FACULTY OF SCIENCE
White matter in Asperger syndrome using
diffusion tensor magnetic resonance imaging
Literature Thesis: Research Master Brain and Cognitive Sciences - Neuroscience
10/5/2012
Esther D. A. van Duin BSc. (0522074)
Supervisor: Prof. T. van Amelsvoort MD PhD (Department of psychiatry AMC-UvA and Maastricht University) Co-assessor: J. Zinkstok MD PhD (Institute of psychiatry, King’s college London UK)
1
2 Contents
List of Abbreviations ... 3
Abstract ... 6
1. Introduction ... 7
1.1 Clinical characteristics of Asperger syndrome ... 8
1.2 Prevalence ... 8
1.3 Etiology ... 9
1.4 Thesis aim and outline ... 10
2. Neuroanatomical differences ... 11
2.1 Functional imaging in Asperger syndrome ... 12
2.2 Structural imaging in Asperger syndrome ... 12
2.2.1 Brain volume abnormalities ... 12
2.2.2 Grey matter abnormalities in Asperger syndrome ... 16
2.2.3 White matter abnormalities in Asperger syndrome ... 17
2.2.4 Conclusion ... 18
3. Diffusion Tensor Imaging (DTI) ... 20
3.1 General principle of DTI ... 20
3.2 DTI quantitative measures ... 22
3.3 Summary ... 23
4. White matter DTI findings ... 24
4.1 DTI of the normal developing brain ... 24
4.2 DTI findings in ASD ... 24
4.2.1 DTI findings on the development of the brain in ASD ... 24
4.2.2 Brain regions with abnormalities found with DTI in ASD ... 25
4.2.3 Conclusion ... 26
4.3 DTI findings in Asperger syndrome ... 26
4.3.5 Methodology ... 26
4.3.2 Frontal, temporal, parietal and occipital lobe white matter abnormalities ... 28
4.3.3 Cingulum abnormalities ... 30
4.3.4 Corpus Callosum abnormalities ... 32
4.3.5 Cerebellum abnormalities ... 33
4.3.6 Local over-connectivity and reduced global connectivity ... 33
4.3.7 Summary ... 35
5. Implications and conclusions ... 36
5.1 Neurobiological correlates with white matter abnormalities ... 36
5.1.1 Neuronal tissue abnormalities underlying DTI measures ... 36
5.1.2 White matter development and post-mortem findings ... 36
5.1.3 Genetics of Asperger syndrome and white matter development ... 37
5.1.4 MRS / SPECT / PET ... 39
5.2 Neurobehavioral correlates with white matter integrity ... 42
5.3 Conclusion ... 42
3
List of Abbreviations
5-HT2a serotonergic receptor 2a
AD axial diffusivity
ADI-(R) Autism Diagnostic Interview (Revised)
AQ Autism Spectrum Quotient
ARNT2 arylhydrocarbon receptor nuclear translocator 2
ASD Autism Spectrum Disorder
ATR anterior thalamic radiation
BA Brodmann area
CC corpus callosum
CG cingulate gyrus
CGMV cerebral grey matter volume
Cho Choline
Cr + PCr Creatine+phosphocreatine
CSF cerebral spinal fluid
DISC1 disrupted in schizophrenia 1
DLPFC dorsolateral prefrontal cortex
DNA Deoxyribonucleic acid
DSM-IV / V the Diagnostic and Statistical Manual of mental disorders fourth / fifth editon
DTI Diffusion Tensor Imaging
DW-MRI / DT-MRI Diffusion Weighted/Tensor Magnetic Resonance Imaging
EQ Empathy Quotient
FA fractional anisotropy
FSIQ Full Scale Intelligence Quotient
GWAS genome-wide association study
HFA high functioning autism
ICD-10 International Statistical Classification of Diseases and related health problems, 10th revision
IFO inferior fronto occipital fasciculi
IGF1 Insuline-like growth factor
ILF inferior longitudinal fasciculi
IQ Intelligence Quotient
LFA low functioning autism
MD mean diffusivity
MRI magnetic resonance imaging
MRS magnetic resonance spectroscopy
(H)-MRS (proton) magnetic resonance spectroscopy
MS multiple sclerosis
NAA N-acetyl aspartate
NLGN3 Neuroliging-3
NTF3 Neurotrophin factor 3
NTRK1 / 3 Neurotrophic Tyrosine Kinase Receptor, type 1 / 3
PDD pervasive developmental disorder
4
PET positron emission tomography
PIQ Performance Intelligence Quotient
PPI pre pulse inhibition
RD Ral diffusivity
ROI region of interest
SHANK3 SH3 and multiple ankyrin repeat domains 3
SL streamlines
SLF superior longitudinal fasciculi
SNP's Single nucleotide polymorphisms
SPECT single photon emission computed tomography
UK United Kingdom
UNC uncinate fasciculi
VBM voxel based morphometry
6
Abstract
Asperger syndrome is a pervasive neurodevelopmental disorder and part of the highly genetic and increasingly diagnosed Autism Spectrum Disorder (ASD). Asperger syndrome is characterized by impairments in social behavior, communication and repetitive patterns of behavior and interests. Although the etiology of Asperger syndrome is still unclear, there is consensus that Asperger syndrome and ASD in general have a neurobiological basis with pathological brain abnormalities. ASD is proposed to be a “connectivity disorder” and connections between different brain areas are also thought to be altered and differently organized in the Asperger brain. The development of the neuroimaging technique of diffusion tensor imaging (DTI) facilitates the research to abnormal brain connectivity in Asperger syndrome. In this thesis the current available literature on white matter abnormalities in Asperger syndrome, differences with autism and possible causal factors are discussed in order to gain further insight in the etiology of Asperger syndrome. Widespread abnormalities in white matter microstructural organization in the brain are found to be present in Asperger syndrome and consistent with findings in ASD. Brain area’s showing abnormalities in white matter integrity in Asperger syndrome include frontal- parietal-, temporal- and occipital lobe, and more specifically, the cingulum, the body of the corpus callosum and the right cerebellar region. In general, compared to healthy controls, decreased fractional anisotrophy (FA) along with increased mean and radial diffusivity (MD/RD) has been found in several brain areas in Asperger syndrome. The white matter abnormalities are suggested to be caused by different underlying risk factors, including variations in candidate genes encoding proteins involved in neuronal growth, migration and myelination. Reduced white matter integrity found in Asperger syndrome, results in abnormal connectivity between brain regions implicated in several Asperger’s and ASD characteristics. Among these are problems with theory of mind, social behavior, understanding of emotions and attention. The findings in Asperger syndrome support the hypotheses of local over-connectivity contrasted with reduced long range connections in ASD. The investigation of this type of neurobiological endophenotypes can help to explain the etiology of Asperger syndrome and of ASD in general. Knowledge on underlying neuroanatomical differences could ultimately lead to better diagnostics and treatment.
