion in proximity of
the periaqueductal gray region of the midbrain [12±
14]. If the patient was left on her own, especially if
she was lying in bed, she tended to fall asleep. If she
was sitting and her eyes were held open, she was able
to interact with the hospital staff using simple forms
of non-verbal communication (nodding for `yes' and
`no', smiling). She showed a prominent facial hypomimia. During the months she spent in the neurological rehabilitation department she exhibited a low
level of emotional response, mainly lying in bed or
remaining seated.

Mutism after periaqueductal gray lesion

FIG. 1. CT scan without contrast obtained 2 days after G.M. developed
a complete mutism with bilateral paralysis of the third bilateral cranial
nerve (ptosis, large ®xed pupils and divergent strabismus) (June 1997).
The number on the upper corner of each scan marks the caudal rostral
progression. The CT scan shows a hypodense (ischemic) area located in
the brain stem (the midbrain lesion is indicated by black arrows in the
2nd and 3rd slice); a hypodense image in the right thalamic region (5th
slice), and; an old ischemic area related to the ®rst stroke (December
1994) affecting the left lenticular nucleus, part of the head of the caudate
nucleus and the left internal capsule (3rd±6th slice).

NeuroReport
Three months after the second stroke (September
1997) a series of instrumental examinations were
carried out to assess the extent of the mesencephalic
lesion. The EEG exhibited a regular activity in the
frequency range 7±9 Hz in the occipital region. A
rich polymorphic theta and delta activity mainly on
the left hemisphere was observed. The auditory
evoked potentials of the brain stem (BERA) were
normal. The blink re¯ex level was normal, thus
showing that no serious disorders had affected the
activating reticular formation. A SPECT (performed
using [99m Tc]bicisate) highlighted a pathological left
frontal and temporal cortical hypoactivity, while the
left basal nuclei were not observable.
More than 7 months after the second stroke the
clinical condition of the patient remains unchanged.
She is not able to sit upright, lies in bed or sits on a
couch by means of appropriate support. She is
hypersomnic, sleeps most of the day, yet if she has to
perform tasks (for example, eat or perform motor
and cognitive rehabilitation tasks) and her eyelids are
held open, she can stay awake for a few hours. She
can swallow liquids and chew food. Her sphincter
control is preserved. She exhibits spontaneous movements in her lower limbs and her left arm. Her tactile
discrimination and pain sensitivity are preserved. She
is oriented as to time and space (she knows she is in a
hospital and what year it is), she can recognize her
relatives. She is completely mute, does not speak, nor
does she produce any type of vocalization, not even
in reaction to painful stimuli. Over more than 5
months' hospitalization she has not produced any
type of vocalization, nor has she been heard crying.
Verbal comprehension is good, if she is called, she
turns her head towards her interlocutor. She understands and correctly performs all simple commands
given to her by hospital staff. Upon request she can
lift her lower and upper limbs correctly. However, a
mild right hemiparesis is still observable. She is able
to smile, show her teeth and open her mouth, if she is
asked to show her tongue, she opens her mouth and
tries to stick out her tongue without succeeding. At
times when she gives an af®rmative answer to a
question (for example, `Are you thirsty?') she moves
her lips in the correct position in order to pronounce
the ®rst phoneme of the word `oui', but she is not
able to produce any sound. She performs cerebellar
tasks correctly but slowly. If her eyelids are held
open she is able to perform simple verbal tasks (for
example, show two ®ngers with her left hand, three
®ngers, etc. upon request). She is able to perform the
pointing test correctly [15]. A magnetic resonance
study performed in December 1997 enabled a more
detailed description of the extent of the lesion which
affected the periaqueductal structures of the midbrain
(Fig. 2A,B).
Vol 10 No 4 17 March 1999

683

NeuroReport

A. Esposito et al.

FIG. 2. Transverse (A) and coronal (B) MRI scans obtained 5 months
after G.M. developed the second stroke (December 1997). The number
on the upper right corner of each scan marks the caudal to rostral
progression of the transverse slices, and the anterior to posterior
progression of the coronal slices. The midbrain lesion is indicated by
black arrows in the transverse (2nd and 3rd slice) and coronal (2nd slice)
scans.

Discussion
The cases of complete mutism in humans following
a lesion to the PAG are extremely rare in the
literature [1,7,9,16]. Numerous electrostimulation
and experimental lesion studies on various animal
species highlighted the fundamental role of this
structure in initiating vocalization and in the control
of pain±analgesia, fear±anxiety, autonomic regula684

Vol 10 No 4 17 March 1999

tion and sexual behavior components [6,8,17].
Lesions to the PAG in numerous animal species and
in humans provoke complete mutism without affecting voluntary motility and cognitive functions. Patients are oriented as to time and space and are able
to understand and perform verbal tasks [10,11].
Lesions to other structures of the nervous system,
e.g. the bilateral anterior cingulate cortex, may also
provoke mutism syndromes which, however, are

