Sage Journals: Discover world-class research

Abstract

Background. Drastic functional reorganization was observed in the ipsilateral primary motor cortex (M1) of a Paralympic long jumper with a unilateral below-knee amputation in our previous study. However, it remains unclear whether long-term para-sports are associated with ipsilateral M1 reorganization since only 1 athlete with amputation was investigated. Objective. This study aimed to investigate the relationship between the long-term para-sports and ipsilateral M1 reorganization after lower limb amputation. Methods. Lower limb rhythmic muscle contraction tasks with functional magnetic resonance imaging and T1-weighted structural imaging were performed in 30 lower limb amputees with different para-sports experiences in the chronic phase. Results. Brain activity in the ipsilateral primary motor and somatosensory areas (SM1) as well as the contralateral dorsolateral prefrontal cortex, SM1, and inferior temporal gyrus showed a positive correlation with the years of routine para-sports participation (sports years) during contraction of the amputated knee. Indeed, twelve of the 30 participants who exhibited significant ipsilateral M1 activation during amputated knee contraction had a relatively longer history of para-sports participation. No significant correlation was found in the structural analysis. Conclusions. Long-term para-sports could lead to extensive reorganization at the brain network level, not only bilateral M1 reorganization but also reorganization of the frontal lobe and visual pathways. These results suggest that the interaction of injury-induced and use-dependent cortical plasticity might bring about drastic reorganization in lower limb amputees.

Keywords

magnetic resonance imaging lower extremity motor cortex prefrontal lobe amputees para-sports

Introduction

Structural and functional brain reorganization occurs following peripheral lesion such as amputation.^1,2 For instance, motor-related areas such as the premotor cortex (PMC), precentral gyrus, and postcentral gyrus exhibited decreased cortical thickness after lower limb amputation (LLA).^3,4 Additionally, gray matter (GM) volume in the vision motor-related and subcortical areas was also reduced following LLA.^4-6 Moreover, functional connectivity between motor-related areas across hemispheres was reduced, accompanied by structural changes as represented by weakening of the corpus callosum fibers.^4,7 These changes are interpreted as degenerative maladaptations of the brain after LLA and have been associated with phantom limb pain.^8,9

Evidence of maladaptation after LLA is consistent across various studies; however, several studies have shown exceptional results regarding the reorganization of the ipsilateral primary motor cortex (M1) of the amputated limb. For example, increased corticospinal excitability ipsilateral to the amputated side has been demonstrated.^10-12 Moreover, the drastic expansion of the M1 areas innervating lower-limb muscles and increased signal intensity in the ipsilateral M1 was observed with functional MRI (fMRI) during contraction of the residual limb in a long jumper with a unilateral LLA who holds the current long jump world record (8.62 m) but not in non-athletes with a unilateral LLA nor able-bodied long jumpers.¹³ Undoubtedly, higher-level control of the athletic prosthesis is needed to achieve higher performance; therefore, it was inferred that ipsilateral M1 involvement would be required to optimize the motor function for precise prosthesis control.

In the corticospinal tract (CST), it is anatomically known that most parts of the fibers descend to the opposite side via the spinal cord, while up to 15% of intrinsic fibers could descend to the ipsilateral spinal cord directly in non-human primates.^14,15 Although not fully understood in humans, studies using transcranial magnetic stimulation (TMS) have suggested the existence of an ipsilateral pathway, especially in the proximal muscles, even though functional significance has not been clearly defined.¹⁶ Recently, several studies have reported that the ipsilateral motor pathway plays a crucial role in compensating for motor impairment in both upper and lower limbs after subcortical brain injury.^17-21 This phenomenon can also be seen after spinal cord injury^22,23; therefore, ipsilateral M1 activation might help compensate for the functional deficit after movement disorders.

Based on these studies, the human ipsilateral pathway descending from the ipsilateral M1 has intrinsic potential to contribute to the improvement of motor function. Although people with LLA do not suffer central nervous system injury, ipsilateral M1 activation could be considered a positive adaptation for the optimization of a residual limb as with other movement disorders. In the case of upper limb amputees, most researchers have clarified an enlargement and lateral displacement of the motor representation of residual limb,¹⁰ while some studies found robust ipsilateral cortical activity when controlling their intact hands, regardless of the unknown mechanism of inducement.²⁴ Therefore, it is assumed that factors other than just amputation might be required to induce ipsilateral M1 involvement. Specifically, as para-sports require intense use and precise control of the prosthesis, this may very likely cause reorganization of the ipsilateral M1. It is well known that intense limb use induces structural and functional plasticity (ie, use-dependent/learning-induced plasticity), and such plasticity has been associated with motor performance.^25-27 Thus, it is hypothesized that intense limb use after amputation would induce the involvement of the ipsilateral motor pathway descending from the ipsilateral M1 on the amputated side to optimize the motor function for precise prosthesis control. Nonetheless, whether long-term para-sports are associated with ipsilateral M1 reorganization remains unclear since only 1 amputee athlete and 4 amputee non-athletes were evaluated in our previous study.¹³ Therefore, the purpose of the present study was to investigate the effects of long-term para-sports training on post-LLA brain activation. Additionally, to analyze whether GM volume was related to sports years, we also measured T1 images and performed voxel-based morphometry (VBM).

Methods

Participants

This study included 30 LLA individuals in the chronic phase, which is characterized by the stable condition of the amputated parts for more than a year since hospital discharge (age, 39.7 ± 14.5 years [range: 15–70]; 24 men and 6 women; 21.1 ± 14.8 post-injury years [range: 1–60], 12.6 ± 10.8 years of para-sports experience [range: 0–52], 14 transtibial [TT] and 16 transfemoral [TF] amputations, 14 had been amputated on the left body side). Table 1 shows the demographics and relevant clinical characteristics of all participants. The participants were recruited between April 2018 and February 2021 through referrals from prosthetists and a sports club for amputees using the following eligibility criteria: (1) chronic phase after unilateral LLA, (2) no phantom pain and abnormal peripheral sensation in the amputated limbs, (3) lacking medical or neurological diseases, and (4) no experiences of neurotrauma, psychotropic medication, or MRI contraindication. We excluded the T1 image of Sub. 5 and fMRI images of Sub. 20 due to abnormal machine noise and excessive head movement (ie, >2 mm for translation or 2° for rotation), respectively. This study was conducted under the tenets of the Declaration of Helsinki (2000 revision) with the approval of the human ethics committee of the Graduate School of Arts and Sciences, The University of Tokyo (2018, approval no. 581-2). Before the experiment, each participant received a detailed explanation of the experimental procedures and signed a consent form.

