Results of our study reveal a significant age-related decline of fiber tract integrity within CB and UF, but not IOFF in schizophrenia. Of note, such a decline was not observed in age-matched controls. In addition, the decline in fiber tract integrity observed in schizophrenia was not associated with either illness duration, medication effects, or total positive or negative symptoms (though a trend was noted toward a statistically significant negative correlation between positive symptoms and FA in schizophrenia).
Results of our study suggest the possibility of progressive changes in white matter fiber integrity in schizophrenia over time. This association, however, while suggestive, has not been extensively studied in schizophrenia. Among the small number of studies that have addressed this issue, one structuralMRI imaging study by Mathalon et al. (2001)
demonstrated fronto-temporal gray, but not white matter age-related volume decline in chronic schizophrenic patients. In a separate electrophysiological study using ERP, O’Donnell et al. (1995)
observed a greater increase of P300 latency in older chronic schizophrenic patients. Another longitudinal study (van Haren et al., 2007
) observed increased cortical gray matter atrophy in schizophrenic patients before the age of 45 in a 5-year-follow-up. This group also found white matter to increase excessively before the age of 32 as compared to normal healthy controls, which showed a steadier trajectory of white matter volume growth. Finally, the only study by Jones et al. (2006)
utilizing DTI and fiber tractography, demonstrated, on a small sample of younger schizophrenics, an age-related increase of tract integrity within the Superior Longitudinal Fasciculus. The difference between our results and those reported by Jones et al. might be explained by different DTI acquisition methods, difference in sample sizes, as well as differences in subjects’ age distribution and illness duration (our study includes twice as many subjects as Jones’s paper, moreover our subjects were much older, and medicated/sick for much longer period of time). Nevertheless, taken together, these findings suggest that there may be a neurodegenerative process in pathological aging (Arnold and Trojanowski, 1996
), which affects the white matter but does not start right after psychotic onset, but instead might be delayed because of prolonged brain maturation (Benes, 1989
Of further note, we found no significant decline in FA with age in healthy control subjects, which differs from the findings of Sullivan and Pfefferbaum (2006)
, where their findings showed a continuous decline of FA starting at the age of 20. This difference between our findings and theirs might be accounted for by the fact that their study was designed to look at lifespan and not at the kind of constriction of age that characterizes the current study. Finally, a study by Raz et al. (2005)
reported findings that suggested that there was a decline in white matter integrity that was accelerated in older adults. According to their findings, the mid-fifties appear to be the inflection point for an age trend in volume changes of the brain for the normal healthy population, which is consistent with our findings. It is of further interest to longitudinally follow-up our study cohort to determine whether or not we find such continued changes.
Possible limitations of the study should be noted. First, this study was not designed to longitudinally follow up patients over the course of their illness. Thus cohort effects such as severity of illness (even though overall total positive and negative symptoms were not associated with FA decline), might be a possible confound in the finding of FA and age correlations in schizophrenia. Second, we observed chronic schizophrenic patients who have been medicated for years. We note also that even though we did not demonstrate correlations between diffusion measures and age of onset, duration of illness, or medication, we can not totally rule them out as possible confounds. Further, even though there is some evidence of white matter loss after treatment with antipsychotic medications (Christensen et al., 2004
), this evidence is weak, and requires further investigation. We also realize that since we study chronic schizophrenics, duration of illness is closely associated with age. However, since we observed correlations between diffusion measures and patients’ age, and did not observe such correlations with age of onset, or with duration of illness, we hypothesize that there is another factor, probably related to brain development differences in schizophrenia that plays a role in white matter pathology. Finally, additional studies examining prodrome subjects at high risk for developing schizophrenia and patients diagnosed with a first episode of schizophrenia will help to determine the impact of these factors on DTI measures in schizophrenia.
Additionally, the study cohort consisted only of male schizophrenic patients. Since theories of a possible neuroprotective potential for estrogen have been reported (Bardenstein and McGlashan, 1990
), further investigations are needed to reveal possible gender differences in age effects on brain connectivity. Finally, an additional study limitation is the fact that the sequence used in this study was a relatively low resolution (especially out-of-plane) and only 6 directions of diffusion were collected. Nevertheless, we note that the findings of our study were sufficiently robust. Even at lower resolution the differences between groups and correlational changes with age were found.
Another potential limitation, often discussed in DTI reviews, is posed by the fact that tractography requires an FA threshold to be set at a certain value, which again limits the possibility of measuring voxels below a certain FA value. Kanaan et al. (2006)
demonstrated using FA histograms, however, that the FA differences between schizophrenia patients and controls fall above the 0.2 FA values. Thus to eliminate possible bias, lower stopping threshold should be chosen. We note that we chose to set the threshold at 0.15 (Jones et al., 2006
), which should robustly encompass any pathological white matter voxels in the schizophrenia brain.
Further, we note that diffusion tensor tractography is not free of its own limitations. With today’s scanning techniques, results from fiber tractography do not describe the spatial extent of individual axons but instead describe the average diffusion properties in white tissue at the scale of a millimeter. Thus the diffusion tensors at this scale should be regarded as estimators of local diffusion properties, and tractography as a visualization of features in this field. Validation of the features captured in the tensor field is thus very challenging since the definition of the tissue properties corresponding to the diffusion field is not well understood. However, similarities between DT-MRI tractography and real white matter fiber architecture in the human brain are compelling even at today’s limited resolution. We also note that this method has been used to generate white matter atlases (Mori et al., 2007
). Further, in this study, and as noted previously, to make sure that there were no group differences in tract appearance, all tracts were visually inspected, and also their mean length, fiber number, and mean distribution angle were reported, and compared between groups, where no group differences were found.
One imaging parameter that could be viewed as a limiting factor in fiber tractography is the small number of diffusion directions collected (6 in our case). The literature regarding the accuracy of tensor estimation in relation to the number of directions is large, and no agreement has been reached. Some believe that since 6 directions are sufficient to generate tensors, it should be sufficient to track the main diffusion direction, and thus to perform accurate tractography. Others believe that 12 or more directions are needed to increase tensor estimation accuracy. Our 6-direction acquisition has been used before in surgical planning (Talos et al., 2007
), and confirms the accuracy of tractography. However, with newer, faster MR systems, where it is not too costly to use higher directional resolution, one should certainly consider doing so. Additionally, we note that tractography stopping criteria of 20° of curvature change within 1 mm, are quite common in the literature, and have been introduced in order to deal with potential artifacts in the data that could influence the tractography output (see review in Kubicki et al., 2007
In summary, this study underscores the need for further investigation of progressive changes in schizophrenia, as well as underscores the importance of including age as a possible confounding factor in future white matter investigations.