Mutism after periaqueductal gray lesion
transient [3], whereas lesions to the periaqueductal
gray matter provoke absolute and irreversible mutism [9,10].
The localization of the lesion in the patient we
studied was carried out both at neuroanatomical and
functional level. Mutism in this patient was coupled
with excessive sleepiness, bilateral ptosis, bilateral
divergent strabismus and large ®xed pupils. This
picture is due to a lesion localized in the rostral
brain stem, which affects the nucleus of the third
cranial nerve, the Edinger±Westphal nucleus and the
periaqueductal gray matter, following occlusion of
the rostral basilar artery [13,14]. This lesion did not
affect the tactile, proprioceptive, auditory and pain
discrimination pathways, or the pyramidal tract.
The mutism patterns following a lesion to the
PAG are to be distinguished from the `locked-in'
syndrome, apallism and global aphasia. Since G.M.
was able to understand verbal commands and perform voluntary movements with her limbs, the
mutism pattern she exhibited cannot be due to the
®rst two clinical syndromes (locked-in syndrome
and apallism), nor to global aphasia. In fact, after the
®rst stroke (December 1994), which had provoked a
lesion to the basal ganglia of the left hemisphere, the
patient had suffered Broca's aphasia but had shown
a good recovery of verbal functions. Ten months
after the ®rst stroke the patient had shown a good
recovery of her verbal expression and the only
aphasic symptoms left were reduced ¯uency and
some phonemic paraphasias. The second stroke
(June 1997) not only provoked a mesencephalic
lesion but also a thalamic lesion mainly affecting the
right thalamus. Whereas lesions to the left thalamus
are known to provoke aphasic syndromes, no thalamic lesions have so far been described in correlation
with clinical patterns of complete and permanent
mutism [18±21].

Conclusions
Experimental neuroanatomical and neurophysiological studies suggest that the PAG, and in particular
the dorso-lateral portion, is the phylogenetically
oldest structure responsible for the generation of
vocalization [5,7]. This critical structure receives
different pathways from numerous subcortical structures (midline thalamus, mediobasal amygdala, sev-

NeuroReport
eral hypothalamic areas) and from the anterior
cingulate cortex and sends projections to the reticular formation of pons and medulla [17]. The hypothesis has been put forward that the PAG does not
participate in the motor coordination of vocalization
itself, but is allegedly responsible for the initiation
of vocalization and speech following appropriate
emotional stimuli [7]. This suggestion seems to be
corroborated by some aspects of G.M.'s behavior.
Indeed, the patient could not produce any moaning
sound or vocalization, not even in reaction to
painful stimuli, whereas she could chew, swallow
and open her mouth upon request. This fact allows
for putting forward the hypothesis that the PAG
plays a crucial role in the initiation of vocalization
and speech, since it acts as a link between other
emotional-motivational structures and the periambigual reticular formation, a structure of the reticular
substance functioning as motor coordinator of cranial nerves and respiratory centers, two structures
accounting for vocalization and speech.

References
1. Lebrun Y. Mutism. London: Whurr Publisher, 1990.
2. Ersahin Y, Mutluer S, Saydam S and BarcËin E. Clin Neurol Neurosurg 62,
130±134 (1997).
3. JuÈrgens U and Von Cramon D. Brain Lang 15, 234±248 (1982).
4. Cumming JL, Benson F, Houlihan JP and Gosenfeld LF. J Nerv Mental Dis 171,
255±259 (1983).
5. JuÈrgens U. Anatomical and functional cerebral organisation of phonation in
animals. In: Creutzfeldt O, Scheich H and Schreiner Chr., eds. Hearing
Mechanism and Speech. Berlin: Springer-Verlag, 1979: 171±182.
6. de Lanerolle NC and Lang FF. Functional neural pathways for vocalisation in the
cat. In: Newman JD, ed. The Physiological Control of Mammalian Vocalization.
New York: Plenum Press, 1988: 21±41.
7. JuÈrgens U. Behav Brain Res 62, 107±117 (1994).
8. Ploog D. Human Neurobiol 6, 227±238 (1988).
9. Botez M and Barbeau A. Int J Neurol 8, 300±320 (1971).
10. Steriade M, Botez MI and Petrovici I. Psychiat Neurol 141, 38±58 (1961).
11. Botez M. IdeggyoÂgyaÂszati Sz 1, 13±29 (1964).
12. Segarra JM. Arch Neurol 22, 408±418 (1970).
13. Caplan LR. Neurology 30, 72±79 (1980).
14. Caplan LR. Vertebrobasilar system syndromes. In: Vinken PJ, Bruyn GW and
Klawans HL, eds. Handbook of Clinical Neurology, vol. 53. Amsterdam: Elsevier,
1988: 371±409.
15. Paradis M. The Assessment of Bilingual Aphasia. Hillsdale, NJ: Erlbaum, 1987.
16. Lebrun Y. Mutism. In: Asher RE, ed. The Encyclopedia of Language and
Linguistics. Oxford: Pergamon Press, 1994: 2655±2658.
17. Behbehani MM. Prog Neurobiol 46, 575±605 (1995).
18. Crosson B. Subcortical Functions in Language and Memory. New York: The
Guilford Press, 1992.
19. Crosson B, Zawacki T, Brinson G et al. J Neurolinguistics 10, 277±300 (1997).
20. Wallesh C-W. J Neurolinguistics 10, 267±276 (1997).
21. Fabbro F, Peru A and Skrap M. J Neurolinguistics 10, 347±367 (1997).
ACKNOWLEDGEMENTS: We are grateful to Professor Y. Lebrun, Professor M.
Paradis and Dr V. DaroÁ for improving the manuscript.

Received 2 December 1998;
accepted 4 January 1999

Vol 10 No 4 17 March 1999

685