Table 1.

Population demographics and relevant clinical information from the participants.

Subject	Sex	Age (years)	Age at amputation (years)	Amputation at left/right	Amputation at tibia/femur	Time since amputation (years)	Time spent sports (years)	Sports
1*	M	21	5	L	TT	16	6	Athletic sports
2*	M	28	14	R	TT	14	12	Athletic sports
3*	M	36	7	L	TT	29	23	Athletic sports
4	M	37	17	L	TF	20	17	Athletic sports
5*	M	32	2	R	TT	30	25	Amputee soccer
6*	M	41	4	L	TT	37	24	Tennis
7	M	23	2	L	TF	21	10	Badminton
8	M	38	18	R	TT	20	18	Athletic sports
9	M	46	45	R	TF	1	1	Para power lifting
10	M	32	26	R	TT	6	5	Wheelchair basketball
11*	M	54	25	R	TF	29	20	Swimming
12	M	66	17	R	TF	49	0	none
13	M	49	21	R	TT	28	10	Baseball
14*	M	61	7	R	TF	54	52	Mountain climbing
15	M	49	32	R	TT	17	15	Athletic sports
16	M	38	20	L	TF	18	10	Athletic sports
17*	M	45	39	L	TF	6	6	Mountain climbing
18*	F	70	10	L	TT	60	30	Mountain climbing
19	M	69	66	L	TF	3	1	Athletic sports
20	F	15	14	L	TT	1	1	Athletic sports
21	M	26	3	L	TF	23	13	Athletic sports
22	M	46	26	R	TF	20	4	Amputee soccer
23	M	30	20	L	TF	10	6	Paratriathlon
24	M	24	17	R	TT	7	3	Nordic Skiing
25*	F	20	1	R	TF	19	15	Swimming
26	F	53	47	R	TT	6	3	Athletic sports
27	M	39	16	R	TF	23	10	Sitting volleyball
28	M	37	24	R	TF	13	7	Golf
29*	F	23	9	L	TT	14	10	Athletic sports
30*	F	44	5	L	TF	39	20	Athletic sports
Mean ± SD		40 ± 15	19 ± 15			21 ± 15	13 ± 11

The participants who showed significant activation of the ipsilateral primary motor cortex during the amputated knee task were shown in bold font and added an asterisk beside the number of subjects. In addition, 22 of the 30 participants used their prostheses during sports, and 8 did not. These 8 participants were underlined in the sports section. Note, 6 of those 8 participants also used the residual limb a lot in sports scenes such as swimming, soccer, and sitting volleyball (Sub. 5, 9, 11, 22, 25, 27). One person (Sub.10) used a wheelchair in the sports scene but used his prostheses a lot in his sports training. Overall, 29 of the 30 participants used the residual limb a lot with or without a prosthesis in the sports scene, except for Sub.12 (no sports history).

Experimental Procedures

All participants performed rhythmic isometric muscle contraction tasks that comprised 6 motor tasks involving both legs (right and left) and 3 joints (ankle, knee, and hip) in an MRI scanner. The all participants were instructed to perform voluntary isometric contractions of the knee (rectus femoris) and hip joint (gluteus maximus) at constant rhythm of 1 Hz. The ankle joint on the non-amputated side performed the plantarflexion movement at 1 Hz, while an imagination task of cyclic plantarflexion from the first-person perspective without muscle contraction was assigned on the amputated side.²⁸ Before MRI measurement, the participants completed questionnaires, including 1) medical history: all previous diseases; 2) injury date: date of amputation and post-amputation duration, and 3) sports history: sport played by the participants at least once a week and number of years spent playing. For fMRI measurements, participants performed the contraction task using each lower limb joint. As a yellow fixation point was flashed on the screen at 1 Hz, instructions such as “L Knee” or “R Hip” were presented underneath. Subsequently, the participants contracted these muscles at a 20% maximum voluntary contraction (MVC) following the blinking yellow fixation point (Figure 1A).

Figure 1.

Experimental setup and group level brain activity. (A) The participants lay on their backs in the MRI gantry and were instructed to observe the screen and perform a contraction task on each lower limb joint during fMRI. (B) Brain activity in each task, as indicated by second-level group whole-brain analysis.

Before entering the MRI scanner, the participants underwent a practice session of contracting each muscle at 20% MVC without contracting other muscles during the task using surface electromyogram (EMG). Specifically, we conducted the following 2 measures to prevent unnecessary muscle contractions during the task. First, during the practice session, EMG was used to ensure that unnecessary muscle contractions for body stabilization did not occur. For correct learning of the task, the participants were asked to complete 20 consecutive sessions. Actual muscle activity is described in the supplementary material, and there was no unnecessary muscle contraction. Second, during the actual scan, 2 monitors, 1 on the side of the participant and the other on the parietal side of the head, showed whole body and checked constantly that the task was being performed correctly and without unnecessary muscle contractions. We monitored their bodies and confirmed that there was no visible movement due to the unnecessary muscle contraction during the task. As it is difficult to see muscle contractions of hip muscles directly under the supine position, we carefully checked and confirmed that there was no movement of the pelvis or femur accompanied with muscle contractions. Detailed methods and results are described in Supplementary file 1.

