24th November 2023

Neuroimaging and Genetics of Development

Embryology

The period of brain development least well informed by human neuroimaging is the embryological period – that is, the first 8 weeks following conception. Instead, much of what is known about brain maturation during this time comes from rodent and non-human primate histology (and relatively few post-mortem human studies). 

During weeks 1 and 2 of development, the fertilised egg divides repeatedly to become the morula (a solid 16- to 32-celled mass) and then the blastocyst (a fluid-filled 50- to 150-celled mass), which subsequently embeds within the uterine wall. During week 3, the two-layered embryo becomes a three-layered structure through the process of gastrulation. Gastrulation gives rise to different stem cell lines, such as neural stem cells – so-called neural progenitor cells – arranged as the neural plate. 

Neurulation

During days 20–27, the outer ridges of the neural plate fold inwards and fuse to form the hollow neural tube in a process termed neurulation. Meanwhile, the release of molecular signals from elsewhere in the embryo encourages the rostral–caudal (head-to-toe) organisation of this primitive nervous system. Neural tube closure begins medially and extends axially, with the rostral (anterior) and caudal (posterior) neuropores closing on days 25 and 27, respectively. 

Just prior to closure, however, the rostral end of the neural tube forms 3 brain vesicles: the prosencephalon, the rhombencephalon and the mesencephalon. The prosencephalon further divides to form the telencephalon and diencephalon, which give rise to the forebrain; the rhombencephalon splits into metencephalon and myelencephalon, which later become hindbrain; and the mesencephalon becomes the midbrain (without subdivision). This process is completed during week 7. 

From days 25 to 42, rapid, symmetrical division of neural progenitor cells significantly increases their number. Thereafter, asymmetrical division gradually increases the number of new neuronal cells relative to neural progenitor cells. Cell division occurs in ‘proliferative zones’ such as the ventricular zone and subventricular zone (so-called due to their anatomical proximity to the embryological origins of the lateral ventricles).

Foetal Development 

The period of foetal development spans from week 9 of gestation until term. 

Migration

During weeks 12 to 20, vast numbers of proliferated neurons migrate from the ventricular and subventricular zones to different areas of the growing cortex. The earliest neurons may migrate short distances via somal translocation, extending long processes and rearranging intracellular structures to ‘reach’ their target destination. However, migration mostly proceeds along scaffolding structures provided by glial (supporting) cells. 

The earliest neurons to migrate generally constitute the deepest layers of the forming cortex, with later neurons forming more superficial layers. Four transient layers of neurons are initially formed – the preplate, marginal zone, subplate and cortical plate. The subplate and cortical plate are of particular importance: migrating neurons will initially synapse with subplate neurons, where they receive connections from neurons in immature thalamic and brainstem areas but later form permanent synapses with cortical plate neurons and break away from the subplate. The redundant subplate eventually degenerates, whilst the cortical plate gives rise to layers I-VI of the neocortex. 

Structural MRI (sMRI) has been used to confirm such migration events. You will recall that sMRI utilises a large magnet and radiofrequency pulses to influence the natural magnetic properties of protons in tissues in order to create detailed anatomical images. T1 sequences enhance fat signal, whereas T2 sequences enhance water signal. When applied to the examination of the foetus, sMRI is sometimes simply known as ‘foetal MRI’. 

By week 20, foetal MRI can distinguish the cortical plate (low T2 signal) from subplate (high T2 signal). Studies show that the subplate is at its thickest (approximately four-times that of the cortical plate) at week 22. After week 30, however, the subplate gradually disappears, with accompanying reductions in proliferative zone volumes on MRI. The cortical plate grows; meanwhile, driving increases in cortical thickness. Helpfully, one research group recently published a full ‘spatiotemporal atlas’ of brain changes between gestational weeks 21 and 39, available to view online (see Gholipour et al. 2017, Scientific Reports). 

