UNITY: A low-field magnetic resonance neuroimaging initiative to characterize neurodevelopment in low and middle-income settings.

Measures of physical growth, such as weight and height have long been the predominant outcomes for monitoring child health and evaluating interventional outcomes in public health studies, including those that may impact neurodevelopment. While physical growth generally reflects overall health and nutritional status, it lacks sensitivity and specificity to brain growth and developing cognitive skills and abilities. Psychometric tools, e.g., the Bayley Scales of Infant and Toddler Development, may afford more direct assessment of cognitive development but they require language translation, cultural adaptation, and population norming. Further, they are not always reliable predictors of future outcomes when assessed within the first 12-18 months of a child's life. Neuroimaging may provide more objective, sensitive, and predictive measures of neurodevelopment but tools such as magnetic resonance (MR) imaging are not readily available in many low and middle-income countries (LMICs). MRI systems that operate at lower magnetic fields (< 100mT) may offer increased accessibility, but their use for global health studies remains nascent. The UNITY project is envisaged as a global partnership to advance neuroimaging in global health studies. Here we describe the UNITY project, its goals, methods, operating procedures, and expected outcomes in characterizing neurodevelopment in sub-Saharan Africa and South Asia.

Chromatin remodeling and Rb activity.

Progression of cells through the cell cycle is central to normal cell proliferation, and checkpoints that regulate this cycle are targets of tumorigenic mutations. One of these checkpoints is the Rb family of proteins that seems to regulate exit of cells from both G(1) and S phase of the cell cycle. Recent studies have linked the function of the Rb family to chromatin remodeling enzymes.

Rb function in cell-cycle regulation and apoptosis.

Loss of cell-cycle control is a hallmark of neoplastic cells. One regulator of the critical G1 to S-phase transition in the cell cycle is the retinoblastoma tumour suppressor protein Rb, which interacts with the E2F family of cell-cycle transcription factors to repress gene transcription required for this transition. Through its interaction with E2F, Rb also regulates genes that control apoptosis. Here we review the roles of Rb in regulating the cell cycle and apoptosis and discuss recent results linking these Rb functions to chromatin-remodelling enzymes.

Regulation of fibronectin biosynthesis by dexamethasone, transforming growth factor beta, and cAMP in human cell lines.

The regulation of fibronectin (FN) biosynthesis by dexamethasone (a synthetic glucocorticoid), forskolin (an activator of adenylate cyclase), and transforming growth factor beta (TGF-beta) was examined in six human cell lines. Dexamethasone treatment produced the largest increase in FN biosynthesis in the fibrosarcoma cell line, HT-1080 (approximately 45-fold). This seems to result from a dexamethasone-mediated increase in FN mRNA stability which increases the message half-life from approximately 11 to 26 h. The relative instability of FN mRNA in the fibrosarcoma (t1/2 11 h) compared to normal fibroblasts (70 h) appears to result from the particular transformed phenotype of the HT-1080 cells. Forskolin and TGF-beta increase the rate of FN gene transcription in most of the cell lines. These effects (four- to six-fold) occur rapidly and do not require protein synthesis in the responsive cell lines which include normal fibroblasts. However, in the fibrosarcoma (HT-1080), a surprisingly large induction (20-30-fold) is observed and this induction is different from that in the normal fibroblasts and the other cell lines in that both protein synthesis and a lag period are required. Synergism is seen with dexamethasone and either forskolin or TGF-beta in HT-1080 cells increasing the rate of FN biosynthesis approximately 200-fold to a level similar to normal fibroblasts. This seems to result from a combination of FN mRNA stabilization (dexamethasone) and increased transcription (forskolin and TGF-beta).

Interferon regulatory factor-2 is a transcriptional activator in muscle where It regulates expression of vascular cell adhesion molecule-1.