Subsequently, the participants entered the MRI machine while lying on their backs and watched the monitor while performing the contraction task as previously practiced over 3 fMRI scan sessions (Figure 1A). We intensively monitored their limbs during the task using 2 cameras to ensure proper task performance and that no other muscles were contracted, as mentioned above. Each session included an 8-s pre-trial dummy scan, 6 alternate task repetitions (6 movement types × 1 repetition), and rest periods. The 6 movements were randomly ordered to impede the prediction of the subsequent task. This order was consistent across all participants. Both task- and rest-periods lasted 20 s; therefore, each session lasted 268 s (Figure 1B), which was repeated thrice. During the tasks, T2* weighted (T2w) images were obtained using a 3.0T MRI (MAGNETOM Prisma, Siemens, Munich, Germany). Finally, high-resolution T1 weighted (T1w) structural images were acquired.

Scanning Procedure

All MRI data were acquired using a 3.0-T MRI scanner with a 64-channel head coil. We obtained functional T2w echo-planar images reflecting the BOLD signal using the following parameters: repetition time (TR) = 2000 ms, echo time (TE) = 25 ms, flip angle = 90°, field of view (FOV) = 192 mm, 39 contiguous axial slices acquired in interleaved order, thickness = 3.0 mm, in-plane resolution = 3.0 × 3.0 mm and bandwidth = 1776 Hz/pixel. These image acquisition parameters were chosen based on previous studies using fMRI during the same contraction task.¹³ Siemens auto-align was run at the start of each session. For structural analysis, high-resolution T1w structural images were acquired using a 3D magnetization-prepared rapid acquisition with gradient echo pulse sequence using the following parameters: TR = 2000 ms, TE = 2.9 ms, flip angle = 9.0°, FOV = 256 mm, 176 contiguous axial slices, thickness = 1.0 mm, and in-plane resolution: 1.0 × 1.0 mm.

MRI Scans Flipping

Initially, we flipped all MRI scans of the participants with left-side amputations in the sagittal plane to normalize all amputated sides to the right.^29,30 Therefore, the right and left hemisphere were considered as ipsilateral and contralateral to the amputated leg, respectively. The images were subsequently analyzed.

Task fMRI Analysis

For all sessions, the first 4 scans were excluded to allow for magnetization equilibrium. The remaining data were analyzed using Statistical Parametric Mapping 12 (SPM12, Wellcome Department of Imaging Neuroscience, London, UK; www.fil.ion.ucl.ac.uk/spm) software implemented in MATLAB (Mathworks, Sherborn, MA) using the standard preprocessing pipeline. Preprocessing was performed as follows: 1) Realignment: all images were realigned to the first volume; 2) Coregistration: the T1w structural and average echo-planar imaging (EPI) images from each fMRI session were aligned to superimpose the head-position information; 3) Normalization: all EPI scans were normalized to the Montreal Neurological Institute (MNI) space (TPM.nii in SPM12) by referencing normalization parameters of segmentation of the structural scan; and 4) Smoothing: EPI scans were smoothed using a Gaussian kernel of 8 mm.

After preprocessing, statistical analyses were performed using 2 steps. First, we performed single-participant (first-level) analyses using a general linear model to depict the general brain activation features of each participant. Head motion parameters obtained during the preprocessing stage were included as nuisance regressors in each session to minimize the effects of head motion artifacts. We created the following 6 contrasts: (1) right (amputated) ankle contraction > rest, (2) left (non-amputated) ankle contraction > rest, (3) right (amputated) knee contraction > rest, (4) left (non-amputated) knee contraction > rest, (5) right (amputated) hip contraction > rest, and (6) left (non-amputated) hip contraction > rest. Subsequently, to determine the activated areas in each task of the group effect (second-level), we performed a one-sample t-test for each task based on a random-effects model. Statistical thresholds in both first and second-level analyses were set at a voxel and cluster level of P < .001 (uncorrected) and P < .05 (family-wise error [FWE] corrected).

To identify the activated brain regions correlated with injury and sports years, we used multiple regression models to determine the correlation of brain activity during each joint contraction with injury/sports years. When performing a multiple regression analysis, age, injury, and sports years were added as covariates. Statistical thresholds were set at a voxel and cluster level of P < .001 (uncorrected) and P < .05 (family-wise error [FWE] corrected), respectively. Furthermore, to verify whether the results differ depending on the amputated part, the TT and TF amputees were compared by drawing an approximate line based on the least-squares method in each group and calculating its slope. Finally, to assess whether correlations were amputated part-specific, we evaluated brain activity during the same task in the non-amputated side (ie control conditions and areas) using the inverted x-coordinate of MNI locations that were significantly correlated with sports years.

T1 Structural Image Analysis

The GM volume of each participant was calculated using the SPM12 standard preprocessing pipeline. T1w images were initially segmented for GM, white matter (WM) and cerebrospinal fluid, based on the International Consortium of Brain Mapping template for East Asian populations. Next, a normalization procedure was conducted, where GM images were registered to a standard MNI space to interpolate differences in each brain shape using high-dimensional DARTEL normalization. In the final preprocessing phase, a Gaussian kernel of 12 mm was used to smooth the normalized T1w structure images.

To identify whether the regional GM volume correlated with injury and sports years, we performed a correlation analysis of GM parameter estimates with injury or sports years using multiple regression models. Age, injury, and sports years were added as covariates. Further, we used total intracranial volume as a nuisance regressor to minimize the effect of each brain size. The statistical threshold was set at a voxel and cluster level of P < .001 (uncorrected) and P < .05 (FWE corrected). Furthermore, between-group differences were verified by drawing an approximate line based on the least-squares method in each group and calculating its slope.

