Lesion characteristics driving right-hemispheric language reorganization in congenital left-hemispheric brain damage
Introduction
In adults, damage to the left cerebral hemisphere commonly results in specific aphasic symptoms associated with the lesion site (Gottesman and Hillis, 2010, Kreisler et al., 2000). Satisfying recovery from aphasia is often associated with left-hemispheric perilesional activation (Lidzba et al., 2012, Saur et al., 2006, Szaflarski et al., 2013), with right-hemispheric language activation more commonly seen in the more impaired patients, i.e., those with large lesions (Sims et al., 2016) or incomplete recovery (Saur et al., 2006, Szaflarski et al., 2013). In contrast to this pattern, children with left hemispheric brain damage (i.e., acquired pre- or perinatally) do not usually show persisting and substantial language deficits. Language development after congenital brain lesions may be delayed (Chilosi et al., 2005), but by school-age, such patients are perceived as competent speakers of their native language (Bates and Roe, 2001, Ilves et al., 2014). Only when confronted with linguistically complex tasks do children and adolescents with early left-hemispheric brain lesions show deficits in grammar processing (Knecht and Lidzba, 2016, Lidzba et al., 2013, Schwilling et al., 2012) and discourse (Reilly et al., 1998, Reilly et al., 2013). The neural mechanisms underlying language recovery in adult aphasics and language acquisition in patients with perinatal stroke seem to differ with respect to the brain tissue involved: Directly comparing children with perinatal left-hemispheric stroke with recovered adult aphasics, Szaflarski and colleagues demonstrated that the pediatric group recruited a bilateral network for language production, while the adult group was indistinguishable from healthy controls regarding lateralization (Szaflarski et al., 2014). That many (but not all) patients with left-hemispheric pre-/perinatal pathology may rely on a right-hemispheric language network has repeatedly been demonstrated for patients with and without epilepsy (Liegeois et al., 2004, Rasmussen and Milner, 1977, Tillema et al., 2008). The reorganized, right-hemispheric activation pattern encompasses areas homotopic to those involved in left-hemispheric language organisation, as investigated in patients with congenital white-matter lesions (Staudt et al., 2002), and in patients with left-hemispheric brain damage after epilepsy surgery (Tivarus, Starling, Newport, & Langfitt, 2012). Interhemispheric reorganization seems to affect networks connecting various regions of the brain, including the cerebellum (Lidzba, Wilke, Staudt, Krägeloh-Mann, & Grodd, 2008). This mirrored topography in cortical representation is also reflected in a corresponding reorganization of the subcortical fibre tracts in the right as compared to the left hemisphere (Duffau, Peggy Gatignol, Mandonnet, Capelle, & Taillandier, 2008).
The driving force of typical or atypical language lateralization both in healthy subjects and patients with early left-hemispheric brain lesions has been a relevant research topic since Rasmussen's and Milner's (1977) seminal report on a large series of Wada tests. Not all patients with early left-hemispheric lesions have right-sided language dominance (Liegeois et al., 2004, Rasmussen and Milner, 1977), and thus, more specific factors must be expected to play a role. Perinatal infarctions of the left middle cerebral artery (predominantly affecting perisylvian cortex including the temporal lobe) seem to be more often associated with reorganization of language comprehension (Jacola et al., 2006) than pre- or perinatally acquired periventricular lesions affecting fronto-parietal white matter only (Brizzolara et al., 2002, Staudt et al., 2001). Small developmental lesions (like vascular malformations) or slowly evolving pathologies (like tumours) are not commonly associated with atypical language (Gaillard et al., 2007, Pataraia et al., 2004). In children with intractable epilepsy due to early lesions, lesion location at or near the typical language areas were not commonly associated with language reorganization in one study (Liegeois et al., 2004), but the likelyhood for atypical language organisation was increased with left frontal malformations of cortical development in another sample (Wilke et al., 2011). Epileptic seizures and even interictal epileptic activity have the potential to interfere with the representation of language function, presumably even independently of a structural lesion (Janszky, Mertens, Janszky, Ebner, & Woermann, 2006). In the absence of epileptic activity, strategically-located lesions affecting the articulatory motor tract may be sufficient for the reorganization of language production in patients with periventricular brain damage and intact cortex (Staudt et al., 2008, Staudt et al., 2001). Thus, the specific topography, extent and timing of a lesion, but also accompanying factors such as epilepsy seem to be important factors driving language (re-)organisation.
When studying language reorganization, it is important to consider that numerous functional imaging studies have shown that language comprehension and language production recruit overlapping, but dissociable neural networks in the brain (for a review see (Price, 2012)). Language production and language comprehension can be hemispherically dissociated not only in children with epilepsy (Kurthen et al., 1992, Wilke et al., 2010) and patients with prenatally acquired periventricular lesions (Staudt et al., 2001), but also in healthy children and adolescents (Lidzba, Schwilling, Grodd, Krageloh-Mann, & Wilke, 2011). Current neuroanatomical models propose that language processing is organized in a dorsal and a ventral language stream (‘dual-route model’), where both streams participate both in language comprehension and production (Dick and Tremblay, 2012, Friederici, 2015, Hickok and Poeppel, 2015). In the domain of language comprehension, however, the bilaterally represented ventral stream seems more important. This pathway of temporal and frontal language regions, connected by the extreme capsule fibre system and the uncinate fascicle, is involved in bottom-up language processing, such as phonological word form detection, morphosyntactic and lexical-semantic categorization, and lexical access (Friederici, 2015, Hickok and Poeppel, 2015, Skeide and Friederici, 2016). Consequently, in adult aphasics, comprehension deficits are associated with damage to the ventral extreme capsule fibre system (Kummerer et al., 2013). In the course of typical language acquisition in healthy children, the first two or three years are dominated by bottom-up language processing, relying mainly on the ventral stream (Skeide & Friederici, 2016). In contrast, the strongly left-lateralized dorsal stream is engaged in functions seemingly more crucial for language production. It encompasses the posterior frontal lobe, anterior insula, and the temporo-parietal junction, and it is relevant for the integration of sensory-motor information and the processing of complex syntax (Friederici, 2015, Hickok and Poeppel, 2015). Analogously, impairments in speech repetition are associated with lesions to the dorsal superior longitudinal and arcuate fascicle pathway in aphasic patients (Kummerer et al., 2013).
