Quantitative MRI evaluation of the effects of rest and different intensities of running on the recovery of thigh muscle microdamage after a half marathon

Abstract

Objective

To explore the application value of quantitative magnetic resonance multi-parameter imaging technology, Diffusion Tensor Imaging (DTI) combined with T2 mapping, in evaluating the effects of rest and different intensities of running on the recovery of thigh muscle microdamage after a Half Marathon (HM).

Methods

Twenty-four amateur marathon runners were enrolled and randomly divided into three groups to compare different post-marathon recovery strategies: the High-Intensity Running (HIR) group, the Moderate-Intensity Running (MIR) group, and the Rest (REST) group. MR scans were performed on the participants before the race (pre-run) and at 3 h (post-3 h), 24 h (post-24 h), 48 h (post-48 h), and 72 h (post-72 h) after the race. The Fractional Anisotropy (FA), Apparent Diffusion Coefficient (ADC), and T2 relaxation time (T2 value) of the thigh muscle groups were measured.

Results

Compared with pre-run, all groups showed significantly decreased FA (p < 0.001) and increased T2 (p < 0.001) at post-3 h, indicating HM-induced microdamage (characterized by myofiber disruption and interstitial edema). REST group FA recovered to baseline at post-72 h, while MIR group recovered partially and HIR group showed no recovery. T2 trends aligned with FA. No significant ADC differences were observed.

Conclusion

Our findings suggest that quantitative MR multi-parameter imaging can sensitively detect muscle microdamage and the recovery process after HM. The REST group shows the fastest recovery, while moderate- and high-intensity running delay recovery. Therefore, it is not recommended to use moderate- or high-intensity running as an intervention within 3 days after the race.

Introduction

As an extreme sport, the marathon has gained increasing popularity among the general public. However, the high-intensity and long-duration repetitive contraction and relaxation of muscles can lead to microdamage in skeletal muscle fibers, known as Exercise-Induced Muscle Micro-Injury (EIMI) [1]. This condition is characterized by a series of symptoms such as delayed onset muscle soreness (DOMS), accompanied by muscle stiffness, decreased muscle strength, and reduced range of joint motion [2, 3]. Helping runners resume normal training as soon as possible after the race without increasing the risk of injury is a crucial issue in post-marathon recovery [4–8]. There are various post-marathon recovery strategies [9], among which running and rest are the most simple and feasible options, and are more suitable for the actual situation of long-distance runners.

Currently, there is no clear consensus on whether rest or continued running after a marathon is beneficial for the recovery of muscle injury and function, and the conclusions drawn are also inconsistent. Existing literature on post-exercise muscle recovery mainly relies on indirect reflections of muscle status, such as subjective pain scale assessment [10], biochemical parameter detection [11], and biomechanical indicators [12]. Quantitative magnetic resonance technologies (DTI, T2 mapping) can realize real-time observation of in vivo muscle tissue under non-invasive conditions. They can not only accurately locate specific muscle groups but also quantitatively evaluate changes in microstructures, showing great application potential in the fields of muscle microdamage diagnosis and rehabilitation monitoring. This study used DTI and T2 mapping techniques to evaluate the microstructural changes of thigh muscles within 3 days after a half marathon, as well as the effects of rest and running at different intensities on muscle microdamage and its recovery, aiming to provide imaging evidence for running sports and post-exercise muscle recovery strategies.

We hypothesized that: (i) Quantitative MRI parameters (FA, T2 values) can sensitively detect early muscle microdamage and dynamic recovery processes after a half-marathon.(ii) Compared with moderate-intensity and high-intensity running, passive rest will promote faster recovery of thigh muscle microdamage within 3 days after a half-marathon.

Materials and methods

Study subjects

To eliminate the impact of gender on the study results, this study recruited male amateur marathon runners in Hangzhou. With an effect size of f = 0.41, a significance level (α) of 0.05, and a power of 1-β = 0.80, the total sample size required for a two-level single-factor design was calculated to be 30, with 10 cases per group. A total of 40 volunteers were recruited, among whom 2 were excluded due to unqualified MRI images (motion artifacts, n = 1; low signal-to-noise ratio, n = 1), 7 were excluded due to poor compliance, and 4 opted out by informed consent. After matching based on body mass index (BMI), 3 participants were excluded, leaving a final cohort of 24 subjects (all completed the Health Status Questionnaire). No participants were excluded due to injury during the study period. All subjects had no experience of formal training and did not engage in marathon running as a profession.

