Diffusion tensor imaging (DTI)
Water diffusion can occur equally in all directions (isotropic diffusion), for example in cerebrospinal fluid where diffusion is not restricted or in brain tissue where water diffusion is restricted similarly in all directions (e.g., gray matter tissue which has a complex cellular structure). Water diffusion is called anisotropic (preferentially diffusing in one direction) where the brain tissue microstructure contains fibres that are aligned (e.g., white matter fibre tracts); in that case, water diffusion will preferentially occur along the axis of the fibre tracts. An ellipsoid model of anisotropic water diffusion tensor (describing linear associations between vectors) can be calculated for each anatomical voxel (the smallest volumetric unit of brain images).
DTI is thus an in vivo brain imaging tool that provides an index of the micro-structural integrity of white matter tissue . Mean diffusivity (MD) provides a general measure of water diffusion without differentiating the direction of diffusivity. Another measure called fractional anisotropy (FA), when found to be high, will indicate a preferred direction of water diffusion in the region of interest [22, 23]; when found to be reduced relative to normative data, it broadly suggests reduced white matter integrity . Ultrastructural studies directly comparing DTI parameters with tissue pathology have associated changes in DTI water diffusivity measures with dysmyelination of white matter tracts; other tissue alterations that influence water diffusivity are axonal pathology and changes in cell densities .
There have been over 60 studies using DTI to evaluate white matter integrity in schizophrenia . The majority of these studies have focussed on chronic schizophrenia and have reported evidence of multiple areas of white matter disruption most notably in the corpus callosum, prefrontal white matter, SLF, and cingulum bundle [26, 27]. There have been fewer DTI studies of early phase schizophrenia. This cohort is however extremely important, as it allows for the investigation of pathology core to the illness, with minimal impact of confounders such as medication, age, and length of time with illness.
DTI in early phase schizophrenia
DTI studies of early phase schizophrenia have yielded inconsistent findings. In a review of this literature (2010), it was observed that for each white matter fibre tract that was found abnormal in the clinical sample relative to the normative sample, there was at least one other, negative research report . In more recent studies (2010–2013), the pattern of mixed findings remains, however with fewer reports of negative findings [28, 29] than positive findings [30–38]. Altogether, several different white matter tracts have been reported as disrupted in schizophrenia at different stages of the illness, supporting the hypothesis that white matter deficits could possibly be widespread throughout the whole brain . In early phase schizophrenia more specifically, the three white matter tracts most often implicated are the SLF [31, 34, 37, 39–42], the splenium of the corpus callosum [31, 33, 34, 39, 40, 43–45], and the fronto-occipital fasciculus [25, 31, 34, 37, 39, 41–43, 46].
DTI in cannabis users without schizophrenia
Different lines of evidence support the assumption that early cannabis use in a developing (adolescent) brain could be markedly more damaging than in a more mature brain (with ‘early users’ defined as those below age 17 years; [4, 47]). In healthy volunteers, a greater detrimental impact of early initiation of regular cannabis use (relative to a later initiation) has been reported for visual reaction times , cognitive performance , and volumetric brain tissue abnormalities .
DTI studies have reported that early adolescent regular cannabis use in otherwise healthy young adults is associated with reduced FA values in white matter tracts involving fronto-temporal connections , and with increased mean diffusivity (MD) in the prefrontal section of the corpus callosum . MD quantifies water diffusion in each voxel and is increased when there is reduced white matter integrity. It is thus possible that early onset cannabis use in adolescence might decrease white matter structural integrity in otherwise healthy individuals. The white matter fibre tracts most often reported as abnormal in cannabis users include the SLF [51, 53–56], the corpus callosum [52, 53, 57, 58], and more broadly defined temporal [51, 53, 56, 58] and frontal regions [53, 54, 57].
DTI in cannabis users with schizophrenia
A recent review of epidemiological evidence found that onset of cannabis use in early adolescence is associated with a particularly increased risk of developing schizophrenia , while the lifetime rate of cannabis use use in adults with schizophrenia is associated with earlier onset of the illness .
In early adolescent onset of schizophrenia, cannabis-positive patients showed reduced FA values relative to cannabis-negative patients in several white matter tracts including the SLF . In people with first episode psychosis, cannabis-naive patients had reduced FA in the corpus callosum, relative to patients with early onset of cannabis use and healthy controls . On the other hand, patients with recent onset schizophrenia and early adolescent cannabis use had increased FA values in temporal and frontal regions, relative to healthy controls; no differences were found between controls and patients without early adolescent cannabis use .
Although some findings go in opposite direction (decreased and increased FA values), altogether the empirical DTI data supports the assumption of a greater detrimental effects of cannabis on an immature brain in both healthy volunteers and patients with recent onset schizophrenia. In addition to being a potential “second hit” for psychosis in a vulnerable brain, failure to control for this confounding variable could underlie the inconsistent findings in previous DTI studies of early phase schizophrenia and would demand for this variable to be factored into future studies of white matter abnormalities in psychosis.
Proton magnetic resonance spectroscopy (1H-MRS)
Another neuroimaging technique that will be used in this study is 1H-MRS, which will be acquired from the same targeted brain region (the left SLF). The neurochemical of interest is N-acetylaspartate (NAA), a free amino acid that produces the most prominent resonance in 1H-MRS of the human brain: a peak located at 2.02 ppm on the spectral profile .
In vivo concentration levels of NAA are slightly higher in white matter relative to gray matter tissue . Post-mortem studies have demonstrated that NAA is synthesized in neurons, transported into white matter and then catabolized into aspartate and acetate in oligodendrocytes via aspartoacylase . NAA catabolism is therefore closely linked to myelin lipid metabolism, as it provides a very important source of acetate which is crucial for myelin lipid production and maintenance [64, 65].
