Structural changes in eloquent cortex secondary to glioma in sensorimotor area

Abstract

This study aims to investigate the structural reorganization in the sensorimotor area of the brain in patients with gliomas, distinguishing between those with impaired and unimpaired strength. Using voxel‐based morphometry (VBM) and region of interest (ROI) analysis, gray matter volumes (GMV) were compared in the contralesional primary motor gyrus, primary sensory gyrus, premotor area, bilateral supplementary motor area, and medial Brodmann area 8 (BA8). The results revealed that in patients with right hemisphere gliomas, the right medial BA8 volume was significantly larger in the impaired group than in the unimpaired group, with both groups exceeding the volume in 16 healthy controls (HCs). In patients with left hemisphere gliomas, the right supplementary motor area (SMA) was more pronounced in the impaired group compared to the unimpaired group, and both groups were greater than HCs. Additionally, the volumes of the right medial BA8 in both the impaired group were greater than HCs. Contralateral expansions in the gray matter of hand‐ and trunk‐related cortices of the premotor area, precentral gyrus, and postcentral gyrus were observed compared to HCs. Furthermore, a negative correlation was found between hand Medical Research Council (MRC) score and volumes of the contralateral SMA and bilateral medial BA8. Notably, our findings reveal consistent results across both analytical approaches in identifying significant structural reorganizations within the sensorimotor cortex. These consistent findings underscore the adaptive neuroplastic responses to glioma presence, highlighting potential areas of interest for further neurosurgical planning and rehabilitation strategies.

Exploring neuroplasticity in sensorimotor gliomas: This study reveals how gliomas affect brain eloquent areas, highlighting increased cortical volume in specific regions as a compensatory response to maintain motor function. By comparing 59 patients with sensorimotor gliomas, we demonstrate significant cortical changes correlating with motor strength, emphasizing the dynamic brain's adaptability in the face of motor impairment.

1. INTRODUCTION

Motor functions are fundamental for maintaining normal human activity. As gliomas develop in sensorimotor areas, they frequently result in motor dysfunction, substantially diminishing the patient's quality of life (Amidei & Kushner, 2015). Interestingly, previous studies have shown that paresis resulting from brain damage can undergo partial recovery (Sawner & LaVigne, 1970). Brain reorganization and neuroplasticity may be attributed to the change in the volume of gray matter in the motor function area compared to its state before the dysfunction onset (Almairac et al., 2018). Gray matter volume (GMV) can diminish because of tumor infiltration (Lamichhane et al., 2023). Contrarily, it may expand because of functional compensation mechanisms (Almairac et al., 2018). In patients with gliomas, any functional deficiency stemming from direct damage to the motor function area can be restored through cortical reconfiguration or functional compensation (Huang et al., 2022). Given that this reconfiguration process is time‐intensive, this particular occurrence is predominantly seen in patients with glioma and is rarely observed in swiftly progressing lesions such as strokes (Varona et al., 2004). We may gain deeper insights into the structural alterations that underlie motor function compensation by examining movement‐associated cortical structures in patients with glioma with varying mobility states.

Duffau et al. associated this adaptation with brain neuroplasticity, defined as modifying neural connections in the central nervous system (Duffau, 2005, 2006). It encompasses four tiers of compensation: first, functional reconfiguration within and second, around the tumor; third, involvement of distant ipsilateral hemisphere areas; and finally, engagement of contralesional regions. This sequence aligns with the extent of damage (Duffau, 2008). Given this neuroplastic mechanism, tumors can also influence the functionality and cortical structure of the remote sections (Nakajima et al., 2020), paralleling changes in neurological evaluations. These observations indicated a potential link between structural and functional compensation in these patients.

Voxel‐based morphometry (VBM) is a premier analytical neuroimaging tool used to quantify cortical volumetric alterations (Ashburner & Friston, 2000). It facilitates gray matter examination at the voxel level and discerns anatomical variance throughout the brain. Despite its extensive use in diverse neurological and medical conditions, its application in glioma research is limited, primarily because of the glioma's mass effect and imprecision in aligning magnetic resonance imaging (MRI) data with brain templates because of surrounding edema. Despite the infrequency of glioma cases in the motor region and the constraints of current analytical methods, comprehensive studies exploring the relationship between glioma‐induced changes in motor function and cortical alterations are lacking. To investigate structural changes in the sensorimotor cortex secondary to glioma and explore neuroplasticity, an important step is removing the tumor effects during the analysis. In this study, we removed the tumor‐affected area before assessing cortical changes. This method reduces the impact of tumor presence and surrounding tissue infiltration, thereby ensuring a more accurate analysis. Additionally, we explored the correlation between changes in cortical volume and functional impairment following tumor invasion of motor function regions.