Inclusion Criteria [1]:1. having completed one full marathon or two half-marathons in previous running experience;2. Regular long-term running practice: running ≥ 3 times per week, with each running distance ≥ 10 km; or monthly running mileage ≥ 100 km;3. No history of heart or kidney diseases, and no history of regular medication use;4. Voluntary participation in the study and signing of the informed consent form.

Exclusion Criteria:1. History of lower limb muscle trauma, surgery, or lower limb muscle injury during the study period;2. Presence of systemic diseases affecting the muscular system, such as diabetes and muscular dystrophy;3. Images that fail to meet the evaluation requirements;4. Presence of contraindications for magnetic resonance (MR) examination;5. Use of medications that may affect muscle metabolism or imaging results (e.g., statins, non-steroidal anti-inflammatory drugs [NSAIDs]) within 7 days prior to the HM and MRI scans.

Grouping Criteria: A total of 24 volunteers were enrolled in this study and randomly divided into three groups: the High-Intensity Running (HIR) group, the Moderate-Intensity Running (MIR) group, and the Rest (REST) group. The HIR and MIR groups conducted running on asphalt roads or plastic tracks. The volunteers’ own sports watches were used to monitor heart rate during exercise, and exercise intensity was determined based on the volunteer’s heart rate reserve [13]. For the HIR group: The target heart rate was calculated as resting heart rate + 70% × (maximum heart rate - resting heart rate); For the MIR group: The target heart rate was calculated as resting heart rate + 50% × (maximum heart rate - resting heart rate). Maximum heart rate was calculated using the formula: Maximum heart rate = 220 - age [14]. Resting heart rate refers to the volunteers’ heart rate in a quiet and resting state. During running, volunteers in the HIR and MIR groups were required to maintain their heart rates within their respective target heart rate ranges as mentioned above.

This study was approved by the Research Ethics Committee of the Affiliated Hospital of Hangzhou Normal University (No.: 2024 (ES)-KS-011), in accordance with the Declaration of Helsinki for human research. Written informed consents were obtained from all included participants, which clearly stated their right to withdraw from the study at any time without penalty.

MRI scanning protocol

All participants completed a personal exercise status questionnaire and informed consent form prior to the first magnetic resonance imaging (MRI) examination to collect information on age, body mass index (BMI), running duration, pace, monthly running volume, number of races, and injury history. All participants had the right to withdraw from the examination at any time. Intense exercise and medications affecting muscle function were avoided for 7 days prior to the MRI scan.

After completing the half marathon (HM), the participants underwent the second MRI scan (post-3 h) within 3 h. The third (post-24 h), fourth (post-48 h), and fifth (post-72 h) MRI scans were conducted sequentially at 24 h, 48 h, and 72 h after the HM, respectively. Participants in the HIR group and MIR group performed 30 min of high-intensity or moderate-intensity running, respectively, before undergoing the third, fourth, and fifth MRI scans. They then rested in a supine position for 30 min prior to the MRI scans [15].

The REST group received no intervention post-HM, relying on natural recovery mechanisms, and refrained from any warm-up or muscle relaxation activities post-HM.

MRI scanning equipment and parameters

A Siemens Avanto 1.5 T superconducting MRI scanner (Germany) with an 8-channel PA Matrix coil was used for magnetic resonance scanning. The scanning sequences and parameters were set as follows:

1. Axial T1-weighted imaging (T1WI) of bilateral thighs: Field of View (FOV) = 400.0 mm × 400.0 mm, matrix = 256 × 256, slice thickness = 1.6 mm, slice gap = 0.8 mm, number of slices = 60, voxel size = 1.6 mm × 1.6 mm × 1.6 mm, Repetition Time (TR) = 2,200.00 ms, Echo Time (TE) = 3.06 ms, number of excitations (NEX) = 3, acquisition time = 2 min 49 .

2. Axial Turbo Spin Echo (TSE)-T2-weighted imaging (T2WI) of bilateral thighs: FOV = 400.0 mm × 400.0 mm, matrix = 256 × 256, slice thickness = 4.0 mm, slice gap = 0.8 mm, number of slices = 40, voxel size = 1.0 mm × 1.0 mm × 4.0 mm, TR = 4,000.00 ms, TE = 83.00 ms, NEX = 2, acquisition time = 2 min 48s.

