Neurodevelopment Research Paper


It is well recognised that there are substantial differences in the amount of research activity concerned with different medical conditions, but the reasons for disparities are not fully understood. Within the field of neurology, Al-Shahi and colleagues found that the amount of research into a condition was not proportionate to the frequency of the disease [1]. Common conditions such as stroke or migraine were under-researched relative to rare diseases such as Wilson's disease or Creutzfeldt-Jakob disease, when assessed using ‘publication ratio’, a measure of number of publications divided by disease frequency. They suggested that as well as political, economic and scientific interest, research activity might also be influenced by the fashionable nature of some conditions, and recommended that similar analyses should be conducted in other areas. However, their analysis did not take into account the severity of disease, which might be expected to play some part in research priorities. A review of grants awarded by the National Institutes of Health [2] confirmed that research funding was not related to how common a disease was, but it was related to disease burden, as indicated by mortality, years of life lost, and disability-adjusted life years. Thus severity of the impact of a condition seemed to play a larger role in determining research priorities than disease frequency.

The field of neurodevelopmental disorders is a microcosm in which these issues can be further explored. The term neurodevelopmental disorder is used in two ways [3]. The first refers to conditions affecting children's neurological development with a known genetic or acquired etiology, such as Fragile X syndrome or fetal alcohol syndrome. The second use refers to conditions of presumed multifactorial etiology in which certain aspects of neurodevelopment are selectively impaired; this includes such conditions as autistic disorder, developmental dyslexia, and attention-deficit hyperactivity disorder (ADHD). Both types of condition were included in the current analysis. In this article, counts are provided for publications in this field to consider how far the amount of a research on a condition can be explained by (a) frequency of disorder, (b) severity of disorder or (c) the disciplines involved in research. In addition, research activity is evaluated in terms of funding by the National Institutes of Health (NIH) relating to these conditions.


Conditions included in the analysis

The conditions included in this review are shown in Table 1. These were identified from Rutter's Textbook of Child and Adolescent Psychiatry, 5th edition [4] and a review of behavioural phenotypes [5]. Those from the latter source were included only if an estimate of prevalence was available. The category of ‘intellectual disability’ (ID) posed problems, because this can be both a symptom of a known disorder, and a nonsyndromal condition of unknown etiology. Furthermore, in the UK, the term ‘learning disability’ is used to refer to intellectual disability, whereas elsewhere ‘learning disability’ is used for specific difficulties in a child of normal IQ. Despite these problems, it was decided to attempt an estimate of publications focusing on ID, but data on this condition need to be treated with particular caution.

Number of people affected

Estimates of prevalence per 100 were computed from Rutter et al. [4] or Udwin and Dennis [5], with the mean being used if a range was given.

Severity of condition

Measures of disease burden used in mainstream medicine focus mainly on mortality and morbidity and are not in general appropriate for neurodevelopmental disorders. It was therefore necessary to derive an ad hoc measure for this study. This was a 4-point scale representing the extent to which an affected individual could be expected to obtain educational qualifications and live independently in adulthood (see Table 2). Nine expert clinicians who saw children with neurodevelopmental disorders were asked to estimate severity for conditions shown in Table 1 on this scale. For some conditions, this scale was difficult to apply because of the wide variation in severity. For instance, a person with autistic disorder is unlikely to fall in category 1, but could fall into category 2, 3 or 4. In such a case, raters were asked to give their best estimate of the average level of severity. The data in Table 3 shows the average across all raters. For rare conditions, few clinicians had sufficient experience of cases to be able to make a judgement, and in those cases, the author provided a rating based on the description of the phenotype in Udwin and Dennis [5].

Table 3. Prevalence per 100, severity rating, number of publications (1985–2009), rate of increase in publications, % publications with genetic/animal model content, estimated N affected children in the UK, and publication index for each condition, ordered by prevalence.

Estimates of number of publications

Web of Knowledge was the basis for research counts; the search terms used are shown in Table 1; for a paper to be counted, the search term had to appear in the title. This exercise revealed that some disorders lacked consistent terminology and were correspondingly hard to search. Total number of articles for each condition was obtained for 5-year bands from 1985 to 2009. Note that coding of articles was not mutually exclusive, so that if search terms for two conditions were included in the title, the same article would be counted towards toward the total for both conditions. The word ‘Down’ is not a valid search term in Web of Knowledge, so it was necessary to estimate the number of articles on Down syndrome. From a Google scholar search it was established that 9.9% of articles on Down syndrome include ‘trisomy 21’ in the title. Thus the number of articles on Down syndrome in Web of Knowledge was estimated by multiplying the number of articles with ‘trisomy 21’ in the title by 10.1.

Publication index

This was derived in a method similar to that used by Al-Shahi et al [1] by dividing the total number of articles by prevalence. This was scaled so that it could be readily interpreted as a count of the N papers in 25 years per 100 affected cases in a population of 11.56 million, which is an estimate of number of children in the UK [6].

Rate of increase in publications

This measure was derived by computing the slope of the line linking the number of publications in 5-year time bins across the period 1985–2009.

Estimates of amount of genetic/animal research

Preliminary analyses suggested that the amount of research on a condition was partly determined by the disciplines involved; for some conditions, a high proportion of studies involved genetics and/or mouse models. To gain an estimate of how much research on each condition was of this type, searches were re-run to find the proportion of papers on a condition that had the words ‘gene’, ‘genetic’, ‘mouse’, ‘mice’ or ‘chromosome’ in the topic field.

Amount of funding by the National Institutes of Health (NIH)

Data on NIH funding were obtained using the Research Portfolio Online Reporting Tools (RePORT) ( For each disorder, the database was interrogated to identify funded projects for the period 2000–2010, and the total funding in thousands of dollars was computed, together with the NIH institute that provided the funds. It is important to note that the RePORT software identifies projects on the basis of keywords, and may include projects that do not have a central focus on the disorder in question. This means that, on the one hand, the amount of funding focused on a specific disorder is likely to be over-estimated, and on the other hand, that the same project is likely to be included in counts for more than one disorder. For this analysis, it proved impossible to obtain realistic estimates of funding for studies of intellectual disability, as the search terms from Table 1 identified numerous projects concerned with a wide range of disorders. This category was therefore omitted from this analysis. For the remaining conditions, a measure of rate of increase in funding between 2000–2010 was computed from the slope of the function for funding by year.


