Our main finding is that patients with schizophrenia reveal a significant decrease in FA and an increase in trace in the AC compared with healthy controls. To the extent that FA is related to axonal integrity, density, caliber and myelination (Beaulieu, 1994
), our findings of subnormal levels of FA in the AC suggest the presence of microstructural abnormalities in this fasciculus in patients with schizophrenia. Our results are consistent with those of Highley et al. (1999)
, who reported reduced fiber density in the AC in schizophrenia patients postmortem
. With respect to our finding of higher trace in the schizophrenia patients, compared with controls, several previous DTI studies have suggested that increased trace is associated with immaturity of the brain (Neil et al., 1998
) and structural brain disorders, such as cerebral ischemia (Kim et al., 2005
) or hydrocephalus(Gideon et al., 1994
). The changes in trace and FA in our study thus suggest that there are structural abnormalities in the AC in schizophrenia.
Our correlational findings between age and integrity of the AC in schizophrenia are similar to results of previous investigations with the cingulum, the uncinate fasciculus (Mandl et al., 2008
; Rosenberger et al., 2008
), and the whole-brain WM (Jones et al., 2006
; Mori et al., 2007
). Our finding, and that of others, that suggest negative correlations between age and white matter integrity support the hypothesis that the pathophysiology of schizophrenia may reflect progressive neurodegeneration (de Haan and Bakker, 2004
; DeLisi, 1997
). However, we observed no significant correlations between duration of illness and FA, suggesting that it is likely age rather than duration of illness that is relevant to the observed correlations. Longitudinal studies are, however, needed to clarify further whether or not white matter integrity reduction reflects disease progression. The latter is particularly important given that findings of a correlation between age and FA have been reported in normal controls (Kubicki et al., 2008
; Sullivan and Pfefferbaum, 2006
). Thus more research needs to be done that can tease apart what is age related effects that are seen in both normal controls and patients with schizophrenia, and to address the question of whether or not the age effects observed are different in schizophrenia and suggest disease progression. Furthermore, additional studies will be needed to clarify whether medication or disease progression affect the altered integrity of AC. Longitudinal and first-episode studies of AC integrity will help to clarify these issues.
Significant positive correlations were observed between patients' FA in the AC and severity of positive symptoms, as measured by total scores on SAPS. This is not the first time that positive correlations have been observed between FA and positive symptom severity in schizophrenia patients. On the contrary, significant positive correlations have been reported in the corpus callosum (Hubl et al., 2004
; Rotarska-Jagiela et al., 2008
), cingulum bundle (Hubl et al., 2004
), superior longitudinal fasciculus (Seok et al., 2007
; Shergill et al., 2007
), arcuate fasciculus (Hubl et al., 2004
), and inferior fronto-occipital fasciculus (Szeszko et al., 2008
). However, the present study is the first (to our knowledge) to report a significant positive correlation between positive symptom severity and FA in the AC in schizophrenia patients. With regards to the cause of this seemingly paradoxical yet consistently reported correlation, Whitford et al. (2010)
have suggested that the answer may lie in the extent of the conduction delays that would be expected to arise from mild as opposed to severe white-matter damage. Specifically, Whitford et al. (2010)
suggest that mild conduction delays
(such as might be expected from mild FA reductions) could result in the aberrant neural synchronization (Andreasen et al., 1999
) that has been proposed to underlie the symptoms of schizophrenia (Andreasen et al., 1999
; Bartzokis, 2002
). In contrast, severe conduction delays
(such as might be expected from more severe FA reductions) could result in neural signals that are sufficiently disjointed so as to be unincorporable into a coherent phenomenological framework such as necessary for a coherent, systematized delusion, for example. Our results showed severe positive symptoms were related to minimally dysfunctional FA. However, we found negative symptoms were not related to FA, though AC FA reduction predicted higher negative symptom score as we predicted. Nonetheless, we need more functional studies about the relationship between specific symptoms of schizophrenia and white-matter tracts. Testing this speculation may provide a worthwhile avenue for future research.