7
1. Introduction
In 1944 the pediatrician and child psychiatrist Hans Asperger was the first to describe a few cases of boys in the age between 6 and 11 years old which he described as little ‘professors’, because of their extreme special circumscribed interests and abnormal use of formal language (Asperger 1944). The boys had impairments of social interaction, communication and behavior, resulting in difficulties with peer relationship. In contrast to Leo Kanner’s classification of autism – where a typical delay in language development is an additional symptom - there was no specific attention for Asperger syndrome and the term was not used in mental health care for up to fifty years after Hans Asperger’s publication (Kanner 1943). Lorna Wing essentially introduced the term Asperger syndrome, by describing 34 cases with clinical features similar to the cases Hans Asperger had described and considering that there was need for a new diagnostic subgroup in addition to typical autism (Wing 1981). She suggested additionally that symptoms of Asperger syndrome could be recognized within the first two years of life by, for instance, the absence of imaginative play. Not until 1992, ten years after Wing had reintroduced the term, Asperger syndrome became a standard diagnosis when it was descripted in the tenth edition of the International Classification of Diseases (ICD-10(WHO 1993) and two years later when the classification was added to the fourth edition of the Diagnostic and Statistical manual of mental disorders (DSM-IV(APA 2000) (Table 1).
Table 1 (Diagnostic criteria of Asperger syndrome compared between three sets of clinical definitions. Conditions in bold represent symptoms defined as necessary for the presence of the syndrome (Table adapted from Klin & Volkmar 1997)
8 1.1 Clinical characteristics of Asperger syndrome
The diagnosis of Asperger syndrome is based on several criteria, including pervasive abnormalities in social behavior, impairments in reciprocal communication and an intense, sometimes obsessional interest in a particular subject (Attwood 2007). People with Asperger syndrome have normal to high intellectual abilities (IQ => 70) but lack the ability to understand emotions and have deficits in empathy (Woodbury-Smith & Volkmar 2009; Baron-Cohen 2009). Asperger syndrome is a pervasive neurodevelopmental disorder (PDD) and part of the Autism Spectrum Disorder (ASD) (DSM-IV). In addition to Asperger syndrome, the spectrum includes (atypical) autism (also known as ‘classic/childhood’ autism) and pervasive developmental disorder not otherwise specified (PDD-NOS) with different clinical phenotypes, making it a very heterogeneous spectrum. The first signs of ASD are present in early childhood and symptoms are clinically characterized by difficulties with social interactions, limited communication, abnormal use of language and stereotyped, repetitive patterns of behavior, interests and activates. In contrast to PPD-NOS and some cases of autism, children with Asperger syndrome have a moderate to high IQ and they often ‘talk before they walk’, whereas in autism and PPD-NOS speech development is typically delayed and not of great strength. The diagnostic criteria of impaired language development is used to distinguish Asperger syndrome from the clinically similar high functioning autism (HFA), which is diagnosed as autistic people with a total IQ score greater than or equal to 70. Phrase language development before 3 years of age is apparent in children with Asperger syndrome in contrast to (high functioning) autistic children where a delay in language development is typically seen (DSM-IV 1994). The scores on the performance IQ (PIQ) in autism is therefore often higher than the scores on verbal IQ (VIQ), in contrast to Asperger syndrome where the VIQ score is in general higher than the PIQ score (Lincoln et al. 1997).
1.2 Prevalence
The prevalence of ASD has dramatically increased over the past decades (Baird et al. 2006). Whereas the first survey of autism estimated a prevalence of 4:1000, recent research in the UK has estimated the population prevalence of ASD to be almost 1% (Baron-Cohen 2009). In the
9
Netherlands similar numbers are found (Roelfsema et al. 2011). The prevalence of Asperger syndrome depends on the diagnostic criteria used by epidemiological studies. A combination of six studies investigating the prevalence of Asperger syndrome compared to autism, estimates a median prevalence of 2.6/10.000, which was lower than the median prevalence of autism of 13/10.000 (Woodbury-Smith & Fred R Volkmar 2009). However a study by Ehlers and Gillberg, with other diagnostic criteria, came to a much higher number of 28.5/10.000 (Ehlers & C Gillberg 1993). The mean age of onset for autism is a little younger than that of Asperger, with only a few months difference. However, this finding was not strongly significant (Howlin 2003).
The sex ratio of affected individuals is 9:1 (boys : girls) for Asperger syndrome, whereas in autism this ratio is indicated as 4:1(Scott et al. 2002). Because of the high prevalence of Asperger syndrome in boys and the symptoms that suit male type behavior, some researchers suggest Asperger syndrome to be an ‘extreme male brain’ disorder (Baron-Cohen 2002; Baron-Cohen & Hammer 1997; Baron-Cohen 2009). This is supported by the observation Hans Asperger made, finding similar Asperger syndrome traits of the probands in other - particularly male - family members. Therefore genetic risk factors in male specific genes are expected to play a role in the etiology of Asperger syndrome. Moreover the strong family history of ASD amongst first-degree relatives implicates a high heritability rate which is confirmed by genetic studies showing that ASD is highly genetic (Folstein & Rutter 1988).
1.3 Etiology
Although the etiology of Asperger syndrome is still unclear, there is consensus that ASD in general and Asperger’s in specific have a neurobiological basis with pathological brain abnormalities. Because of the heterogenic manifestation of core symptoms in ASD, research is shifting towards the investigation of specific endophenotypes - heritable neuropsychological or neurobiological traits that lie on the pathway from genes to clinical phenotypes - in more homogenous patient populations including only Asperger patients (Gottesman et al. 2003). Several endophenotypes are suggested to underlie the disease phenotype of Asperger syndrome. The often young age of onset and the high heritability rate indicate that neurobiological
10
abnormalities possibly caused by genetic risk factors, influence the developing brain. A number of studies suggested specific brain areas to be involved in Asperger’s and neuroanatomical abnormalities have frequently been reported in this syndrome compared to healthy controls of the same age (see section 2 of this chapter).
ASD is proposed to be a “connectivity disorder” and it is thought that the connections between different areas in the Asperger brain are also altered and differently organized (Catani & Ffytche 2005; Courchesne & Pierce 2005; Geschwind & Levitt 2007; Frith 2004; Just et al. 2004; Just et al. 2007). Abnormal neuronal connectivity leads to impaired information processing between brain regions, which could be one of the underlying factors causing the characteristic Asperger’s symptoms. Therefore more interest is arising for the investigation of the integrity of these specific white matter connections in the brain as an endophenotype of Asperger syndrome. The investigation of this type of neurobiological endophenotypes can help to explain the etiology of Asperger syndrome, which could ultimately lead to better diagnostics and treatment.
1.4 Thesis aim and outline
The aim of this literature thesis is to review current literature (using PubMed) and give an overview of the knowledge on white matter abnormalities in the brain of people with Asperger syndrome and possible differences with ASD. First a brief description of the neuroanatomical differences between healthy controls, autism and Asperger syndrome will be outlined. In addition the focus will lie on white matter integrity in ASD and Asperger syndrome, investigated with diffusion tensor imaging (DTI), a relatively new imaging technique which can visualize connectivity in the brain. Finally, in line with findings of other neurobiological correlates, different theories and ideas on implications and possible causal factors of altered white matter integrity will be discussed.
11
2. Neuroanatomical differences
Several areas in the brain that have been implicated in ASD are also important for the core Asperger symptoms (figure 1. Amaral 2008). With the use of animal models, lesion studies, post-mortem brain material and brain imaging techniques including (functional) magnetic resonance imaging (MRI) and (proton) magnetic resonance spectroscopy (MRS), abnormal brain function and neuroanatomy has been characterized in ASD (Amaral et al. 2008; Toal et al. 2005). Most of these studies have investigated ASD or autism and highlighted regions including the frontal lobe, temporal lobe, amygdala and cerebellum to be pathological.