Table 3 shows the estimates of prevalence, severity of condition, number of publications, rate of increase in publications, proportion of genetic/animal papers for each condition, and publication index, with disorders ranked by prevalence. Most variables were skewed, and so nonparametric Spearman correlations were used when assessing association between variables.

The first point to note is that the publications count has only a weakly significant relationship to prevalence of a condition (rs = .34, N = 35, p = .046). To explore predictors of publications, the publication index was used, i.e. a measure that reflects the number of publications relative to estimated number of affected individuals. It is evident on inspection that there is a general trend for a higher publication index for the rarer conditions; the Spearman correlation between publication index and prevalence is rs = −.91, N = 35, p<.001. However, when we consider the common conditions, we see that most of them have a relatively low severity index; indeed, there is a significant negative correlation between severity and prevalence, rs = −.61, N = 35, p<.001. Furthermore, the severe, rare conditions are more likely than common, milder conditions to be monogenic disorders, consistent with the fact that they are likely to have a high proportion of publications including terms indicative of genetic studies; the correlation between % genetic papers and prevalence was rs = −.52, N = 35, p = .001. The correlation between % genetic papers and severity was also significant: rs = .41, N = 35, p = .015.

Further analysis was conducted to consider how far the publication index for each condition related to severity and genetic content. A regression analysis was conducted to find the best-fitting function relating severity to publication index, and a log-log relationship gave the best fit, with R = .648, p<.001. Figure 1 shows the scatterplot relating mean log severity to log publication index, together with the line showing predicted log publication index.

Figure 1. Regression of log publication index on log severity, with 95% confidence interval shown with dotted lines.

A constant of 4 is added to log publication index to avoid negative numbers. Abbreviations: ADHD: attention deficit hyperactivity disorder; ASD: autism spectrum disorder; CP: cerebral palsy; DCD: developmental co-ordination disorder; de Lange: Cornelia de Lange syndrome; FraX: fragile X; ID: intellectual disability (shown in brackets to indicate that the publication index is overestimated); NF1: neurofibromatosis type 1; PKU: phenylketonuria; SLI: specific language impairment; T. sclerosis: tuberous sclerosis; VCF: velocardiofacial syndrome.

Adding the proportion of genetic papers as an independent variable led to a significant increase of fit to give R = .728; change in fit: F (1,32) = 7.51 p = .01. Adding prevalence to the model after including severity and genetic index, accounted for a further 24.3% of the variance, change in fit: F (1,31) = 33.24, p<.001, bring the multiple R up to .879.

The rate of increase in publications across the five 5-year age bands is logically independent from the publication index. This measure revealed two clear outliers, autistic spectrum disorder, and ADHD (see Table 3). Autism started from a high baseline, with 1215 publications during 1985–1989, increasing 5-fold by 2005–2009. The increase for ADHD was even more dramatic, from 356 publications in 1985–1989 to 6158 in 2005–2009. In general, rate of increase in publications was positively correlated with N publications in 1985–1989, rs = .35, N = 35, p = .042, but a high initial level of publications did not guarantee rapid growth. For instance, Down syndrome started in 1985–1989 with 2121 publications, more than either ADHD or autism, but this had risen only to 3474 by 2005–2009. It is noteworthy also that Lesch-Nyhan syndrome, which had a remarkably high publication index, showed a slight decrease in number of publications over a 25 year period.

Table 4 shows data on NIH funding per condition by year, and the slope of the function. There are striking parallels with the data on publications. The Spearman correlation between the total N publications from Table 3 and the total funding from Table 4 is .883 (N = 34, p<.0001) and the correlation between the slope indicating increase in publications from 1985 to 2009, and the slope of increase in NIH funding from 2000 to 2010 is .522 (N = 34, p = .002). In a final analysis, the proportion of funding from different NIH institutes was computed for the most common neurodevelopmental disorders. These data are shown in Figure 2.

Figure 2. Proportion of grant income from different NIH institutes for the most common neurodevelopmental disorders.

Abbreviations: NIMH: National Institute of Mental Health; NINDS: National Institute of Neurological Disorders and Stroke; NIDCD: National Institute on Deafness and Other Communication DIsorders; NICHD: Eunice Kennedy Shriver National Institute of Child Health and Human Development.


Before discussing the results, it is important to note that the range of conditions investigated is not comprehensive, and the approach adopted here could be extended to include a wider range of disorders. For instance, psychiatric conditions such as antisocial behavior disorder and schizophrenia were not included here, though a case can be made for treating them as neurodevelopmental disorders [3]. Reliance on a textbook of behavior phenotypes as a source book also meant that many neurological disorders without clear genetic etiology, such as congenital hypothyroidism or spina bifida, were also excluded.

It is also important to note that estimates of all of the main variables in this analysis are imprecise. A frustrating feature of this study was that is was hard to be confident that search terms identified all relevant papers without including papers on other conditions. Subsets of identified papers were scrutinised to ensure that they reflected the intended content, but relevant articles would be missed if they did not include the search terms in the title. For some conditions, the concern was not of missing relevant papers but of using search terms that would yield false positives. Thus, the acronym SLI (used here for specific language impairment) is associated with 16 meanings in Wikipedia, as well as referring to genes in potatoes and worms. Attempts were made to restrict searches to exclude such usages, but it is nevertheless likely that the estimate for N publications is inflated by occasional errors. Another condition that was difficult to search was XXX trisomy, because XXX is used widely with a range of meanings. Conversely, the estimate for velocardiofacial syndrome will be slightly deflated because it was decided to exclude searches for papers with only the acronym VCF in the title, given that this acronym refers to a host of things other than velocardiofacial syndrome, and only a small minority of publications including it in the title were relevant. As noted above, identifying papers on intellectual disability was particularly problematic, and estimates of research on this topic are inflated because of the difficulty excluding articles on specific learning disabilities. It is noteworthy that nevertheless the amount of research on intellectual disability was below the level predicted from severity and prevalence (see Figure 1).

The second key variable in the analysis is prevalence. Estimates were taken from reputable published sources, but are bound to be imprecise for very rare disorders. Furthermore, for some disorders, especially those with multifactorial etiology, prevalence can vary substantially depending on the severity cutoff used. For instance, the prevalence of fetal alcohol syndrome disorder is estimated at 1 per 1000, but this condition is far less commonly identified in the UK than in North America, where a milder phenotype of ‘fetal alcohol spectrum disorder’ is recognised and estimated to be nine times as common [7].