Previously, we have suggested that vascular cell adhesion molecule-1 (VCAM-1) and its integrin receptor alpha4beta1 mediate cell-cell interactions important for skeletal myogenesis. Expression of the receptors subsequently subsides in muscle after birth. Here, we examine the mechanism regulating VCAM-1 gene expression in muscle. An enhancer located between the TATA box and the transcriptional start site is responsible for VCAM-1 gene expression in muscle-this element is inactive in endothelial cells where VCAM-1 expression is dependent on nuclear factor kappaB sites and inflammatory cytokines. We identify interferon regulatory factor-2 (IRF-2), a member of the interferon regulatory factor family, as the enhancer-binding transcription factor and show that expression of IRF-2 parallels that of VCAM-1 during mouse skeletal myogenesis. IRF-2 is not dependent upon cytokines for expression or activity, and it has been shown to act as a repressor in other nonmuscle cell types. We show that the basic repressor motif located near the COOH-terminal of IRF-2 is not active in muscle cells, but instead an acidic region in the center of the molecule functions as a transactivating domain. Although IRF-2 and VCAM-1 expression diminishes on adult muscle fiber, they are retained on myogenic stem cells (satellite cells). These satellite cells proliferate and fuse to regenerate muscle fiber after injury or disease. We present evidence that VCAM-1 on satellite cells mediates their interaction with alpha4beta1(+) leukocytes that invade the muscle after injury or disease. We propose that VCAM-1 on endothelium generally recruits leukocytes to muscle after injury, whereas subsequent interaction with VCAM-1 on regenerating muscle cells focuses the invading leukocytes specifically to the sites of regeneration.

A receptor in pituitary and hypothalamus that functions in growth hormone release.

Small synthetic molecules termed growth hormone secretagogues (GHSs) act on the pituitary gland and the hypothalamus to stimulate and amplify pulsatile growth hormone (GH) release. A heterotrimeric GTP-binding protein (G protein)-coupled receptor (GPC-R) of the pituitary and arcuate ventro-medial and infundibular hypothalamus of swine and humans was cloned and was shown to be the target of the GHSs. On the basis of its pharmacological and molecular characterization, this GPC-R defines a neuroendocrine pathway for the control of pulsatile GH release and supports the notion that the GHSs mimic an undiscovered hormone.

Assessing and minimizing time-dependent inhibition of cytochrome P450 3A in drug discovery: a case study with melanocortin-4 receptor agonists.

1-[(2R)-2-([[(1S,2S)-1-amino-1,2,3,4-tetrahydronaphthalen-2-yl]carbonyl]amino)-3-(4-chlorophenyl)propanoyl]-N-(tert-butyl)-4-cyclohexylpiperidine-4-carboxamide (1) is a potent melanocortin-4 receptor agonist that exhibited time-dependent inhibition of cytochrome P450 (P450) 3A in incubations with human liver microsomes. In incubations fortified with potassium cyanide, a cyano adduct was identified by liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis as a cyanonitrosotetrahydronaphthalenyl derivative. The detection of this adduct suggested that a nitroso species was involved in the formation of a metabolite intermediate (MI) complex that led to the observed P450 inactivation. Further evidence supporting this hypothesis derived from incubations of 1 with recombinant P450 3A4, which exhibited a lambda(max) at approximately 450 nm. The species responsible for this absorbance required the presence of beta-nicotinamide adenine dinucleotide phosphate reduced form (NADPH), increased with increasing incubation time and decreased following the addition of potassium ferricyanide to the incubation mixture, suggestive of an MI complex. Similar results were obtained with rat liver microsomes and with recombinant P450 3A1. When rats were dosed with indinavir as a P450 3A probe substrate, plasma exposure to indinavir increased three-fold following pretreatment with 1, consistent with drug-drug interaction projections based on the k(inact) and K(I) parameters for 1 in rat liver microsomes. A similar approach was used to predict the magnitude of the corresponding drug-drug interaction potential in humans dosed with a drug metabolized predominantly by P450 3A, and the forecast area under the curve (AUC) increase ranged from four- to ten-fold. These data prompted a decision to terminate further evaluation of 1 as a development candidate, and led to the synthesis of the methyl analogue 2. Methyl substitution alpha to the amino group in 2 was designed to reduce the propensity for formation of a nitroso intermediate and, indeed, 2 failed to exhibit time-dependent inhibition of P450 3A in human liver microsomal incubations. This case study highlights the importance of mechanistic studies in support of drug-discovery and decision-making processes.