3. Axial Echo Planar Imaging (EPI)-based DTI of bilateral thighs: FOV = 400.0 mm × 400.0 mm, matrix = 128 × 128, slice thickness = 4.0 mm, no slice gap, number of slices = 60, voxel size = 3.1 mm × 3.1 mm × 4.0 mm, TR = 6,800.00 ms, TE = 60.00 ms, NEX = 2, b-value = 450 s/mm², 12 diffusion directions, acquisition time = 15 min 32 s. Anatomical images of the bilateral thigh muscles were obtained via T1WI, while Diffusion-Weighted Imaging (DWI) maps, ADC maps, FA grayscale maps, and FA color-coded maps (Fig. 1) were acquired using DTI.

4. T2 mapping scanning parameters: TE = 13.8 ms, TR = 974 ms, FOV = 400 mm, slice thickness = 6 mm, Average = 1, voxel size = 1.6 × 1.6 × 6 mm, Bandwidth = 610 Hz/Px.

Fig. 1.

Axial MRI Images of Bilateral Thigh Muscles in an Amateur Marathon Runner (In each image, the left side represents the right thigh of the amateur marathon runner, and the right side represents the left thigh). A 43-year-old male amateur marathon runner with 4 years of running experience and a monthly running mileage of 300 km. A Axial T1-weighted image of bilateral thigh muscles, showing the anatomical structure of the thigh muscles. B Axial T2-weighted fat-saturated Dixon (T2-fs-Dixon) image of bilateral thigh muscles, used to observe whether injuries such as edema occur in the thigh muscles; no obvious muscle edema is seen in this case. C, D DWI map and ADC map, respectively. The combination of these two maps is used to detect whether there are diffusion-restricted lesions in the lower limb muscles and evaluate the degree of diffusion restriction; no diffusion-restricted lesions are observed in this subject. E Color-coded FA map of magnetic resonance DTI. Blue represents muscle fiber bundles running in the craniocaudal direction, red represents those running in the left-right direction, and green represents those running in the anteroposterior direction. Most thigh muscles run in the craniocaudal direction, so most areas in the color-coded map appear blue. F Grayscale FA map of magnetic resonance DTI. The gradient from black to white in the image indicates a gradual increase in the degree of water molecule anisotropy

Image analysis

Post-processing of images was performed on a Siemens Neuro 3D DTI post-processing workstation. Image analysis and measurement of Regions of Interest (ROI) were conducted via double-blind assessment by 2 radiologists specialized in musculoskeletal imaging (with 4 years and 30 years of experience in musculoskeletal imaging diagnosis, respectively). On the fused axial T1WI and DTI images, ROI of the anterior thigh muscle group (rectus femoris, vastus intermedius, vastus medialis, and vastus lateralis) and posterior thigh muscle group (long head of biceps femoris, semitendinosus, and semimembranosus) on both sides were delineated at the level where the adductor longus muscle disappears, for the measurement of ADC values and FA values (Fig. 2). T2 values were measured by delineating ROI on T2WI images using syngo MR Workplace (syngo MR E11, Siemens). When manually drawing ROIs, the observers were required to avoid fascia, subcutaneous adipose tissue, and large blood vessels. The measured FA, ADC, and T2 values of each individual muscle were summed up and averaged to obtain the overall magnetic resonance parameter values of the thigh muscles.

Fig. 2.

Schematic diagram of Region of Interest (ROI) Delineation for Bilateral Thigh Muscles in a Marathon Runner (In the figure, the left side represents the right thigh of the amateur marathon runner, and the right side represents the left thigh). On the fused image of T1WI and DTI map, when observing from the proximal to the distal thigh, the ROI of the rectus femoris (RF), vastus medialis (VM), vastus intermedius (VI), vastus lateralis (VL), semimembranosus (SM), semitendinosus (ST), and long head of biceps femoris (BFL) were delineated at the level where the adductor longus (AL) muscle disappears

Statistical analysis

This study employed the SPSS 31.0 statistical software (IBM Corp, Armonk, NY, USA). The intraclass correlation coefficient (ICC) was used to assess the consistency of measurement results obtained by two physicians. A repeated-measures analysis of variance (ANOVA) model was adopted, with “intervention modality” (HIR, MIR, REST) as the between-group variable and “time point” (pre-run, post-3 h, post-24 h, post-48 h, and post-72 h) as the within-group variable, to analyze the thigh muscle microinjury and its recovery before and after the marathon. When significant main effects or interactions were detected, the Bonferroni method was performed for post-hoc comparisons. Independent-samples t-tests were used for pairwise comparisons of FA, ADC, and T2 values of the thigh muscle among all groups and at each time point. For ease of comparison, the pre-exercise T2, FA, and ADC values of each subject were standardized to 100%, and all subsequent measurements were normalized relative to this baseline level. The statistical significance level was set at p < 0.05, and measurement data were presented as mean ± standard deviation (x ± s).

