Introduction
 Autism spectrum disorder (ASD) is a heterogeneous and persistent neurodevelopment disease that affects 1 in 150 children (1,2). It is associated with various symptoms: 1) severe and sustained impairment in social interaction, 2) reduced and impaired communication abilities, and 3) restricted and/or stereotyped patterns of behavior and interest. These behaviors typically present by the time the child reaches the age of 3 years old (3,4).
There is a general agreement that children who have autism tend to exhibit signs of developmental delay within the first 18 months. However, 25-45% of children with autism initially demonstrate near-normal development until 18-24 months or beyond (2). On the whole, the disease emerges prior to the point at which children reach the age of 3; however, clinical presentation of the disorder is complex and can vary from one patient to the next (3,5).
Although the causes of autism remain relatively unknown, some researchers have suggested that genetic, developmental, and environmental factors could contribute to the development of autism (4). While the exact etiopathological mechanisms of ASD are still not comprehensively understood, a variety of data indicates that the condition can be attributed to disturbances in brain growth and maturation as a significant pathomechanism (3).
Over the course of the last three decades, several magnetic resonance imaging (MRI) studies have examined the brain anatomy of patients with autism in order to identify structural abnormalities and enlargement in some parts of the brain. Through these studies, frontal lobes, amygdala, and cerebellum were highlighted as pathological in autism; as such, these areas of the brain have attracted significant attention in studies that investigate the morphology of the brain (2-4). MRI is considered to represent the gold-standard technique for the identification of structural brain alterations in patients with neuropsychiatric disorders such as autism, and it is widely regarded as a valuable method through which a neuroanatomic model of pathophysiology can be developed, and ultimately, effective therapeutic interventions can be identified (4). Through the use of imaging technology, researchers have observed that the overall brain weight of children (2- 4.5 years) with autism is statically heavier than that of age and gender-matched controls (6). Furthermore, there is also evidence from child and adolescent samples to suggest that much can be learned by examining the impact of early variation in the trajectories of anatomical brain development and their influence on later outcomes (9).
While a great deal of research and discussion has focused on whether the brain size of people with autism is abnormal, the majority of research studies have not mentioned age-related changes in brain development. Those that do, have raised the possibility that the brain of younger children with ASD is enlarged (5). Although there is a common perception that the brain development of children with ASD is abnormal, there is little evidence of this disorder in the published literature (13).
In a word, a further understanding of the brain size and development of children with ASD during the first five years of life is of utmost importance in supporting early clinical diagnosis and developing early behavioral interventions. To contribute to knowledge in this area, this study aimed to review and summarize existing morphometric brain investigations that employed MRI with patients with autism as a means of examining the structural brain anatomy and development of autistic children who were aged 1 to 5 years.

Methods
To review the literature, we followed the PRISMA guidelines to query the PubMed and Scopus databases using the following search terms: (magnetic resonance imaging OR MRI) AND ((Autism spectrum disorder (ASD) OR Autism) AND brain structure) AND (magnetic resonance imaging” OR MRI AND “Autism spectrum disorder” (ASD) OR Autism AND brain size) AND (brain MRI AND Autism AND “brain volume”).
The last search was conducted between the 21st and 26th March 2015. The title and abstracts of all identified articles were studied and then irrelevant articles were omitted. The full text of the remaining articles was then reviewed, and any articles that were deemed to be irrelevant in terms of the objectives of this systematic review were also omitted. The reference list of the relevant articles was screened to detect any further articles that had not been identified during the database search process.
There were no constraints in terms of the dates of the research papers employed within the research; however, only English-language articles were considered as part of the systematic review. Reviews were not included in the selection, but were used to identify any additional original studies. We searched all the selected papers to include studies in which the mean age of the ASD sample was below five years of age. Studies that focused only on ASD subjects with known etiology (e.g., Fragile-X, syndrome, tuberous sclerosis) were not included. Any studies that involved less than 10 patients and all case reports were excluded.
Data about the authors, publication data, country, patient characteristics (age, gender, duration of the diseases), and the results of the MRI were extracted.
The initial search strategy yielded a total of 757 records. After the removal of duplicates, 63 remained. Of these, only 10 met the inclusion criteria and were case-control studies. The search procedure was summarized in a PRISMA Chart and can be seen in Figure 1.

Data synthesis
The evidence table presented in Table 1 presents a summary of the data that was extracted from the eligible studies. The results were organized according to the data that provided an association between the incidence of ASD patients and MRI study parts, and information on sample demographics, authors, and the brain size are reported in Table 1.

Results
Based on the methodology, a total of 757 articles were obtained in our initial search. After deleting the duplicate and irrelevant articles, only 10 studies remained that were deemed to be relevant to our inclusion criteria.
The population size of the included studies ranged between 492 and 338 ASD/control and the mean age of the patients and controls ranged between 1 and 5 years. Overall, 492 ASD patients were considered in this systematic review. Their brain size and volume as measured by MRI are presented in Table 1.