Figure 1 Brain areas implicated in mediation of three core symptoms impaired in autism: social behavior, language and communication, and repetitive and stereotyped behaviors (adapted from Amaral et. al. 2008)
12 2.1 Functional imaging in Asperger syndrome
The few functional imaging studies that included separate Asperger groups in addition to autism and control groups, have demonstrated abnormal activation patterns in the inferior temporal sulcus during social cognition tasks (Schultz et al. 2000), deactivation of frontal lobe areas during several neuropsychological tests, including theory of mind tasks (Happe et al. 1996), and abnormal functional integration of e.g. the amygdala and parahippocampal gyrus, during incidental processing of facial expressions was found in Asperger’s compared to controls (Welchew et al. 2005). Implicit and explicit processing of emotional facial expressions was also associated with abnormal brain activity in a mixed group of autism and Asperger’s, compared to controls, including mesolimbic, cerebellar and temporal lobe brain regions (Critchley et al. 2000).
2.2 Structural imaging in Asperger syndrome
2.2.1 Brain volume abnormalities
In children with ASD one of the neuroanatomical endophenotypes consistently found is megalencephaly (enlarged brains) and macrocephaly (large head size), which is also often observed in first degree relatives without ASD (for a review see Herbert 2005). The prevalence rate of macrocephaly in ASD is around 20% compared to 3% in the general population. The theory of the ‘extreme male brain’ phenotype in ASD and Asperger’s in particular, supports the assumption that megalencephaly is a feature of Asperger’s since 90% of the Asperger’s is male which are known to have larger brains than females (Baron-Cohen & Hammer 1997). Results of MRI studies showed that especially very young children with autism have abnormal enlargements in total brain volume (figure 2) in the range of 5-10% (Hazlett et al. 2006; Sparks et al. 2002).
13
Figure 2 Three phases of growth pathology in autism. Red line=ASD. Blue line=age-matched typically developing individuals. Early brain overgrowth in autism is followed by arrest and possible decline. In some regions and individuals, the arrest of growth may be followed by degeneration, indicated by the red dashes that slope slightly downward. (Adapted from Courchesne et al., 2007)
In a study measuring head circumference after birth and again after one year, macrocephaly appeared to be more common in the Asperger syndrome group compared to the –relatively low functioning- autism group ( Gillberg & de Souza 2002). They suggested that both conditions should be treated separately in research but also that subgrouping of patients should be based on etiology rather than level of functioning.
One study using structural magnetic resonance imaging (sMRI) in children and adolescence with Asperger syndrome, high functioning autism (HFA) and low functioning autism (LFA) compared to controls, investigated neuroanatomical differences between these groups (Lotspeich et al. 2004). They found no significant differences in total, grey and white matter volume between Asperger syndrome compared to HFA and controls, which was partly in line with another MRI study in adults showing no differences in total cerebral volume between Asperger syndrome and controls (Mcalonan et al. 2002). However Lotspeich and colleagues suggest that Asperger syndrome is on the mild end of ASD, based on the non-significant finding of Asperger’s cerebral grey matter volume being intermediate between HFA and controls. Moreover, they did find a significant positive correlation between PIQ and white matter in the Asperger group in contrast to the HFA and control group, where this correlation was significantly smaller since it was absent (figure 3). Although previous studies showed no association between IQ and white matter volume, these results do indicate functional white tissue differences between autism and Asperger syndrome.
14
Figure 3 Correlation between cerebral white tissue volume and performance IQ. Groups included 21 subjects with Asperger syndrome (ASP), 18 with high-functioning autism (HFA), and 21 control subjects. Between-group differences, ASP vs HFA p=0.03, ASP vs controls p=0.004 (adapted from Lotspeich et. al. 2004)
Several studies found that brain enlargement in ASD is generally a result of white matter (see Box 1.) increase, rather than grey matter. Especially early in life, white matter contributes disproportionally to the total volume increase, where white matter was 15% larger in 6 to 12 year old autistic boys compared to controls, the contribution of white matter to the total volume increase was 65%. White matter enlargement was found in young autistic children (Courchesne et al. 2001), contrasted to less white matter concentration in older autistics compared to their age-matched healthy controls (Chung et al. 2004; Waiter et al. 2005). Where grey matter increases are found to persist into adulthood, this seems not the case with white matter increase and total brain volume enlargement in autism (review by Amaral et al., 2008). Whether this abnormal uniform development of white matter is also present in Asperger syndrome is not as clear, since there are only a few studies performed investigating white matter development in a separate group of Asperger patients.
15 BOX 1. White matter (Kandel et al. 2000)
In the central nervous system there are two types of tissue, grey- and white matter, named after the color postmortem brain tissue obtains as a result of formaldehyde preservation. White matter is mainly composed of myelinated axons which form the main transmitting element of neurons (figure 4). The myelin is a fatty shade around the axon which facilitates the fast transition of action potentials – the conducting signal - along the axon. The speed and distribution of signal transition depends on the structure of the myelinated axon. In addition the myelination of axons is, to some extent, regulated by electrical activity which indirectly influences the gross geometry of axons (Demerens et al. 1996). The communication between grey matter regions therefore depends on the quality of white matter connections and the integrity of the myelinated axons.
Different types of white matter fiber bundles or tracts form the main connection between grey matter regions in the brain (Catani & Ffytche 2005). There are (1) association tracts, which connect distal cortical areas in one hemisphere, like the superior/inferior longitudinal fasciculus and the cingulum, (2) projection tracts, connecting cortical to subcortical structures, for instance the corona radiate and the fornix, and (3) commissural tracts connecting the left and right hemisphere, including the corpus callosum and the anterior/posterior commissures. (Figure 5)
Figure 5. Mynert’s classification of white matter tracts visualized with diffusion tensor imaging and superimposed on medial and lateral views of the brain surface (adapted from Catani et al., 2005)
Figure 4. Schematic overview of two neurons and the direction and transition of electrical signals (spikes).
16
2.2.2 Grey matter abnormalities in Asperger syndrome
There are a number of structural magnetic resonance imaging studies comparing the Asperger brain structure with autistic- and/or healthy control brains. McAlonan and colleagues investigated grey matter differences with voxel based morphometry and found only regional grey matter volume differences, but no total volume differences between HFA and Asperger syndrome (Mcalonan et al., 2008). The grey matter volume of subcortical, posterior cingulate and precuneus regions (including the thalamus and pallidum) was significantly larger in Asperger syndrome compared to HFA. In addition a recent anatomic likelihood meta-analysis of MRI studies comparing grey matter differences between Asperger, autism and healthy controls indicated widespread grey matter abnormalities in several regions of the Asperger brain (Yu et al. 2011). The abnormal regions included the bilateral amygdala, hippocampal gyrus and prefrontal cortex where, compared to controls, lower grey matter volumes were found. In a limited number of regions, including bilateral inferior parietal lobe and left middle temporal/fusiform gyrus, greater volumes of grey matter were found, whereas in the autistic group several regions were found to have enlarged grey matter volumes compared to controls. The total ALE meta-analysis revealed that grey matter increases were less present in the Asperger group compared to the autism group. Interestingly more bilateral enlargements of grey matter volume was found in autism, whereas in Asperger syndrome cluster of enlarged grey matter volume were found in the language dominant-left hemisphere in contrast to lower grey matter volume in the right hemisphere (Yu et al. 2011). An additional recent sMRI study investigated the total volume (grey and white matter) of the amygdala and hippocampus in Asperger’s compared to controls. A larger right and total volume of the amygdala was found in Asperger’s ( Murphy et al., 2012). There was also an age-related increase in left amygdala volume present in the control group but not in the Asperger group, reflected by an early amygdala enlargement in Asperger’s. This is consistent with previous studies reporting larger amygdala volumes in children with ASD compared to controls, which is in line with the overgrowth theory of enlarged brains of children with ASD (Sparks et al. 2002).