Research results

General information of subjects and consistency test

Forty (n = 40) volunteers were recruited. After exclusion criteria, post BMI matching, n = 24 male volunteers (20–50 yrs) were enrolled. All were injury free prior and during study, and completed five MR scans. Table 1 reflects volunteer general information.

Table 1.

General information of subjects

Basic Information	Height (cm)	Weight (kg)	Pace (km/h)	BMI (kg/m²)	Age (years)
Overall	171.2 ± 5.2	64.4 ± 6.4	11.1 ± 0.7	23.0 ± 1.6	31.5 ± 4.5
HIR	173.2 ± 4.3	54.5 ± 3.4	11.2 ± 0.6	20.0 ± 1.0	28.1 ± 4.5
MIR	165.3 ± 1.7	67.3 ± 3.5	10.4 ± 0.3	22.5 ± 1.2	29.2 ± 4.8
REST	169.1 ± 3.7	67.3 ± 3.7	11.6 ± 0.3	24.0 ± 1.2	37.2 ± 4.2

Descriptive statistics are presented as mean ± standard deviation

The Intraclass Correlation Coefficients (ICC) of FA, ADC, and T2 values measured by the two observers were all greater than 0.75, with 95% Confidence Intervals (95%CI) of 0.760–0.880, 0.757–0.821, and 0.703–0.801 respectively. This indicates good consistency of the measurements (Table 2).

Table 2.

Measurement of inter - observer reproducibility

Parameter	ICC	95%CI	p value
FA	0.820	0.760–0.880	< 0.001
ADC(×10-3mm2/s)	0.789	0.757–0.821	< 0.001
T2	0.752	0.703–0.801	< 0.001

Bold values indicate statistically significant differences

FA Fractional Anisotropy, ADC Apparent Diffusion Coefficient, T2 Transverse Relaxation Time, ICC Intraclass Correlation Coefficient

Variations in Fractional Anisotropy (FA), Apparent Diffusion Coefficient (ADC), and T2 values of global thigh muscle pre- and post-HM

Across Groups Global thigh muscle FA, ADC, and T2 values at all time points are summarized in Table 3. Repeated-measures ANOVA results for within- and between-group comparisons are presented in Table 4, with pairwise comparisons of FA, ADC, and T2 values across time points detailed in Tables 5, 6, 7 and 8.

Table 3.

FA, ADC and T2 values of overall thigh muscle in each group at different time point

Parameter	Group	pre-run	post-3 h	post-24 h	post-48 h	post-72 h
FA	HIR	0.254 ± 0.008	0.246 ± 0.020	0.245 ± 0.016	0.245 ± 0.016	0.248 ± 0.011
		100.00%	96.97%	96.51%	96.51%	97.69%
	MIR	0.274 ± 0.035	0.263 ± 0.014	0.264 ± 0.022	0.268 ± 0.016	0.272 ± 0.017
		100.00%	95.99%	96.45%	97.99%	99.27%
	REST	0.262 ± 0.072	0.253 ± 0.017	0.258 ± 0.031	0.259 ± 0.026	0.263 ± 0.028
		100.00%	96.48%	98.51%	99.01%	100.30%
ADC (×10-3mm2/s)	HIR	1.743 ± 0.047	1.748 ± 0.034	1.753 ± 0.074	1.765 ± 0.076	1.736 ± 0.050
		100.00%	100.29%	100.57%	101.24%	99.60%
	MIR	1.773 ± 0.088	1.772 ± 0.077	1.781 ± 0.083	1.773 ± 0.057	1.779 ± 0.061
		100.00%	99.93%	100.48%	99.98%	100.32%
	REST	1.739 ± 0.169	1.745 ± 0.040	1.761 ± 0.055	1.748 ± 0.071	1.756 ± 0.073
		100.00%	100.35%	101.27%	100.52%	100.98%
T2 (ms)	HIR	32.987 ± 2.257	34.042 ± 1.279	33.942 ± 1.458	33.804 ± 1.484	33.811 ± 1.842
		100.00%	103.20%	102.89%	102.48%	102.50%
	MIR	33.463 ± 1.371	34.486 ± 1.776	33.963 ± 1.689	33.794 ± 1.134	33.841 ± 1.616
		100.00%	103.06%	101.49%	100.99%	101.13%
	REST	32.704 ± 4.464	33.953 ± 4.266	33.381 ± 4.261	32.993 ± 3.715	32.337 ± 4.812
		100.00%	103.82%	102.07%	100.88%	98.88%

Table 4.