A third key variable is severity, which was estimated according to the criteria from Table 2, averaging ratings across a range of experts. All raters agreed that this was difficult to do for those conditions with variable phenotype, such as tuberous sclerosis or autistic spectrum disorder. For a condition such as intellectual disability, severity will also depend on prevalence, with milder forms being more common than more severe. Because it was in general not possible to determine the severity of cases included in specific studies, average estimates were taken, but results for specific conditions need to take into account unreliability that is inevitable when using these fuzzy estimates. Furthermore, the severity scale was devised to assess how far the condition affected educational attainment and adult independence, but did not take into account physical symptoms that characterise some of the rarer conditions. Thus, the majority of females with Turner syndrome do not have significant intellectual impairment and live independent adult lives, and hence this condition is rated as mild. However, the syndrome involves cardiac problems and infertility, which are not trivial symptoms.

Despite these limitations, the analysis gave a striking replication of the findings on neurological conditions [1], namely that, when prevalence is taken into account, the number of publications on rare conditions is greatly in excess of that for common conditions. At the extremes of the distribution considered here, between 1985 and 2009 there were 428 publications on Lesch-Nyhan syndrome, which affects an estimated 57 children in the UK, compared with 340 publications on speech sound disorder, which affects over 1 million. However, these two extremes suggest a reason why this is so: more research is done on more severe conditions. Speech sound disorder can be marked and persistent, but it is usually a relatively mild condition that is often transient and may have little impact on a child's social or educational prospects, whereas Lesch-Nyhan syndrome has severe neurological and cognitive consequences, including self-mutilation as a striking symptom, and is associated with death from renal failure or hypotonia in childhood or young adulthood. Overall, it was possible to show that a high proportion of variance in number of publications could be accounted for in terms of severity.

Furthermore, the amount of research on a given condition will in part depend on the disciplines involved in investigating it. Given the large number of identified publications, it was not possible to classify the discipline of each one, but a rough index was obtained to estimate the proportion of publications that involved genetic investigations. Not surprisingly, these were more common for single-gene disorders, most of which were rare. The amount of genetic research on a disorder was also a significant independent predictor of publication index.

There were some intriguing exceptions where severity failed to predict research activity, which can be seen as outliers in Figure 1. Within the conditions with presumed multifactorial etiology, ADHD and autism spectrum disorder fell within the range of predicted research, with Tourette syndrome slightly above predicted levels. However, the publication indices for dyslexia, dyscalculia, developmental coordination disorder and speech sound disorder fell well below prediction. Dyslexia and SLI are comparable to ADHD in prevalence and severity, yet the number of publications for dyslexia was 4 times lower than for ADHD and that for SLI was 16 times lower than ADHD. It is also interesting to note that there were three times as many studies on developmental dyslexia as on SLI, and nearly ten times as many as on speech sound disorder, although these conditions are similar in terms of rated severity and prevalence. It is perhaps relevant that these latter two conditions posed particular problems when searching for publications, because of the variable terminology used to refer to cases, suggesting that diagnostic uncertainty might be a factor leading to low research rates. This is a field where there is considerable debate about how to diagnose and categorise cases, whether there are disorders that are distinct from normal variation, and where the same terminology that is used to refer to a disorder is also used to refer to symptoms in other disorders. Even though both disorders feature in the diagnostic classifications of ICD-10 [8] and DSM-IV [9], few researchers adopted definitions or nomenclature from the diagnostic manuals.

Another intriguing contrast is between Turner syndrome and Klinefelter syndrome, both caused by sex chromosome aneuploidies, and both affecting fertility. Table 3 indicates that there were nearly two times as many publications on Turner syndrome as on Klinefelter syndrome, even though Klinefelter syndrome is more than three times as common. Here the discrepancy is likely to relate to the ease of identification of affected individuals. The prevalence rates for Klinefelter syndrome are derived from prenatal and/or newborn surveys, but because the phenotype is not striking, many affected individuals are not identified in childhood. In contrast, the physical phenotype is far more obvious in Turner syndrome. This kind of explanation could also account for the dearth of studies on XYY and XXX trisomies; the majority of affected individuals are not identified because the phenotype is mild [10].

Even after taking into account severity and rates of genetic research, prevalence still was a significant predictor, with higher rates of publication for the rarer disorders. In considering reasons for the higher publication ratios in rare vs. common diseases, Al-Shahi et al.[1] noted that if the publication ratio for stroke were equal to that of variant Creutzfeldt­Jakob disease, clinicians and researchers interested in stroke would have had to read about 10 000 papers per week! One can turn this logic on its head and argue that if rare disorders had a publication index comparable to that for common disorders, there would be too few papers to give a critical mass of evidence in the area. For instance, if we extrapolate from the publication index for SLI (0.13 papers per 100 affected children over the 25 year period), for rare conditions where the national number of affected cases runs into hundreds rather than thousands, there would be no effective research literature. For the even rarer ‘orphan’ conditions with only a handful of affected individuals, lack of research is recognised as a serious problem [11]. The current analysis indicates that research activity on these disorders would have to occur at a rate well in excess of that predicted by extrapolation from more common disorders, in order to have a critical mass of publications and build research capacity.

The data reported here also suggest a positive effect of critical mass, insofar as the rate of growth in research on a disorder was related to the number of publications recorded in the first time interval, 1985–1989. This is compatible with the idea that a field develops by skilled researchers training their students and postdocs, who are likely to then study the same condition. If each mentor trains several others, then growth will be exponential rather than linear. Nevertheless, there were two conditions – ADHD and autism spectrum disorder – where the growth in research was unexpectedly high. For autism, there are two factors that could be implicated: first, broadening of the diagnostic criteria has led to a dramatic increase in diagnosis [12], and second, research funding has been directed to this area, for instance, by the Combating Autism Act passed in 2006 by the United States Congress, which authorized nearly 1 billion dollars over five years to combat autism and related disorders. For the more classic forms of autism, both prevalence and severity are comparable to Down syndrome, yet funding for autism is six times greater, and the slope showing increase of NIH funding over time is dramatically higher than for any other condition. It seems likely that government initiatives play a large role in explaining the extraordinary rise of publications in autism, though there are in addition several private foundations that provide substantial funding for autism research.

