Keywords
Landau–Kleffner syndrome - aphasia - dual-stream model - epileptogenic zone
Introduction
Landau–Kleffner syndrome (LKS) is a rare neurological disorder characterized by acquired
aphasia. LKS presents with distinctive electroencephalography (EEG) findings, including
diffuse continuous spike and wave complexes (CSW), particularly during sleep. There
has been little research on the mechanisms of the aphasia and its origin within the
brain. CSW occurs bilaterally in 80% of LKS patients and unilaterally in only 20%;
however, even unilateral cases exhibit spikes of LKS in both hemispheres synchronously.[1]
We diagnosed LKS in a 4-year-old female with an epileptogenic zone located primarily
in the right STG (nondominant side). In the course of her illness, she had early signs
of motor aphasia recovery but was slow to regain language comprehension and recover
from hearing loss. We suggest that the findings from our patient's brain imaging and
the disparity between her recovery from expressive and receptive aphasias are consistent
with the dual-stream model of speech processing in which the nondominant hemisphere
also plays a significant role in language comprehension.
Case Presentation
A 4-year-old, right-handed female presented with frequent (one to five times an hour)
paroxysmal blinking and impaired consciousness. Her health and psychomotor development
history had been normal until age three; however, her verbal articulation had begun
to slowly regress 6 months before the onset of blinking episodes. An EEG revealed
diffuse frequent Rolandic spike and wave synchronously in both hemispheres or independently
in the right hemisphere mainly in the central and parietal region ([Fig. 1A]). The seizures with rapid eye blinking and impaired consciousness during awake were
soon disappeared by administration of levetiracetam (25 mg/kg/d); however, the EEG
remained unchanged. The patient's speech ability gradually worsened, and she developed
aphasia within 3 months.
Fig. 1 (A) The scalp EEG during an acute phase revealed diffuse Rolandic CSW at central and
parietal bilaterally synchronized or right-hemisphere independently. (B,C) ECD-SPECT at relapse phase (B) showed hyperperfusion in the bilateral anterior temporoparietal area with right-hemisphere
dominance. This hyperperfusion disappeared in the recovery phase (C). (D) The equivalent current dipoles for MEG spikes (white dots) clustered mainly in the bilateral superior temporal gyri with right-hemisphere dominance
(AEF) to 1-kHz tone burst (N100m) was absent at the right auditory cortex, whereas
it remained normal at the left (large white circle, arrow).
A long-term EEG revealed diffuse CSW during both waking and sleeping states. CSW index
was 86.1%. No abnormalities were seen during brain magnetic resonance imaging, and
auditory brainstem responses were normal. Consequently, we diagnosed her with LKS.
The exome sequencing test was negative including GRIN2A. Based on the results of the
neuroimagings, we expect that the etiology of the syndrome is not structural, but
functional, with genetic factors yet to be proven.
The Wechsler Preschool and Primary Scale of Intelligence-III (WIPPSI-III) test at
age 4 years and 3 months was 80 for Full scale intelligence quotient (IQ), 55 for
verbal IQ, 116 for performance IQ, and 80 for processing speed IQ.
Immediately after the diagnosis of LKS, we treated her with two courses of steroid
pulse therapy (intravenous methylprednisolone [IVMP]) and clonazepam (0.04 mg/kg/d).
Fortunately, the CSW was effectively suppressed by these treatments.
For the evaluation of language function over time, we used the Sign-Significate method
(S-S method) of the National Rehabilitation Center for Persons with Disabilities,
which is commonly used in Japan. Before starting treatment, most of her vocalizations
were nonsensical and it was difficult to understand only auditory instructions, however,
within a month of the CSW being suppressed, it became possible to call some of the
items depicted on the picture cards. (One out of 16 tasks was answered correctly in
terms of accurate object names, and five tasks were comprehensible to the examiner
using infant words such as “meow” for cat and “vroom-vroom” for car). On the contrary,
there was no noticeable improvement in auditory instruction comprehension, and she
did not respond to listening task at all.
Three months later, small spikes again appeared in the right hemisphere, and the EEG
examination showed a recurrence of CSW. Her expressive speech disorder was less severe
at relapse compared with the first episode.
Examination by 99 mTc-ethylcysteine dimer single-photon emission computed tomography (ECD-SPECT) at relapse
showed hyperperfusion in the bilateral anterior temporoparietal area with right-hemisphere
dominance ([Fig. 1B]).
The patient's ability to respond to sound was tested with the use of an MEG at relapse,
which showed equivalent current dipoles of spikes to be concentrated in the bilateral
temporal gyrus, predominantly in the right hemisphere ([Fig. 1D]). An auditory evoked field from a 1 kHz tone burst (N100m) was absent in the right
auditory cortex but remained normal in the left ([Fig. 1D] and [Supplementary Figure S1] [online only]). The CSW and small spikes in the right hemisphere were suppressed
by one course of IVMP and increased doses of clonazepam (0.08 mg/kg/d).