Repeated-measures ANOVA for within-group and between-group differences in each parameter value

Parameter		df	Mean square	F	Sig.
FA	intervention modality	2	0.001	1.434	0.026
	time point	2.618	0	0.368	0.035
	intervention modality×time point	5.236	0.001	1.061	0.018
ADC(×10-3mm2/s)	intervention modality	2	0.0000473	0.018	0.482
	time point	2.369	0	0.038	0.977
	intervention modality×time point	4.738	0.004	0.429	0.817
T2	intervention modality	2	251.836	30.711	< 0.001
	time point	2.986	5.373	3.554	0.019
	intervention modality×time point	5.972	0.674	0.446	0.045

Bold values indicate statistically significant differences

Table 5.

Post-hoc power test comparisons of overall thigh muscle FA, ADC and T2 values in each group at each time point

Parameter		post-3h	post-24h	post-48h	post-72h
FA	HIR vs REST(p)	>0.99	<0.001	0.027	<0.001
	MIR vs REST(p)	>0.99	0.013	0.145	0.435
	HIR vs MIR(p)	>0.99	0.073	0.468	>0.99
ADC (×10-3mm2/s)	HIR vs REST(p)	>0.99	>0.99	>0.99	>0.99
	MIR vs REST(p)	>0.99	>0.99	>0.99	>0.99
	HIR vs MIR(p)	>0.99	>0.99	>0.99	>0.99
T2 (ms)	HIR vs REST (p)	0.572	0.312	<0.001	<0.001
	MIR vs REST(p)	0.791	0.205	>0.99	<0.001
	HIR vs MIR(p)	>0.99	0.196	0.824	<0.001

Bold values indicate statistically significant differences

Table 6.

Independent-samples t-test of each parameter value at each post-half marathon time point vs. pre-run

Parameter		post-3h vs pre-run (pvalue)	post-24h vs pre-run (pvalue)	post-48h vs pre-run (pvalue)	post-72h vs pre-run (pvalue)
FA	HIR	<0.001	<0.001	<0.001	<0.001
	MIR	<0.001	0.018	0.032	0.045
	REST	<0.001	0.034	0.046	>0.99
ADC(×10-3mm2/s)	HIR	>0.99	0.946	0.954	>0.99
	MIR	0.934	0.810	>0.99	>0.99
	REST	>0.99	0.846	0.975	0.934
T2	HIR	<0.001	0.026	<0.001	<0.001
	MIR	<0.001	0.021	0.035	0.017
	REST	<0.001	0.017	0.047	0.782

Bold values indicate statistically significant differences

Table 7.

Independent-samples t-test of each parameter value at each post-half marathon time point vs. post-3h

Parameter		pre-run vs post-3h (pvalue)	post-24h vs post-3h (pvalue)	post-48h vs post-3h (pvalue)	post-72h vs post-3h (pvalue)
FA	HIR	<0.001	0.912	>0.99	>0.99
	MIR	<0.001	0.966	0.032	<0.001
	REST	<0.001	0.031	<0.001	<0.001
ADC(×10-3mm2/s)	HIR	>0.99	0.691	0.502	>0.99
	MIR	0.934	0.972	>0.99	>0.99
	REST	>0.99	0.629	0.755	0.934
T2	HIR	<0.001	>0.99	0.947	0.852
	MIR	<0.001	0.275	0.394	0.031
	REST	<0.001	0.181	<0.001	<0.001

Bold values indicate statistically significant differences

Table 8.

Independent-samples t-test: post-half marathon time points vs. previous time point parameter values

Parameter		pre-run vs post-3h (pvalue)	post-24h vs post-3h (pvalue)	post-48h vs post-24h (pvalue)	post-72h vs post-48h (pvalue)
FA	HIR	<0.001	0.912	0.461	0.128
	MIR	<0.001	0.966	0.048	0.174
	REST	<0.001	0.031	0.629	0.119
ADC(×10-3mm2/s)	HIR	>0.99	0.691	0.826	>0.99
	MIR	0.934	0.972	0.843	0.891
	REST	>0.99	0.629	>0.99	0.878
T2	HIR	<0.001	>0.99	0.287	>0.99
	MIR	<0.001	0.275	0.390	>0.99
	REST	<0.001	0.181	0.296	0.026

Bold values indicate statistically significant differences

FA value changes in global thigh muscle pre- and post-HM

Repeated-measures ANOVA revealed significant main effects of “time point” and “intervention modality” on thigh muscle FA values (both p < 0.05), as well as a significant interaction between these factors (p < 0.05).(Table 4).