The assumption of abnormal myelin biosynthesis in schizophrenia, strongly supported by several different lines of evidence [8, 9, 66], can thus be examined by 1H-MRS studies as long as the targeted brain region involves a single tissue type (white matter) that allows a meaningful interpretation of findings in terms of the catabolic cycle of NAA.
Noteworthy for this study, the cannabinoid receptors CB1 are present on astrocytes and oligodendrocytes and may thus be implicated in the detrimental impact of early adolescent cannabis use by affecting the trajectory of white matter development in the critical period of adolescence [4, 53].
1H-MRS: technical limitation
Due to the low concentration (mM) of the neurochemicals detected by 1H-MRS, localized spectroscopy studies generally sample a relatively large brain volume and require averaging several acquisitions to build up a reasonable signal-to noise ratio from the brain volume of interest. As such, most previous 1H-MRS clinical studies have reported on NAA signals originating from both gray and white matter tissues taken together as a whole; unfortunately, this approach has prevented the interpretation of findings in regards to the specific anabolic and catabolic activities of NAA.
In this proposed study, we will sample a large brain volume comprised of 95% white matter. Our previous 1H-MRS data acquired in the same brain region has demonstrated that across more than 150 brain scans acquired with these anatomical landmarks, the mean (SD) fractional content of white matter was 95(2.8)% (unpublished). These 1H-MRS data will thus provide insight about the specific catabolic cycle of NAA in early phase schizophrenia and consequently, about the regional availability of acetate which is required for biosynthesis of myelin. In this context, regional levels of NAA can be considered a marker of myelin integrity.
Previous relevant 1H-MRS studies
In adolescent chronic cannabis users, reductions in NAA concentration levels were reported, relative to non-user controls, in the anterior cingulate region encompassing mainly gray matter . Levels of NAA might also be altered in schizophrenia but findings are inconsistent across studies, thus inconclusive. Indeed, if we compile previous 1H-MRS studies of schizophrenia while selecting studies with the best contemporary methods (those that referenced neurochemical levels to internal water and that used a sample size of 20 subjects or more in each group in order to decrease probabilities of “noise discoveries”) [68, 69], no consensus can be reached in the current literature .
In the frontal/prefrontal regions of the brain, the focus of this study, some 1H-MRS studies have reported that concentrations were reduced in established schizophrenia relative to healthy controls for levels of NAA [71–74], and age-adjusted NAA . On the other hand, several other studies have reported normal levels of NAA in the same frontal/prefrontal regions in never treated first episode psychosis , medicated first episode psychosis [77, 78], and established schizophrenia [79–86]. These studies, for the most part, sampled a brain volume of interest encompassing both gray matter and white matter tissue types, consequently precluding any specific interpretation of findings in terms of the precise anabolic or catabolic cycle of NAA. Obviously there is a need to search for confounding variables that might impact on the current mixed 1H-MRS findings reported across population samples and laboratories. Early adolescent onset of regular cannabis use certainly has the potential to be one such factor [3, 5].
Transverse (T2) relaxation time constants
Another neuroimaging modality involved in this study targets the same white matter brain region, while maintaining the focus at the cellular level. The transverse relaxation time constants of regional tissue water (involving both intracellular and extracellular tissue water) and of NAA (intracellular) provide an index of the integrity of the microcellular environment of the brain region studied. In fact, T2 relaxation time constants are dependent on the morphological parameters of cell size and cell packing density in the brain region studied; they also reflect intracellular molecular mobility as they are dependent on the frequency of molecule-microenvironment interactions . As such, prolonged T2 time constants are associated with reduced cell densities.
Transverse time constants of NAA in the context of ~95% white matter tissue will provide an index of intracellular density of cells in this brain region, oligodencrodytes being an important target among these cells . Transverse time constants of water will provide an index of intracellular plus extracellular cell packing densities without differentiation of which type of cells are present in the specific region of interest.
Different from DTI and 1H-MRS studies of schizophrenia, the studies assessing T2 relaxation time constants in this illness have, to this date, reported consistent findings. The T2 time constants of water were found to be prolonged relative to healthy controls in prefrontal white matter [75, 88] and anterior corpus callosum , in adults with established schizophrenia [75, 88, 89] and first episode psychosis . No differences were observed between adults with first episode psychosis and those with established schizophrenia . These findings altogether support the assumption of abnormal axonal milieu and myelin structures in schizophrenia.
Transverse relaxation time constants of NAA, on the other hand, were found to be shortened relative to healthy controls in prefrontal white matter of adults with established schizophrenia [75, 88] and first episode psychosis , thus supporting the assumption of increased intracellular (oligodendrocytes) cell density. In gray matter (anterior cingulate cortex), T2 time constants of NAA were also shortened in adults with schizophrenia relative to healthy controls but the group difference did not reach statistical significance .
Because of the consistency of reported findings with these particularly sensitive neuroimaging indices of cell packing densities, T2 relaxation time constants acquired in this proposed study will be used as an anchor point against which DTI and 1H-MRS measures will be interpreted. The aim is to generate, from these in vivo data, a plausible interpretation of the specific cellular abnormalities associated with schizophrenia, which might differ from those associated with early adolescent onset of regular cannabis use.
Noteworthy for this project, the T2 relaxation time constants of NAA and tissue water have never been used to help differentiate the detrimental impact of cannabis use from the white matter cellular abnormalities associated with schizophrenia. We expect that these sensitive measures of intracellular and extracellular cell packing densities will be related to DTI (FA) values, as water diffusivity is also influenced by cell densities .