In this longitudinal investigation, we measured developmental changes in brain volume and in CT in a group of children aged 8–12 years with autism and healthy controls. Decrease in grey matter volume was observed over time and appeared to be related, at least partially, to a decrease in CT. However, these findings disappeared when analyses were controlled for FSIQ which could be related to the small sample size. Associations between developmental changes in CT and several clinical features were also observed. Despite the small sample size and the narrow age range, these preliminary findings emphasize that the time course of brain development rather then the final product is most disturbed in autism (29
Over the last two decades several neuroimaging structural investigations have reported an increase in brain size in autism (2
). More recently, additional evidence has emerged pointing to the presence of brain enlargement in individuals with autism and more importantly to the existence of age-related volumetric changes involving total brain as well as specific cerebral lobes (1
). Our findings in the present investigation are consistent with some but not all of these observations. No differences in brain volume were observed here which is inconsistent with most previous studies of brain volume. This discrepancy could be related to the small sample size and/or to the age of participants. However, our result of age-related changes in grey matter volume is concordant with several cross-sectional studies providing indirect evidence of abnormal developmental trajectories in autism (1
In the present investigation, decrease in CT over time was observed in the autism and control groups. This decrease involved several brain regions including the frontal, parietal, and occipital lobes. These observations are consistent with previous investigations in typically developing children, between the ages of 5 and 11 years, reporting cortical thinning over large brain areas such as the right dorsolateral frontal and bilateral occipitoparietal cortices (32
). Interestingly, increase in CT over time was observed here in the temporal lobe in healthy controls but not in the autism group which is concordant with findings of grey matter thickening in left and right posterior perisylvian regions of Wernicke’s area in neurotypical children (32
). These structural alterations in the temporal lobe suggest the existence of an aberrant cortical development in autism and might explain the previously reported functional abnormalities in this structure (33
Several clinical features including social deficits and repetitive behaviors were associated with changes in CT overtime in different brain regions in the autism group. Neuroanatomical measurements in the frontal lobe were associated with social functioning. This observation is consistent with recent evidence implicating the mirror neuron system in autism (18
) and with investigations reporting on abnormalities in the frontal lobe circuitry in this disorder (18
). In the present study, relationships were also observed between repetitive/stereotyped patterns of behaviors and changes in CT the temporal lobe in the autism group. This association is concordant with a structural MRI study which found partial correlations between repetitive behavior measures and volumes of multiple regions including those in the temporal lobe (36
). This relationship is also supported by recent evidence from neuroimaging studies examining individuals with obsessive-compulsive disorder, which is characterized by repetitive/ritualistic behaviors, and reporting on structural abnormalities in the temporal lobe (37
). Finally, all the correlations described above appear to indicate that excessive thinning of the cortex was associated with more severe symptomatology independent of the clinical feature and the brain region involved. Elevated scores on the ADI-R subdomains correlated with more decreases in CT. These observations suggest that changes in CT over time might serve as an indicator of illness severity, but further studies are definitely warranted.
Our preliminary observations point to an excessive decrease in grey matter volume and CT in individuals with autism during a period of development when competitive elimination of synapses and dendritic/axonal arborization represent the main neuronal structural activity. These findings should be examined in light of the recent evidence suggesting that changes in grey matter volumes, may, at least in part, reflect intracortical myelination and the resulting partial volume effect (39
). Changes in CT overtime observed in the present investigation could reflect alterations in grey/white matter boundary related to myelination and changes in the cortical mantle itself. Hence, increase in white matter volume could lead to cortical thinning as a consequence of increased myelination between grey and white matter in the periphery of the cerebrum. However, in the present investigation no differences in white matter volumes in the two groups were observed at baseline, at follow-up and over time. This observation suggest that the excessive decrease in CT observed in the autism group is related to decrease in grey matter possibly related to alterations of the cortical grey matter during the peri-adolescence period. An excessive decrease in any of the different elements of the cortical mantle (cell bodies, neuropil, and synapses of neurons) in isolation or in combination could lead to the findings observed here. Consequently, neuropathologic studies of individuals with a wide age range are therefore warranted to assess the basic histologic features of the cortex in autism in different age groups to determine the exact ultrastructural underpinnings of these alterations. In the absence of white matter volume changes, these alterations suggest the existence of alterations intrinsic to grey matter in autism during the peri-pubertal period and could be related to synaptic development and maturation. This observation is supported by evidence from genetic studies reporting a link between several genes implicated in synapse formation and dendritic spine maturation, such as SHANK3, and autism (40
The neurobiologic underpinnings of cortical changes observed in the present investigations remains to be elucidated. Previous research has implicated several neurotrophins, including brain-derived neurotrophic factor (BDNF), in neuronal development (41
). Several studies have reported on abnormal serum levels of BDNF in individuals with autism throughout the lifespan (42
). Interestingly, age-related differences have been reported with a recent study describing lower levels of BDNF in the serum of children with autism under 9 years of age when compared to adolescents, adults and control subjects (45
). These developmental changes might be related to the altered peri-pubertal cortical organization observed in the present study. Additionally, the serotoninergic system has have also been implicated in autism (46
). This system might be playing a role in the abnormal brain development in autism either directly or indirectly by regulating neuronal development signaling factors for neurotrophins (47
). However, while mounting evidence is emerging linking abnormalities in neurotrophins and serotonin to autism, it is not clear whether these alterations are the primary cause or secondary to the altered cortical changes or just an associated process. Finally, structural and neurobiologic abnormalities should also be examined in light of the fundamental role that genetic and environmental factors play in influencing brain structure and function (48
). This is particularly true for autism where recent evidence have been reported linking gene polymorphisms to cortical enlargement (49
) and grey matter overgrowth (50
Findings from this preliminary study should be examined in light of several methodological limitations. First, lower functioning individuals and females were not included in this study which might limit the generalizability of its findings. Future investigations of this nature will have to address possible gender related differences in the development of cortical structures and associated severity of clinical features. Second, only two MRI scans were obtained from each participant and additional scans obtained over time would have provided additional data point to allow more accurate conclusions. Third, the exclusion of participants younger than 8 years of age is limiting since most brain development occurs prior to this age. Therefore, final interpretations from the present study should not be made before findings from studies of very young children are completed. Fourth, the morphometric program utilized generated global and lobar measures of volume and CT. The lack of topographic specificity limits the interpretations of the findings and warrant future studies applying more advanced programs to allow the examination of structural alterations at the gyral or sulcal level or on a voxel-by-voxel basis. Finally, adjustment of the p-value might be needed in light of the number of comparisons conducted.
In summary, this longitudinal MRI study is to our knowledge the first investigation to examine longitudinal changes in brain volume and CT in children with autism. Developmental structural changes were observed and relationships were found with clinical measurements. This investigation provides preliminary evidence of the age-related changes in grey matter volume and CT during the crucial peri-pubertal period of development in males with autism. However the neurobiologic underpinnings of these changes and the genetic control of these mechanisms remain to be elucidated. Future longitudinal studies applying multimodal imaging techniques while examining genetic influences by assessing gene polymorphisms are needed to evaluate developmental trajectories across the lifespan in this disorder.