Myelination

During myelination, oligodendrocytes and Schwann cells ensheath central and peripheral nervous system axons, respectively, with fatty myelin. The primary function of the myelin sheath is to speed up the axonal propagation of action potentials via saltatory conduction. 

Myelination begins by the fifth month of gestation and is mostly complete by the second year of life but continues into the third decade of life in some areas (e.g. corpus callosum). Myelination begins in the spinal cord / at exiting motor roots and progresses rostrally. Sensory pathways generally myelinate earlier than motor pathways, and association areas myelinate last. Some fibres will never undergo myelination. 

Again, much of what is known about myelination comes from non-human and post-mortem human histology. That said, diffusion tensor imaging (DTI) – an MRI technique useful for the assessment of white matter microstructure – has been used to study myelination as early as 18 weeks gestation. Within DTI, fractional anisotropy (FA) describes the extent to which water molecules diffuse along axons (versus freely in all directions), where FA values (between 0 and 1) are directly proportional to axon/tract density. DTI (by FA) studies in foetuses aged >23 weeks corroborate that myelination likely begins in corticospinal tracts before progressing to the optic radiations and corpus callosum. Neurite orientation dispersion and density imaging (NODDI) is a diffusion MR technique specifically developed for brain MRI and may image human neurodevelopment with greater sensitivity than standard DTI.

Gyrification

Gyrification describes developmental folding of the cortex that results in characteristic gyri (peaks) and sulci (troughs). Gyrification is best studied in vivo by sMRI. The degree of cortical folding in a given brain is described by the gyrification index: a ratio of the lengths of the long, folded cortical contour and the shorter, smooth brain surface. The greater the gyrification index, the higher the degree of folding, and vice versa.

Using foetal MRI, cortical folding patterns have been identified from 20 weeks gestation. Several large sMRI studies have examined gyrification cross-sectionally (that is, imaging many individuals of different ages at one point in time) or longitudinally (i.e. repeatedly imaging (typically fewer) individuals at numerous time points). Both cross-sectional and longitudinal sMRI studies find that the gyrification index increases after birth, peaks at around 2 years of age, and then decreases throughout adulthood (at a rate of approximately 1% per year). Infantile increases in gyrification are faster in the first than the second year of life and are faster in association areas than in visual, auditory and sensorimotor cortex. Subsequent reductions in folding are most pronounced in frontal and occipital cortices and might be slightly slower in males than in females. 

Imaging Synaptic Density

Seminal work from the 1980s shows that synaptic densities follow a trajectory similar to that of the gyrification index. Through histological staining of synaptic proteins in post-mortem human tissue, Huttenlocher and colleagues found that synapse density rapidly increased during the first year of life, whereafter it gradually fell with increasing age. 

This overlay of synaptic and gyrification dynamics is no coincidence. Rather, it is believed that synaptic events determine local patterns of cortical folding. Migrating neurons synapsing with each other likely pull neighbouring areas of cortex together into gyri (with intervening sulci), whereas synaptic pruning may release such tethering connections, reducing folding. 

The particularly pronounced decrease in synaptic density during childhood and adolescence is termed synaptic pruning and is considered central to appropriate neurodevelopment. Preclinical and post-mortem work has long implicated aberrant synaptic pruning in some psychiatric illnesses (such as schizophrenia and autism), but until recently, only indirect in vivo imaging approaches have been possible.

FDG PET

One such example is fluorodeoxyglucose positron emission tomography (FDG PET). Here, a radiolabeled glucose analogue is injected intravenously and crosses the blood–brain barrier, then is taken-up by highly metabolising cells before the isotope decays and resultant radiation is detected by a surrounding scanner. Thus, FDG PET signal may be considered a proxy for neuronal metabolic rate and – as neurotransmission accounts for approximately 40% of brain energy expenditure – synaptic density, too. 