At 3 h post-HM: FA values in the HIR, MIR, and REST groups were all significantly reduced compared with pre-run levels (all p < 0.001), with reductions exceeding 3% in all groups (Table 6).

At 24 h post-HM: FA values remained significantly lower than pre-run levels in all three groups (all p < 0.05; Table 6). Relative to 3 h post-HM, FA slightly decreased in the HIR group and increased in the MIR group (both p > 0.05), while it significantly increased in the REST group (p = 0.031; Table 7). Between-group comparisons showed significant differences in FA between HIR vs. REST and MIR vs. REST (p < 0.001 and p = 0.013, respectively), but no difference between HIR and MIR (p = 0.073; Table 5).

At 48 h post-HM: FA values were still significantly lower than pre-run levels in all groups (HIR: p < 0.001; MIR: p = 0.032; REST: p = 0.046; Table 6). Compared with 3 h post-HM, FA showed a non-significant decrease in the HIR group (p > 0.99) but significant increases in the MIR and REST groups (p = 0.032 and p < 0.001, respectively; Table 7). Relative to 24 h post-HM, FA significantly increased in the MIR group (p = 0.048) but showed a non-significant rise in the HIR and REST group (p = 0.416 and 0.629; Table 8). A significant between-group difference was observed only between the HIR and REST groups (p < 0.05; Table 5).

At 72 h post-HM: FA values remained significantly lower than pre-run levels in the HIR and MIR groups (p < 0.001 and p = 0.045, respectively), while the REST group’s FA was comparable to pre-run levels (p > 0.99; Table 6), indicating full recovery in the REST group but not in the HIR or MIR groups. Compared with 3 h post-HM, FA showed a non-significant increase in the HIR group (p > 0.99) and significant increases in the MIR and REST groups (both p < 0.001; Table 7). Relative to 48 h post-HM, FA increased non-significantly in all three groups (all p > 0.05; Table 8). Between-group comparisons revealed a significant difference only between the HIR and REST groups (p < 0.001; Table 5).

Recovery Trend Summary: The REST group initiated FA recovery as early as 24 h post-HM, followed by the MIR group at 48 h post-HM. In contrast, the HIR group showed no signs of FA recovery throughout the 72-hour observation period. By 72 h post-HM, only the REST group’s FA value had returned to pre-run levels (Fig. 3-①).

Fig. 3.

Changes in FA, ADC, and T2 values in HIR, MIR, and REST groups at different time points.* indicates a statistically significant difference within the HIR group compared to pre-run (p < 0.05), ** indicates a statistically significant difference within the HIR group compared to pre-run (p < 0.01); #indicates a statistically significant difference within the MIR group compared to pre-run (p < 0.05), ## indicates a statistically significant difference within the MIR group compared to pre-run (p < 0.01); †indicates a statistically significant difference within the REST group compared to pre-run (p < 0.05), †† indicates a statistically significant difference within the REST group compared to pre-run (p < 0.01)

ADC values changes in global thigh muscle pre- and post-HM

Statistical analysis revealed no significant differences in thigh muscle ADC values across “time point” or between “intervention modality” (Table 4). Furthermore, pairwise comparisons showed no significant differences among the MIR, HIR, and REST groups (Table 5).

T2 values changes in global thigh muscle pre- and post-HM

Repeated-measures ANOVA revealed significant main effects of “time point” and “intervention modality” on thigh muscle T2 values (both p < 0.05), as well as a significant interaction between these factors (p < 0.05) (Table 4).

At 3 h post-HM: T2 values in the HIR, MIR, and REST groups were significantly elevated relative to pre-run levels (all p < 0.001). No statistically significant intergroup differences in T2 values were detected at this time point (HIR vs. REST:p = 0.572;MIR vs. REST:p = 0.791; HIR vs. MIR:p > 0.99; Table 5), confirming the efficacy of randomization in subject grouping.