The other condition showing a very high increase in publications over time is ADHD, again with an associated high level of NIH funding. ADHD is comparable in prevalence and severity to SLI and dyslexia, and the proportion of genetic research on ADHD is relatively low, yet it attracts 19 to 21 times as much funding as these two conditions. Figure 2 suggests an explanation in terms of the professional disciplines with primary responsibility for these conditions. ADHD, like autism, is funded through several NIH institutes, but primarily by NIMH, suggesting most researchers in this field are psychiatrists. Dyslexia is funded mainly through NICHD and is the domain of psychologists, whereas SLI is funded mainly through NIDCD and is investigated principally by speech-language pathologists. This analysis suggests that inequities between professional disciplines in access to research training and funding may do a disservice to children who are affected by common yet under-researched neurodevelopmental disorders.

In summary, it is misleading to focus only on prevalence when comparing research activity for different conditions, as this suggests a massive excess of research on rare disorders. If severity is taken into account, the excess reduces considerably, although it is still present. This may, however, have to do with critical mass of research: there have to be proportionately more publications per affected individual for rare than common disorders, to ensure there is a body of work to build on. Nevertheless, even among conditions of similar frequency and severity there are some intriguing discrepancies in levels of research activity. It is suggested that these may have to do with extent of involvement of genetic researchers, specific initiatives in research funding, and lack of research funding in some disciplines working with neurodevelopmental disorders.


Thanks to Prisca Middlemiss and Courtenay Frazier Norbury for advice and comments on this paper, and to Andrew Whitehouse for reviewing the manuscript. Also, many thanks to those who assisted with ratings of severity: Gillian Baird, Charles Hulme, Paul Hutchins, Anne O'Hare, Daniel Pine, Emily Simonoff, David Skuse, Margaret Snowling, and Anita Thapar.

Author Contributions

Analyzed the data: DB. Wrote the paper: DB.


  1. 1. Al-Shahi R, Will RG, Warlow CP (2001) Amount of research interest in rare and common neurological conditions: bibliometric study. Br Med J 323: 1461–2.R. Al-ShahiRG WillCP Warlow2001Amount of research interest in rare and common neurological conditions: bibliometric study.Br Med J32314612
  2. 2. Gross CP, Anderson GF, Powe NR (1999) The relation between funding by the National Institutes of Health and the burden of disease. New Engl J Med 340: 1881–7.CP GrossGF AndersonNR Powe1999The relation between funding by the National Institutes of Health and the burden of disease.New Engl J Med34018817
  3. 3. Bishop D, Rutter M Rutter M, Bishop D, Pine D, Scott S, Stevenson J, Taylor E, Thapar A, editors. (2008) Neurodevelopmental disorders: conceptual approaches. Rutter's Child and Adolescent Psychiatry Oxford: Blackwell. 32–41.D. BishopM. Rutter2008Neurodevelopmental disorders: conceptual approaches.M. RutterD. BishopD. PineS. ScottJ. StevensonE. TaylorA. ThaparRutter's Child and Adolescent PsychiatryOxfordBlackwell3241
  4. 4. Rutter M, Bishop D, Pine D, Scott S, Stevenson J, et al., editors. (2008) Rutter's Child and Adolescent Psychiatry. Oxford: Blackwell. M. RutterD. BishopD. PineS. ScottJ. Stevenson2008Rutter's Child and Adolescent Psychiatry.OxfordBlackwell
  5. 5. Udwin O, Dennis J (1995) Psychological and behavioural phenotypes in genetically determined syndromes: a review of research findings. In: O'Brien G, Yule W, editors. Behavioural phenotypes. London: Mac Keith Press. pp. 90–208.O. UdwinJ. Dennis1995Psychological and behavioural phenotypes in genetically determined syndromes: a review of research findings.G. O'BrienW. YuleBehavioural phenotypesLondonMac Keith Press90208
  6. 6. Office for National Statistics (2008) Office for National Statistics2008Mid-2008 Estimates of Household Population for England and Wales (experimental). Mid-2008 Estimates of Household Population for England and Wales (experimental).
  7. 7. Harris J (2008) Brain disorders and their effect on psychopathology. In: Rutter M, Bishop D, Pine D, Scott S, Stevenson J, Taylor E, Thapar A, editors. Rutter's Child and Adolescent Psychiatry. Oxford: Blackwell. pp. 459–73.J. Harris2008Brain disorders and their effect on psychopathology.M. RutterD. BishopD. PineS. ScottJ. StevensonE. TaylorA. ThaparRutter's Child and Adolescent PsychiatryOxfordBlackwell45973
  8. 8. World Health Organization (1993) The ICD-10 classification for mental and behavioural disorders: Diagnostic criteria for research. Geneva: WHO. World Health Organization1993The ICD-10 classification for mental and behavioural disorders: Diagnostic criteria for research.GenevaWHO
  9. 9. American Psychiatric Association (1994) Diagnostic and statistical manual of mental disorders, 4th edition. Washington, D.C: American Psychiatric Association. American Psychiatric Association1994Diagnostic and statistical manual of mental disorders, 4th edition.Washington, D.CAmerican Psychiatric Association
  10. 10. Leggett V, Jacobs P, Nation K, Scerif G, Bishop DVM (2010) Neurocognitive outcomes of individuals with a sex chromosome trisomy: XXX, XYY, or XXY: a systematic review. Dev Med Child Neurol 52: 119–29.V. LeggettP. JacobsK. NationG. ScerifDVM Bishop2010Neurocognitive outcomes of individuals with a sex chromosome trisomy: XXX, XYY, or XXY: a systematic review.Dev Med Child Neurol5211929
  11. 11. Ayme S, Schmidtke J (2007) Networking for rare diseases: a necessity for Europe. Bundesgesundheitsblatt-Gesundheitsforschung-Gesundheitsschutz 50: 1477–83.S. AymeJ. Schmidtke2007Networking for rare diseases: a necessity for Europe.Bundesgesundheitsblatt-Gesundheitsforschung-Gesundheitsschutz50147783
  12. 12. Rutter M (2005) Incidence of autism spectrum disorders: Changes over time and their meaning. Acta Paed 94: 2–15.M. Rutter2005Incidence of autism spectrum disorders: Changes over time and their meaning.Acta Paed94215

Citation: Gao Y, Sheng C, Xie R-h, Sun W, Asztalos E, Moddemann D, et al. (2016) New Perspective on Impact of Folic Acid Supplementation during Pregnancy on Neurodevelopment/Autism in the Offspring Children – A Systematic Review. PLoS ONE 11(11): e0165626.