Results

Task-Related Brain Activity

Figure 1B and Table 2 summarizes brain activity during each task. The contralateral M1 and ipsilateral cerebellum were significantly activated in the non-amputated ankle task (contralateral M1: 4105 voxels; ipsilateral cerebellum: 1295 voxels). The bilateral M1, cerebellum, and motor-related areas were activated in the non-amputated hip task (bilateral M1 extending into motor-related areas: 9971 voxels; ipsilateral cerebellum: 1859 voxels; operculum: 975 voxels). As in the ankle task, the contralateral M1 and ipsilateral cerebellum were significantly activated in the non-amputated knee task (contralateral M1: 1776 voxels; ipsilateral cerebellum: 553 voxels). However, compared with the non-amputated knee task, extensive motor-related regions were activated during the amputated knee task (contralateral M1 extending into ipsilateral M1: 3378 voxels; ipsilateral cerebellum: 2295 voxels; operculum: 2708 voxels). The activation areas in the amputated hip task were like that of the non-amputated side (bilateral M1 extending into motor-related areas: 5102 voxels; ipsilateral cerebellum: 628 voxels; operculum: 2351 voxels) (Table 2). In the amputated ankle task, wider motor-related areas were activated, including the contralateral central operculum, bilateral M1, contralateral supplementary motor area (SMA), ipsilateral cerebellum and bilateral supramarginal gyrus (Figure 1B and Table 2).

Table 2.

Brain Regions Activated During the Muscle Contraction Tasks in All Participants.

Area	Non-amputated Ankle					Area	Amputated Ankle
	Cluster size	MNI coordinates					Cluster size	MNI coordinates
	Cluster size	x	y	z	z value		Cluster size	x	y	z	z value
C_Precentral gyrus	4105	6	−18	68	6.76	I_Central opercluum	24 825	−46	4	10	6.35
I_Cerebellum	1295	−14	−38	−26	6.65	C_Anterior insula		−44	8	−2	5.94
C_Parietal operculum	1025	46	−28	24	5.63	C_Supplementary motor cortex		−6	−12	64	5.84
I_Parietal operculum	1065	−54	−28	22	4.63	I_Middle frontal gyrus		42	44	2	5.75
						C_Precentral gyrus		−4	−18	70	5.72
						I_Cerebellum	6040	12	−42	−24	6.01
						C_Supramarginal gyrus	3118	−64	−34	32	5.50
						I_Supramarginal gyrus	2085	62	−26	38	6.34
	Non-amputated Knee						Amputated Knee
	Cluster size	MNI coordinates					Cluster size	MNI coordinates
	Cluster size	x	y	z	z value		Cluster size	x	y	z	z value
C_Precentral gyrus	1776	8	−18	64	5.96	C_Postcentral gyrus	3378	−10	−34	70	7.24
I_Cerebellum	553	−20	−38	−24	5.64	C_Precentral gyrus		−8	−18	66	6.46
						C_Supplementaly motor cortex		−8	−4	50	6.68
						I_Cerebellum	2295	12	−42	−22	6.55
						C_Central opercluum	2049	−48	2	6	5.63
						I_Parietal opercluum	659	52	−30	28	4.95
	Non-amputated Hip						Amputated Hip
	Cluster size	MNI coordinates					Cluster size	MNI coordinates
	Cluster size	x	y	z	z value		Cluster size	x	y	z	z value
C_Supplementaly motor cortex	9971	2	−14	62	5.96	C_Precentral gyrus	5102	−4	−16	72	6.86
I_Supplementaly motor cortex		−2	−8	62	5.87	C_Supplementaly motor cortex	5102	−6	−12	64	5.97
C_Precentral gyrus		14	−26	68	5.54	C_Central opercluum	890	−44	0	12	4.82
I_Precentral gyrus		−12	−28	66	5.26	C_Parietal opercluum	978	−52	−30	20	5.85
I_Cerebellum	1859	−24	−36	−28	6.27	I_Cerebellum	628	26	−38	−28	5.06
I_Parietal operculum	975	−54	−26	22	5.43	I_Parietal opercluum	483	50	−32	26	4.05

C = contralateral, I = ipsilateral.

Notably, individual (first-level) analysis revealed that 12 of the 30 participants exhibited significant ipsilateral M1 activation during the amputated knee task. The median (25^th and 75^th percentile values) number of sports years for participants who showed and did not show significant ipsilateral M1 activation were 20.0 (11.5, 24.3) and 6.5 (3.0, 12.3) years, respectively. Figure 2 shows representative brain activity during the amputated and non-amputated knee task in 6 typical participants. These 6 individuals were relatively similar in age and injury years but differed in sports years. The 3 participants on the left with fewer sports years showed only contralateral M1 activation during the knee task. Contrastingly, participants with relatively longer sports years showed distinct bilateral M1 activation only during the amputated knee task.

Figure 2.

Representative brain activities of participants with relatively short and long sports years history. Representative brain activities during both amputated and non-amputated knee contraction in 6 typical participants. The 3 participants on the left with fewer sports years showed only contralateral M1 activation during the knee task. Contrastingly, participants with relatively longer sports years showed distinct ipsilateral M1 activation during only the amputated knee task.

Association Between Brain Activity and Sports and Injury Years

Brain activity in the contralateral dorsolateral prefrontal cortex (DLPFC), primary motor and somatosensory areas (SM1), inferior temporal gyrus (ITG), and ipsilateral SM1 during the amputated knee task was positively correlated with sports years at a cluster level of P < .05 (FWE corrected) (Figure 3A). The correlated cluster of the M1 was defined as the SM1 because it exceeded the central sulcus and spanned the primary somatosensory area (S1). All data points of clusters correlated with sports years are plotted in Figure 3B. The correlations between brain activity during the non-amputated knee task and sports years as control conditions using the above clusters (inverted x-coordinate of MNI locations) are shown in Figure 3B. Regarding the non-amputated knee task, there were no significant correlations between sports years and brain activity in the bilateral SM1, DLPFC, and ITG. As shown in Figure 3B, brain activity in each region of one TF amputee (Sub.14, the right-most plot in each graph) was relatively larger and might have affected the findings. Thus, a reanalysis was performed after excluding this participant, but the results were not affected.

Figure 3.