Motor functions involve a network of brain regions that control voluntary movement. The primary motor cortex (M1) is essential for initiating movements, containing a map of the body where different regions correspond to specific muscle groups, in an arrangement often illustrated as the “motor homunculus” (Gordon et al., 2023). The supplementary motor area (SMA), located on the midline surface of the frontal lobe, anterior to the primary motor cortex, aids in the planning and sequencing of complex movements (Sheets et al., 2021). These areas work in a complicated fashion, with premotor region planning and SMA sequencing, and ensure precise movement by modulating the activation of M1 (D'Aleo et al., 2022; Ding et al., 2023; Yip & Lui, 2023). In addition, the adjacent BA 8 within the frontal lobe may be responsible for motor planning (Dadario et al., 2023). This study investigated alterations in the GM volume in motor‐related areas and the relationship between alterations in the contralateral GM volume and hand strength grading in patients with unilateral sensorimotor glioma. We hypothesized that contralateral motor‐related areas compensate for damaged hand strength and that the elevated volume of certain brain areas may be the structural basis of motor function compensation.

2. METHODS

2.1. Participants

Glioma patients who were treated at the Fifth Ward at Department of Neurosurgery of Beijing Tiantan Hospital between January 2018 and October 2021 were reviewed. Firstly, we performed an initial screening of clinical and imaging data from 146 patients who met our research criteria for glioma. Subsequently, the patients underwent further selection based on inclusion criteria. The inclusion criteria were as follows: (1) pathologically confirmed glioma according to the 2021 World Health Organization central nervous system tumor classification (Louis et al., 2021); (2) preoperative MRI‐confirmed unilateral tumor involving the precentral gyrus and/or postcentral gyrus with a midline shift <1 cm; (3) age >18 years; (4) no neurological or psychiatric disorders; (5) no history of neurosurgical treatment, radiotherapy, or chemotherapy; and (6) right‐handed according to the Edinburgh Handedness Inventory. Selecting right‐handed patients aimed to minimize potential variability introduced by handedness (Knecht et al., 2000). The exclusion criteria were as follows: (1) head motion >1 mm; (2) head rotation >1°; (3) slice thickness >1.0 mm during MRI; (4) no molecular histopathological report. Finally, a total of 59 patients (lesion in the left hemisphere, n = 32) were included in the study for the subsequent analysis. The patients were categorized into four groups according to the tumor side (left tumor group and right tumor group) and the score of hand muscle strength controlled by the lesional hemisphere (an impaired group with a score of 0–4 and an unimpaired group with a score of 5).

Another 16 healthy controls (HCs) were included in the study after matching for demographic characteristics. The criteria for enrolling HCs were as follows: (1) age > 18 years; (2) no neurological or psychiatric disorders; (3) no history of neurosurgical treatment, radiotherapy, or chemotherapy; and (4) right‐handed according to the Edinburgh Handedness Inventory. The exclusion criteria were as follows: (1) head motion >1 mm; (2) head rotation >1° and (3) slice thickness >1.0 mm in T1‐weighted images. The Ethics Committee of Beijing Tiantan Hospital approved this study.

2.2. MRI scanning

All images were acquired in a 3.0 T Magnetic Resonance Imaging scanner (Siemens, Munich, Germany) at the Beijing Neurosurgical Institute. For high‐resolution three‐dimensional (3D) T1‐weighted images, the parameters were as follows: repetition time (TR) = 2300 ms, flip angle (FA) = 8°, echo time (TE) = 2.30 ms, field of view (FOV) = 240 mm × 240 mm, and voxel size = 0.9 × 0.9 × 0.9 mm³. Regarding the T2 weighted images, the parameters were as follows: TR = 5000 ms, FA = 150°, TE = 105.00 ms, FOV = 230 mm × 230 mm, and voxel size = 0.5 × 0.5 × 4.0 mm³.

2.3. Tumor mask

The masks of gliomas were defined based on regions with a high signal on the T2 weighted images，determined and drawn as a tumor mask. Two neuroradiologists manually drew tumor masks independently using the MRIcron software (http://mccauslandcenter.sc.edu/crnl/tools). If the difference between the masks in the same patient was >5%, a third neuroradiologist with >15 years of experience in glioma diagnosis confirmed the final mask (Figure 1).

FIGURE 1.

The tumor overlapping map of all enrolled patients. The color of each voxel indicates the number of overlap (1–20).

2.4. Preprocessing for VBM analysis

The 3D T1‐weighted images were preprocessed for VBM analysis based on SPM8 package (http://www.fifil.ion.ucl.ac.uk/spm). The preprocessing pipeline was as follows: (1) tumor masks were overlapped with the corresponding individual 3D T1 images using MRIcron; (2) the overlapped areas of the tumor mask and individual T1 images were removed to avoid the influence of the tumor using MRIcron software; and (3) the images without tumors were segmented into gray matter, white matter, and cerebrospinal fluid. This step aims to minimize the registration inaccuracies caused by the mass effect of the tumor. The Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra algorithm was used to normalize the segmented images spatially (Goto et al., 2013). Subsequently, the processed T1 images were spatially normalized, then these fully normalized images were resliced through trilinear interpolation to a final voxel size of 1.5 × 1.5 × 1.5 mm³. (4) Using the tool of checking sample homogeneity using covariance in VBM8 (Farokhian et al., 2017), we checked and confirmed the overall homogeneous quality of the normalized data. (5) Acquire total intracranial volumes (TIVs) using the sum of segmented GM, white matter, and cerebrospinal fluid templates. (6) Smooth the GM templates using an 8‐mm full width at half maximum with a Gaussian kernel. The details of the “estimate and write” and “smooth” options are listed in Appendix S1.

