In this study, a bi-tensor model of DTI was used to reveal the CC degenerative pattern of PD patients with different sides of onset and its relationship with motor and non-motor symptoms. Our main findings were as following: (1) In LPD patients, FW and FAT were decreased in bilateral hemispheres of CC, while MDT was increased in RH. (2) FAT decreased significantly in all CC subregions except for subregions 1, 3 and 6, and MDT increased significantly in anterior and posterior CC (CC subregions 1, 5, 6, 7 and 10) in LPD group. (3) The LI analysis of CC and its segments reflected the significant interhemispheric FAT asymmetry of the anterior and middle part of CC in RPD group, with a more significant reduction in the right hemisphere of CC. (4) Injuries in each subregion of CC were associated with motor and non-motor symptoms (e.g., cognition and emotion). In summary, extensive damage of CC was observed in LPD patients than in RPD patients, while relative limited and asymmetrical damage was found in RPD patients. These findings suggest that abnormalities in the callosal white matter may be one of the potential mechanisms in PD with unilateral onset and linked to the presence of non-motor symptoms and prognosis in PD.
Two hemispheres of human brain are asymmetrical in structure and function. Interhemispheric divergent processing can improve the ability to perform multiple tasks at the same time and prevent response competition between the two hemispheres[11]. CC is the largest white matter fiber bundle connecting the two hemispheres of the brain, as shown in Fig. 1. Different segments of CC are connected to different brain regions, thereinto, the anterior and middle parts of CC transmit cognitive, emotional, and motor information between the frontal and parietal lobes, while the posterior fibers pass through the central posterior body, isthmus, and spleen to transmit somatosensory, auditory, and visual information[11, 13, 33]. As can be seen from this study, almost all three diffusion parameters (FW, FAT, and MDT) showed damage in bilateral hemispheres of CC in LPD patients, even though MDT of LH was not significantly increased (p = 0.099, uncorrected). Inflammatory response may be reflected by decreased FW, which is the result of hyperplasia of inflammatory cells [34]. Likewise, decreased FAT and increased MDT also indicate damage in the fiber bundle. So extensive damage was found in LPD patients in our study. Previous studies have shown that patients with left side lateralization was associated with faster symptom progression[35] and worse outcomes[36, 37]. Hence CC may be one of the potential mechanisms influencing prognosis in PD. Correlation analysis performed in LPD group showed that the injury of CC, especially in LH, was related to motor symptoms and emotion, and the injury of subregion 5 was related to motor symptoms and cognition. In RPD group, the MDT of RH was only higher than HC, which was specifically manifested as the damage of subregion 7, reflecting the preferable integrity of CC in RPD group. The results of LI analysis showed no obvious asymmetry in the structure of CC and its segments (the LI value ≥ 0.1 or ≤-0.1). However, in RPD patients, an obvious asymmetry in the anterior and middle parts of CC was observed, which is called pseudo-left lateralization[38], that is, the damage was relatively more serious in the right hemisphere of CC. Correlation analysis showed the degree of lateralization of segment III was negatively correlated with the anxiety scores in RPD patients, indicating that more complete the LH structure was, the lower anxiety level patients would be.
Previous studies have revealed that dopamine levels distributed asymmetrically in healthy brain, which were higher in the left side of the striatum than in the right side. The neurobiological asymmetry made the right striatum relatively more susceptible to dopaminergic denervation and then caused disease on the left side of the body[39]. They suggested that dominant side of body had a protective effect due to enhanced exercise which may in turn cause better synaptic plasticity in the dominant hemisphere (contralateral to the dominant hand) of brain[40–42]. As all patients enrolled in our study were right-handed subjects, which means they had left hemispheric dominance. Better synaptic plasticity of left hemisphere somehow increased the possibility to overcome pathological alteration, therefore patients predominating with right-side motor symptoms can perform milder severity compared with those predominating with left-side[35]. Moreover, more distribution of striatal dopamine in left hemisphere enhanced connections to the cortex to make it more robust in face of damage and therefore followed positive consequences in terms of symptoms[9]. In the future, it might be possible to combat disease progression by strengthening exercises in the more affected side of the body. Therefore, in this study, the injury in right hemisphere of CC was more obvious in RPD patients due to the protective effect of dominant hemisphere. In LPD patients, the pathological injury was easy to reach the limit due to the nerve vulnerability of right hemisphere and therefore resulted in extensive damage in bilateral hemispheres with the spread of pathology[43, 44]. In accord with other clinical studies, patients with motor symptoms occurring on the dominant side had fewer motor defects than those on the non-dominant side, besides, the latter had more severe motor symptoms[8]. In addition, cognitive decline was faster in patients with left-sided onset. Given all this, we speculate that the extensive damage of CC may be one of the underlying mechanisms causing poor motor performance and cognitive function in LPD patients.
There are some advantages and limitations in this study. Firstly, our sample size was relatively large, so the results would be more reliable. However, it is worth mentioning that the demographic data was not balanced in order to maximize the sample size. There were more male patients in RPD group, and the education level of PD patients is lower than that of HC. Previous literature has reported the influence of sex and education on brain structure alterations in PD[45], so we put these two factors as covariates in the subsequent comparison of neuroimage parameters. Secondly, we adopted a bi-tensor model of DTI which may eliminate the influence of water outside the cell and increase the sensitivity of the diffusion measures to the tissue structure. Compared with the traditional diffusion measures that relied on the single tensor model, the use of the bi-tensor model may be a more sensitive and reliable marker of PD and its progression[28, 46]. To date, it has not been reported in neuroimaging studies related to CC in PD. In this study, the single b1000 value was used to obtain FW, while multi-shell (multi b value) will be required to fit FW more accurately in the future. Thirdly, three-dimensional segmentation was performed to evaluate CC, which reduced manual deviation and improved signal-to-noise ratio compared with two-dimensional analysis with single plane.