Editor: Cheryl S. Rosenfeld, University of Missouri Columbia, UNITED STATES

Received: February 27, 2016; Accepted: October 14, 2016; Published: November 22, 2016

Copyright: © 2016 Gao et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All of the relevant sources are cited within the paper and its tables, and all relevant data are within the paper.

Funding: This study was funded by the Canadian Institutes for Health Research (CIHR) (MOP-142723).

Competing interests: The authors have declared that no competing interests exist.


There are major emotional, societal, and economic implications of impaired neurodevelopment and/or autism in children [1, 2]. These children will often require specialized schooling and other community resources. Although the survival/life-span of these infants may not be seriously affected, many of them may need treatments throughout their lifetime, and the cost to the public health care system could be huge. When they reach adulthood, productivity is often lower than those with normal development, indirectly increasing the societal burden.

It has been established that supplementation with folic acid around the time of conception reduces the risk of neural tube defects (NTDs) in the offspring [3, 4, 5, 6]. However, whether folic acid has a similar effect on impaired neurodevelopment and/or autism remains elusive. This article therefore focuses on assessing the role of folic acid supplementation during pregnancy and folate metabolism on neurodevelopmental outcomes including autism spectrum disorders (ASDs), other than NTDs.

Materials and Methods

Search strategy

We conducted an online search of relevant literature compiled by the National Library of Medicine from Medline and EMBASE (searched on December 31, 2014 of the site: and, with restriction to human studies. We first created 3 files (restricting our search to English literature) using the following key words: 1) folate or folic acid (171322 papers identified by this search); 2) maternal or pregnancy or pregnant or gestation or gestational or prenatal or antenatal or periconception or periconceptional (1349219 papers identified by this search); and 3) autism or autism spectrum disorders or developmental delay or development or neurodevelopment or mental or cognitive or language or personal-social or gross motor or fine motor or behaviour or intellectual or intelligence or Bayley Scale (8268145 papers identified by this search). We then merged the 3 files. All abstracts of the papers identified by merging the 3 files were screened by two independent reviewers in our group to exclude irrelevant studies (such as those on NTDs); because the causation between folic acid supplementation in pregnancy and NTDs has been established, our review is interested in outcomes other than NTDs.

Study selection

We included randomized controlled trials (RCTs), cohort studies, and case control studies that examined the association between folic acid supplementation during pregnancy and neurodevelopment/autism in the offspring children. Data extraction was conducted independently and screened all records at the title level by two reviewers (Chao Sheng and Ri-hua Xie). To enhance sensitivity, records were only removed if both reviewers excluded at the title level. The second level of review was at the abstract level followed by another round of review at the full-text level. Two independent reviewers abstracted data using a standardized form. When there was a disagreement it was resolved by discussion with a third reviewer (Yunfei Gao). Corresponding authors were contacted via e-mail at least three times to obtain data if the outcome of the neurodevelopment/autism in the offspring children could not be readily abstracted from the publication.

Study quality assessment

The quality of included cohort and case control studies was assessed with the Newcastle Ottawa Scale [7]. Using this checklist, Yunfei Gao evaluated each of the included articles, with additional inputs from Chao Sheng and Ri-hua Xie. The details are shown in Table 1. Divergent views were resolved by consulting a third reviewer.

Data extraction and synthesis

Data extracted from each study included the first author’s last name, publication year, main outcome, sample size, study design, age of children, effect of folic acid, and comments on the study. Because of major heterogeneity in original studies in terms of study design and outcome and exposure measurements, no attempt to summarize the effect by meta-analysis was made. When studies demonstrated conflicting findings on different outcome measures, the study was defined as “no association”.


Literature search

A total of 3,348 papers were identified. Sixty five full-text articles were assessed for eligibility after screening. Most frequently the removed papers were animal studies or reviews or commentary/discussion in the interpretation of the study findings on other pregnancy outcomes (e.g., NTDs) or studies in humans but the effects of folic acid supplementation during pregnancy on neurodevelopment/autism in the offspring children was not examined. See details of selection in Fig 1.

A total of 22 original papers that looked at the association between folic acid supplementation in pregnancy and neurodevelopment/autism were identified after the screening in Table 2. Because of major heterogeneity in study design, exposure measurement, and outcome measurement, no attempt was made for quantitative synthesis of effect by meta-analysis. The 43 full-text excluded articles with reasons of exclusion were listed as supplement information in S1 Appendix.

Study characteristics

The 22 eligible studies include 2 RCTs, 18 cohort studies, and 2 case control studies. The baseline characteristics and further summarized information are outlined in Table 2. The main outcomes include ASDs, autism, developmental delay, cognition, attention function, neurodevelopment, emotional problems, and behavioural problems. The children range in age from 12 months to 11 years. Among the 21studies, 7 studies included more than 1000 children.


Fifteen studies showed a beneficial effect of folic acid supplementation on neurodevelopment/autism, 6 studies found no statistically significant result, while one study found a harmful effect at high dose of folic acid supplementation (see Table 2). There were 3 studies that had ASD as the main outcome measure. The first was a cohort study of 85,176 children aged 3.3 to 10.2 years [8]. The rate of ASD in children whose mothers took folic acid was 0.10%, whereas the rate for mothers who did not take folic acid was 0.21%, with adjusted odds ratio (OR) of folic acid users 0.61 (95% confidence interval (CI), 0.41–0.90). Another study was the Childhood Autism Risks from Genetics and Environment, a case-control study in the United States [9]. In the 837 mother-child pairs, the mean folic acid intake in the first month of pregnancy was significantly greater for mothers of normally developing children than for mothers of children with a confirmed diagnosis of ASD. A mean daily folic acid intake of ≥ 600ug during the first pregnant month was associated with reduced ASD risk (adjusted OR: 0.62; 95% CI: 0.42–0.92; P = 0.02). This finding was consistent with another case-control study by the same author [10], which showed that mean folic acid intake in early pregnancy was significantly higher for mothers of normally developing children than for mothers of children with ASD.