2.5. Region of interest

We focused on the precentral gyrus (PrG), postcentral gyrus (PoG), supplementary motor area (SMA), and premotor area (PMA) to investigate volumetric changes in brain areas. The medial BA 8 was also analyzed because of its proximity to tumors and its possible function in motor planning (Dadario et al., 2023). In our study, we employed brain masks of the superior occipital gyrus (SOG) as part of our VBM analysis to serve as a control measure. This approach was inspired by methodologies in existing literature, notably the study conducted by Almairac et al. (2018), which emphasized the importance of considering control regions to discern disease‐specific changes from normal anatomical variability. The inclusion of the SOG aimed to exclude potential confounding effects attributable to individual differences in brain structure, ensuring that the observed volumetric changes in regions of interest (ROIs) are directly related to the pathological impact of gliomas. The masks of the thalamus were not subjected to comparison because of the possibility of volumetric alterations within the thalamus, which could result from positional shifts and invasion by tumors in the vicinity of the thalamus observed in several patients, and thus were not suitable for controls. We bilaterally selected six subareas of the precentral gyrus (head and face region, caudal dorsolateral area 6, upper limb region, trunk region, tongue and larynx region, and caudal ventrolateral area 6), four subareas of the postcentral gyrus (upper limb, head and face region, tongue and larynx region, area 2, and trunk region), SMA (medial area 6), PMA (dorsolateral area 6 and ventrolateral area 6), medial area 8, and SOG brain masks from the BrainNetome atlas tool (http://atlas.brainnetome.org) (Fan et al., 2016). Then, we resliced and registered the masks of the brain areas to the normalized patient brain images using SPM8 (Table 2).

TABLE 2.

Brain ROI coordinates.

Gyrus	Subregions	Anatomical descriptions	Left hemisphere MNI(X,Y,Z)	Right hemisphere MNI(X,Y,Z)
Precentral gyrus	PrG_L(R)_6_1	Head and face region	−49, −8, 39	55, −2, 33
	PrG_L(R)_6_2	Caudal dorsolateral area 6	−32, −9, 58	33, −7, 57
	PrG_L(R)_6_3	Upper limb region	−26, −25, 63	34, −19, 59
	PrG_L(R)_6_4	Trunk region	−13, −20, 73	15, −22, 71
	PrG_L(R)_6_5	Tongue and larynx region	−52, 0, 8	54, 4, 9
	PrG_L(R)_6_6	Caudal ventrolateral area 6	−49, 5, 30	51, 7, 30
Postcentral gyrus	PoG_L(R)_4_1	Upper limb, head and face region	−50, −16, 43	50, −14, 44
	PoG_L(R)_4_2	Tongue and larynx region	−56, −14, 16	56, −10, 15
	PoG_L(R)_4_3	Area 2	−46, −30, 50	48, −24, 48
	PoG_L(R)_4_4	Trunk region	−21, −35, 68	20, −33, 69
Supplementary motor area	SFG_L(R)_7_5	Medial area 6	−6, −5, 58	7, −4, 60
Premotor area	SFG_L(R)_7_4	Dorsolateral area 6	−18, −1, 65	20, 4, 64
	MFG_L(R)_7_6	Ventrolateral area 6	−32, 4, 55	34, 8, 54
Medial BA8	SFG_L(R)_7_1	Medial area 8	−5,15, 54	7, 16, 54

Note: We use MNI (Montreal Neurologic Institute) template coordinates to report the locations of subregions.

Abbreviations: BA 8, Brodmann area 8; MFG, Middle frontal gyrus; PoG, Postcentral Gyrus; PrG, precentral gyrus; SFG, superior frontal gyrus; SMA, supplementary motor area.

2.6. Muscle strength assessment

Strength measurements of the hand governed by the hemisphere affected by the lesion were obtained from medical records. Muscle strength was assessed within 2 days before tumor resection using the UK Medical Research Council (MRC) muscle strength test. Each patient's hand MRC score was tested accordingly:

Score 0: No muscle contraction detected.

Score 1: Trace of muscle contraction, but no movement at the joint.

Score 2: Movement at the joint, but not against gravity.

Score 3: Movement against gravity but not against added resistance.

Score 4: Movement against some resistance but weaker than normal.

Score 5: Normal strength.

Spearman's correlation analysis and Student's t‐test were performed to investigate further the relationship between hand MRC score and GMVs in motor‐related areas. In our study, only a few individuals exhibited a decline in lower limb strength, and the grading of lower limb strength was either higher or the same as that of hand muscle strength. Therefore, we opted for hand MRC score for our analysis because it represents more severe impairment and provides stronger representativeness.