At 24 h post-HM: T2 values in all three groups remained significantly higher than pre-run levels (HIR: p = 0.026; MIR: p = 0.021; REST: p = 0.017; Table 6). Compared with 3 h post-HM, T2 value decreased across all groups without significant differences (HIR: p > 0.99; MIR: p = 0.275; REST: p = 0.181; Table 7), and no intergroup differences were observed (HIR vs. REST: p = 0.312; MIR vs. REST: p = 0.205; HIR vs. MIR: p = 0.196; Table 5). At 48 h post-HM: T2 values were still significantly elevated relative to pre-run levels in all groups (HIR: p < 0.001; MIR: p = 0.035; REST: p = 0.047; Table 6). Relative to 3 h post-HM, T2 values decreased non-significantly in the HIR and MIR groups (p = 0.947,0.394 respectively), whereas the REST group exhibited a statistically significant reduction (p < 0.001; Table 7), indicating that T2 values in the HIR and MIR groups had not initiated recovery by this time point, while the REST group’s thigh muscle T2 values had begun to recover. Compared with 24 h post-HM, T2 values decreased across all groups with no significant differences (HIR: p = 0.287; MIR: p = 0.390; REST: p = 0.296; Table 8). Intergroup comparisons at post-48 h revealed a significant difference between the HIR and REST groups (p < 0.001), but no differences between HIR vs. MIR (p = 0.824) and MIR vs. REST (p > 0.99; Table 5). At 72 h post-HM: T2 values remained significantly higher than pre-run levels in the HIR and MIR groups (p < 0.001 and p = 0.017, respectively), whereas the REST group’s T2 values were not significantly different from pre-run values (p = 0.782; Table 6). Relative to 3 h post-HM, T2 values decreased non-significantly in the HIR group (p = 0.852) , but significantly in the MIR and REST groups (p = 0.031 and p < 0.001, respectively; Table 7). Compared with 48 h post-HM, T2 values increased non-significantly in the HIR and MIR groups (p > 0.99), while the REST group showed a significant reduction (p = 0.026; Table 8). All pairwise intergroup comparisons at 72 h post-HM yielded significant differences (all p < 0.001; Table 5).

Recovery Trend Summary: T2 values demonstrated a consistent trend of immediate elevation post-HM, followed by gradual reduction over time. The REST group’s T2 values initiated recovery by 48 h post-HM and normalized to pre-run levels by 72 h post-HM; the MIR group’s T2 values began to recover at 72 h post-HM but had not yet returned to pre-run levels; and the HIR group failed to exhibit any recovery throughout the 72-hour observation period (Fig. 3-③).

Discussion

Reliability of image analysis

With ICC consistently > 0.75 for FA, ADC and T2 values from independent measurements by two radiologists, excellent inter-observer reliability was confirmed. This validates the rationality and feasibility of our technical approach for quantitative muscle injury assessment, and underpins the clinical implementation of DTI and T2 mapping for diagnosing and monitoring such injuries.

Currently, muscle injury diagnosis (contusion, laceration, degeneration included) relies heavily on conventional MRI and clinical symptoms, which are limited in evaluating injury severity and repair dynamics. High inter-observer reliability ensures these quantitative metrics are applicable to multi-center studies and routine clinical work, avoiding decision-making bias from inter-operator measurement inconsistency. It also facilitates establishing standardized FA, ADC and T2 reference ranges in normal and injured muscles, advancing standardized application of DTI and T2 mapping in muscle injury diagnosis.

Sensitivity and dynamic monitoring value of quantitative magnetic resonance imaging for subclinical injuries

In this study, DTI combined with T2 mapping techniques were employed to assess microstructural alterations in the thigh muscles subsequent to a half marathon. During the 3-day period following half marathon completion, conventional magnetic resonance (MR) sequences failed to identify any muscle damage, yet quantitative MR-derived parameters demonstrated evident alterations. At 3 h after the half marathon, FA values of the thigh muscles decreased to their lowest levels while T2 values increased to their highest levels; subsequently, FA values presented a gradual upward trend and T2 values a gradual downward trend over time, and these parameters essentially returned to their pre-run baseline levels by 3 days after the run.