Several studies found similar beneficial effects of folic acid supplementation on other areas of neurodevelopment. For example, a study in Massachusetts [11] showed that for each 600 ug/day increment in total folate intake during the first trimester, Peabody Picture Vocabulary Test-III score at age 3 years was 1.6 points (95% CI 0.1–3.1; p = 0.04) higher. Forns et al found that omission errors (defined as the number of targets to which the individual did not respond) were lower in those whose mothers took dietary supplementation with folic acid and vitamins during pregnancy [12]. In a cohort study [13] involving 553 mother-child pairs in Greece, neurodevelopment at 18 months was assessed using the Bayley Scales of Infant and Toddler Development (3rd edition). Compared with non-users, daily intake of 5 mg supplemental folic acid was associated with a 5-unit increase on the scale of receptive communication and a 3.5-unit increase on the scale of expressive communication. Roth et al assessed severe language delay (defined as minimal expressive language: only 1-word or unintelligible utterances at the age of 3 years) in a cohort of 38,954 children, and found that adjusted ORs for 3 patterns of exposure to maternal dietary supplements (no supplement as the reference) were 1.04 (95% CI, 0.62–1.74) for other supplements but no folic acid; 0.55 (95% CI, 0.35–0.86) for folic acid only; and 0.55 (95% CI, 0.39–0.78) for folic acid in combination with other supplements, demonstrating a clear protective effect of folic acid supplementation during pregnancy [14]. A study by Steenweg-de et al found a higher risk of emotional problems in 3 year old children using the Child Behavior Checklist (CBCL) whose mothers did not use supplements or started folic acid supplements late in pregnancy (OR: 1.45; 95% CI: 1.14, 1.84) compared to children whose mothers started folic acid supplement in early pregnancy [15]. Similarly, in a prospective cohort study, Roza et al examined the association between folic acid supplement use during the first trimester and behavioural and emotional problems identified by the CBCL in 4,214 toddlers at the age of 18 months. This study found that folic acid supplement use protected both from internalizing (OR of no use 1.65; 95% CI 1.24, 2.19) and externalizing problems (OR of no use 1.45; 95% CI 1.17, 1.80), after adjusting for maternal characteristics, birth weight, and fetal head size [16]. In another prospective cohort study, Julvez et al found that folic acid supplement during pregnancy was associated with improved neurodevelopment in children after adjusting for a number of sociodemographic and behavioural factors (results obtained from linear regression models): higher scores on verbal (b (regression slope) = 3.98, SE (standard error of regression slope) = 1.69), motor (b = 4.54, SE = 1.66), verbal-executive function (b = 3.97, SE = 1.68) scores, social competence (b = 3.97, SE = 1.61), and lower rate of inattention symptom [OR = 0.46; 95% CI 0.22, 0.95] [17].


A total of 22 original papers that looked at the association between folic acid supplementation in pregnancy and neurodevelopment/autism were identified after the screening, with 15 studies showing a beneficial effect of folic acid supplementation on neurodevelopment/autism, 6 studies found no statistically significant effect, while one study found a harmful effect at high dose of folic acid supplementation [18]. Two papers that suggested an adverse effect of folic acid on ASDs were not included in our review because no data on individual subjects were available in these two studies [19, 20]. Both papers used ecological data to support their hypothesis: prenatal folic acid supplementation and autism diagnoses in the United States since the 1980s in King’s study, and published autism incidence rates and prescriptions for folic acid in Rochester, Minnesota from 1976 to 1997 in Beard’s study. Beard’s study found a correlation coefficient of 0.87 (95% CI 0.19–0.99) between autism rates and the prescription prenatal vitamins containing folic acid and a correlation coefficient of 0.62 (95% CI 0.38–0.95) between autism rates and pediatric folic acid. However, during the same period, major changes in other factors such as diagnostic criteria, public awareness, disease surveillance, and screening efforts have all played an important role in the increased rates of diagnosed ASDs, so ecological data may not be suitable to analyze the association between folate and ASDs. Data from recent ASDs surveillance in the United States revealed a major increase in ASDs prevalence during a period with no change in policies regarding prenatal folic acid supplementation or folic acid food fortification (2002 to 2008), suggesting that ecological analyses are seriously flawed [21]. On the other hand, in a small sample of children (77) born to mothers used folic acid supplementation >5 mg/day during pregnancy had a statistically significantly lower mean psychomotor scale score (difference, -4.35 points; 95% CI, -8.34 to -0.36) than children whose mothers used a recommended dosage of folic acid supplements (0.4–1.0 mg/day) [18]. The finding from a single study with small sample needs to be replicated. Castro et al conducted a systematic review of studies involving on relationship between folic acid and ASD. All 11 papers included in Castro’s review met the inclusion criteria in our review. It concluded that although lower folate levels were associated with increased risk of ASD, the effects of folate-enhancing interventions on the clinical symptoms of ASD have yet to be confirmed [22].

Folic acid, or folate (vitamin B9) is an essential nutrient that is required for DNA replication and as a substrate for a range of enzymatic reactions involved in amino acid synthesis and vitamin metabolism. Demands for folate increase during pregnancy because of increased demands from the fetus. It has been conclusively established that folic acid deficiency prior to and during early pregnancy (up to 12 weeks of gestation) causes increased risk of NTDs, and periconceptional supplementation of folic acid can dramatically lower this risk (as much as 70%) [3, 4, 5, 6]. If folic acid deficiency prior to and during early pregnancy can cause NTDs, it may also cause milder forms of fetal brain damage that could be expressed as impaired neurodevelopment/autism in early childhood, and this effect may not be restricted to pre-conception and early gestation (<12 weeks of gestation, as the neural tube closes at that time so no NTDs after that). Laboratory investigations in animals and humans have shown that folate plays an important role in early brain development. In humans, there are active placental transports of folate and fetal brain folate levels are higher than adult levels [6]. In rats, the concentrations of many folate-dependent enzymes were substantially higher during early development than adult levels [23]. Dams and developing pups fed with diets eliminating folic acid 1 week prior to birth were less viable and had lower brain weights, lower activity level, and lower level of S-adenosyl-L-methionine concentrations in brain tissue of surviving offspring than animals reared on normal diets [24]. Ferguson et al examined whether gestational dietary folate deficiency not producing severe NTDs could lead to other functional impairments in mice [25]. They found that prenatal folate deficiency in mice led to an increase in anxiety-related behaviours. Worthy of our attention is that Padmanabhan et al found a hypomorphic mutation of the mouse MTRR gene, which results in developmental delay, as well as congenital malformations, including neural tube, heart, and placental defects, showing that folate metabolism has distinct transgenerational epigenetic functions that are responsible for specific developmental processes [26]. Even with normal dietary folate, the hypomorphic mutations in the MTRR gene associated with reduced expression may still lead to congenital abnormalities.