17
2.2.3 White matter abnormalities in Asperger syndrome
Structural MRI studies have previously reported widespread abnormalities in white matter (BOX 1.) volume in Asperger syndrome (Chen, Jiao, & Herskovits, 2011; Mcalonan et al., 2002). Manual tracing of white matter regions in structural magnetic resonance images, revealed bilateral white matter excesses in areas concentrated around the basal ganglia in Asperger’s compared to controls (figure 6).
Figure 6 Relative deficits clusters (blue) and excesses clusters (red) in white matter volume in people with Asperger syndrome compared with controls. Right side of brain is shown on the left side of each panel. The Z coordinate for each axial slice in standard space (MNI) is given in millimeters. (Adapted from McAlonan et. al. 2002)
18
In addition, two small clusters part of the uncinate fasciculus and anterior commissure in the left hemisphere also showed significant white matter excess compared to controls. White matter deficits appeared to be dominated in the left hemisphere, concerning the brain stem and frontal, temporal and occipital tracts. The left hemisphere normally has a delayed development compared to the right one, which is suggested to be caused by superior temporal lobe speech areas reaching maturation later than pathways linking lower order regions (Paus et al. 1999). The areas with white matter volume deficits included the left superior temporal lobe speech area (BA 22) perhaps responsible for language problems found in autism. The deficit area also included clusters of the inferior longitudinal and occipitofrontal fasciculus which are main connections in the limbic pathways (Pugliese et al. 2009). The mean white matter volume excesses found in the Asperger group, differed for 42% from controls and the mean white matter volume deficits differed for 21% from controls (Mcalonan et al. 2002). White matter located in frontal regions showed the greatest enlargement of 27% over controls.
Interestingly this study also investigated sensorimotor gating using pre pulse inhibition of the startle response (PPI) and found reduced mean PPI in Asperger compared to controls. Since PPI is thought to depend party on the connectivity of frontostriatal pathways, the combined anatomical and sensorimotor findings of McAlonan and colleagues suggest that abnormalities in frontostriatal white matter tracts, perhaps resulting in defective sensorimotor gating, can lead to Asperger characteristics including difficulties with the inhibition of repetitive thoughts, speech and actions (Mcalonan et al. 2002). Frontostriatal abnormality is often implicated in several developmental disorders (van Ewijk et al. 2012) including ASD (McAlonan et al. 2005; Toal et al. 2005; Bradshaw & Sheppard 2000).
2.2.4 Conclusion
In summary neuroanatomical abnormalities in total-, grey- and white matter volume are present in Asperger syndrome in which white matter is reported to contribute disproportionally to the brain enlargements found predominantly early in life. White matter abnormalities are found to
19
be present in several brain regions in Asperger syndrome and since it is suggested that ASD is a connectivity disorder (Catani & Ffytche 2005; Courchesne & Pierce 2005; Geschwind & Levitt 2007; Frith 2004; Just et al. 2004; Just et al. 2007), the investigation of specific white matter tracts as an endophenotype for Asperger syndrome seems plausible. However, only a few studies so far have investigated specific white matter anatomy in separate Asperger groups (Pugliese et al. 2009; Bloemen et al. 2010; Catani et al. 2008; Mcalonan et al. 2002). Nevertheless the development of the neuroimaging technique of diffusion tensor imaging (DTI) - allowing the investigation of specific white matter tracts in the brain - facilitates the research to abnormal brain connectivity in Asperger syndrome and preliminary results look promising (Bloemen et al. 2010).
20
3. Diffusion Tensor Imaging (DTI)
3.1 General principle of DTI
Diffusion tensor imaging (DTI), also known as diffusion weighted/tensor magnetic resonance imaging (DW-MRI / DT-MRI) is a relatively new MRI technique which can be used for in vivo (the living brain) research into the anatomical connectivity between brain regions (Alexander et al., 2007). In contrast to structural voxel based morphometry (VBM) - MRI, where white matter structures can only be investigated on a coarse level, DTI allows a more detailed investigation of the white matter anatomy; the white matter tracts between cortical areas can be visualized and the integrity of the fiber can be measured.
The technique is based on the random diffusion of water molecules (Le Bihan 2003) and measures the strength and directionality of this diffusion. The directional information on fiber orientations is used for specific white matter ‘tractography’ where the primary eigenvector’s ( λ ) of diffusion from every voxel in the brain is used to visualize and characterize major fiber bundles (Basser et al. 2000). DTI (indirectly) reflects white matter integrity in the brain, since diffusion of water is more readily along the orientation of axons (fiber tracts) that are restricted by cell membranes and myelin (Figure 7 and 8). The microscopic hindrances and restrictions that the molecules experience during diffusing through tissue represent a quantitative measure for the integrity of the white matter in that region. In grey matter and pure water, for instance the cerebral spinal fluid (CSF), the diffusion coefficient will be the same in every direction (isotropic) since diffusion is unrestricted. However, in white matter tissue the diffusion is restricted by the tight package of cellular axons and the myelinated sheath around them, resulting in a more anisotropic – i.e. longer along the axons than perpendicular to it - direction of diffusion.
21
Figure 7 A1) SMI32-stained pyramidal neurons in cerebral cortex and A2) transmission electron micrograph of a myelinated neuron. B) Schematic overview of neuron and direction of free/restricted diffusion. C) Detailed schematic overview of myelinated axon and direction of diffusion (partly adapted from http://www.fmrib.ox.ac.uk/fsl/).
Figure 8 Quantification of white matter integrity using diffusion tensor imaging. A) Local and B) Colour coded direction of diffusion tensor used to quantify white matter microstructure. The tensor represents the three different diffusion directions (Eigenvectors 1,2,3). C) Map of Fractional Anisotropy (FA) values in the brain. More isotropic diffusion (dark) corresponds with a low FA value and is present in grey matter tissue. More anisotropic (light) diffusion corresponding with high FA values (closer to 1), is more present in white matter tissue (adapted from http://www.fmrib.ox.ac.uk/fsl/)
22 3.2 DTI quantitative measures
There are several different quantitative values that can be measured with DTI (figure 7 and 8). The most commonly used are fractional anisotropy (FA) and mean diffusivity (MD). FA describes the direction (or degree) to which water molecules diffuse in space (isotropic or anisotropic, ranging from 0 to 1, respectively) and is related to the presence and coherence of the surrounding structures. (Le Bihan 2003; Alexander et al., 2007). MD characterizes the amount of diffusion and the overall presence of obstacles like (myelinated) membranes (MD: (λ1 + λ2 + λ3 ) / 3). Reduced FA is associated with less axonal organization or integrity and in general found in regions of crossing fibre tracts, since the direction of diffusion is greater. In one single tract reduced FA could implicate more isotropic diffusion due to deficits in the axonal integrity or demyelination. However the exact interpretation of the FA value and underlying factors remains ambiguous. Increased MD represents more free diffusion of water molecules in one brain voxel probably due to less restriction by membranes or myelinated axons. Mean diffusivity can be used for clinical application, for instance in the case of an ischemic accident or inflammation, where it can visualize changes in the empty spaces between brain regions (Tievsky et al. 1999; Mintorovitch et al. 1991). Although the interpretation of MD and especially FA measures remains complex, it is suggested that in a single fiber bundle, increased MD and reduced FA might reflect more isotropic diffusion of water molecules due to less restricted movement possibilities (Alexander et al., 2007).
