I MPORTANT WM AND GM MATURATIONAL CHANGES OCCUR IN THE PRETERM INFANT ’ S BRAIN

Preterm birth offers a unique window of opportunity to study early brain development. Using the recently established FBA approach applied to longitudinal dMRI data of very preterm infants, we reveal the important whole-brain WM and GM maturational changes occurring from the 33th week GA of early brain development to TEA.

FBA has the advantage of being able to characterize multiple fiber orientations within the voxel, providing microstructural metrics that are more fiber-specific than the ones obtained with the DTI model, and to perform whole-brain comparisons of the fiber-specific properties. Furthermore, we have acquired longitudinal data, what strengthens the statistical power of our analysis.

64 Regarding WM maturation, from the 33th week GA of preterm brain development to TEA, we show a significant longitudinal increase in fiber density and in fiber cross-section, and thus an increase in micro and macroscopic maturation of all major cerebral WM fibers, bilaterally, in line with the important axonal growth, neural organization and maturational processes known to occur during this developmental period (Batalle et al., 2017; Kostovic and Jovanov-Milosevic, 2006; Kostovic and Judas, 2010a; Kostovic et al., 2014; van den Heuvel et al., 2015). Increases in WM fiber cross-section throughout this period can reach 50%, while fiber density augments by 20%, what suggests that, beyond the augmentation in intra-axonal volume/packing, there is a particularly important augmentation in fiber bundles spatial extent. Such can be due to the maturational processes taking place during this period, such as myelination, leading to an important increase of fiber bundles diameter (Baumann and Pham-Dinh, 2001;

Volpe, 2001a). Projection fibers are the ones exhibiting the highest increase in fiber density and, specially, in fiber cross-section, which supports the hypothesis that the observed increases in these metrics might be related to myelination processes, known to have a particular spatial-temporal pattern, starting in the central projection fibers (Kinney et al., 1988).

Fiber density and fiber cross-section metrics increase significantly also in the brainstem, cerebellum and cerebellar peduncles, where projection fibers pass and early myelination occurs. Furthermore, also deep GM undergoes important maturational changes. In fact, there is a significant increase regarding FBA metrics in the thalamus, from where main ascending projection fibers originate, leading to the establishment of functional thalamo-cortical connections during this critical developmental period (Batalle et al., 2017; Kostovic and Jovanov-Milosevic, 2006;

Kostovic and Judas, 2010a; Krsnik et al., 2017; Pouchelon and Jabaudon, 2014).

In addition, important maturational changes occur in the cortical GM, as is unveiled by this longitudinal FBA analysis. From the 33th week GA to TEA, there is a significant increase of fiber cross-section, accompanied by a decrease of fiber density in various cortical regions, distributed among frontal, parietal, temporal, insula, cingulate and occipital cortices. In fact, in terms of cortical brain development, the third trimester of pregnancy is characterized by a significant increase in cortical plate complexity, marked by an important growth of the dendritic tree structure, with basal dendrites cross-connections running in parallel to the cortical surface, an increasing number of neuronal connections, including in-growth of thalamo-cortical afferents, transformation of the radial glial cells into astrocytes and proliferation of glial cells (Bystron et al., 2008). There is, therefore,

65 an important configurational change, from a highly structured cortical plate, with the apical dendrites and radial glia fibers creating a directional and columnar microstructure setting (Ball et al., 2013b; McKinstry et al., 2002; Ouyang et al., 2019a), to a more complex multidirectional microstructural environment, what could justify the observed decrease of fiber density. Fiber density is a directionally-dependent metric, being proportional to the volume of the fibers aligned in that direction (Raffelt et al., 2017), which, in GM, may comprise multiple cellular components sharing the same orientation, namely apical dendrites, axons and astrocytes processes (Budde et al., 2011; Budde and Annese, 2013). Therefore, its calculation might be affected by the increase in cortical geometric complexity, similarly to what is observed when evaluating cortical FA changes during early brain development, which is known to diminish throughout early cortical maturation (Ball et al., 2013b; Batalle et al., 2019; Eaton-Rosen et al., 2015; McKinstry et al., 2002).

Furthermore, this decrease of fiber density in cortical GM is also in agreement with the observed decrease in cortical NDI, which has been described during early brain development until TEA (Batalle et al., 2019). This reduction of NDI in cortical GM has been related to the cortical expansion occurring during this period (Huttenlocher, 1990), as well as cell death or apoptosis, what could explain a reduction of the neuronal density in the cortical plate (Chan and Yew, 1998; Lossi and Merighi, 2003). The observed regional significant increase in cortical FC might be related to an augmentation of the number of basal dendrites and glial cells in cortical GM areas, as well as myelination of intracortical axons. Indeed, a decreased FC has been related to decreased myelination (Gajamange et al., 2018), what has been confirmed by histological staining (Malhotra et al., 2019). Furthermore, the highest FC increase was observed in primary sensorial and motor processing areas, which have been proven to be the first cortical regions to demonstrate the presence of intra-cortical myelin (Huttenlocher and Dabholkar, 1997b; Rowley et al., 2017).

The observed longitudinal increase of fiber cross-section accompanied by a decrease of fiber density in various cortical GM regions was complemented by a NODDI analysis, showing that, from the 33th week GA to TEA there is a significant ODI increase in cortical GM, what has been related to increased dendritic branching and cortical complexity (Batalle et al., 2019).

The fact that the primary motor, somatosensory, visual and auditory cortices are among the various cortical regions undergoing significant maturational changes is in line with the establishment of the activity-dependent thalamo-cortical connectivity during this period, transmitting the environmental extrinsic input (from sensory periphery) through the sensory thalamic nuclei to the cerebral cortex,

66 holding therefore the potential to modulate cortical maturation (Kostovic and Judas, 2007; Milh et al., 2007).

Our data acquired longitudinally throughout this early period of preterm brain development has allowed thus, for the first time, to measure cortical FBA metrics, providing new in vivo imaging biomarkers of major maturational microstructural cortical changes.

Our results, specifying the major brain WM and cortical GM maturational changes occurring typically during the third trimester of pregnancy, complement and are in line with previous findings. In WM fibers, in particular, previous findings have revealed mainly an increase of NDI and FA, in parallel to a decrease of MD and RD during early brain development, consequently to an increase in fiber organization and preliminary myelination (Aeby et al., 2009; Batalle et al., 2017;

Dean et al., 2017; Dubois et al., 2008; Huppi et al., 1998a; Kunz et al., 2014;

Mukherjee et al., 2002; Nossin-Manor et al., 2013; Shim et al., 2012). On the other hand, in cortical GM, studies have shown mostly a decrease of FA and an increase in ODI during early brain development, in relation with the increased dendritic density and general increase in cortical complexity (Ball et al., 2013b; Batalle et al., 2019; McKinstry et al., 2002). Such significant maturational processes may underlie the important brain growth and significant increase of both GM and WM volumes observed during the third trimester of pregnancy and brain development (Brody et al., 1987; Dubois et al., 2008; Huang et al., 2006; Huppi et al., 1998b; Nossin-Manor et al., 2013).

3.3. Music enhances maturation of brain correlates of emotional