Correlation of brain activity during the amputated knee contraction with sports and injury years. (A) Brain regions that are correlated with sports and injury years. Regions significantly correlated with sports years are shown in red, and regions significantly correlated with injury years, in green. (B) All brain activity values in all major areas that were correlated with sports years are plotted as circles. Transtibial and transfemoral values were indicated by black and white circles, respectively. The approximate lines for the transtibial and transfemoral groups were shown as solid and dotted lines, respectively. The 4 graphs in the bottom row are plotted as a control condition for comparison. Cont = contralateral, Ipsi = ipsilateral, DLPFC = dorsolateral prefrontal cortex, ITG = inferior temporal gyrus, SM1 = primary motor and somatosensory areas, OFG = occipital fusiform gyrus.

In contrast, two regions, the contralateral hippocampus and occipital fusiform gyrus, showed significant positive correlations with injury years (Figure 3A and Table 3).

Table 3.

Activated brain regions during amputated knee contraction correlated with sports years and the result of VBM analysis.

Positive correlation between brain activity and sports years
Area	Amputated Knee
	Cluster size	MNI coordinate				Cluster level		Peak level
	Cluster size	x	y	z	z value	corrected P	uncorrected P	corrected P	uncorrected P
C_Dorsolateral prefrontal cortex	403	−30	42	40	5.49	.016	.003	.001	<.001
I_Pre/postcentral gyrus	325	22	−30	66	4.01	.034	.006	.342	<.001
C_Precentral gyrus	310	−24	−12	30	4.7	.04	.007	.033	<.001
C_Inferior temporal gyrus	293	−44	−8	−26	4.44	.048	.009	.088	<.001
Positive correlation between brain activity and injury years
Area	Amputated Knee
	Cluster size	MNI coordinate				cluster level		peak level
	Cluster size	x	y	z	z value	corrected P	uncorrected P	corrected P	uncorrected P
C_Occipital fusiform gyrus	219	−24	−86	−20	4.15	.122	.026	.210	<.001
C_Hippocampus	116	−18	−18	−12	4.07	.368	.091	.266	<.001
Positive correlation between gray matter volume and sports years, although not significant (uncorrected P < .001)
Area	Amputated Knee
	Cluster size	MNI coordinate				cluster level		peak level
	Cluster size	x	y	z	z value	corrected P	uncorrected P	corrected P	uncorrected P
I_Postcentral gyrus	184	8	−35	77	3.62	.980	.577	.436	<.001
C_Postcentral gyrus	93	−11	−47	77	2.93	.992	.705	.984	.001
I_Precentral gyrus	77	23	−26	72	2.86	.993	.735	.980	.001
Positive correlation between gray matter volume and injury years, although not significant (uncorrected P < .001)
Area	Amputated Knee
	Cluster size	MNI coordinate				cluster level		peak level
	Cluster size	x	y	z	z value	corrected P	uncorrected P	corrected P	uncorrected P
I_ Superior parietal lobule	478	17	−60	71	3.28	.909	.349	.767	.001
Negative correlation between gray matter volume and injury years, although not significant (uncorrected P < .001)
Area	Amputated Knee
	Cluster size	MNI coordinate				cluster level		peak level
	Cluster size	x	y	z	z value	corrected P	uncorrected P	corrected P	uncorrected P
C_Middle cingulate gyrus	532	−8	−9	39	3.08	.891	.323	.910	.001
I_Middle prefrontal cortex	378	50	−12	−20	3.12	.939	.407	.888	.001
C_Middle temporal gyrus	78	−29	0	42	3.09	.945	.730	.908	.001

C = contralateral, I = ipsilateral.

GM Volume in Association With Sports and Injury Years

Table 3 summarizes the statistical results of VBM analysis. No significant results were obtained from the VBM analysis.

Muscle Activity During the Task

The results of the EMGs in the practice sessions showed that EMG of only the primary target muscle increased during the task, and contraction of other muscles was not observed in any participant (Supplementary file 1).

Discussion

This study suggested that long-term, routine para-sports participation after LLA can induce functional reorganization in the ipsilateral M1. Twelve of the 30 participants who had a relatively longer history of para-sports participation (median: 20.0 years) exhibited significant ipsilateral M1 activation during the amputated knee task. Additionally, the number of sports years was positively correlated with brain activity in the contralateral DLPFC, SM1, and ITG of the amputated limb during knee muscle contraction on the amputated side. No significant correlation was found in the structural analysis. Moreover, there were no differences between the amputated part, that is, TT and TF, of all results, likely because the thigh muscles in both TT and TF amputation are key in precisely controlling the prosthesis. These results indicated an interaction between amputation-induced and use-dependent brain reorganization in individuals with LLA who continue to participate in long-term para-sports.

Brain Activity During Contraction Tasks in LLA and Contribution of Prosthetic Sports Participation

The contralateral M1 was mainly activated during the ankle and knee task of the intact leg, consistent with the results of a similar evaluation in able-bodied individuals.^13,31–33 Contralateral activity in the M1 is associated with the lateral CST.³² The lateral CST originates from the contralateral M1 and is crucially involved in controlling the movements of the distal extremities.^33–35 Therefore, contralateral M1 activation is assumed to contribute to unilateral limb contraction via the lateral CST. Compared with the non-amputated knee task, the amputated knee task showed higher activation in wider regions, including the contralateral M1, bilateral cerebellum, and bilateral operculum. Notably, the averaged image showed no significant brain activity in the ipsilateral M1 (Figure 1B). However, in the individual level (1st level) analysis, 12 of the 30 participants who showed significant activation in the ipsilateral M1 during the amputated knee task had a relatively longer history of para-sports participation (median sports years: 20.0 years). Contrastingly, the median sports years for the remaining 17 participants (no ipsilateral activation) was 6.5 years. The typical brain activities in Figure 2 suggested that the number of sports years was associated with activation of the ipsilateral SM1. Participants who had a relatively longer sports participation history and who are represented on the right side of each representative brain activation in Figure 2 showed significant ipsilateral SM1 activation. Conversely, participants who had a relatively shorter sports participation history and who are represented on the left side in Figure 2 showed only contralateral SM1 activation. All participants, on the left and right sides, who showed different activation patterns, were equal in age and injury years but differed in sports years. In the analysis of the correlation between sports years and brain activity, there was a significant correlation of sports years with brain activity in not only the hemisphere contralateral to the amputated limb but also in the ipsilateral SM1 (Figure 3A). This result was not observed in the control area (inverted x-coordinate of MNI locations) on the non-amputated side (Figure 3B), indicating that this functional reorganization was specific to the amputated parts. Although below the statistical threshold, the ipsilateral SM1 tended to adapt structurally and with increasing sports years (Table 3). Accordingly, it was inferred that long-term para-sports participation could induce functional and structural reorganization of the ipsilateral SM1 to improve the motor function of the amputated limb.