Human amygdala functional network development: A cross-sectional study from 3 months to 5 years of age.

Although the amygdala's role in shaping social behavior is especially important during early post-natal development, very little is known of amygdala functional development before childhood. To address this gap, this study uses resting-state fMRI to examine early amygdalar functional network development in a cross-sectional sample of 80 children from 3-months to 5-years of age. Whole brain functional connectivity with the amygdala, and its laterobasal and superficial sub-regions, were largely similar to those seen in older children and adults. Functional distinctions between sub-region networks were already established. These patterns suggest many amygdala functional circuits are intact from infancy, especially those that are part of motor, visual, auditory and subcortical networks. Developmental changes in connectivity were observed between the laterobasal nucleus and bilateral ventral temporal and motor cortex as well as between the superficial nuclei and medial thalamus, occipital cortex and a different region of motor cortex. These results show amygdala-subcortical and sensory-cortex connectivity begins refinement prior to childhood, though connectivity changes with associative and frontal cortical areas, seen after early childhood, were not evident in this age range. These findings represent early steps in understanding amygdala network dynamics across infancy through early childhood, an important period of emotional and cognitive development.

Regulation of vascular cell adhesion molecule-1 expression by IL-4 and TNF-alpha in cultured endothelial cells.

Interaction between vascular cell adhesion molecule-1 (VCAM-1) on endothelial cells and alpha 4 integrins on leukocytes is thought to mediate the selective recruitment of eosinophils and lymphocytes that occurs in allergic diseases. IL-4 is associated with allergic conditions, and it has been shown to selectively increase expression of VCAM-1 on endothelial cells in vivo, suggesting that it could be responsible for VCAM-1 expression in allergic disease. Using a combination of immunofluorescence, flow cytometry, and Northern analysis, we compared the effect of TNF-alpha and IL-4 on VCAM-1 expression. TNF-alpha is also associated with allergic diseases, and it rapidly increases transcription of the VCAM-1 gene. The effect of IL-4 was relatively modest with prolonged kinetics: VCAM-1 was not detected until 72 h after treatment with IL-4. However, when TNF-alpha and IL-4 were combined, there was a synergistic increase in VCAM-1 expression and a dramatic prolongation of the appearance of VCAM-1 on the cell surface. This synergy results from a combination of transcriptional activation by TNF-alpha and the stabilization of resulting transcripts by IL-4. We propose that IL-4 allows subthreshold concentrations of TNF-alpha (concentrations that would not normally activate expression of adhesion molecules on the endothelium) to selectively increase VCAM-1 expression and to prolong its appearance on the surface of cells in allergic disease.

Domains A and B in the Rb pocket interact to form a transcriptional repressor motif.

The retinoblastoma protein (Rb) is a tumor suppressor that regulates progression from the G1 phase to the S phase of the cell cycle. Previously, we found that Rb is a transcriptional repressor that is selectively targeted to promoters through an interaction with the E2F family of cell cycle transcription factors--when Rb is tethered to a promoter through E2F, it not only blocks E2F activity, it also binds surrounding transcription factors, preventing their interaction with the basal transcription complex, thus resulting in a dominant inhibitory effect on transcription of cell cycle genes. Here we examine the repressor motif of Rb. The two domains in the Rb pocket, A and B, which are conserved across species and in the Rb-related proteins p107 and p130, are both required for repressor activity. The nonconserved spacer separating A and B is not required. Although neither A nor B alone had any repressor activity, surprisingly, repressor activity was observed when the domains were coexpressed on separate proteins. Transfection assays suggest that one domain can recruit the other to the promoter to form a repressor motif that can both interact with E2F and have a dominant inhibitory effect on transcription. Using coimmunoprecipitation and in vitro binding assays, we show that A and B interact directly and that mutations which disrupt this interaction inhibit repressor activity. The Rb pocket was originally defined as the binding site for oncoproteins from DNA tumor viruses such as adenovirus E1a. We present evidence that E1a interacts with a site formed by the interaction of A and B and that this interaction with A and B induces or stabilizes the A-B interaction.