2.7. ROI statistical analysis

The gray matter volumes under brain masks were extracted from the smoothed gray matter of the T1 images using the REST tool (v1.8) “extract ROI signal” (Song et al., 2011). We then performed an ANOVA test and post‐hoc analysis with Bonferroni correction to compare the volume of masks in the unimpaired, impaired, and HCs groups using GraphPad V8.3.1. We further compared the contralesional superior occipital gyri between the three groups to exclude individual brain differences.

2.8. Voxel‐wise statistical analyses

The VBM8 toolbox (http://dbm.neuro.uni-jena.de/vbm.html) was used to estimate GMV. The GMVs of patients with glioma and HCs were analyzed in SPM 8 using two‐sample t‐test models. We compared patients with glioma and decreased hand muscle strength to HCs, patients with glioma and normal hand muscle strength to HCs, and patients with glioma and decreased hand muscle strength to normal hand muscle strength using a two‐sample t‐test (Figure S1). All two‐sample t‐tests were performed after correcting for covariates, including age, tumor volume, and TIV. Effect size was calculated using Cohen's d calculation. The “explicit mask” was set as the resliced corresponding brain mask files. After automatically analyzing the results, we identified regions with different voxels between the two groups. The p‐value threshold was set at .05, with family‐wise error correction with 20 contiguous voxels. Volume mapping to surface was visualized by BrainNet Viewer (Xia et al., 2013). Additionally, cortical thickness was indirectly inferred from volumetric assessments within predefined cortical masks. This approach assumes a correlation between cortical volume within these masks and cortical thickness, with the rationale being that greater volume indicates thicker cortex in the specified areas. We utilized VBM and ROI analysis for volumetric measurements. These measurements were taken as proxies for assessing cortical thickness in the premotor area, precentral gyrus, and postcentral gyrus.

2.9. Statistical analysis

For all participants, we performed a Student's t‐test for TIV and tumor volume. Chi‐square tests or Fisher's exact tests were performed for histopathology, tumor side, hand MRC score, tumor grade, chromosome 1p/19q status, and mutations in the isocitrate dehydrogenase gene.

3. RESULTS

3.1. Demographic and neurocognitive characteristics

A total of 59 patients with glioma (21 in impaired group, 38 in unimpaired group) and 16 healthy controls matched for basic demographic characteristics were enrolled in this study. There are 11 cases of gliomas located in the left hemisphere of the brain, and 10 cases on the right side. The tumor grades according to the WHO classification are as follows: Grade II—30 cases, Grade III—6 cases, Grade IV—23 cases. Among the three groups of impaired, unimpaired, and HCs, there was no difference in total intracranial volume between any two groups (Table 1). Meanwhile, no significant differences were seen with regard to tumor side (p = .831), WHO grade (p = .331), pathology (p = .512), 1p/19q chromosome status (p = .309), isocitrate dehydrogenase 1 (IDH1) mutation status (p = .312), and tumor volume (p = .357) between impaired group and unimpaired group. The location and frequency of selected gliomas were displayed in Figure 1. The lesions were predominantly located in the frontal and parietal regions, with a focus on the motor functional area.

TABLE 1.

Clinical characteristics of glioma impaired group, unimpaired group, and HCs.

Clinical characteristics	Impaired (n = 21)	Unimpaired (n = 38)	HCs (n = 16)	p value
				Impaired vs. unimpaired	Impaired vs. HCs	Unimpaired vs. HCs
Total intracranial volume (cm3)	1415.72 ± 132.29	1381.17 ± 141.53	1418.06 ± 118.27	.362 a	.956 a	.364 a
Tumor side			—	.831 b
L	11	21
R	10	17
Hand MRC score
0	1	0	0	<.001 c	<.001 c	NaN
3	3	0	0
4	17	0	0
5	0	38	16
WHO grade, n			—	.331 c
II	8	22
III	3	3
IV	10	13
Pathology			—	.512 b
Astrocytoma, IDH‐mutant	6	11
Oligodendroglioma, IDH‐mutant, and 1p/19q‐codeleted	5	14
Glioblastoma, IDH‐wildtype	10	13
1p/19q chromosome status			—	.309 b
Non‐codeletion	16	24
Codeletion	5	14
IDH mutation status			—	.312 b
Wildtype	10	13
Mutation	11	25
Tumor volume (cm3), median (IQR)	57.82 (28.81, 86.89)	39.55 (24.24, 61.08)	—	.357 a

Abbreviation: MRC, Medical Research Council.

^a Results of Student's t‐test.

^b Results of Chi‐square test.

^c Results of Fisher's exact test.