The patient was taught sign language, which improved her communication skills and
allowed her to learn new words and sentences. Six months after the recurrence of CSW,
her verbal comprehension began to recover; she regained the ability to have simple
conversations, repeat sentences, and learn new words. (At this point,14 out of 16
tasks were answered correctly in terms of accurate object names, and she also got
the instructional comprehension task right at 14 out of 16 tasks). She continues to
recover, and the ECD-SPECT showed almost normal perfusion pattern in both hemispheres
a year after recovery ([Fig. 1C]). At the 2-year follow-up from the onset of the disease, she has been able to recite
simple sentences and carry-on daily conversations of preschool age, but it still takes
time for her to learn complex words. The Wechsler Intelligence Scale for Children-IV
(WISC-IV) test at the age of 6 was 76 for FSIQ, 76 for verbal IQ, 93 for performance
IQ, 76 for working memory IQ, and 80 for processing speed IQ.
In addition to language functioning, psychological factors such as difficulty in concentrating,
hyperactivity, and impulsivity may also have an impact on these problems. She will
enroll in a school for the deaf.
Discussion
We treated a 4-year-old girl with LKS who suffered rapid language regression. Her
epileptogenic zone was located in the right temporal cortex, including the primary
auditory cortex. In the acute phase, the EEG showed spikes that spread synchronously
in both hemispheres, suggesting that the dysfunction, including the left hemisphere,
caused rapid language regression. On the other hand, after IVMP treatments, MEG and
ECD-SPECT showed that the epileptogenic area was mainly in the right hemisphere, while
the AEF waves from the auditory cortex remained on the left side. These findings indicated
that after the first treatment, the dysfunction of the right hemisphere was strong,
but the left probably retained some function.
We hypothesized that this difference in time between left and right functional recovery
was related to the difference between recovery of speech and recovery of language
comprehension and new word learning.
Brain dysfunction was localized by trauma and infarction studies in adults, which
showed that the nondominant hemisphere lesion had no significant effect on language
function. The question arising at this time is whether the entire function of language
can remain normal in early childhood if only the right language center is impaired.
Our patient showed early signs of recovery from expressive aphasia, but why did she
continue to have receptive aphasia and diminished language comprehension? Unlike aphasia
in adult, the right-hemisphere disorder has been reported to cause delays in language
comprehension and gestures in early childhood.[2] In the period of language acquisition, it requires a process of understanding what
the words mean by integrating and understanding the visual, auditory, and contextual
information. It is thought that the right hemisphere works predominantly with respect
to its integrating role. This right-hemisphere integration processing advantage is
less noticeable once word acquisition is completed to some extent, replaced by a rapid,
efficient, and automatic process of mapping well-known sounds onto well-known semantic
patterns. Presumably, patients with early childhood onset LKS also require this right-hemisphere
function during the recovery process of lost language.
Advances in brain network research, such as diffusion tensor imaging, have led to
the language acquisition theory called the dual-stream model ([Fig. 2]).[3]
[4]
[5] The first step is for the input language information to be analyzed by the bilateral
dorsal STG and superior temporal sulcus for frequency-time and phonological domains.
Fig. 2 The dual-stream model of speech processing, with some simplification, based on the
study by Hickok and Poeppel.2 The early stage of speech processing occurs bilaterally
in the auditory cortex on the dorsal STG and STS (surrounded by an ellipse), then
divides into ventral and dorsal streams. IFG, inferior frontal gyrus; ITS, inferior
temporal sulcus; MTG, middle temporal gyrus; PM, premotor; SPT, Sylvian parietal-temporal;
STG, superior temporal gyrus; STS, superior temporal sulcus.
The input language is further analyzed in two parallel routes: the dorsal and the
ventral pathways. The dorsal pathway, with its strong left hemispheric dominance,
centers on grammar in higher-order language functions. The ventral pathway, which
has a slightly weaker left hemispheric dominance, plays a role in accessing lexical
memory and processing linguistic meaning after the recognition of phonemes as language.
In other words, it can be said that the input language information is initially processed
in both hemispheres, and that language acquisition and generation involve collaboration
between the two hemispheres, not just the dominant hemisphere.
Our patient's symptoms of auditory agnosia and language persisted even after the spike
spreading to the left hemisphere had subsided, suggesting that the frequency and phonological
analysis originally performed bilaterally may have been inadequate because the right
hemisphere was still damaged. In fact, follow-up ECD-SPECT 1 year after relapse showed
that the hyperperfusion of the right hemisphere has disappeared ([Fig. 1C]), and by this time her receptive aphasia was getting much better. This course of
ECD-SPECT recovery and language recovery supports our considerations. We speculate
that our case, course, and neuroimaging are consistent with the dual-stream model
of speech processing in which the non-dominant hemisphere also plays a major role
in language comprehension, especially in early phonological processing and new word
acquisition.