A few clinical studies have compared metabolites or cofactors of the folate-methionine pathway in children with autism [27–37]. While results from these studies have not been consistent a dysfunctional folate-methionine pathway has been identified that may have an impact on developmental delays including autism [38]. This pathway is crucial for DNA synthesis, DNA methylation, and cellular redox balance.

Clinical case series have also linked folic acid deficiency to other types of fetal brain damage such as intracranial calcification [39]. Del Campo et al reported several cases of patients who, in addition to the structural anomalies typical of maternal methotrexate exposure, have significant developmental delay, and suggested that prenatal exposure to folic acid antagonists increases the risk of mental retardation [40]. Arakawa et al observed that the EEG maturation was delayed in children born to mothers with low serum folate [41]. A recent study evaluated the nutritional intake in 111 Chinese children with autism (aged 2 to 9 years) and compared with the national Dietary Reference Intakes (DRI) [42]. They found that the children with autism had adequate or exceeded intakes in energy, protein, vitamins B1, B2, E, niacin, magnesium, and iron, but had inadequate intakes in folic acid, vitamins A, B6, C, calcium, and zinc, nutrients known to be important for brain development and function [42].


To our knowledge, this is the first systematic review examining the impact of folic acid supplementation during pregnancy on neurodevelopment/autism in the offspring children. We did an extensive search of relevant literature and selected studies strictly. After merging and analyzing the selected studies, we provided a preliminary conclusion.


Our study has some limitations. First, the magnitude of the protective effect of folic acid supplementation observed in most of the included studies was quite small as compared with the known effect of folic acid supplementation on NTDs. We speculate that compared with NTDs, the diagnosis of neurodevelopmental disorders is more subtle and requires a much longer observation period. As a result, the potential effect of folic acid supplementation may have been offset by measurement errors or loss of follow ups. Second, most of the included studies were observational. However, one RCT [43] showed beneficial effects of folic acid supplementation, consistent with a majority of the observational studies, which adds weight to the evidence. Third, because of major heterogeneity in original studies in terms of study design and outcome and exposure measurements, no attempt to summarize the effect by meta-analysis was made. Finally, there may be studies with negative results that were not published in the searchable databases because of potential publication bias. Although the major heterogeneity of the included studies prevented us from a formal assessment of publication bias, global inspection of all included studies did not find any systematic trends in terms of positive/negative findings.

Implications for research

The limited data identified suggests that folic acid supplementation in pregnancy protects against impaired neurodevelopment including ASDs in children, and may improve cognitive function and intellectual and motor function. However, it is hard to draw a conclusion due to the limitations of the identified studies. Large scale RCTs with validated diagnosis and high follow up rate are needed in order to produce robust evidence regarding the effects of folic acid supplementation in pregnancy on fetal neurodevelopment.


In summary, our review of the literature suggests that folic acid supplementation in pregnancy may protect against impaired neurodevelopment including ASDs in children, and may improve cognitive function, intellectual, and motor function.


This study was funded by the Canadian Institutes for Health Research (CIHR) (MOP-142723).

Author Contributions

  1. Conceptualization: SWW YFG.
  2. Data curation: YFG CS RHX WS.
  3. Formal analysis: YFG CS RHX WS.
  4. Funding acquisition: SWW.
  5. Investigation: YFG CS RHX WS EA DM LZ MW SWW.
  6. Methodology: SWW YFG CS RHX WS.
  7. Project administration: SWW.
  8. Resources: SWW.
  9. Supervision: SWW.
  10. Validation: YFG CS RHX WS.
  11. Visualization: YFG CS RHX WS EA DM LZ MW SWW.
  12. Writing – original draft: YFG CS.
  13. Writing – review & editing: SWW WS EA DM LZ MW.