Other less frequently used DTI measures are axial (parallel) (AD) and radial (perpendicular) diffusivity (RD), which represent a decomposition of the diffusion coefficient into a component (AD: principal axis=λ1) in the axon’s average principal- and average perpendicular direction (RD:
minor axes = (λ2 + λ3) / 2 ) and 4) number of streamlines (SL), generated by the tractography software to visualize fibers according to the principal diffusion directions defined by the largest eigenvector of the diffusion tensor (Mori et al. 2002; Goodlett et al. 2005). Number of streamlines is suggested to be a possible surrogate of tract volume (Pugliese et al. 2009). Axial and radial diffusivity are valuable measures to combine with FA, since they provide more specific insight into the neurobiological nature of abnormal integrity. For instance an increased isotropic or decreased anisotropic diffusion of water molecules (probably reflected by reduced FA) can be due
23
to increased radial diffusivity, decreased axial diffusivity or a combination of both. Demyelination is shown to be associated with increased radial diffusivity and decreased axial diffusivity (Harsan et al. 2006; Song et al. 2005). In addition regions with lower degree of neuronal branching are associated with decreases in radial diffusivity (Suzuki et al. 2003).
3.3 Summary
Altogether, the different DTI measures can reflect abnormal white matter organization and integrity, which could result in deficient signal communication. Therefore DTI is a very promising and important tool for the identification of specific neuropathology and applications of DTI in psychiatric research are quickly emerging (Thomason&Thomason 2011).
24
4. White matter DTI findings
4.1 DTI of the normal developing brain
White matter measures, including volume, mean diffusivity and diffusion tensor anisotropy, normally follow a nonlinear trajectory with increasing age (Hasan et al. 2007). Mean diffusivity decreases and fractional anisotropy increases during the first few decades of live. Thereafter the trajectory of FA and MD plateaus and eventually FA gradually declines whereas MD increases with age. This increase and reduction in FA and MD during the first phase of live is probably the result of axonal and neuronal pruning: reduction in brain grey matter and the formation of new neuronal networks (Thomason & Thompson 2011).
4.2 DTI findings in ASD
4.2.1 DTI findings on the development of the brain in ASD
As described previously the white matter development in ASD has been found to be different with enlarged white matter volume especially early in live (see section 2). One of the causal factors that might underlie the regional volume differences of white matter is thought to be abnormal myelination. DTI could serve as a useful tool in the investigation of these differences. In general reduced FA and increased MD is consistently found to be associated with autism (Alexander et al. 2007; Barnea-Goraly 2004; Catani et al. 2008; Noriuchi et al. 2010). Studies investigating specific white matter maturation and development in ASD found that age-dependent white matter changes in both children and adolescents (age range from 8 to 20 years) with ASD is similar in direction compared to children and adolescents with a typical development (Shukla, Keehn & Müller 2011). However they suggest that the increased FA decreased MD and decreased RD are more diminished and affected by greater variability in the ASD group compared to those in typical development. There is one DTI study done in ASD toddlers (ranging from 1.8-3.3 years) showing differences in FA that indicate an increase in white matter maturity compared to typical developing toddlers. Increased FA values were measured in several areas, including the genu and
25
splenium of the corpus callosum, the left posterior limb of internal capsule, external capsule and forceps minor (Ben Bashat et al. 2007). In addition a significant lower level of FA was found only in the cortico-spinal tracts, predominantly on the left side. This supports the previously reported white matter overgrowth seen in young children with ASD. Although this finding is inconsistent in older children, where predominantly reduced FA is found in several areas of the brain, it does support the structural volumetric MRI findings showing increased white matter volumes in young children with autism followed by decreased white matter volume in autistic adolescence compared to controls (Chung et al. 2004; Waiter et al. 2005; Courchesne et al. 2001).
4.2.2 Brain regions with abnormalities found with DTI in ASD
A study investigating white matter integrity in autistic children, found specific regions with reduced FA compared to controls (Barnea-Goraly 2004). These regions included white matter (in the region of) the corpus callosum (important for effective bihemispheric communication), the fusiform gyrus and superior temporal sulcus (both important for the processing of socially relevant material like faces), and the ventromedial prefrontal cortices, anterior cingulate gyri, temporoparietal junctions and the amygdala (all regions implicated in theory of mind processing or the awareness of mental states and emotions). Some of these findings where replicated in a recent DTI study investigating a mixed sample of high-functioning autism and Asperger syndrome (diagnosed as ASD) children, which found widespread white matter alterations in children with high-functioning ASD especially in areas implicated in social cognition and information integration (Noriuchi et al. 2010). They reported significantly lower FA values in regions including white matter around the left dorsolateral prefrontal cortex, the amygdala, superior longitudinal fasciculus, anterior corpus callosum and cingulate cortex. Reduced FA in similar regions was found in studies investigating groups with both children and adolescence with ASD compared to controls: the superior temporal gyrus and stem (Lee et al. 2007), the internal and external capsule (Keller et al. 2007; Brito et al. 2009; Cheung et al. 2009), frontal regions (Sundaram et al. 2008) and areas around and within the corpus callosum (Alexander et al. 2007;
26
Keller et al. 2007) were found to have significantly lower FA values in ASD compared to healthy controls.
4.2.3 Conclusion
These results add to the general idea that information processing is altered in the ASD brain and that the connections between grey matter regions implicated in ASD are abnormal and differently organised. However, most studies combine different ASD diagnostic subtypes, leading to a heterogeneous patient group suggested to have different underlying neurobiology causing their clinical characteristics. Therefore the investigation of white matter integrity in homogeneous populations is suggested to get a better understanding of the etiology and neurobiological differences between the diagnostic subgroup in the Autism Spectrum Disorder.
4.3 DTI findings in Asperger syndrome
So far there have only been a few studies investigating white matter connectivity in Asperger syndrome (Bloemen et al. 2010; Pugliese et al. 2009; Catani et al. 2008; Beacher et al. 2012). Additionally there have been studies investigating mixed groups of children and/or adults with autism and Asperger syndrome, compared to controls. Since there are only four studies that have been investigating separate Asperger groups, the main findings will be summarized per brain region and possible similarities or differences with DTI findings in the mixed groups will be outlined.
4.3.5 Methodology
Different DTI methodology has been used to identify differences in the microstructural integrity of white matter tracts connecting brain areas implicated in Asperger syndrome. Two recent studies used the combined approach of voxel-based-morphometry (VBM) and diffusion tensor imaging to analyse white matter connectivity in the Asperger brain (Bloemen et al. 2010; Beacher et al. 2012).
