Evidence For Reciprocal Structural Network Interactions Between Bilateral Crus Lobes And Broca's Complex

Objective: Medulloblastoma is the most common malignant brain tumor in children.Cerebellar mutism syndrome(CMS)after medulloblastoma surgery is a troublesome clinical problem with high incidence,non-transient symptoms and unclear neural circuit pathogenesis.The association between the transvermian approaches(TrA)and this syndrome has been questioned.Previous studies suggested that Damage to the dentatothalamocortical tracts in their section around the fourth ventricle was the most likely cause,but the origin of such tracts in the cerebellum and the end point in the cerebral cortex were not clear,leading to the occurrence of a variety of clinical prevention strategies.The speech disorders in CMS are motor speech disorders.Recent literature has suggested that it can be manifested not only in language executive disorders but also language planning disorders.In the field of neuroscience,the role of the cerebellum in language execution has been recognized,and its role in language planning has been gradually recognized in recent years.The functional areas of the cerebellum involved in language are centered in bilateral Crus lobes,but the mechanism of neural circuitry is unclear.The left(dominant)inferior frontal gyrus,especially Broca's area and the ventral premotor area,plays an important role in the motor language network,which performs the functions of language execution and planning from posterior to anterior anatomically.Whether the cerebellum directly innervates the left inferior frontal gyrus through fibers has been controversial.In recent years,some studies based on diffusion magnetic resonance imaging(d MRI)tractography have incidentally reported the presence of the tracts originating from the cerebellum and targeting into the left inferior frontal gyrus or the opposite tracts in healthy subjects,but none of them have made a clear explanation of these tracts' trajectory.The quality of clinical MR data is generally inferior to research data and the acquisition of clinical data is a complex and time-consuming systematic process,suggesting that rigorous clinical study design is a key prerequisite.Exploration of this structural connection using publicly available multimodal MRI data from healthy subjects will undoubtedly facilitate the process of solving these two problems.As the only noninvasive fiber-tracking technique available in vivo,d MRI tractography faces a challenge to reconstruct polysynaptic,long-range and contain-crossing-fiber tracts between the cerebral cortex and the cerebellar cortex.Deterministic tractography and probabilistic tractography have their respective advantages in tracking long-range and cross-fiber tracts,so this study used three types of fiber-tracking software(see Methods section),that have been proven to be able to track these fibers,to reconstruct this structural connection.Furthermore,fiber-tracing fails to give information of specific function to the constructed tracts.we roughly labeled the function of the fibers originating from the cerebellum and targeting into the left inferior frontal gyrus by analyzing the spatial relationship between the language network obtained by analysis of resting-state functional MRI(rsf MRI)and the end points of these tracts at the subject level.Methods: Publicly available and high-quality structural MRI(s MRI),d MRI and rsf MRI data from all subjects of the HCP Lifespan Pilot project-1a were used in the present study.Part 1: Human Connectome Project multi-modal parcellation 1.0 atlas and spatially unbiased atlas for the cerebellum and brainstem was respectively applied on each subject's s MRI to segment the left inferior frontal gyrus(including Broca's complex)and the cerebellums(including bilateral Crus lobes).After MRtrx3 software was used to track the tracts originating from the cerebellum and targeting into the left inferior frontal gyrus,indicating these tracts mainly originated from bilateral Crus lobes and mainly targeted into Broca's complex(this part of the tracts were named as Crus-Broca tracts),we used probabilistic tractography(MRtrix3 software and FSL software)and deterministic tractography(DSI Studio)to track reciprocal structural network interactions between bilateral Crus lobes and Broca's complex respectively with the same d MRI data set.The consistency among related tractography results derived from each software was analyzed.Part two: Each subject's rsf MRI data were analyzed using independent component analysis built in Pre Surg Mapp Toolbox to obtain Language-related resting-state