  1. 1. Baker BL, McIntyre LL, Blacher J, Crnic K, Edelbrock C, Low C. Pre-school children with and without developmental delay: behaviour problems and parenting stress over time. J Intellect Disabil Res. 2003; 47: 217–230. pmid:12787154
  2. 2. Zwaigenbaum L. Screening, risk and early identification of Autism Spectrum Disorders, D. 1 st Eds. New York: Oxford University Press; 2011.
  3. 3. Berry RJ, Li Z, Erickson JD, Li S, Moore CA, Wang H, et al. Prevention of neural-tube defects with folic acid in China. China-U.S. Collaborative Project for Neural Tube Defect Prevention. N Engl J Med. 1999; 341:1485–1490. pmid:10559448
  4. 4. Czeizel AE, Dudás I, Paput L, Bánhidy F. Prevention of neural-tube defects with periconceptional folic acid, methylfolate, or multivitamins? Ann Nutr Metab. 2011;58: 263–271. pmid:21865678
  5. 5. Prevention of neural tube defects: results of the Medical Research Council Vitamin Study. MRC Vitamin Study Research Group. Lancet. 1991, 338:131–137. pmid:1677062
  6. 6. Greenblatt JM, Huffman LC, Reiss AL. Folic acid in neurodevelopment and child psychiatry. Prog Neuropsychopharmacol Biol Psychiatry. 1994, 18:647–60. pmid:7938557
  7. 7. Wells GA, Shea B, O'Connell D, Peterson J, Welch V, Losos M, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Available from:
  8. 8. Surén P, Roth C, Bresnahan M, Haugen M, Hornig M, Hirtz D,et al. Association between maternal use of folic acid supplements and risk of autism spectrum disorders in children. JAMA. 2013; 309:570–577. pmid:23403681
  9. 9. Schmidt RJ, Hansen RL, Hartiala J, Allayee H, Schmidt LC, Tancredi DJ, et al. Prenatal vitamins, one-carbon metabolism gene variants, and risk for autism. Epidemiology. 2011, 22: 476–485. pmid:21610500
  10. 10. Schmidt RJ, Tancredi DJ, Ozonoff S, Hansen RL, Hartiala J, Allayee H, et al. Maternal periconceptional folic acid intake and risk of autism spectrum disorders and developmental delay in the CHARGE (CHildhood Autism Risks from Genetics and Environment) case-control study. Am J Clin Nutr. 2012; 96: 80–89. pmid:22648721
  11. 11. Villamor E, Rifas-Shiman SL, Gillman MW, Oken E. Maternal intake of methyl-donor nutrients and child cognition at 3 years of age. Paediatr Perinat Epidemiol. 2012; 26:328–335. pmid:22686384
  12. 12. Forns J, Torrent M, Garcia-Esteban R, Cáceres A, Pilar Gomila M, Martinez D, et al. Longitudinal association between early life socio-environmental factors and attention function at the age 11 years. Environ Res. 2012;117:54–59. pmid:22608140
  13. 13. Chatzi L, Papadopoulou E, Koutra K, Roumeliotaki T, Georgiou V, Stratakis N, et al. Effect of high doses of folic acid supplementation in early pregnancy on child neurodevelopment at 18 months of age: the mother-child cohort 'Rhea' study in Crete, Greece. Public Health Nutr. 2012;15:1728–1736. pmid:22314109
  14. 14. Roth C, Magnus P, Schjølberg S, Stoltenberg C, Surén P, McKeague IW, et al. Folic acid supplements in pregnancy and severe language delay in children. JAMA. 2011; 306:1566–73. pmid:21990300
  15. 15. Steenweg-de Graaff J, Roza SJ, Steegers EA, Hofman A, Verhulst FC, Jaddoe VW, et al. Maternal folate status in early pregnancy and child emotional and behavioral problems: the Generation R Study. Am J Clin Nutr. 2012;95:1413–21. pmid:22572645
  16. 16. Roza SJ, van Batenburg-Eddes T, Steegers EA, Jaddoe VW, Mackenbach JP, Hofman A, et al. Maternal folic acid supplement use in early pregnancy and child behavioural problems: The Generation R Study. Br J Nutr. 2010;103:445–452. pmid:19772683
  17. 17. Julvez J, Fortuny J, Mendez M, Torrent M, Ribas-Fitó N, Sunyer J. Maternal use of folic acid supplements during pregnancy and four-year-old neurodevelopment in a population-based birth cohort. Paediatr Perinat Epidemiol. 2009;23:199–206. pmid:19775381
  18. 18. Valera-Gran D, García de la Hera M, Navarrete-Muñoz EM, Fernandez-Somoano A, Tardón A, Julvez J, et al. Folic acid supplements during pregnancy and child psychomotor development after the first year of life. JAMA Pediatr. 2014 Nov;168(11):e142611 pmid:25365251
  19. 19. Beard CM, Panser LA, Katusic SK. Is excess folic acid supplementation a risk factor for autism? Med Hypotheses. 2011;77:15–7. pmid:21454018
  20. 20. King CR. A novel embryological theory of autism causation involving endogenous biochemicals capable of initiating cellular gene transcription: a possible link between twelve autism risk factors and the autism 'epidemic'. Med Hypotheses. 2011;76:653–660. pmid:21388746
  21. 21. Autism and Developmental Disabilities Monitoring Network Surveillance Year 2008 Principal Investigators; Centers for Disease Control and Prevention: Prevalence of autism spectrum disorders—Autism and Developmental Disabilities Monitoring Network, 14 sites, United States, 2008. MMWR Surveill Summ. 2012; 61:1–19.
  22. 22. Castro K, Klein LD, Baronio D, Gottfried C, Riesgo R, Perry IS. Folic acid and autism: What do we know? Nutr Neurosci. 2014 Aug 2.
  23. 23. Boetz MI, Reynolds EH. Biochemical functions in the CNS in folic acid. In: Folic Acid In Neurology, Psychiatry and Internal Medicine, 1 st ed. New York: Raven Press; 1979.
  24. 24. Middaugh LD, Grover TA, Blackwell LA, Zemp JW. Neurochemical and behavioral effects of diet related perinatal folic acid restriction. Pharmacol Biochem Behav. 1976; 5:129–134.
  25. 25. Ferguson SA, Berry KJ, Hansen DK, Wall KS, White G, Antony AC. Behavioral effects of prenatal folate deficiency in mice. Birth Defects Res A Clin Mol Teratol. 2005; 73:249–252. pmid:15744731
  26. 26. Padmanabhan N, Jia D, Geary-Joo C, Wu X, Ferguson-Smith AC, Fung E, et al. Mutation in folate metabolism causes epigenetic instability and transgenerational effects on development. Cell. 2013; 155:81–93. pmid:24074862
  27. 27. Adams JB, Holloway C. Pilot study of a moderate dose multivitamin/mineral supplement for children with autistic spectrum disorder. J Altern Complement Med. 2004; 10:1033–1039. pmid:15673999
  28. 28. Adams JB, George F, Audhya T. Abnormally high plasma levels of vitamin B6 in children with autism not taking supplements compared to controls not taking supplements. J Altern Complement Med. 2006; 12:59–63. pmid:16494569
  29. 29. Arnold GL, Hyman SL, Mooney RA, Kirby RS. Plasma amino acids profiles in children with autism: potential risk of nutritional deficiencies. J Autism Dev Disord. 2003; 33: 449–54. pmid:12959424
  30. 30. D'Eufemia P, Finocchiaro R, Celli M, Viozzi L, Monteleone D, Giardini O. Low serum tryptophan to large neutral amino acids ratio in idiopathic infantile autism. Biomed Pharmacother.1995; 49:288–292. pmid:7579010
  31. 31. James SJ, Melnyk S, Jernigan S, Cleves MA, Halsted CH, Wong DH, et al. Metabolic endophenotype and related genotypes are associated with oxidative stress in children with autism. Am J Med Genet B Neuropsychiatr Genet. 2006; 141B:947–956. pmid:16917939
  32. 32. Lowe TL, Cohen DJ, Miller S, Young JG. Folic acid and B12 in autism and neuropsychiatric disturbances of childhood. J Am Acad Child Psychiatry. 1981; 20:104–11. pmid:7217543
  33. 33. Moretti P, Sahoo T, Hyland K, Bottiglieri T, Peters S, del Gaudio D, et al. Cerebral folate deficiency with developmental delay, autism, and response to folinic acid. Neurology. 2005; 64:1088–90. pmid:15781839
  34. 34. Moretti P, Peters SU, Del Gaudio D, Sahoo T, Hyland K, Bottiglieri T, et al. Brief report: autistic symptoms, developmental regression, mental retardation, epilepsy, and dyskinesias in CNS folate deficiency. J Autism Dev Disord. 2008; 38:1170–1177. pmid:18027081
  35. 35.

0 Thoughts to “Neurodevelopment Research Paper

Leave a comment

L'indirizzo email non verrà pubblicato. I campi obbligatori sono contrassegnati *