Independent repressor domains in ZEB regulate muscle and T-cell differentiation.

ZEB is a zinc finger-homeodomain protein that represses transcription by binding to a subset of E-box sequences. ZEB inhibits muscle differentiation in mammalian systems, and its Drosophila orthologue, zfh-1, inhibits somatic and cardiac muscle differentiation during Drosophila embryogenesis. ZEB also binds to the promoter of pivotal hematopoietic genes (including those encoding interleukin-2, CD4, GATA-3, and alpha(4)-integrin), and mice in which ZEB has been genetically targeted show thymic atrophy, severe defects in lymphocyte differentiation, and increased expression of the alpha(4)-integrin and CD4. Here, we demonstrate that ZEB contains separate repressor domains which function in T lymphocytes and muscle, respectively. The most C-terminal domain inhibits muscle differentiation in mammalian cells by specifically blocking the transcriptional activity of the myogenic factor MEF2C. The more N-terminal domain blocks activity of hematopoietic transcription factors such as c-myb, members of the ets family, and TFE-III. Our results demonstrate that ZEB has evolved with two independent repressor domains which target distinct sets of transcription factors and function in different tissues.

Interaction of a common factor with ATF, Sp1, or TATAA promoter elements is required for these sequences to mediate transactivation by the adenoviral oncogene E1a.

The adenovirus protein E1a stimulates transcription of both viral and cellular genes. Unlike most other transcription factors, it induces transactivation through several different promoter elements. The mechanism by which elements of diverse sequence mediate the effect of E1a is the focus of this study. Three E1a-responsive elements (an ATF site, an Sp1 site, and a TATA box containing the sequence TATAA) were studied to determine whether their interaction with a common factor is necessary for transactivation. In transfection assays, each element was used as a competitor against promoter constructs containing the other elements. The elements as competitors had no effect on basal transcription, but each competitor completely inhibited transactivation by E1a. Competitors that were not E1a responsive failed to inhibit transactivation. Therefore, either E1a itself or an E1a-inducible factor interacts with each of the elements to cause transactivation, most likely though an association with each element's specific binding protein.

The Rb family contains a conserved cyclin-dependent-kinase-regulated transcriptional repressor motif.

Progression through the cell cycle is dependent on the sequential expression of cyclins, which combine with cyclin-dependent kinases (cdks) to form active kinases. The transition from G1 to S phase is dependent on D cyclins in complex with cdk4 or cdk6 and cyclin E complexed with cdk2. One target of G1 cyclins is the retinoblastoma susceptibility protein (Rb). Rb is a transcriptional repressor that is selectively targeted to genes through interaction with the E2F family of cell cycle transcription factors. Rb is a member of a family of proteins that include p107 and p130. The three proteins share a region known as the pocket that is important for binding E2F and is also the binding site for oncoproteins from DNA tumor viruses that inactivate Rb. We have found that two conserved domains within the Rb pocket (A and B) interact to form a transcriptional repressor motif (K. N. B. Chow and D. C. Dean, Mol. Cell. Biol. 16:4862-4868, 1996). Here we demonstrate that p107 also has an A-B repressor motif, and using domain swapping and coimmunoprecipitation assays, we compare A and B from Rb and p107. Finally and most importantly, we demonstrate that the A-B interaction which forms the repressor motif is blocked by G1 cdk phosphorylation, thereby blocking repressor activity. This A-B repressor motif is then the first example of a cdk-regulated transcriptional repressor.