27
Altough VBM analyses do not give specific information on tracts or fibers, it is possible to indicate regions with possible altered tracts. Bloemen and colleagues investigated white matter integrity in adults with Asperger syndrome versus healthy controls using VBM-DTI to measure FA, MD and RD throughout the whole brain (Bloemen et al. 2010). The (very recent) study by Beacher and colleagues focused on sex differences in brain structure between males and females (adults) with Asperger and healthy controls (Beacher et al. 2012). With VBM-DTI they measured FA and MD more specifically using manually drawn regions of interest (ROI) brain volume with structural MRI, and MD and FA in specific regions of interest (ROI) using VBM-DTI. A different fairly new approach besides VBM-DTI is tractography applied to DTI (DTI-tractography) which allows a more detailed investigation of white matter integrity than the VBM approach (Basser et al. 2000). It is the only technique that can quantify both white matter volume and microstructural integrity of specific fibre tracts in vivo, whereas VBA-DTI allows for whole-brain or ROI analysis of white matter integrity, instead of specific white matter tracts. The DTI-tractography approach has recently been applied in two studies focusing on the Asperger brain anatomy in contrast to healthy controls. Pugliese and colleagues used DTI-tractography to measure the length of streamlines (SL) and both FA and MD of white matter tracts in the limbic pathways (figure 9) in children and adults with Asperger syndrome and compared them with controls (Pugliese et al. 2009). DTI tractography was previously also used by Catani and colleagues to investigate specifically the cerebellar feedback projections with volumetric, MD and FA measures, in male adults with Asperger syndrome (Catani et al. 2008).
Figure 9 Limbic pathways containing association and projection fibers, visualized with diffusion tensor imaging tractography (adapted from Pugliese 2009).
28
4.3.2 Frontal, temporal, parietal and occipital lobe white matter abnormalities
In male adults with Asperger syndrome decreased frontal white matter integrity was found with the VBA whole brain approach, reflected by reduced FA in frontal regions (figure 10) that contained the UNC (uncinate fasciculi), IFO (inferior fronto occipital fasciculi) and ATR (anterior thalamic radiation) (Bloemen et al. 2010). Decreased FA in the right UNC and bilateral IFO was also found in the DTI tractography study, in boys and males with Asperger syndrome and additionally increased MD in the IFO and increased number of SL in the UNC was reported in the same study (Pugliese et al. 2009). However all these frontal abnormalities were not significant after correction for multiple comparisons (Pugliese et al. 2009).
Besides frontal region findings, significant white matter differences in the internal capsule, temporal lobe and perisylvian areas, were found in adults with Asperger syndrome (Bloemen et al. 2010). Bilateral reduction in FA was found in regions containing the IFO, ILF and SLF (inferior and superior longitudinal fasciculi). Furthermore under-connectivity in the parietal and occipital lobe, represented by reduced FA in regions containing the IFO, ILF and SLF, was found in male adults with Asperger syndrome compared to controls. Moreover in some areas in the frontal and temporal lobe (containing the UNC, IFO, ILF and CG), reduced FA was overlapped by significantly higher RD in Asperger’s (Bloemen et al. 2010).
Figure 10 Clusters of decreased FA in people with Asperger syndrome compared to healthy controls. (Radiological convention; image left = subject’s right). Slices are 2 mm and Z-coordinates range from -24 to + 54 (Adapted from Bloemen et al., 2010).
29
These findings are consistent with studies investigating combined groups of Asperger’s and autism subjects, and ASD in general. Abnormal frontal white matter connectivity has consistently been found in studies investigating Asperger syndrome as part of ASD (including besides Asperger’s also autism and/or PDD-NOS subjects) (Sundaram et al. 2008; Shukla, Keehn & Müller 2011; Shukla et al. 2011; Noriuchi et al. 2010; Cheng et al. 2010). Specifically, FA was shown to be lower in frontal short range fibres in children with ASD (mixed samples including Asperger’s, autism and PDD-NOS) compared to typically developing children (Sundaram et al. 2008). In addition reduced FA in white matter around the left dorsolateral prefrontal cortex (DLPFC) in a mixed group of children with HFA and Asperger syndrome compared to healthy controls (Noriuchi et al. 2010). Moreover decreased FA and increased MD and RD was found in the SLF, ILF and IFO in a combined group of children and adolescents with autism and Asperger syndrome compared to typical developing participants (Shukla et al., 2011). Inconsistent with previous studies, Cheng and colleagues found greater FA and reduced RD in frontal regions in adults with ASD (12 Asperger’s, 11 autism, 2 PDD-NOS), instead of reduced FA and increased RD (Cheng et al. 2010). Although the authors state that the FA imbalance observed in ASD may be the result of developmental aberrance of white matter skeletons, this could also be due to the heterogeneous population of different ASD subgroups included in the studies. In another study investigating a mixed group of ASD subtypes, children with ASD (11 Asperger syndrome and 15 autism) had compromised short distance tracts in the frontal, temporal and parietal lobe (reduced FA and increased MD and RD) compared to typically developing children (Shukla et al. 2011). Interestingly there was no significant difference in (lobe-specific) short- or long-distance tracts between children with autism and Asperger’s (ASD subtypes).
Frontal disconnectivity is suggested to be associated with several behavioural characteristics of Asperger syndrome, including problems with executive functioning, attentional processes, emotion and theory of mind (Courchesne & Pierce, 2005; Happe, 1996; Kandel et al., 2000). The temporo-parietal areas where reduced white matter integrity was found, have previously been indicated to be important for emotional processing, language and social cognition and found to be implicated in ASD (Barnea-Goraly 2004; Keller et al. 2007; Just et al. 2004). In addition the
30
occipital lobe, were abnormalities where also present, is involved in the processing of facial and emotionally expression (Deeley 2007). Abnormalities in the regions found to have abnormal white matter integrity, can therefore be associated with connectivity problems between brain regions implicated in Asperger’s and ASD symptoms.
4.3.3 Cingulum abnormalities
Several studies in separate Asperger groups reported abnormal white matter integrity in the cingulum (Bloemen et al. 2010; Pugliese et al. 2009; Beacher et al. 2012), an association fibre that runs within the cingulate gyrus around the corpus callosum and part of the limbic pathway. In males with Asperger syndrome, lower FA combined with higher RD was found in regions (Cuneus, Cingulate gyrus) including parts of the cingulum, compared to controls (Bloemen et al., 2010). In addition increased MD and number of SL in the right cingulum was reported in a study investigating Asperger males compared to controls (Pugliese et al., 2009) (figure 11). These results are consistent with findings in mixed groups of
ASD including Asperger syndrome (besides autism) were decreased FA and increased MD and RD where found in ASD compared to controls (Shukla et al., 2011).