Since it has been reported that the orbitofrontal cortex plays an important role in decision making processes (Krawczyk, 2002
), we investigated the relationship between FA in the AC, which interconnects the orbitofrontal cortices (Di Virgilio et al., 1999
; Peuskens et al., 2004
), and its abnormalities in schizophrenia, and the Iowa Gambling Task. Our results indicate that such a relationship indeed exists, suggesting the notion that changes in interconnectivity of the AC affect decision making processes in schizophrenia. It is possible that altered interconnectivity of the AC may be related to the same pathology producing orbitofrontal cortex volume reduction, which has been shown to be associated with thought disorder in schizophrenia (Nakamura et al., 2008
). It is unclear why FA is higher in patients with severe decision making impairment, however a similar result was found in an impulsivity study of schizophrenia (Hoptman et al., 2004
), where FA was positively correlated with impulsivity in the left postcentral gyrus, right superior/middle temporal gyrus, and bilateral fusiform gyrus, which might compromise a fronto-temporal-limbic circuit. Additionally, patients with schizophrenia showed a pattern of compromised decision making that is somewhat different from orbitofrontal cortex lesion patients (Shurman et al., 2005
). We therefore suspect that the AC plays a role in the patterns of decision making impairment in schizophrenia. Of further note, equivocal results have been found in terms of differences in Iowa Gambling Task performance between SZ and NC, with some studies showing significant differences (Kester et al., 2006
; Ritter et al., 2004
), but others not (Cavallaro et al. 2003
; Evans, et al. 2005
; Rodriguez-Sanchez et al. 2005). Regarding our study, we observed that patients with a large FA reduction had minimal decision making impairments, which could potentially account for the minimal Iowa Gambling Task differences between patients and controls.
While AC FA was associated with decision making and positive symptoms in our study, it is unclear whether schizophrenia symptoms and decision making are related. Some studies suggest decision making impairment of schizophrenia and negative symptoms are associated (Shurman et al., 2005
), while some studies do not (Evans CE et al., 2005
). Another study suggests that some negative symptoms are associated with frontal medial cortex dysfunction, as measured with a Theory of Mind test, but that there was not a significant correlation between the Iowa Gambling Task and negative symptom scores (Martino DJ et al., 2007
). To our knowledge, there is no evidence to suggest a relationship between decision making and positive symptom scores. However, regardless of whether decision making and clinical symptoms are related, AC integrity seems to be related to multiple symptoms in schizophrenia.
There is a considerable overlap between the participant sample in the current paper and the sample presented in the Whitford et al., paper. Specifically, 16 (61.5%) of SZ patients and 11 (42.3%) of the HC participants were also investigated in the Whitford et al., paper. There is some evidence from the literature that the CC and AC are closely related in terms of their structural integrity. In a previous study about callosal agenesis (Fischer et al., 1992
), there was hypertrophy of AC in patients with callosal agenesis, suggesting that AC may compensate for damage to the CC. Regarding the relationship between AC and CC in the present study, we correlated the AC Fractional Anisotropy results with the CC1 (genu) Fractional Anisotropy results from the Whitford et al., paper and observed a non-significant trend for a positive correlation (r=0.32, p=0.10), across both groups.
Our study is the first to measure AC integrity using tractography in schizophrenia. Tractography has shown higher sensitivity and specificity compared to ROI or voxel-based morphometry analysis (Kanaan et al., 2006
). Though the AC plays an important role in interconnection of the two hemispheres, there are few studies that investigate its changes in schizophrenia. By using tractography to analyze AC fibers, we were able to define, extract and directly measure AC integrity and its relationship with clinical symptoms and age. We note, however, that a diffusion tensor model is not always adequate, especially in voxels containing complex crossing tracts from differently oriented fibers(Alexander et al., 2002
; Tuch etl al., 2002). Nonetheless DTI is the most popular approach to tractography and fiber crossings are not such an issue with AC. Further studies should include a multi-tensor approach that may provide more information about AC tract and integrity. Since some studies have found that changes in FA might be related to medication dosage (Minami et al., 2003
; Okugawa et al., 2004
), and our schizophrenia subjects are all chronically medicated, an additional studies are needed to clarify whether the relationship between age and FA is due to a medication effect. Sex differences in fiber numbers or cross sectional areas in the AC have also been reported (Allen and Gorski, 1991
; Highley et al., 1999
). In the current study, however, we only investigated changes in male subjects. We therefore need a larger, mixed-sex cohort to obtain a full interpretation of AC pathology in schizophrenia.