Bilateral M1 activity was observed for both the left and right hip tasks, which is a well-known attribute of the proximal muscles receiving stronger projections of the ipsilateral M1 than the distal muscles via the corticoreticular pathway.^17,36 Moreover, activated areas during the amputated ankle imagery task differed from those during other tasks since no muscle contraction was involved. Consistent with previous findings, the DLPFC, SMA, and cerebellum were activated during motor imagery.^28,37,38

Possible Mechanisms of Involvement of the Ipsilateral M1 for Amputated Leg Contraction

The neural mechanism underlying the involvement of the ipsilateral M1 in amputated leg contraction can be presumed in 2 steps. First, a decrease in GABA levels in the ipsilateral and contralateral M1 after mitigating interhemispheric inhibition caused by post-LLA-induced compensational plastic changes, as inferred from WM structural analysis by Li et al.⁸ Reduced microstructure in the corpus callosum (inhibitory interhemispheric pathways) in LLA was reported, which could induce plastic changes in both hemispheres by decreasing GABA levels. GABA is considered an important inhibitory neurotransmitter^39,40 with decreased GABA inhibition unmasking pre-existing synaptic connections and easily causing synaptic plasticity⁴⁰; therefore, this key transmitter for short-term expansion of cortical representation could induce long-term reorganization similar to that observed in the present study through intense amputated limb use. Hordacre and Bradnam¹² reported that in the M1 contralateral to the amputated side, there were decreased GABA receptors and mitigated interhemispheric inhibition during the first walk after LLA. Furthermore, there was an earlier reduction in GABA receptors, which mediated inhibition, in the M1 ipsilateral to the amputated side than in the contralateral M1. Therefore, LLA-induced interhemispheric inhibition could be mitigated by reduced GABA levels in the bilateral hemisphere, which would induce bilateral M1 reorganization.

The second step is that intense post-LLA use of the amputated limb promotes bilateral reorganization, which we termed the “sports-induced reorganization” step. Strength training could uniquely modulate GABA-mediated inhibition between hemispheres and ipsilateral corticospinal excitability in healthy subjects.⁴¹ Even in an LLA case, Hordacre et al^11,12 revealed enhanced ipsilateral M1 excitability of amputated limbs using TMS, accompanying the progression of acute-phase post-LLA prosthetic rehabilitation. Moreover, there were longitudinal changes in interhemispheric inhibition over time accompanied by changes in GABA levels during the acute rehabilitation phase. Our correlation analysis of injury years and brain activity did not show significant results in ipsilateral brain regions (Figure 3A), suggesting that para-sports training to be a necessary component for facilitation of the ipsilateral motor-related pathways. In summary, this evidence of adaptation after amputation suggests that, initially, altered microstructure in the corpus callosum, and decreased GABA levels occur in both hemispheres (ie, first step [injury-induced reorganization]) after amputation. Subsequently, continuing para-sports overcomes this reorganization as a natural response in the ipsilateral M1 (ie, second step [sports-induced reorganization]). It is important to note that GABA was not measured in our study; therefore, careful interpretation is needed. Future studies should clarify how GABA is altered during the second step of this adaptation.

Overall, this ipsilateral M1 functional reorganization was not seen in the able-bodied athletes in a similar experiment¹³ and was considered to be a unique result that revealed new properties of human brain plasticity.

Reorganization in the Frontal Lobe and Visual Pathways Induced by Long-Term Sports Activities

Correlation analysis of sports years showed significant correlations with the ipsilateral SM1 as well as the contralateral SM1, DLPFC, and ITG. Reorganization in the contralateral SM1 appeared to be a natural consequence because it was previously reported that cortical expansion accompanies intense limb use and motor learning.²⁷ In contrast, although there is no direct anatomical connection between the DLPFC and M1,^42,43 DLPFC activation decreases ipsilateral M1 corticospinal excitability.^42,44 In fact, the DLPFC is known to be involved in motor control, for example, in force control⁴⁵ and motor memory maintenance,⁴⁶ which might have contributed to the task of regulating force to 20% MVC in this study. Conversely, both M1s inhibit the other M1 via transcallosal inhibition.⁴⁷ Therefore, we assumed that activation of the DLPFC specifically caused the suppression of the inhibitory function of the M1 on the contralateral side to the amputated side (ie, the same side as the DLPFC) and resulted in the disinhibition of the ipsilateral SM1.

The ITG is known as a core region of the ventral stream related to visual information processing.⁴⁸ Moreover, cortical plasticity of the ventral stream was also shown to accompany motor learning.⁴⁹ In fact, several studies have demonstrated that ITG hypertrophy and increased brain connectivity occur in able-bodied athletes.^50,51 Therefore, it is possible that visuomotor learning was enhanced by the increase in sports years which in turn induced the reorganization of the ITG in the current study; however, this assumption is speculative, and further studies are needed.

Limitations of the Study

First, we used sports years as an index to quantify the approximate activity level over a long period of time. However, we did not assess physical activities in detail; therefore, careful interpretation of our results is required. Since activity level is an important point in the discussion of use-dependent plasticity, a future longitudinal study would be necessary to specifically identify the characteristics of individual physical activities, including the sports events, intensity, frequency, and other activity levels in daily life.