3.2. Comparison of supplementary motor area

The information of ROI masks was summarized in Table 2. For the left hemispheric glioma, voxel‐wise statistical analysis showed that the GMVs of the contralateral SMA were greater in the impaired group than in the unimpaired group (Figure 2a; k = 26, p = .008; with peak at (3, −4, 54), T = 4.19, Z = 3.65, p = .031) and HCs (Figure 2b; k = 272, p < .001; with peak at (15, −7, 73), T = 8.63, Z = 3.62, p < .001). Moreover, the GMVs of both the ipsilateral SMA (Figure 2c; k = 47, p = .014, with peak at (−14, −12, 73), T = 5.85, Z = 4.79, p < .001) and the contralateral SMA (Figure 2d; k = 346, p < .001; with peak at (18, −10, 73), T = 10.59, Z = 3.62, p < .001) were greater in the unimpaired group than in the HCs. Similar results were observed in patients with right hemispheric gliomas. The GMVs of both the ipsilateral SMA (Figure 2e; k = 229, p = .001; with peak at (14, −7, 75), T = 9.42, Z = 6.27, p < .001) and the contralateral SMA (Figure 2f; k = 65, p = .002; with peak at (−14, −12, 73), T = 8.05, Z = 3.62, p < .001) were greater in the unimpaired group than in the HCs. The original T maps of these results were shown in Figure S1.

FIGURE 2.

Brain regions showing significant changes in GMV. Comparisons of GM volume in different brain areas between HCs, glioma patients with impaired hand muscle strength and glioma patients with unimpaired hand muscle strength with the corresponding mask after controlling effects for age, tumor volume, and total intracranial volume. The two‐sample t test map (volume mapping to surface) of (a) Right SMA between the impaired group and the unimpaired group with left‐side gliomas. (b) Right SMA between the impaired group with left‐side gliomas and HCs. (c) Left SMA between the unimpaired group with left‐side gliomas and HCs. (d) Right SMA between the unimpaired group with left‐side gliomas and HCs. (e) Right SMA between the unimpaired group with right‐side gliomas and HCs. (f) Left SMA between the unimpaired group with right‐side gliomas and HCs. (g) Right medial BA8 between the impaired group with left‐side gliomas and HCs. (h) Right medial BA8 between the unimpaired group with left‐side gliomas and HCs. (i) Left medial BA8 between the impaired group and the unimpaired group with right‐side gliomas. (j) Right medial BA8 between the impaired group and the unimpaired group with right‐side gliomas. (k) Right medial BA8 between the unimpaired group with right‐side gliomas and HCs.

Regarding the ROI analysis the projection of the masks of the ROIs on the brain surface was shown in Figure 3a. For the left hemispheric glioma, the GMV of the contralateral SMA increased in the impaired group (0.483 ± 0.037 cm³) compared with the unimpaired group (0.438 ± 0.037 cm³) and HCs (0.394 ± 0.040 cm³) (Figure 3b; impaired vs. unimpaired: p = .009, Cohen's d = 1.22; impaired vs. HCs: p < .001, Cohen's d = 2.31; unimpaired vs. HCs: p = .003, Cohen's d = 1.14). However, for right hemispheric glioma, no difference in the GMV of the ipsilateral or contralateral SMA was found among the three groups (Figure S2).

FIGURE 3.

Comparison of GMV among different groups. (a) Localization of the regions of interest in the brain. Comparison of GMV in the supplementary motor area (b), medial Brodmann area 8 (c), superior frontal gyrus (d), middle frontal gyrus (e), postcentral gyrus (f), postcentral gyrus (g), and superior occipital gyrus (h). *p < .05; **p < .01; ***p < .001; ns, no significance.

3.3. Comparison of medial Brodmann area 8

For left hemispheric glioma, voxel‐wise statistical analysis showed that the GMVs of both the contralateral medial BA8 in the impaired group (Figure 2g; k = 68, p = .003; with peak at (14, 21, 66), T = 5.70, Z = 4.42, p = .002) and contralateral medial BA8 (Figure 2h; k = 383, p < .001; with peak at (9, 15, 69), T = 6.17, Z = 4.97, p < .001) in the unimpaired group were greater than those in the HCs. For right hemispheric gliomas, the GMVs in the bilateral medial BA8 (Figure 2i,j; k = 83, p < .001 contralaterally and k = 36, p = .003 ipsilaterally, with peak at (−6, 17, 46), T = 5.36, Z = 4.24, p = .006 contralaterally and (5, 17, 57), T = 5.30, Z = 4.21, p = .007 ipsilaterally) were greater in impaired group compared to unimpaired group. The GMVs in the ipsilateral medial BA8 (Figure 2k; k = 60 and 63, p = .002; with peak at (14, −7, 75), T = 5.44, Z = 6.27, p < .001, (8, 6, 72), T = 5.29, Z = 4.37, p < .001) were greater in unimpaired group compared to HCs.