Figure 11 Differences between Asperger syndrome (ASP=orange) and healthy comparison (bleu) group in the number of streamlines in the Cingulum. * Differences are significant at p<.05 ** p<.01 (Adapted from Pugliese et al., 2009)
31
Figure 12 Correlation between age and number of streamline in the right cingulum for Asperger syndrome (green) and healthy comparison group (blue). There is a medium positive correlation r=.364 between the two variables with a statistical significant p=.018. (Adapted from Pugliese 2009)
Interestingly the white matter integrity of the cingulum has also been associated with age and sex differences between Asperger’s and controls (Pugliese et al. 2009). A strong correlation between age and MD in several regions of the limbic pathway, including the bilateral cingulum, is found in both Asperger’s and controls. Moreover the age-related differences between Asperger’s and controls in the number of SL of the right cingulum indicate a specific overgrowth in childhood in Asperger’s followed by an apparent growth arrest during early adulthood (figure 12). This is in line with the enlarged white matter volumes reported in ASD during early childhood (Herbert 2005). A recent study investigating sex differences in the Asperger brain, found that FA in the bilateral cingulum was significantly higher in males than females (Beacher et.al. 2012). An interaction effect between sex and diagnose was found, showing that the difference between males and females was significant in the control group, but not in the Asperger’s group. There was also a trend towards lower FA in bilateral cingulum in Asperger males compared to control males, but this was not significant. Additional analyses revealed a sex and diagnosis interaction to be predominently present in the anterior portion of the cingulum rather than the posterior part (Beacher et al. 2012).
32
The cingulate gyrus/cortex is important for social behavior, empathic cognition and pain perception, and implicated in ASD with fMRI showing abnormal activation patterns (Di Martino 2009). White matter abnormalitites in the cingulum could therefore be associated with typical Asperger symptoms and are consistent with previous findings in ASD (Shukla, Keehn & Müller 2011; Lo et al. 2011).
4.3.4 Corpus Callosum abnormalities
VBM DTI studies reported altered oragnisazation of the corpus callosum (CC) in Asperger’s compared to controls. Asperger males had lower FA and higher RD in clusters of voxels including the CC compared to controls (Bloemen et al. 2010). The three separate parts of the CC, the genu-body- and splenium, measured with DTI-tractography, also showed lower FA and higher RD in a mixed sample of children with ASD (11 Asperger syndrome and 15 autistim) (Shukla, Keehn & Müller 2011). In the same study, age was found to be positively associated with FA and negatively with MD and RD in the corpus callosum, but this was only significant in typically developing children and not in the ASD group (Shukla, Keehn & Müller 2011). In addition an interaction effect between sex and diagnose was found when comparing FA values of the body of the CC, between males and females in Asperger´s and controls (Beacher et al. 2012). A trend towards lower FA in Asperger males compared to control males was found and higher FA values in the body of the CC were present in males compared to females although this latter finding was only significant in the control group and not in the Asperger´s group.
The results in separate Asperger syndrome DTI studies, regarding the corpus callosum are consistent with DTI (Alexander et al. 2007; Keller et al. 2007; Barnea-Goraly 2004) and MRI - both structural and functional - findings in ASD (Chung et al. 2004; Just et al. 2007; Vidal et al. 2006). For example a 14% reduction in corpus callosum volume was found to be associated with reduced FA in the genu and splenium of the CC in ASD. These findings suggest abnormal CC neuroanatomy in Asperger’s and ASD in general, resulting in lower interhemispheric connectivity. This reduced connection combined with problems in information integration and lower IQ are
33
thought to lead to specific emotional and social deficits implicated in Asperger syndrome (Paul et al. 2007).
4.3.5 Cerebellum abnormalities
A DTI-tractography study in males revealed reduced cerebellar FA values in Asperger´s compared to controls (Catani et al. 2008). Specificly reductions in FA in the right superior cerebellar peduncle an the right short intracerebellar fibres was found in people with Asperger syndrome. FA in another white matter region in the cerebellum, the left superior cerebellar peduncle, was interestingly associated with social behaviour as assessed by the ADI (autism diagnostic interview). In contrast to these results, lower RD was found in Asperger males compared to controls, in clusters of voxels located in the left cerebellum, possibly containg the superior-, middle- and inferior cerebellar peduncle (Bloemen et al. 2010). These contradictory findings are also present in MRI studies in ASD, finding inconsistent results regarding the cerebellum .
The role of the cerebellum in Asperger’s and ASD in general is not yet clear and futher research is necesarry to replicate and confirm the white matter abnormalities previously found. However suggestions have been made about the implications of alteration in the cerebellar pathways from and to the cerebral cortex. The white matter abnormalities in the outflow pathway could for instance influence the cerebellar feedback to the cortex necessary for the successfull adaptation of social behavior (Catani et al. 2008).
4.3.6 Local over-connectivity and reduced global connectivity
Most DTI findings in Asperger syndrome add to the increasing evidence that people with ASD have abnormal local and global white matter tracts (figure 13). Local over-connectivity contrasted with reduced long range connections are thought to be associated with ASD (Herbert 2005). Long range white matter tracts, including the corpus callosum and cingulum, are found to have reduced FA which was, at least in the case of the CC, associated with reduced volume (see 4.3.3 and 4.3.4). The assumption can therefore be made that in Asperger’s reduced global (long range) connectivity
34
is present. In addition widespread abnormalities in local (short rage) connections is also present in Asperger syndrome, including compromised tracts in the frontal, temporal, parietal and occipital lobes (see 4.3.2). This is party in line with studies in ASD. There has been one study in children with autism and Asperger syndrome showing predominantly abnormalities in frontal short range fibres rather than long range tracts (Sundaram et al. 2008). However another study did support the theory of both short and long range abnormalities. In this recent DTI study in a combined ASD group of children with autism and Asperger syndrome, lower FA in long-distance tracts was found in ASD compared to the typical developing group and compromised short-distance tracts compared to controls in the frontal lobe, where found represented in lower FA and higher MD and radial diffusivity (Shukla et al., 2011). Significant abnormalities of short distance white matter tracts was also found in the temporal and parietal lobe, reflected in increased MD and radial diffusivity (Shukla et al. 2011). Interestingly there was no significant difference in the microstructural abnormalities in short distance tracts between the autism and Asperger syndrome group.
Figure 13 Short-distance (red) and long-distance tracts (blue) obtained from the mean FA skeleton (green), overlaid on the standard brain (MNI) template. (Shukla et al 2011 supplementary figure)
35
4.3.7 Summary
In summary these preliminary findings suggest that people with Asperger syndrome have widespread differences from controls in the anatomy and microstructural integrity of white matter tracts throughout their brain both in short- and in long range tracts. In general, decreased FA along with increased MD, RD and number of SL, has been found in several brain regions in Asperger syndrome compared to healthy controls. Brain areas showing abnormalities in white matter integrity in Asperger syndrome include frontal- parietal-, temporal- and occipital lobe (Bloemen et al. 2010), and in addition more specifically, the internal capsule (Bloemen et al. 2010), the cingulum (Bloemen et al. 2010; Pugliese et al. 2009; Beacher et al. 2012), the body of the corpus callosum (Bloemen et al. 2010; Beacher et al. 2012) and the right cerebellar region (Catani et al. 2008). These findings are consistent with DTI results in ASD and studies in separate autism groups (Alexander et al. 2007; Barnea-Goraly et al., 2004; Cheng et al., 2010; Keller et al. 2007; Noriuchi et al., 2010; Shukla et al., 2011; Sundaram et al., 2008)).
36
5. Implications and conclusions
5.1 Neurobiological correlates with white matter abnormalities
The exact underlying mechanism resulting in lower FA and higher MD, RD and number of SL found in several brain areas in Asperger syndrome, is still unclear. However there are some suggestions to make based on postmortem investigations of white matter abnormalities, risk factors in genes encoding proteins involved in neuronal development and results of other biochemical correlates in Asperger’s disorder measured with different imaging techniques.