Transcriptional repression by RB-E2F and regulation of anchorage-independent survival.

Mutations that lead to anchorage-independent survival are a hallmark of tumor cells. Adhesion of integrin receptors to extracellular matrix activates a survival signaling pathway in epithelial cells where Akt phosphorylates and blocks the activity of proapoptotic proteins such as the BCL2 family member Bad, the forkhead transcription factor FKHRL-1, and caspase 9. Insulin-like growth factor 1 (IGF-1) is a well-established epithelial cell survival factor that also triggers activation of Akt and can maintain Akt activity after cells lose matrix contact. It is not until IGF-1 expression diminishes (~16 h after loss of matrix contact) that epithelial cells deprived of matrix contact undergo apoptosis. This suggests that IGF-1 expression is linked to cell adhesion and that it is the loss of IGF-1 which dictates the onset of apoptosis after cells lose matrix contact. Here, we examine the linkage between cell adhesion and IGF-1 expression. While IGF-1 is able to maintain Akt activity and phosphorylation of proapoptotic proteins in cells that have lost matrix contact, Akt is not able to phosphorylate and inactivate another of its substrates, glycogen synthase kinase 3beta (GSK-3beta), under these conditions. The reason for this appears to be a rapid translocation of active Akt away from GSK-3beta when cells lose matrix contact. One target of GSK-3beta is cyclin D, which is turned over in response to this phosphorylation. Therefore, cyclin D is rapidly lost when cells are deprived of matrix contact, leading to a loss of cyclin-dependent kinase 4 activity and accumulation of hypophosphorylated, active Rb. This facilitates assembly of a repressor complex containing histone deacetylase (HDAC), Rb, and E2F that blocks transcription of the gene for IGF-1, leading to loss of Akt activity, accumulation of active proapoptotic proteins, and apoptosis. This feedback loop containing GSK-3beta, cyclin D, HDAC-Rb-E2F, and IGF-1 then determines how long Akt will remain active after cells lose matrix contact, and thus it serves to regulate the onset of apoptosis in such cells.

Transcriptional repression and growth suppression by the p107 pocket protein.

p107 is a member of the pocket family of proteins that includes the retinoblastoma tumor suppressor. Overexpression of p107 arrests cells in G1, suggesting that it is important for cell cycle control. This growth suppression is mediated at least in part through the interaction of p107 with a member of the E2F family of cell cycle transcription factors, and this interaction can be disrupted by oncoproteins from DNA tumor viruses such as adenovirus E1a that bind p107. Not only does the binding of p107 to E2F inactivate E2F, but also we show that when p107 is tethered to the promoter through binding to E2F it functions as a general transcriptional repressor. This general repressor activity was also evident when p107 was fused to the DNA binding domain of Gal4 so that it could be directly targeted to the promoter in an E2F-independent fashion. Using p107 mutants, we compared the regions of the protein required for transcriptional repression and cell growth suppression. We found that the pocket domain is sufficient for inactivation of E2F, general repressor activity, and most of the growth suppressor activity. Binding of conserved region 1 from Ela to p107 blocked interaction with E2F, but it did not affect general repressor activity, demonstrating that binding and inactivation of E2F and general repressor activity are distinguishable properties of p107. Within the pocket, two conserved domains, A and B, were sufficient for growth suppression and transcriptional repressor activity. Surprisingly, we found that these two domains were fully functional when they were coexpressed as separate proteins, and we present results suggesting that the domains may interact at the promoter to form an active pocket.

zfh-1, the Drosophila homologue of ZEB, is a transcriptional repressor that regulates somatic myogenesis.