Second, we did not record the EMG during the MRI scan, although it was recorded and confirmed during the practice session, and all participants could contract the agonist muscle at 20% MVC without other muscle activity (Supplementary 1). Therefore, we cannot exclude the possibility of muscle activity differing between the MRI scan and the practice session. Without EMG monitoring, it is difficult to exclude the possibility that isometric contraction (which does not accompany joint movement) of proximal muscles (eg gluteus maximus or paraspinal muscles) influenced the ipsilateral M1 activity. However, we believe that activity level, at least in homologous muscles, was comparable as the participants performed the contraction similarly in both sessions. This was further confirmed by our intensive camera monitoring of their legs during the task.

Finally, we did not recruit age-matched able-bodied participants as controls. Therefore, this study alone cannot infer the possibility of long-term sports practice inducing ipsilateral M1 activation, irrespective of limb amputation. However, we previously measured able-bodied athletes and observed no ipsilateral M1 activation during the same contraction tasks.¹³ This result could support our conclusion that both amputation and long-term para-sports are needed for further reorganization in the ipsilateral M1.

Conclusion

Our fMRI study revealed post-LLA functional brain reorganization in the ipsilateral M1. Specifically, 12 of the 30 participants who had significant activation in the ipsilateral M1 during the amputated knee task had a relatively longer history of para-sports participation. Moreover, participants with LLA who continued long-term para-sports also exhibited reorganization in the contralateral DLPFC and ITG. This indicates that long-term para-sports could lead to extensive reorganization at the brain network level, not only of the M1 but also of the frontal lobe and visual pathways. Our findings support the development of neurorehabilitation, including brain-targeted rehabilitation training or non-invasive stimulation after LLA.

Supplemental Material

sj-pdf-1-nnr-10.1177_15459683211056660 – Supplemental Material for Para-Sports can Promote Functional Reorganization in the Ipsilateral Primary Motor Cortex of Lower Limbs Amputee

Supplemental Material, sj-pdf-1-nnr-10.1177_15459683211056660 for Para-Sports can Promote Functional Reorganization in the Ipsilateral Primary Motor Cortex of Lower Limbs Amputee by Tomoya Nakanishi, Nobuaki Mizuguchi, Kento Nakagawa and Kimitaka Nakazawa in Neurorehabilitation and Neural Repair

Footnotes

Acknowledgments

We would like to express our gratitude to the participants for taking part in this experiment. Furthermore, we would like to thank Advanced Bioimaging Support for their advice on MRI data acquisition and analysis techniques.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by a Grant-in-Aid for the Japan Society for the Promotion of Science (JSPS) Fellows to T.N. [grant number 19J21542], and a Grant-in-Aid for Scientific Research (A) from JSPS to K. Nakazawa. [grant number 18H04082].

Data Availability

The dataset generated and analyzed in this study will be available from the corresponding author on reasonable request.

ORCID iD

Tomoya Nakanishi

Supplemental Material

Supplemental material for this article is available online.

References

Chen

Corwell

Yaseen

Hallett

Cohen

. Mechanisms of cortical reorganization in lower-limb amputees. J Neurosci. 1998;18(9):3443-3450.

Ferreri

Guerra

Rossini

. Neurophysiological markers of plastic brain reorganization following central and peripheral lesions. Arch Ital Biol. 2014;152(4):216-238.

Xie

Kane

Dennis

, et al. Case series evidence for changed interhemispheric relationships in cortical structure in some amputees. J Clin Neurosci. 2013;20(4):523-526.

Jiang

Yin

, et al. The plasticity of brain gray matter and white matter following lower limb amputation. Neural Plast. 2015;2015:1-10.

Draganski

Moser

Lummel

, et al. Decrease of thalamic gray matter following limb amputation. Neuroimage. 2006;31(3):951-957.

Jiang

, et al. Progressive thinning of visual motion area in lower limb amputees. Front Hum Neurosci. 2016;10(79):1-6.

Zhang

Wang

, et al. Brain functional connectivity plasticity within and beyond the sensorimotor network in lower-limb amputees. Front Hum Neurosci. 2018;12(10):1-11.

Fan

, et al. Altered microstructure rather than morphology in the corpus callosum after lower limb amputation. Sci Rep. 2017;7:44780.

Makin

Flor

. Brain (re)organisation following amputation: implications for phantom limb pain. Neuroimage. 2020;218:116943.

10.

Schwenkreis

Pleger

Cornelius

, et al. Reorganization in the ipsilateral motor cortex of patients with lower limb amputation. Neurosci Lett. 2003;349:187-190.

11.

Hordacre

Bradnam

. Reorganisation of primary motor cortex in a transtibial amputee during rehabilitation: a case report. Clin Neurophysiol. 2013;124(9):1919-1921.

12.

Hordacre

Bradnam

Barr

Patritti

Crotty

. Ipsilateral corticomotor excitability is associated with increased gait variability in unilateral transtibial amputees. Eur J Neurosci. 2014;40(2):2454-2462.

13.

Mizuguchi

Nakagawa

Tazawa

Kanosue

Nakazawa

. Functional plasticity of the ipsilateral primary sensorimotor cortex in an elite long jumper with below-knee amputation. NeuroImage Clin. 2019;23:101847.

14.

Zaaimi

Edgley

Soteropoulos

Baker

. Changes in descending motor pathway connectivity after corticospinal tract lesion in macaque monkey. Brain. 2012;135(7):2277-2289.

15.

Soteropoulos

Edgley

Baker

. Lack of evidence for direct corticospinal contributions to control of the ipsilateral forelimb in monkey. J Neurosci. 2011;31(31):11208-11219.

16.

Alawieh

Tomlinson

Adkins

Kautz

Feng

. Preclinical and clinical evidence on ipsilateral corticospinal projections: implication for motor recovery. Transl Stroke Res. 2017;8(6):529-540.

17.

Jang

Choi

Kim

Chang

Jung

Yeo

. Injury of the corticoreticular pathway in subarachnoid haemorrhage after rupture of a cerebral artery aneurysm. J Rehabil Med. 2015;47(2):133-137.

18.

Takenobu

Hayashi

Moriwaki

Nagatsuka

Naritomi

Fukuyama

. Motor recovery and microstructural change in rubro-spinal tract in subcortical stroke. NeuroImage Clin. 2014;4:201-208.