For the left hemispheric glioma, ROI analysis indicated that the GMV of the contralateral medial BA8 increased in the impaired group (0.501 ± 0.039 cm³) and the unimpaired group (0.474 ± 0.037 cm³) compared with HCs (0.427 ± 0.046 cm³) (Figure S2; impaired vs. HCs: p < .001, Cohen's d = 0.71; unimpaired vs. HCs: p = .003, Cohen's d = 1.13). For the right hemispheric glioma, the GMV of the ipsilateral medial BA8 increased in the impaired group (0.530 ± 0.044 cm³) compared with the unimpaired group (0.482 ± 0.039 cm³) and HCs (0.427 ± 0.045 cm³) (impaired vs. unimpaired: p = .024, Cohen's d = 1.16; impaired vs. HCs: p < .001, Cohen's d = 2.31; unimpaired vs. HCs: p = .002, Cohen's d = 1.30) (Figures 3c).

3.4. Comparison of premotor area, precentral gyrus, and postcentral gyrus

A comparative cortical thickness analysis across various motor‐related areas revealed notable differences. The premotor area (Figure 3d,e) exhibited increased cortical thickness in subjects with developing gliomas. Compared with HCs, this thickening was statistically significant. Similar observations were made for the precentral and postcentral gyri. The areas of the precentral gyrus associated with hand and trunk motor functions (Figure 3f) also demonstrated a statistically significant increase in cortical thickness (p < 0.001). The postcentral gyrus (Figure 3g), which is involved in sensory processing, followed this pattern of increased thickness (p < .001). Details of the voxel‐based and ROI analysis are provided in Figure S3.

3.5. Comparison of superior occipital gyrus

We selected the contralesional superior occipital gyrus as the control ROI mask based on previous studies (Almairac et al., 2018). No significant results were found for VBM. Furthermore, there was no difference in GMV between groups (Figure 3h), indicating that choosing the superior occipital gyrus as a control was reasonable. The summary of the results of GMV in ROI analysis is presented in Table 3.

TABLE 3.

The alterations of GMV in the regions of interest.

Brain regions	Tumor side	Unimpaired vs. HCs	Impaired vs. HCs	Impaired vs. unimpaired
R SMA	L	↑↑	↑↑↑	↑↑
R Medial BA8	R	↑↑	↑↑↑	↑
L SFG_7_4	R	↑↑↑	↑↑↑	—
R SFG_7_4	L	↑↑↑	↑↑↑	—
L MFG_7_6	R	↑↑↑	↑↑	—
R MFG_7_6	L	↑↑↑	↑↑↑	—
L PrG_6_2	R	↑↑↑	↑↑↑	—
R PrG_6_2	L	↑↑↑	↑↑↑	—
L PrG_6_3	R	↑↑↑	↑↑↑	—
R PrG_6_3	L	↑↑↑	↑↑↑	—
L PrG_6_4	R	↑↑↑	↑↑↑	—
R PrG_6_4	L	↑↑↑	↑↑↑	—
L PoG_4_1	R	↑↑↑	↑↑	—
R PoG_4_1	L	↑	↑↑	—
L PoG_4_4	R	↑↑↑	↑↑↑	—
R PoG_4_4	L	↑↑↑	↑↑↑	↑
L SOG	R	—	—	—
R SOG	L	—	—	—

Note: ↑ (p < .05), ↑↑ (p < .01), and ↑↑↑ (p < .001) mean increment of GMV. R, Right; L, Left.

Abbreviations: BA 8, Brodmann area 8; MFG, middle frontal gyrus; PoG, postcentral gyrus; PrG, precentral gyrus; SFG, superior frontal gyrus; SMA, supplementary motor area; SOG, superior occipital gyrus.

3.6. Correlation of GMV and hand MRC score

Spearman's correlation analysis was performed on the GMVs of the bilateral SMA, bilateral medial BA8, contralateral PMA (dorsolateral area 6, ventrolateral area 6), contralateral precentral gyrus (caudal dorsolateral area 6, upper limb region, and trunk region), and contralateral postcentral gyrus (upper limb, head and face region, tongue and larynx region, area 2, and trunk region). The hand MRC score were significantly negatively correlated with the GMV of ipsilateral medial BA8 (Figure 4a; R = −0.439; p < .001), the GMV of contralateral medial BA8 (Figure 4c; R = −0.395; p = .002), the GMV of contralateral SMA (Figure 4e; R = −0.489; p < .001). For the same three groups of patients, when we compare patients with normal hand muscle strength (Score 5) to those with decreased muscle strength (Score 0–4), the latter groups showed an increase in volume of BA8 or SMA (Figure 4b,d,f).

FIGURE 4.

Volume changes of in sensorimotor area patients with different degrees of hand muscle strength impairment. Correlation of brain mask volumes and hand MRC score in sensorimotor area glioma patients: (a) ipsilateral medial BA8; (c) contralateral medial BA8; (e) contralateral SMA. Volume differences in the sensorimotor area between glioma patients with normal hand muscle strength (Score 5) and those with impaired hand muscle strength (Score 0–4): (b) ipsilateral medial BA8; (d) contralateral medial BA8; (f) contralateral SMA. **p < .01; ***p < .001.