zfh-1 is a member of the zfh family of proteins, which all contain zinc finger and homeodomains. The roles and mechanisms of action of most family members are still unclear. However, we have shown previously that another member of the family, the vertebrate ZEB protein, is a transcriptional repressor that binds E box sequences and inhibits myotube formation in cell culture assays. zfh-1 is downregulated in Drosophila embryos prior to myogenesis. Embryos with zfh-1 loss-of-function mutation show alterations in the number and position of embryonic somatic muscles, suggesting that zfh-1 could have a regulatory role in myogenesis. However, nothing is known about the nature or mechanism of action of zfh-1. Here, we demonstrate that zfh-1 is a transcription factor that binds E box sequences and acts as an active transcriptional repressor. When zfh-1 expression was maintained in the embryo beyond its normal temporal pattern of downregulation, the differentiation of somatic but not visceral muscle was blocked. One potential target of zfh-1 in somatic myogenesis could be the myogenic factor mef2. mef2 is known to be regulated by the transcription factor twist, and we show here that zfh-1 binds to sites in the mef2 upstream regulatory region and inhibits twist transcriptional activation. Even though there is little sequence similarity in the repressor domains of ZEB and zfh-1, we present evidence that zfh-1 is the functional homologue of ZEB and that the role of these proteins in myogenesis is conserved from Drosophila to mammals.

Role of the LXCXE binding site in Rb function.

Oncoproteins from DNA tumor viruses such as adenovirus E1a, simian virus 40 T antigen, and human papillomavirus E7 contain an LXCXE sequence, which they use to bind the retinoblastoma protein (Rb) and inhibit its function. Cellular proteins such as histone deacetylases 1 and 2 (HDAC1 and -2) also contain an LXCXE-like sequence, which they use to interact with Rb. The LXCXE binding site in Rb was mutated to assess its role in Rb function. These mutations inhibited binding to HDAC1 and -2, which each contain an LXCXE-like sequence, but had no effect on binding to HDAC3, which lacks an LXCXE-like sequence. Mutation of the LXCXE binding site inhibited active transcriptional repression by Rb and prevented it from effectively repressing the cyclin E and A gene promoters. In contrast, mutations in the LXCXE binding site did not prevent Rb from binding and inactivating E2F. Thus, the LXCXE mutations appear to separate Rb's ability to bind and inactivate E2F from its ability to efficiently recruit HDAC1 and -2 and actively repress transcription. In transient assays, several of the LXCXE binding site mutants caused an increase in the percentage of cells in G(1) by flow cytometry, suggesting that they can arrest cells. However, this effect was transient, as none of the mutants affected cell proliferation in longer-term assays examining bromodeoxyuridine incorporation or colony formation. Our results then suggest that the LXCXE binding site is important for full Rb function. Mutation of the LXCXE binding site does not inhibit binding of the BRG1 ATPase component of the SWI/SNF nucleosome remodeling complex, which has been shown previously to be important for Rb function. Indeed, overexpression of BRG1 and Rb in cells deficient for the proteins led to stable growth inhibition, suggesting a cooperative role for SWI/SNF and the LXCXE binding site in efficient Rb function.

Forskolin inducibility and tissue-specific expression of the fibronectin promoter.

