There was a significant group difference between the control and the patient group in letter-number sequencing test scores. The multiple regression analysis results showed that there was no track with significantly decreased quantified anisotropy

Table 2 Result of quantitative measurements of diffusion MRI in control, PD and SWEDD

Percentage threshold (%)

Patients with schizophrenia (FDR)

Control subjects (FDR)
















related to letter-number sequencing test scores in healthy controls (p>0.05). When the percentage threshold was increased up to 45%, FDR showed significance but fiber tracking along fiber pathways showed no tracts. However, significantly decreased anisotropy in the arcuate fasciculus (AF), the inferior longitudinal fasciculus (ILF), and the body of the corpus callosum of the patients related to letter- number sequencing test with the percentage threshold of 35% and more (p<0.05) (Table 2).


In this study we applied connectometry, which tracks the differences along the pathways themselves, while conventional connectome analyses are designed to find differences in whole fiber pathways. Therefore, by using connectometry we were able to track statistically meaningful associations to identify the subcomponents of white matter pathways, as well as white matter fascicles as a whole, associated with LNS scores. It is because connectometry does not map the connectome itself but it analyzes differences in the local connectome, associates the local connectome with study variables (here, LNS scores), and then tracks the associations across a pathway [8].

Our study revealed that LNS scores, as a measure of verbal working memory, in recent-onset schizophrenic patients are positively associated with QA in the fibers going through the track of AF, ILF, body of CC, and some fibers of CST, cingulum bundle and IFOF by applying connectometry (Fig. 1). However, LNS scores did not show any significant association with QA in control subjects in lower than 45% percentage threshold. This can be applied as a clinical marker for diagnosis of schizophrenic patients, not observed in healthy people. It has enough sensitivity to capture schizophrenic patients but association of a test score and DTI metrics is a time consuming way for this goal. However, as this marker might be observed in other psychiatric disorders, it may not have enough specificity. So we suppose that conducting studies using patients with other psychiatric disorders may help us check the specificity of this clinical marker.

Our analyses confirm previous findings regarding working memory in schizophrenia to a large extent. However there are some new findings. In addition,

Fiber tractography in one of the schizophrenic patients with working memory deficit

Fig. 1 Fiber tractography in one of the schizophrenic patients with working memory deficit. (a) left and right arcuate fasciculus. (b) left and right inferior longitudinal fasciculus. (c) body of corpus callosum using connectometry has enabled us to visualize the segments of fiber bundles whose density (spin distribution function) is associated with working memory dysfunction.

Individual capacity of working memory function relies on the integrity of frontal and parietal white matter tracts based on previous functional [12] and diffusion imaging [13] evidence. Conducting fMRI in schizophrenic patients while performing working memory tasks, has shown hypoactivation in the DLPFC and medial prefrontal region, involved in task relevant information and maintenance of visuospatial attention during working memory tasks respectively [6].

SLF is a long association fiber tract, commonly said to be composed of five subcomponents, mainly connecting frontal and parietal regions. Most of the white matter fiber tracts detected in this paper to be associated with working memory function, were the association fiber tracts running from superior temporal to middle and inferior frontal regions which mainly form the arcuate fasciculus,which is considered to be one of the subcomponents of the SLF. Also some of the fiber tracts running from DLPFC to superior temporal region were detected. This is known as an important working memory pathway. However, association of working memory function in recent-onset schizophrenia with FA value in superior longitudinal fasciculus using TBSS has been previously identified [7].

Our study showed that the corpus callosum is associated with working memory dysfunction. Schizophrenic patients, like older people, manifest a wide range of cognitive deficits. Therefore, correlation of reductions in the integrity of the body of corpus callosum, which interconnects frontal and parietal regions, in schizophrenic patients with working memory deficit does not seem unexpected. A study regarding cognitive function in older healthy people using TBSS to explore structural connectivity, revealed that working memory function correlated with white matter integrity in the left genu and body of corpus callosum subserving DLPFC, and left IFOF/anterior thalamic radiation with the peak signal intensities near the anterior intraparietal cortex [14].

Our study also indicated ILF and IFOF to be associated with working memory dysfunction. We suppose that the neuropsychological test we have used for measuring working memory (LNS) is dependent not only on working memory but also on other aspects of cognitive function like processing speed and executive function, which ILF and IFOF are shown to be correlated with [4]. So finding an association between test scores and the integrity of ILF and some fibers of IFOF is expectable. However, a study conducted in early-onset schizophrenia showed significant associations between FA values in left ILF and scores on the number- letter switching subtest [15]. Lower ILF FA has been shown to mediate lower working memory performance in overweight adolescents [16].

Association of lower QA in cingulum and working memory deficit is quite understandable due to its major role in memory. But the association of fibers of CST with LNS test seems to be due to its dependence on doing tasks. So, studies using more accurate tests are required.

It would be worthy of note that by administering connectometry approach, QA value has been assessed instead of FA, which is more commonly applied. This may guide us more directly to the pathophysiology of psychoses like schizophrenia. As QA is based on spin distribution function (SDF), measuring the density of diffusion spins, it is thought to be more sensitive to physiological conditions and compactness of bundles. It is while fractional anisotropy (FA) is used to assess diffusion velocity and is a measure of axonal loss, demyelination and pathological conditions.

However, we suppose that further study in our sample using TBSS is required to confirm the anatomical location of fiber tracts we have found and to compare tracking the local connectome and whole fiber tractography. This study further confirms disconnectivity hypothesis in schizophrenia but the temporality of white matter integrity changes and pathology remains a question. In addition, , we should pay attention that although connectometry is quite powerful, it has some limitations because it is relatively insensitive to focal differences in short ranged pathways. Thus, it should always be kept in mind that by implementing connectometry the value of the threshold used to define the set of supra-threshold links is an arbitrary choice.

< Prev   CONTENTS   Source   Next >