5.1.1 Neuronal tissue abnormalities underlying DTI measures
There are different biological factors, e.g. the degree of myelination, volume of tracts, and number of axons that could be associated with DTI measures. It is suggested that in a single fiber bundle, lower FA and higher MD might reflect more isotropic diffusion of water molecules due to less restricted movement possibilities (Alexander 2007). Abnormalities in RD suggest underlying problems of the myelin component of white matter, rather than axial diffusivity problems / axonal degeneration. An increased RD, as found in Asperger syndrome, has previously been associated with increased demyelination (in (e.g.) the corpus callosum of mouse brains) (Song et al., 2005). Reduced FA and higher MD and RD, found in Asperger’s compared to controls, could therefore imply lower than average myelination or higher than average disorganization of fibers in several brain areas, which can result in impaired communication between regions important for functions implicated in Asperger’s characteristics (Le Bihan 2003).
5.1.2 White matter development and post-mortem findings
It is suggested that abnormal white matter integrity visualized with DTI is the result of aberrant maturation of neurons, including abnormal neuronal growth, migration and myelination (Suzuki et al. 2003). Research in ASD described an association between the onset of myelination during brain development and the white matter volume enlargements found per region (Herbert et al.
37
2004). The later myelination in a white matter region finishes and the longer it took, the greater was that region’s volume increase. The frontal and temporal pole have a late myelination onset combined with a long myelination period compared to, for instance, the cingulum. These regions are all found with DTI to have abnormal white matter integrity in Asperger syndrome.
Enlarged white matter volumes in the ASD brain is suggested to be the result of an increased density of cortical mini-columns requiring more short range association fibers (Manuel F Casanova 2004). Evidence from ASD postmortem studies, showing this increased density of minicolumns in the frontal brain regions (Casanova 2002) and poor differentiation of grey-white matter boundaries (Baily 1998), is consistent with the widespread white matter abnormalities found with DTI in Asperger syndrome. This evidence implies that there should be a greater number of neurons in frontal regions, which could result in complex neuronal organization reflected by differences in white matter microstructure between Asperger’s and controls.
One of the few post-mortem Asperger studies compared two Asperger cases with control brains and reported abnormalities in the minicolumnar organization (smaller minicolumns and the neurons were more dispersed than normal) in lamina 3 cortical areas 9,21 and 22 in the right hemisphere (Casanova et al. 2002). In addition other post mortem studies in ASD showed 79% decrease in density of Purkinje cells in the cerebellum and intense cerebellar neuroinflamation extending to white matter, which could possibly relate to the abnormal white matter integrity found in this region in Asperger syndrome (Vegas et al., 2005). This is consistent with post-mortem ASD evidence of malformations in the cerebellum (Bailey 1998, Bauman 1985, 2003).
5.1.3 Genetics of Asperger syndrome and white matter development
A strong family history of Asperger is present among first-degree relatives and several family and twin studies indicate that ASD has a strong genetic basis (Volkmar et al. 1998; Ghaziuddin 2005; Bailey et al. 1995; Folstein & Rutter 1988). Asperger syndrome is therefore thought to be highly
genetic. Moreover, similar regions of reduced FA were found in children with autism and their unaffected siblings compared to controls, suggesting a genetic risk factor underlying this
38
endophenotype (Barnea-Goraly et al. 2010). Therefore the white matter abnormalities found in Asperger syndrome could be the result of specific genotypes, causing differences in protein expression, which could influence neuronal migration and myelin formation. Interestingly, characteristic Asperger traits were already reported by Hans Asperger to be present in family members of probands, particularly in fathers, suggesting male specific genes to be probably effected (Asperger 1944). Although there are no specific risk genes for Asperger identified so far, there are some suggestions possible based on genetic studies, which reported specific chromosomal regions associated with Asperger syndrome.
There have only been a few studies investigating genetic risk factors specifically in families of individuals with Asperger syndrome. The first GWAS (genome-wide association study) in families with a history of Asperger syndrome, revealed nine chromosomal regions with susceptibility loci associated with Asperger syndrome (Ylisaukko-oja et al. 2004). Three regions showed highest significant associations, including two (1q21-22 and 3p14-24) overlapping with classic autism susceptibility loci and one with schizophrenia (13q31-33). The overlap does not necessary mean shared risk genes, but rather implies that these regions probably includes genes associated with higher risk for specific traits effected in these neuropsychiatric disorders. The loci on chromosome 1 (1q21-22) is independently validated in a candidate gene study, finding a strong association with the NTRK1 gene situated in this region (Chakrabarti et al. 2009). This candidate gene study, showed that Asperger syndrome was significantly associated with fourteen genes related to social-emotional behavior, empathy, sex steroids and – interestingly - with neural growth. These are candidate genes to influence white matter integrity in the brain. Some of these genes are interesting candidates to be associated with white matter abnormalities since they were functionally related to neuronal development and connectivity, including IGF1, NTF3, ARNT2, NTRIK1 and NTRK3, all encoding proteins involved in neuronal survival, differentiation, growth and neuronal migration (possibly influencing white matter integrity found to be abnormal in Asperger syndrome with DTI). The last three genes were, besides Asperger, also associated with AQ and/or EQ, scores measuring autistic traits in a normal population sample (Baron-Cohen et al. 2001; Baron-Cohen & Wheelwright 2004).
39
In addition, Asperger syndrome has been associated with different DNA variations (SNP’s) in the DISC1 gene (1q42), encoding a protein involved in several mechanism including neuronal migration and outgrowth (Kilpinen et al. 2008). Moreover a boy with Asperger syndrome was found to have a partial trisomy in chromosome 22, resulting in an additional copy of 22q13 and the SHANK3 gene, important for the maturation and formation of neurons (Durand et al. 2007). Interestingly, SHANK3 is the binding partner of neuroligins; and the X-linked NLGN3 (Xp22.32-p.22.31) member of the neuroligins family, is found to be mutated in Asperger syndrome (Jamain et al. 2003). DISC1, SHANK3 and NLGN3 are also found to be associated with classic autism and mutations in these genes are thought to increase the risk for developing ASD endophenotypes like white matter abnormalities in the brain (Abrahams & Geschwind 2008).
In conclusion, results from genetic studies suggest that predisposing risk factors for Asperger syndrome, encoding proteins important for neuronal migration and connectivity, could possibly influence white matter development in Asperger brain. This could lead to the abnormal white matter integrity found in several brain areas, resulting in problems in signal communication between brain regions implicated in Asperger’s symptoms.
5.1.4 MRS / SPECT / PET
Biochemical correlates could also possibly influence or be related to white matter integrity in Asperger syndrome. Different imaging techniques have previously been used to identify these specific correlates in the Asperger’s brain, including in vivo (proton) magnetic resonance spectroscopy ((H)-MRS), positron emission tomography (PET) and single photon emission computed tomography (SPECT) (Girgis et al., 2011; Kleinhans et al., 2009; Murphy, Critchley, & Schmitz, 2002; Murphy et al., 2006; O’Brien et al., 2010). The DTI findings in frontal regions showing white matter abnormalities is supported by MRS studies measuring neuronal integrity represented by chemical metabolites in Asperger syndrome.