The mechanism of cyclic AMP (cAMP) induction of fibronectin (FN) in HT-1080 and JEG-3 cells differs (D. C. Dean, R. F. Newby, and S. Bourgeois, J. Cell Biol. 106:2159-2170, 1988). In the fibrosarcoma cell line HT-1080, induction requires both protein synthesis and a lag period of 12 to 24 h. In the choriocarcinoma cell line JEG-3, protein synthesis is not required and induction peaks before 24 h, declining thereafter. We show that the FN promoter is transcribed in vitro and that the transcripts initiate at the proper site. Based on transfection experiments with these cells and FN promoter constructions, a cAMP-responsive element (CRE) was identified between -157 and -188 base pairs upstream of the human FN gene. This sequence also conferred cAMP inducibility in both cell lines on the herpesvirus thymidine kinase promoter when it was placed upstream of a thymidine kinase-chloramphenicol acetyltransferase fusion gene. DNase I protection analysis and gel retardation experiments revealed that the CRE was bound by a protein(s) that was present in both HT-1080 and JEG-3 cells as well as in NIH 3T3 cells. Multiple protein-CRE complexes were resolved by gel retardation with extracts of both cell lines. Forskolin treatment of these cells did not alter qualitatively or quantitatively the pattern of CRE-binding proteins that was observed. The FN promoter was at least 10 times more active in HT-1080 than in JEG-3 cells, even though in JEG-3 cells both the rate of FN biosynthesis and the level of accumulated FN mRNA were greater than those in HT-1080 cells. The difference in promoter activity in HT-1080 and JEG-3 cell was mediated by sequences that were located between positions -510 and -56. Deletion of the FN promoter from positions -510 to -56 resulted in an ~30-fold decrease in promoter activity when this construction was transfected into HT-1080 cells, and similar results were observed in NIH 3T3 cells; however, less than a 2-fold effect was observed in JEG-3 cells. Results of these studies suggest that there is some degree of tissue specificity of FN gene expression and reveal that cAMP induction is mediated, in part, by the same element (CRE) in both HT-1080 and JEG-3 cells.

Mapping White Matter Microstructure in the One Month Human Brain.

White matter microstructure, essential for efficient and coordinated transmission of neural communications, undergoes pronounced development during the first years of life, while deviations to this neurodevelopmental trajectory likely result in alterations of brain connectivity relevant to behavior. Hence, systematic evaluation of white matter microstructure in the normative brain is critical for a neuroscientific approach to both typical and atypical early behavioral development. However, few studies have examined the infant brain in detail, particularly in infants under 3 months of age. Here, we utilize quantitative techniques of diffusion tensor imaging and neurite orientation dispersion and density imaging to investigate neonatal white matter microstructure in 104 infants. An optimized multiple b-value diffusion protocol was developed to allow for successful acquisition during non-sedated sleep. Associations between white matter microstructure measures and gestation corrected age, regional asymmetries, infant sex, as well as newborn growth measures were assessed. Results highlight changes of white matter microstructure during the earliest periods of development and demonstrate differential timing of developing regions and regional asymmetries. Our results contribute to a growing body of research investigating the neurobiological changes associated with neurodevelopment and suggest that characteristics of white matter microstructure are already underway in the weeks immediately following birth.

Multivariate characterization of white matter heterogeneity in autism spectrum disorder.

The complexity and heterogeneity of neuroimaging findings in individuals with autism spectrum disorder has suggested that many of the underlying alterations are subtle and involve many brain regions and networks. The ability to account for multivariate brain features and identify neuroimaging measures that can be used to characterize individual variation have thus become increasingly important for interpreting and understanding the neurobiological mechanisms of autism. In the present study, we utilize the Mahalanobis distance, a multidimensional counterpart of the Euclidean distance, as an informative index to characterize individual brain variation and deviation in autism. Longitudinal diffusion tensor imaging data from 149 participants (92 diagnosed with autism spectrum disorder and 57 typically developing controls) between 3.1 and 36.83 years of age were acquired over a roughly 10-year period and used to construct the Mahalanobis distance from regional measures of white matter microstructure. Mahalanobis distances were significantly greater and more variable in the autistic individuals as compared to control participants, demonstrating increased atypicalities and variation in the group of individuals diagnosed with autism spectrum disorder. Distributions of multivariate measures were also found to provide greater discrimination and more sensitive delineation between autistic and typically developing individuals than conventional univariate measures, while also being significantly associated with observed traits of the autism group. These results help substantiate autism as a truly heterogeneous neurodevelopmental disorder, while also suggesting that collectively considering neuroimaging measures from multiple brain regions provides improved insight into the diversity of brain measures in autism that is not observed when considering the same regions separately. Distinguishing multidimensional brain relationships may thus be informative for identifying neuroimaging-based phenotypes, as well as help elucidate underlying neural mechanisms of brain variation in autism spectrum disorders.