Hemispheric asymmetries in cortical gray matter microstructure identified by neurite orientation dispersion and density imaging

Highlights

• NODDI reveals extensive microstructural gray matter asymmetries in the human brain.

• Microstructural asymmetries are in line with previous histological studies.

• NODDI is a promising marker for genetic and behavioral studies on laterality.

Abstract

Histological studies have reported microstructural hemispheric asymmetries in several cortical areas of the human brain, but reliable in vivo assessment methods have been lacking so far. Here, we used neurite orientation dispersion and density imaging (NODDI) to examine microstructural asymmetries in in vivo and determine if findings are in accordance with what has been reported in histological studies. We examined intra-neurite volume fraction (INVF), neurite orientation dispersion (ODI), and isotropic volume fraction (ISO) asymmetries in two independent samples of healthy adults (n = 269 and n = 251). Over both samples, we found greater left-hemispheric INVF in early auditory, inferior parietal and temporal-parietal-occipital areas. In contrast, we found greater right-hemispheric INVF in the fusiform and inferior temporal gyrus, reflecting what has been reported in histological studies. ODI was asymmetric towards the left hemisphere in frontal areas and towards the right hemisphere in early auditory areas. ISO showed less pronounced asymmetries. There were hardly any effects of sex or handedness on microstructural asymmetry as determined by NODDI. Taken together, these findings suggest substantial microstructural asymmetries in gray matter, making NODDI a promising marker for future genetic and behavioral studies on laterality.

Introduction

The left and right hemispheres of the human brain differ in several functional aspects, the most popular being handedness and language lateralization (Güntürkün and Ocklenburg, 2017; Ocklenburg and Güntürkün, 2018). However, the ontogenetic determinants of functional hemispheric asymmetries remain rather unclear (Armour et al., 2014; Brandler et al., 2013; Ocklenburg et al., 2014; Schmitz et al., 2017). In addition to gray matter asymmetries (Ocklenburg et al., 2016b), the structure of inter- and intrahemispheric white matter pathways have been proposed to contribute to functional hemispheric asymmetries (Ocklenburg et al., 2016a). On the microstructural level, neurites, i.e. dendrites and axons, constitute the “building blocks of the computational circuitry of the brain” (Zhang et al., 2012). Neurite orientation dispersion and density imaging (NODDI) is a diffusion MRI technique for in vivo quantification of neurite morphology in humans (Zhang et al., 2012). Specifically, it allows for the estimation of neurite density (intra-neurite volume fraction, INVF), neurite orientation dispersion (ODI), a neurite tortuosity measure, and isotropic volume fraction (ISO). So far, no study has investigated hemispheric asymmetries in microstructure in vivo. However, the detection of microstructural asymmetry is an interesting marker not only for genetic and behavioral studies on hemispheric asymmetries, but also in the context of neurodevelopmental and psychiatric traits. Atypical microstructural asymmetry has been reported in schizophrenia patients (Chance, 2014; Chance et al., 2008; Cullen et al., 2006; Simper et al., 2011) and altered microstructure in autism spectrum disorder (Casanova et al., 2006; van Kooten et al., 2008) has been suggested to underlie atypical functional hemispheric asymmetries (Chance, 2014). In order to provide a reference for potential clinical studies, microstructural asymmetries in healthy subjects have to be established. Thus, the aim of the present study was the first whole brain investigation of gray matter microstructure by applying NODDI to two independent datasets constituting an overall sample of 520 healthy adults.

Several post mortem studies have reported hemispheric asymmetries in microstructure (Schenker et al., 2009) with focus on different brain areas and different markers for microcircuitry (for a comprehensive overview see Table 1). The analysis of cell density has been proposed as an indirect way to estimate the amount of neuropil, i.e. dendrites, axons and synapses. For example, the gray-level index (GLI) estimates the fraction of tissue volume that is occupied by Nissl-stained cell bodies with decreased cell density indicating an increased proportion of neuropil (Amunts et al., 1996). A common feature of mammalian cytoarchitecture is the columnar organization of neurons surrounded by neuropil across cortical layers VI to II. Single vertical rows of neurons with a width of about 60–90 μm are known as minicolumns. Macrocolumns are constituted of several minicolumns, forming thalamocortical or corticocortical columns with a width of about 600–800 μm (Hutsler and Galuske, 2003). Seldon (1981a) suggested greater distance between minicolumns to reflect a higher amount of neuropil. Few studies have investigated hemispheric asymmetries in dendritic branching directly using Golgi impregnations (Seldon, 1982).

Previous histological studies report greater right-hemispheric neuron density in the planum temporale (parts of BA 22, BA 41, and BA 42) (Buxhoeveden et al., 2001; Smiley et al., 2011) as well as greater left-hemispheric average minicolumn width. Moreover, there is a greater average distance between mini- (Buxhoeveden et al., 2001; Chance et al., 2006; Economo and Horn, 1930; Seldon, 1981a) and macrocolumns (Galuske et al., 2000) in the left hemisphere that is compensated for by greater left-hemispheric dendritic spread (Seldon, 1981a, 1981b) and dendrite density (Seldon, 1982). For the mid-posterior fusiform region (BA 37), greater right-hemispheric neuron density and greater left-hemispheric minicolumn width have been reported (Chance et al., 2013). Greater minicolumn width has also been reported for the right-hemispheric anterior fusiform cortex including BA 20 (Di Rosa et al., 2009). In the frontal lobes, Broca's area (BA 44 and 45) is characterized by a greater GLI in the right hemisphere (Amunts et al., 1999; Amunts et al., 2003). In the primary motor cortex (BA 4), right-hemispheric GLI has been found to be greater than left-hemispheric GLI (Amunts et al., 1996; Amunts et al., 1997). In the superior frontal gyrus (BA 9), greater cell density was found in the left hemisphere (Cullen et al., 2006), while no asymmetry was found in a subsequent study only investigating male brains (Smiley et al., 2011). In the occipital and parietal lobes, analysis of the primary visual cortex (V1, BA 17), V2 (BA 18), and area hOc5, an area sensitive to motion, showed a small interhemispheric asymmetry with lower GLI in the left hemisphere (Amunts et al., 2007). In the inferior parietal lobe, greater neuron density was found in the right supramarginal gyrus (BA 40) (Smiley et al., 2012). One important question is if findings from post mortem studies can be replicated in vivo. Thus, the second aim of the present study was to compare the results obtained with NODDI with findings from previous histological studies.