Discussion
Over the last few decades, structural MRI studies in children have increasingly been used to identify relations between brain size and behavioral brain abnormalities like autism (9). In this regard, several studies have found that children with autism exhibit an enlarged brain during childhood (7,8). Within the current study, all 10 selected articles, with the exception of one, demonstrated that the vast majority of ASD patients exhibited an enlarged brain size; however, the extent of this enlargement varied in different studies and for different parts of the brain. Therefore, further studies are required to develop further insights into the areas of the brain that are most likely to be enlarged in the case of children under the age of five who have ASD. In addition, further analysis of the use of MRI as a non-invasive diagnostic technique for diagnosing ASD is also worth further attention.
Several MRI studies involving very young children with autism report a 5-10% abnormal enlargement in total brain volume that persists into early childhood (10-12). Similarly, Hazlett et al. (2005) and Courchesne et al. (2001) found that children with autism exhibit enlargement of the whole brain size (25,10). On the other hand, a study by Hashimoto et al. (1999) found that the whole brain size of children with ASD who were aged between 3-5 years old was smaller than that of the healthy controls (27). To the best of our knowledge, that particular study is the only one to correlate smaller brain size with ASD children in this age range.
The timing of brain enlargement in autism is of particular importance given new evidence from prospective behavioral studies involving ASD infants (15). Some researchers have concluded that an abnormal accelerated rate of growth may serve as an early warning signal that there is a risk for autism (16). In light of this important issue, the study population involved in this review was limited to children who were under the age of 5 years old. Nevertheless, it is overwhelmingly accepted that the brain size of children with autism is slightly reduced at birth and that it dramatically increases within the 4-5 first years of life before eventually reaching a plateau by adulthood (14).
Researchers consistently agree that people with ASD exhibit an accelerated brain volume growth in early childhood of around 10%. This enlargement appears to peak at around 2-4 years of age (18). Ninety percent of 2- and 3-year-old children have brain volumes that are larger than the healthy average (16). Consequently, based on the importance of the first five years of life in terms of brain development as a potential indication of ASD, we selected children aged from 1-5 years as the inclusion criteria for the study selection. Thus, at the age of typical clinical diagnosis of the disorder (i.e., 3-4 years old), the period of pathological growth and arrest has likely already passed, leaving clinicians and researchers with an outcome, rather than a process of pathology by which treatment intervention can be studied (14).
The amygdale plays an important role in language ability, and this region of the brain has heterogeneous structure and function. Moreover, it is intimately connected with the hypothalamus and has dense projections to the brain stem, the frontal and temporal lobes, and to the thalamus (9). As such, the structural evaluation of this region has importance through the diagnostic procedure and, as per the findings of Sparks et al. (2002), Ortiz Mantilla et al, (2010) found that this area of the brain was enlarged in the case of children with ASD who were in this age range (7,9).
As a result, abnormal brain growth was reported in several studies. This abnormal development might be characterized by an increased rate of brain growth from early infancy (2- to 3-year-old) through the preschool period (particularly in the frontal, temporal, parietal lobes, and cerebellum), followed by an abnormally slow cerebral and cerebellar volume increase during late childhood or even puberty and adolescence (19-22). Similarly, the results of the studies by Hazlett et al. (2011), Courchesne et al. (2003), Bloss et al. (2007), and Akshoomoff et al. (2004) that were identified as part of this systematic review, all showed an enlargement in cerebral and cerebellar volume (15,16,28,29).
Additionally, it was found that mostly white matter (WM) increases during childhood and adolescence in the case of children with ASD, while the trajectories of grey matter (GM) volumes follow an inverted U-shaped developmental trajectory. The different developmental curves of WM and GM belie the intimate connections among neurones, glial cells, and myelin, which are fellow components in neural circuits and are linked by lifelong (17). According to our results, Hazeltt et al. (2005), Courchesne et al. (2001), Courchesne (2003) and Bloss et al. (2007) found the WM and GM increased size in ASD children. Similarly, other studies reported that brain size in 4-year-old children with autism exceeded the healthy average, and that excessive brain size was primarily due to increased WM volumes in the cerebellum and cerebrum and increased GM volume in the cerebrum, within which frontal lobes were the most abnormal (10,16,23-25,28).
More broadly, the early and transient period of brain overgrowth and understanding the mechanisms and influences on structural and functional brain development across childhood must be an important factor in the emergence of autistic behavior and may help us to harness the brain’s developmental plasticity to help guide interventions for clinical disorders and to elucidate the path by which optimal development can be promoted (16,17).

Conclusion
On the basis of the existing literature, we can conclude that autism is associated with generalized enlargements of the cerebral hemispheres, the cerebellum, and amygdala. Alternatively, apparent differences in the results of some studies may result from differences in age range selected and image processing. Furthermore, some regions have abnormal developmental trajectories that could point towards particular etiopathological factors and timeframes for possible interventions using trajectories (i.e., morphometric measures by age) because an endophenotype may provide discriminating power where static measures do not. Overall, the findings of the studies included in this systematic review were in agreement with the exception of just one 1991 study, which contradicted the results of the others. Therefore, 9 out of 10 articles demonstrated that the use of MRI and brain imaging can predict autism in children aged between 1-5 years.

Acknowledgement
This study was the result of a medical resident thesis with approval number 930563. It was financially supported by the Research Vice Chancellor of Mashhad University of Medical Sciences.

Conflict of Interest
The authors declare no conflict of interest.