FDG PET studies of age-related changes in brain metabolic rate report inconsistent findings in children and adolescents. Perhaps the largest study to date included 111 individuals aged 5 months to 23 years. Here, the authors found an inverted U-shape relationship between FDG PET signal (normalised to a reference region – the cerebellum) and increasing age in >20 brain regions, broadly mapping the synaptic trajectories discussed above.

NAA MRS

Magnetic resonance spectroscopy (MRS) is the application of nuclear magnetic resonance (NMR) spectroscopy to study tissue biochemistry using MRI. MRS is commonly used to assess brain levels of the metabolite N-acetyl aspartate (NAA) – a marker of neuronal integrity whose exact function remains unclear.

It has been suggested that NAA levels may reflect synapse number and function in vivo. However, half-a-dozen MRS studies have found ongoing increases in NAA levels throughout childhood and adolescence (which may level off during adulthood and remain fairly stable during later life) – i.e. not the curvilinear trajectory of synaptic density described histologically and supported by FDG PET work. Thus, NAA may be a relatively poor proxy measure for synaptic density, whose MRS signal is heavily influenced by other neuronal factors, some of which remain unknown.

UCB-J PET

Fortunately, a recently developed PET tracer now enables us to image synapses directly. In UCB-J PET, UCB-J is labelled with a radioactive isotope of carbon (11C). It is then injected intravenously, selectively binds synaptic vesicle glycoprotein 2A (SV2A) – the cellular target of levetiracetam – and is imaged in a process similar to that described for FDG PET. As SV2A is specifically located in presynaptic axon terminals, the resultant PET signal is directly proportional to total (regional or global) synaptic density. 

Interestingly, UCB-J PET studies to date show no age-related decline in regional synaptic density in healthy volunteers. However, few studies have yet been completed and only include participants >18 years old. Moreover, one UCB-J PET study did find significant reductions in synaptic density with increasing age – but in several neural networks (as defined by a sophisticated data-led approach), not in predetermined regions of interest. Meanwhile, UCB-J PET reliably tracks regional synaptic changes in illness. Consistent with prior FDG PET (and NAA MRS) work, case–control UCB-J PET studies find that cortical synaptic density is considerably reduced in schizophrenia. UCB-J PET studies also show reduced synaptic density in the hippocampus in Alzheimer’s disease and substantia nigra in Parkinson’s disease.

White and Grey Matter Trajectories

We have already discussed that DTI (by FA) is an established MRI technique for imaging myelination events from as early as the second trimester. DTI may also be used then, to study white matter tracts (bundles of myelinated axons connecting different brain regions). This is commonly through DTI tractography – 3D reconstruction of DTI data to visualise specific fibres and tracts. Prior to the development of DTI tractography, T1 and T2 sMRI sequences were commonly manipulated to provide white matter insights. 

Studies taking each approach show that major white matter tracts, including many fibres projecting from subcortical to cortical structures, first develop during the fourth or fifth month of gestation. Thalamocortical and association tracts develop after, during the seventh gestational month or later. All major white matter tracts are established by gestational term. However, total brain white matter volume gradually increases until 30–40 years of age. 

Whilst DTI tractography is useful to map structural connectivity within the brain, functional MRI (fMRI) can tell us about functional connectivity. In both blood oxygen level dependent (BOLD) and arterial spin labelling (ASL) fMRI, changes in blood flow are understood to represent alterations in regional neural activation. In resting state fMRI (rs-fMRI), changes in blood flow are not provoked by activity such as mental or physical tasks. rs-fMRI studies show that canonical sensory, motor and association networks – such as the visual, auditory, somatomotor, default mode and dorsal attention networks – are evident as early as 26 weeks gestation. These networks mature during childhood and adolescence, during which time longer-range connections emerge (e.g. between anterior and posterior regions of the cortex) and shorter-range connections weaken. The network most studied developmentally is the default mode network (DMN), yet the exact age at which the DMN emerges is disputed. 