network in native space,of which activation area located near the left inferior frontal gyrus was called as Broca activated area.Following the principle of functional tractography commonly used for preoperative mapping of functional fibers,the coverage of subregions of Broca's complex and the end points of Crus-Broca tracts in Broca's complex by Broca activated region in each subject were analyzed compared.analyzed.Results:Part one: Qualitative analysis and statistical analysis indicated that the results obtained by these three software were consistent.There was a bidirectional anatomical connection between the cerebellum,especially the bilateral Crus lobes(but not the inferior vermis),and the left inferior frontal gyrus,especially Broca's complex(HCP75,HCP76,HCP77,HCP171 and HCP74).The streamlines,named Crus–Broca tracts,originated from the bilateral Crus lobes,synapsed onto the dentate nucleus,ascended into the superior cerebellar peduncle,went through the tissue forming both the superior wall of the Supratonsillar cleft and the lateral roof of the fourth ventricle,surprisingly bypassed the left red nucleus and the left thalamus,and ended at the subregions of Broca's complex.The streamlines,named Broca–Crus tracts,originated from the subregions of Broca's complex,went far away from the fourth ventricle and ended predominantly at the right Crus lobes.There were reciprocal structural network interactions between each of subregions of Broca's complex representing different language functions and each of bilateral cerebellar Crus ? or ? lobes.HCP75(i.e.Bradman area 45 or Pars triangularis)was the most important subregion contributing to both Crus-Broca tracts and Broca-Crus tracts.Part two: HCP74,HCP75,HCP76,HCP77,and HCP171 were covered by Broca activated regions at the subject level with the average percent(± one standard deviation)of71.0±25.1%,70.1±25.1%,74.1±20.7%,47.0±27.9%,and 61.4±30.3% respectively,suggesting that Broca's complex could reflect the location of Broca activated regions at the subject level to some extent.In the group level,the proportion of streamlines of CrusBroca tracts of each subregion of Broca's complex captured by the region in this subregion covered by Broca activated region was higher than that of volume of this subregion covered by Broca activated region(Wilcoxon's sign rank test for related samples,Z = 3.021,single tail p = 0.0015),and the density of streamlines of Crus-Broca tracts in the region of each subregion covered by Broca activated region was higher than that in the region of this subregion uncovered by Broca active region(Wilcoxon's sign rank test for related samples,Z = 2.957,Single-tailed P =0.0015),indicating that the end points of the Crus-Broca tracts tended to reach Broca activated region with Broca's complex.Although the other subregions showed a same trend,only HCP75 was statistically significant in the latter comparison and could be corrected by multiple comparisons(original P =0.012).At the group level,the density of streamlines of Crus-Broca tracts in the region of HCP75 covered and uncovered by Broca activation region was 0.0167±0.0245 and 0.0096±0.0157/voxel respectively.Conclusion:Part one: The Crus-Broca loop would provide direct evidence to support cerebellum involvement in motor-speech planning.The discovery of Crus-Broca tracts' going through the tissue forming both the superior wall of the Supratonsillar cleft and the lateral roof of the fourth ventricle,both of which is prone to damage in resection of tumors extending to the rostral part of the fourth ventricle by TrA or telovelar approaches(Te A),provides a design idea for clinical studies on CMS prevention: whether TrA or Te A is used,the dentate nucleus and the superior cerebellar peduncle fibers near the superior wall of the Supratonsillar cleft or the lateral roof of the fourth ventricle are the key points of protection.Part two: Crus-Broca tracts could participate in Language-related resting-state network through Broca activated region.In the preoperative mapping of Language-related tracts in patients with language-eloquent brain lesions,the fibers between the cerebellum and Language-related cortical area,especially Crus-Broca loop,should be included in the investigation.In combination with the adjacency of the superior cerebellar peduncle fibers,represented by the Crus-Broca tracts,to the trajectory of approaches for fourth ventricle tumors revealed in Part one,it is further suggested that these tracts should be considered both during operation for CMS prevention and in symptom-tract analysis for exploring the neural circuitry mechanism causing CMS.