4. DISCUSSION

Gliomas in the motor and sensory areas cause structural changes in the cortices of some functional areas. The reorganization of these cortical centers is particularly noticeable in the slow‐progressing process (Duffau et al., 2003). As slow‐growing lesions, low‐grade‐gliomas can cause short‐distance functional (ipsilateral) and long‐distance (contralateral) compensation. In this study, we found that the changes in cortical structure in patients with glioma in the motor area were mainly thickening of the contralateral premotor area, areas related to the hand and trunk in both the postcentral gyrus and the precentral gyrus, and that impaired strength was related to the greater volume of the cortex of the SMA and medial BA8.

4.1. Brain structure reorganization

The brain is a dynamic functional organ in which motor function compensation processes can occur (Fang et al., 2022). However, the compensatory process is still subject to a structural basis, and not every brain region has the same probability of compensation. The motor cortex differs from the rest of the cortex in the cell layer, and motor function is difficult to perform in cortices without abundant pyramidal cells (McColgan et al., 2020). In this study, we selected the SOG as a control, according to previous studies, which may not undergo structural changes, thus excluding differences caused by differences in individual brain volumes (Almairac et al., 2018). When selecting the target brain regions involved in possible compensation, we selected brain regions in the sensorimotor cortex that were closely related to the motor network. In addition, we selected the brain region of the medial BA8 near the tumor, as functional remodeling may occur around the tumor (Duffau, 2014).

Increased volumes in several brain areas were observed secondary to sensorimotor glioma compensation. Our results indicated that the premotor area, areas related to the hand and trunk in both the postcentral and precentral gyri, were thicker, possibly because of the development of glioma. The cortical volumes of the SMA and medial BA8 were greater in the impaired group than in the unimpaired group, which was related to the difference in strength. Finally, as the score of hand MRC decreased, the SMA and medial BA8 cortices showed greater volumes, consistent with previous studies showing that the extent of plasticity was correlated with injury severity (Di Pino et al., 2014; Hu et al., 2020). These results show that the compensatory process may begin in broad sensorimotor areas. As the disease develops until motor dysfunction occurs, the SMA and medial BA8 become thicker to compensate.

Previous studies have mainly focused on patients with a certain functional status, either with or without motor dysfunction, and no research has revealed differences in brain areas in patients with both functional statuses. Our study analyzed various brain areas that may be involved in compensation and found that the SMA and medial BA8 cortex were potential areas that became thicker in the impaired group. Our findings provide a reliable basis for predicting the extent of motor function injury in patients with gliomas.

The SMA plays an important role in initiating and coordinating complex movements, including reaching, grasping, speaking, bilateral hand coordination (Sheets et al., 2021), and cognitive functions (Pinson et al., 2022). Brodmann area 8 is a relatively mysterious brain region, and its specific function remains unclear. It has been found that the lateral side of BA8 is related to motor, cognition, and language functions, and the caudal frontal eye area is responsible for visual attention and eye movement control. Traditionally, the medial area of BA8 and the medial area of BA9, which is also part of the dorsomedial prefrontal cortex, are involved in “theory of mind” and social cognitive functions (Watanabe, 2017). According to recent studies, the medial area of BA8 is a central executive network region located between two SMA regions, which may participate in motor planning by interacting with higher‐order networks (Briggs et al., 2021). In this study, we found that the increased volume of the medial BA8 in the impaired group was negatively correlated with hand MRC score, indicating that the medial BA8 may be involved in regulating motor function.

The reshaping of brain structure and function caused by gliomas is crucial for postoperative functional recovery and even survival in patients. However, current research in this area primarily consists of cross‐sectional studies, lacking longitudinal observations over time. In the future, we will also accumulate data from patients at different time points after surgery to observe the coordination between changes in brain structure and functional recovery. Through these studies, we hope to identify imaging biomarkers from the perspective of brain structural and functional changes, enabling early prediction of patients' functional recovery and their application in clinical settings.

4.2. Consistency of the results

Our study utilized both voxel‐wise and ROI‐based analyses to provide a comprehensive understanding of the neuroplastic changes occurring in patients with gliomas. The consistent results across our analyses highlight the primary motor gyrus and the supplementary motor area as key regions exhibiting significant volumetric changes. These areas likely represent focal points of neuroplasticity, where the brain attempts to compensate for the functional impairments caused by the glioma. The identification of these regions underscores the potential for targeted therapeutic interventions aimed at enhancing motor function and quality of life for patients with gliomas. However, it is also critical to acknowledge and discuss the instances where voxel‐wise and ROI‐wise analyses yielded divergent results. These discrepancies may stem from the inherent differences in sensitivity and specificity between the two analytical methods or from the complex nature of neuroplastic changes themselves. Future studies are encouraged to delve deeper into these inconsistencies, as they may reveal nuanced aspects of brain reorganization in the context of glioma.