The usually small and often heterogeneous samples in which language reorganization after congenital brain lesions can be studied pose a methodological challenge. Nevertheless, it is of both clinical and neuroscientific interest to investigate in more detail (1) the gross lesion characteristics forcing the right hemisphere into language processing, and (2) specific brain regions that are crucial for typical language representation. In order to address these issues, we studied a homogeneous sample of patients with pre- and perinatally acquired left-hemispheric brain lesions. As done previously (Staudt et al., 2001, Staudt et al., 2002), we included only patients with clearly defined lesions involving only the left hemisphere, i.e., strictly unilateral periventricular white matter lesions or middle cerebral artery infarctions. We here employed functional MRI tasks for language production and comprehension in combination with the voxel-lesion symptom mapping (VLSM) approach to identify brain regions that induce reorganization of language comprehension or language production when damaged. Since infarctions of the middle cerebral artery typically affect the frontal and/or temporal grey and white matter, we expected a higher probability of reorganization of language comprehension in patients with middle cerebral artery infarctions than in those with periventricular lesions (Hypothesis 1a). More specifically, we expected that regions within the ventral stream of language processing, namely the middle temporal gyrus and inferior temporal sulcus, and underlying white-matter structures (i.e., the extreme capsule fibre system and the uncinate fascicle) would be the most relevant in driving reorganization of language comprehension (Hypothesis 1b). For language production, we expected that lesions within the dorsal stream of language processing, namely the temporo-parietal junction, premotor cortex, inferior frontal gyrus and anterior insula, and underlying white matter structures would be predictive for reorganization (Hypothesis 1c). Since language production and language comprehension usually recruit overlapping networks of the same hemisphere, we also expected a significant correlation between lateralization of language production and language perception (Hypothesis 2).
Since our sample also contained patients with epilepsy, we were able to test also an alternative hypotheses, namely that language reorganization will be more frequent in patients with epilepsy as compared to patients without epilepsy, irrespective of lesion location or lesion type (Alternative hypothesis 1a∗).
Section snippets
Subjects
Nineteen patients with pre- and perinatally acquired focal lesions to the left hemisphere participated in the study (age range 7–32 years; eight females). Eight patients had infarctions of the middle cerebral artery (MCA), while 11 had unilateral periventricular lesions (PL). Four out of 8 patients with MCA infarction (patients 02, 16, 18, 19) and 2 out of 10 patients with PL (patients 15 and 17) had epilepsy (medicated in all but patient 02; age at manifestation of epilepsy between 0 and 20
Lateralization analyses
In the healthy control sample, mean LI in the Beep Stories task was 0.12 (SD 0.37; range −0.67 to 0.81), indicating, on average across the group, a bilateral activation pattern in the temporal lobe. The lower limit for typical lateralization (mean – 2 SD) was −0.62. Typical lateralization is therefore LI > −0.62 (TYPcomp) and atypical lateralization is LI ≤ −0.62 (REOcomp). Mean LI in the Vowel Identification task in the healthy control sample was 0.69 (SD 0.17; range 0.23 to 0.91), indicating, on
Discussion
In a homogeneous sample of patients with pre-/perinatally acquired, exclusively left-hemispheric lesions, we could identify two predictors, namely lesion type and topography, for the reorganization of language comprehension.
Lesions caused by infarction of the left middle cerebral artery have a higher potential to drive atypical representation of language comprehension than lesions in the left periventricular white matter. The VLSM analysis further specified that this relation is topographical
Funding
This work was supported by the German research council [collaborative research centre 550, to KL, and grant number WI3630, to MW], as well as by the intramural funding programme fortüne [University Tübingen; grant number F1274155] and the Ministry of Science and the Arts, Baden-Württemberg & European Social Fund [“Margarete-von-Wrangell Habilitationsprogramm” to KL].
References (75)
- et al.
Cost function masking during normalization of brains with focal lesions: Still a necessity?
Neuroimage
(2010) - et al.
Dorsal and ventral pathways in language development
Brain and Language
(2013) - et al.
Timing and type of congenital brain lesion determine different patterns of language lateralization in hemiplegic children
Neuropsychologia
(2002) - et al.
A model for social communication and language evolution and development (SCALED)
Current Opinion in Neurobiology
(2014) - et al.
A semi-automatic algorithm for determining the demyelination load in metachromatic leukodystrophy
Academic Radiology
(2012) - et al.
Spatial normalization of lesioned brains: Performance evaluation and impact on fMRI analyses
Neuroimage
(2007) - et al.
The motor somatotopy of speech perception
Current Biology: CB
(2009) - et al.
Assessing language and visuospatial functions with one task: A “dual use” approach to performing fMRI in children
Neuroimage
(2011) - et al.
Language learning and brain reorganization in a 3.5-year-old child with left perinatal stroke revealed using structural and functional connectivity
Cortex
(2016) White-matter pathways for speech and language processing
Handbook of Clinical Neurology
(2015)