Grey matter (regions of neuronal bodies, dendrites and unmyelinated axons) is best imaged by sMRI. Studies show that grey matter volume, thickness and surface area increase early on in childhood but shrink thereafter (from 8–10 years of age). This trajectory is determined by synaptic dynamics – the periods of rapid synaptogenesis followed by synaptic pruning discussed earlier – combined with progressive myelination (decreasing the volume of grey matter relative to white matter).

Genetics

Brain structure is highly heritable. This was first demonstrated by twin studies showing that pairs of monozygotic twins have more closely resemblant brains than dizygotic twins. Total brain volume is more than 80% heritable: consistently more heritable in white than grey matter and in some regions over others (e.g. most of the cerebrum > cerebellum). Interestingly, from 5 to 18 years of age, the heritability of total grey matter volume decreases, whilst the heritability of total white matter volume increases. Waddington suggested that genes ‘buffer’ against environmental influences on development, where highly ‘buffered’ traits – such as brain volume – are said to be highly ‘canalised’. Meanwhile, Gottesman’s ‘range of reactions’ theory posits that one’s genotype specifies limits within which the environment determines a specific phenotype. 

Genes are important in other aspects of neurodevelopment besides brain volumes. For instance, the protein products of the genes EMX2 and PAX6 are transcription factors involved in the foetal migration of neurons. It is perhaps unsurprising that single-gene mutations are sufficient to cause lissencephaly – a pathologically smooth brain, considered a disorder of neuronal migration – for which the genes LIS1 and DCX are causative. 

However, few psychiatric illnesses result from single-gene mutations; most result from the cumulative influence of several heritable and environmental factors, which likely interact. Plomin, DeFries and Loehlin (as later adapted by Scarr and McCartney) proposed that gene–environment interactions may be passive (where the individual’s genotype matches their host environment), reactive/evocative (where others respond in a certain way to an individual’s particular genotype) or active (where the individual seeks an environment that aligns with their genotype). 

Epigenetics encompasses all mechanisms of gene control that do not involve alterations to the DNA sequence and is thought to involve specific chemical changes to structures supporting DNA and chromosome structure (e.g. histone proteins). Epigenetic mechanisms may thus be considered the molecular link between genes and environmental influences in neurodevelopment.