The absence of observed compensatory changes in certain brain regions raises intriguing questions about the underlying neural dynamics and the variable capacity for neuroplasticity in the presence of gliomas. Several factors may contribute to the lack of compensatory mechanisms observed in specific areas. First, the extent and nature of neuroplastic changes are likely influenced by the tumor's location, size, and grade. Tumors in certain brain regions may disrupt local neural networks to an extent that limits the potential for nearby or connected regions to adapt and compensate (Bao et al., 2023). Additionally, the individual variability in brain anatomy and functional organization could result in differing capacities for neuroplasticity, where some patients exhibit robust compensatory changes while others do not. Furthermore, the stage at which our study observations were made could play a critical role. Neuroplasticity is a dynamic process that evolves over time, and the timing of our assessments may have captured a snapshot where compensatory mechanisms had not yet fully developed or were already overwhelmed by tumor progression. For example, the differential findings between left and right hemisphere glioma patients in terms of GMV changes in the SMA may reflect the distinct functional specializations of the two cerebral hemispheres, especially considering the dominance of the left hemisphere for motor functions in right‐handed individuals. The left hemisphere, which is typically dominant for language and motor functions in the majority of right‐handed individuals, may exhibit a greater capacity or necessity for plasticity and reorganization in response to gliomas. This is in part due to its critical role in controlling fine motor skills, which includes hand movements. Moreover, the presence of a glioma in the left hemisphere could disrupt these highly specialized functions more so than a similar lesion in the right hemisphere, necessitating more pronounced compensatory neuroplastic changes in the contralateral (right) SMA or other motor‐related areas to maintain motor function. This interpretation aligns with the concept that brain plasticity and compensatory mechanisms are highly dynamic and context‐dependent, influenced by the specific functions compromised by the glioma, the hemisphere involved, and the individual's pre‐existing neural organization and plasticity potential.

5. LIMITATIONS

This study has certain limitations. First, the sample size was relatively small. Second, hand MRC score was quantified using gradings instead of numerical units, potentially leading to underestimating the strength variations. Further studies on the correlation between volume changes in specific brain regions and functional outcomes may enhance our ability to predict motor function recovery. The discovery of increased volume in the pertinent brain regions in this study raises questions regarding its functional compensatory significance. Further research is imperative to understand the intricate relationship between function, structure, and activation in sensorimotor areas in patients with glioma. Furthermore, the accuracy of GMV results is inherently limited by the VBM method. For example, edema might induce the overestimation in GMV calculations. Region‐based segmentation and intensity normalization help to address this issue. In addition, studies have suggested that combining VBM with surface‐based morphometry (SBM) may improve the accuracy of detecting morphological changes when comparing data from patients and controls (Goto et al., 2022). Two other recent approaches have emerged as potential future directions for accurately estimating cortical morphological changes: a multi‐modal parcellation method that combines structural and functional images, and a synthetic segmentation method that utilizes multi‐contrast images, including T1‐ and proton density‐weighted images. In future studies, we will also explore the use of these novel methods.

6. CONCLUSION

In conclusion, this study elucidates the dynamic nature of the brain's compensatory mechanisms in response to gliomas in the sensorimotor area. These findings underscore the remarkable ability of the brain to undergo structural reorganization, as evidenced by the thickening of the contralateral premotor area and related regions in the postcentral and precentral gyri in response to motor impairment. Specifically, the increased volumes of the SMA and medial BA8 in the presence of motor dysfunction demonstrate that these areas may serve as important centers for functional compensation. This study contributes to the perspective of neurosurgical evaluation by offering insights into the structural changes occurring in both the impaired and unimpaired groups, emphasizing the role of these changes in motor strength variance.

AUTHOR CONTRIBUTIONS

Study concept and design: Yuhao Guo, Shengyu Fang. Data acquisition and analysis: Yuhao Guo, Shengyu Fang, Hongbo Bao. Statistics/verified analytical method: Yuhao Guo and Hongbo Bao. Writing the first draft: Yuhao Guo, Hongbo Bao, Shengyu Fang, Zhishuo Wei and Yinyan Wang. Supervision study: Shengyu Fang, Tao Jiang, and Yinyan Wang. Read and approved final version: All authors.

FUNDING INFORMATION

Beijing Natural Science Foundation, Grant Number: JQ23040; Research Unit of Accurate Diagnosis, Treatment, and Translational Medicine of Brain Tumors, Chinese Academy of Medical Science, Grant Number: 2019‐I2M5‐021; the Public Welfare Development and Reform Pilot Project of Beijing Medical Research Institute, Grant Number: PXM2019_026280_000008.

CONFLICT OF INTEREST STATEMENT

The authors declare that they have no conflict of interest.

Supporting information

APPENDIX S1 Supporting information.

Guo, Y. , Bao, H. , Wei, Z. , Fang, S. , Jiang, T. , & Wang, Y. (2024). Structural changes in eloquent cortex secondary to glioma in sensorimotor area. Human Brain Mapping, 45(8), e26723. 10.1002/hbm.26723

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.