19.

Jang

. A review of the ipsilateral motor pathway as a recovery mechanism in patients with stroke. NeuroRehabilitation. 2009;24(4):315-320.

20.

Pundik

McCabe

Hrovat

, et al. Recovery of post stroke proximal arm function, driven by complex neuroplastic bilateral brain activation patterns and predicted by baseline motor dysfunction severity. Front Hum Neurosci. 2015;9:394.

21.

Cleland

Madhavan

. Ipsilateral motor pathways to the lower limb after stroke: insights and opportunities. J Neurosci Res. 2021;99(6):1565-1578.

22.

Nishimura

Isa

. Cortical and subcortical compensatory mechanisms after spinal cord injury in monkeys. Exp Neurol. 2011;235:152-161.

23.

Lundell

Christensen

Barthélemy

Willerslev-Olsen

Biering-Sørensen

Nielsen

. Cerebral activation is correlated to regional atrophy of the spinal cord and functional motor disability in spinal cord injured individuals. Neuroimage. 2011;54:1254-1261.

24.

Philip

Frey

. Compensatory changes accompanying chronic forced use of the nondominant hand by unilateral amputees. J Neurosci. 2014;34(10):3622-3631.

25.

Nakata

Yoshie

Miura

Kudo

. Characteristics of the athletes’ brain: evidence from neurophysiology and neuroimaging. Brain Res Rev. 2010;62(2):197-211.

26.

Dayan

Cohen

. Neuroplasticity subserving motor skill learning. Neuron. 2011;72(3):443-454.

27.

Callan

Naito

. Neural processes distinguishing elite from expert and novice athletes. Cognit Behav Neurol. 2014;27(4):183-188.

28.

Guillot

Collet

Nguyen

Malouin

Richards

Doyon

. Brain activity during visual versus kinesthetic imagery: an fMRI study. Hum Brain Mapp. 2008;30(7):2157-2172.

29.

Draganski

Moser

Lummel

, et al. Decrease of thalamic gray matter following limb amputation. Neuroimage. 2006;31(3):951-957.

30.

Di Vita

Boccia

Palermo

, et al. Cerebellar grey matter modifications in lower limb amputees not using prosthesis. Sci Rep. 2018;8(1):370-377.

31.

Kapreli

Athanasopoulos

Papathanasiou

, et al. Lower limb sensorimotor network: issues of somatotopy and overlap. Cortex. 2007;43(2):219-232.

32.

Marquis

Muller

Lorio

, et al. Spatial resolution and imaging encoding fMRI settings for optimal cortical and subcortical motor somatotopy in the human brain. Front Neurosci. 2019;13:571.

33.

Krings

Töpper

Foltys

, et al. Cortical activation patterns during complex motor tasks in piano players and control subjects. A functional magnetic resonance imaging study. Neurosci Lett. 2000;278(3):189-193.

34.

Humberstone

Sawle

Clare

, et al. Functional magnetic resonance imaging of single motor events reveals human presupplementary motor area. Ann Neurol. 1997;42(4):632-637.

35.

Welniarz

Dusart

Roze

. The corticospinal tract: evolution, development, and human disorders. Dev Neurobiol. 2017;77(7):810-829.

36.

Jang

Lee

. Corticoreticular tract in the human brain: a mini review. Front Neurol. 2019;10:1188.

37.

Hétu

Grégoire

Saimpont

, et al. The neural network of motor imagery: an ALE meta-analysis. Neurosci Biobehav Rev. 2013;37(5):930-949.

38.

Hanakawa

Dimyan

Hallett

. Motor planning, imagery, and execution in the distributed motor network: a time-course study with functional MRI. Cerebr Cortex. 2008;18:2775-2788.

39.

Jacobs

Donoghue

. Reshaping the cortical motor map by unmasking latent intracortical connections. Science. 1991;251:944-947.

40.

Floyer-Lea

Wylezinska

Kincses

Matthews

. Rapid modulation of GABA concentration in human sensorimotor cortex during motor learning. J Neurophysiol. 2006;95:1639-1644.

41.

Kidgell

Frazer

Rantalainen

, et al. Increased cross-education of muscle strength and reduced corticospinal inhibition following eccentric strength training. Neuroscience. 2015;300:566-575.

42.

Wang

Cao

Lin

Chen

Zhang

. Hemispheric differences in functional interactions between the dorsal lateral prefrontal cortex and ipsilateral motor cortex. Front Hum Neurosci. 2020;14:202.

43.

Miller

Cohen

. An integrative theory of prefrontal cortex function. Annu Rev Neurosci. 2001;24:167-202.

44.

Duque

Labruna

Verset

Olivier

Ivry

. Dissociating the role of prefrontal and premotor cortices in controlling inhibitory mechanisms during motor preparation. J Neurosci. 2012;32(3):806-816.

45.

Jin

Lee

Kim

Yoon

. Noninvasive brain stimulation over M1 and DLPFC cortex enhances the learning of bimanual isometric force control. Hum Mov Sci. 2019;66:73-83.

46.

Kantak

Sullivan

Fisher

Knowlton

Winstein

. Neural substrates of motor memory consolidation depend on practice structure. Nat Neurosci. 2010;13:923-925.

47.

Netz

Ziemann

Hömberg

. Hemispheric asymmetry of transcallosal inhibition in man. Exp Brain Res. 1995;104(3):527-533.

48.

Hickok

Poeppel

. Dorsal and ventral streams: a framework for understanding aspects of the functional anatomy of language. Cognition. 2004;92(1–2):67-99.

49.

Engel

Burke

Fiehler

Bien

Rösler

. Motor learning affects visual movement perception. Eur J Neurosci. 2008;27(9):2294-2302.

50.

Wang

, et al. Motor skill learning induces brain network plasticity: a diffusion-tensor imaging study. PLoS One. 2019;14(2):e0210015.

51.

Duru

Balcioglu

. Functional and structural plasticity of brain in elite karate athletes. J Healthc Eng. 2018;2018:1-8.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.17 MB