References

• Berger A. (2002). Magnetic resonance imaging. BMJ. 324(7328):35. doi: 10.1136/bmj.324.7328.35. 
• Detre J & Wang J. (2002). Technical aspects and utility of fMRI using BOLD and ASL. Clinical Neurophysiology. 113(5):621-34. doi: 10.1016/s1388-2457(02)00038-x.
• Douet et al. (2014). Genetic influences on brain developmental trajectories on neuroimaging studies: from infancy to young adulthood. Brain Imaging and Behaviour. 8(2):234-50. doi: 10.1007/s11682-013-9260-1.
• Fogliarini C et al. (2005). Assessment of cortical maturation with prenatal MRI. Part I: Normal cortical maturation. European Radiology. 15(8):1671-85. doi: 10.1007/s00330-005-2782-1. 
• Gholipour A et al. (2017). A normative spatiotemporal MRI atlas of the fetal brain for automatic segmentation and analysis of early brain growth. Scientific Reports. 7(1):476. doi: 10.1038/s41598-017-00525-w.
• Gottlieb, G. (2003). On Making Behavioral Genetics Truly Developmental. Human Development. 46(6):337–355. https://www.jstor.org/stable/26763768.
• Guimaraes CVA & Dahmoush HM (2022). Fetal Brain Anatomy. Neuroimaging Clinics of North America. 32(3):663-681. doi: 10.1016/j.nic.2022.04.009. 
• Huttenlocher PR et al. (1982). Synaptogenesis in human visual cortex–evidence for synapse elimination during normal development. Neuroscience Letters. 33(3):247-52. doi: 10.1016/0304-3940(82)90379-2. 
• Howes OD et al. (2023). Neuroimaging in schizophrenia: an overview of findings and their implications for synaptic changes. Neuropsychopharmacology. 48(1):151-167. doi: 10.1038/s41386-022-01426-x. 
• Khan YS & Ackerman KM. (2022). Embryology, Week 1. In: StatPearls [Internet], Treasure Island (FL), StatPearls Publishing. Available from: https://www.ncbi.nlm.nih.gov/books/NBK554562/.
• Lebel C & Deoni S. (2018). The development of brain white matter microstructure. Neuroimage. 182:207-218. doi: 10.1016/j.neuroimage.2017.12.097. 
• Li G et al. (2014). Mapping longitudinal development of local cortical gyrification in infants from birth to 2 years of age. Journal of Neuroscience. 34(12):4228-38. doi: 10.1523/jneurosci.3976-13.2014. 
• Marsh R, Gerber AJ & Peterson BS. (2008). Neuroimaging studies of normal brain development and their relevance for understanding childhood neuropsychiatric disorders. Journal of the American Academy of Child and Adolescent Psychiatry. 47(11):1233-51.
• Moffett JR et al. (2007). N-Acetylaspartate in the CNS: from neuro diagnostics to neurobiology. Progress in Neurobiology. 81(2):89-131. doi: 10.1016/j.pneurobio.2006.12.003.
• Plomin R, DeFries JC & Loehlin JC. (1977). Genotype-environment interaction and correlation in the analysis of human behaviour. Psychological Bulletin. 84(2):309-22.
• Raznahan A et al. (2011). How does your cortex grow? Journal of Neuroscience. 31(19):7174-7. doi: 10.1523/jneurosci.0054-11.2011. 
• Sasabayashi D et al. (2021). Anomalous brain gyrification patterns in major psychiatric disorders: a systematic review and transdiagnostic integration. Translational Psychiatry. 11(1):176. doi: 10.1038/s41398-021-01297-8. 
• Shan ZY et al. (2014). Cerebral glucose metabolism on positron emission tomography of children. Human Brain Mapping. 35(5):2297-309. doi: 10.1002/hbm.22328.
• Smyser CD & Neil JJ. (2015). Use of resting-state functional MRI to study brain development and injury in neonates. Seminars in Perinatology. 39(2):130-40. doi: 10.1053/j.semperi.2015.01.006. 
• Stein JL, Hibar DP & Thompson PM. (2016). Genetics of Brain Structure. In: Neuroimaging Genetics: Principles and Practices (New York, 2016; online edn.). ISBN: 9780199920211.003.0010.
• Stiles J. (2010). The basics of brain development. Neuropsychology Review. 20(4):327-48. doi: 10.1007/s11065-010-9148-4. 
• Townsend L et al. (2022). Brain glucose metabolism in schizophrenia: a systematic review and meta-analysis of 18FDG-PET studies in schizophrenia. Psychological Medicine. 1-18. doi: 10.1017/S003329172200174X. 
• Toyonaga T et al. (2022). PET Imaging of Synaptic Density: Challenges and Opportunities of Synaptic Vesicle Glycoprotein 2A PET in Small Animal Imaging. Frontiers in Neuroscience. 16:787404. doi: 10.3389/fnins.2022.787404. 
• Vértes PE & Bullmore ET. (2015). Annual research review: Growth connectomics – the organisation and reorganisation of brain networks during normal and abnormal development. Journal of Child Psychology and Psychiatry. 56(3):299-320. doi: 10.1111/jcpp.12365.
• Zanin E et al. (2011). White matter maturation of normal human fetal brain. An in vivo diffusion tensor tractography study. Brain & Behaviour. 1(2):95-108. doi: 10.1002/brb3.17. 
• Zhao X et al. (2021). Brain Development From Newborn to Adolescence: Evaluation by Neurite Orientation Dispersion and Density Imaging. Frontiers in Human Neuroscience. 15:616132. doi: 10.3389/fnhum.2021.616132.