A half-marathon can cause microstructural changes in the thigh muscles, leading to muscle micro-injuries (subclinical injuries). Running-induced skeletal muscle injuries are mainly characterized by the destruction of muscle fibers, with loss of cellular integrity and tissue edema, resulting in an increase in water content in the intracellular and extracellular spaces. In the study by Froeling et al. [16], only 9 muscles with grade 1 strain were detected by conventional MR after a full marathon. Beyond identifying the aforementioned injuries, DTI further revealed that the eigenvalues and MD of the biceps femoris and semitendinosus muscles remained elevated at 48 h post-marathon. This indicates that DTI has higher sensitivity than conventional MRI sequences, enabling the identification of exercise-induced muscle microdamage at the subclinical stage and providing a reliable basis for early diagnosis.

FA values in DTI reflect the integrity and alignment of muscle fibers—lower FA indicates disrupted myofiber arrangement and reduced water diffusion restriction, which is closely related to muscle microtears and interstitial edema induced by HM. This is a reversible change: as muscle repair proceeds (myofiber regeneration, edema absorption), FA values gradually recover to pre-run levels, as observed in the REST group. In contrast, higher FA values (not seen in this study) may indicate muscle fiber compression or early regeneration, but such scenarios require combined analysis with clinical symptoms and other MRI parameters [1, 17–19]. The increase in T2 value after exercise has long been considered a result of increased extracellular fluid caused by exercise-induced hyperemia [20]; in addition, the increase in T2 value is closely related to the metabolic activity of muscle cells [21] and can be used as an indirect indicator to evaluate the degree of muscle inflammation [22, 23].

Fu et al. [6] conducted a study on the correlation between MR parameters and pathology in an exercise-induced muscle injury (EIMI) mouse model, which verified the above speculation. They found that the FA and T2 values of mice reached their peaks at 48 h after exercise; subsequently, muscle cells entered the regeneration and repair stage. During this stage, with muscle cell regeneration, the reduction of intermuscular spaces, and the increase of macromolecules on the cell surface, the FA value gradually increased while the T2 value gradually decreased. Additionally, the change in T2 value was highly correlated with the expression of fast skeletal troponin, an early-specific injury marker. Lyu et al. [5] carried out a comparative study of magnetic resonance imaging and pathology on post-exercise mice. They found that the change in T2 value of mouse muscle tissue might be related to the concentration of serum myosin heavy chain fragments. This experiment confirmed that T2 mapping can accurately reflect the histopathological changes of muscle injury and its recovery. The aforementioned animal studies, through comparative pathological and imaging research, have conclusively demonstrated that MRI FA values and T2 values can serve as effective quantitative imaging parameters for reflecting the extent of muscle tissue damage and the recovery process. However, the reliability and clinical applicability of such parameters when applied to human models still require further rigorous validation and demonstration.

Thus, DTI combined with T2 mapping can detect muscle microdamage and the subsequent repair of muscle tissue at an early stage, and T2 and FA values can be used as imaging biomarkers for monitoring exercise rehabilitation of muscle tissue.

The impact of daily running at different intensities within three days after a half-marathon on the recovery of thigh muscle micro-injuries

After a half-marathon, runners usually attempt to alleviate the symptoms of muscle injury through running during the recovery period [24, 25]. However, there are discrepancies in the literature regarding the role of running in muscle function recovery. Martin et al. [26] reported that, after a single-leg eccentric contraction cycle exercise, running for 30 min per day at an intensity of 50% of maximum oxygen consumption (VO₂max) for 4 consecutive days did not affect the recovery of muscle function or delayed onset muscle soreness (DOMS). Similarly, in the study by Trevor C et al. [24], within 6 days after downhill running, running for 30 min per day at intensities of 40%, 50%, 60%, and 70% of maximum oxygen consumption (VO₂max) respectively had no beneficial or adverse effects on the recovery of muscle injury indicators and running economy indicators.

Unlike the findings of the aforementioned studies, a study by Wiewelhove, T et al. [27] revealed that adopting an active recovery strategy involving low-to-moderate intensity running (equivalent to 60% of the running speed at the anaerobic threshold) within 1 h after a half-marathon did not promote recovery; instead, it even had a negative impact on the recovery of perceived muscle structural damage and neurohumoral regulation. Similarly, in a study by Sherman et al. [28], 10 marathon runners received either rest or running-based recovery within 6 days after a full marathon. For the running group, the running duration was gradually increased (from 20 min to 45 min) and the running intensity was self-selected. The results showed that the recovery of leg extensor strength was impaired in the running group. Therefore, some scholars [13] argue that running after a marathon, when used as an active recovery strategy, may prolong the muscle recovery time and exacerbate muscle fatigue symptoms.

Nearly all the findings from the aforementioned literature on running as a post-exercise recovery strategy indicate that running either has no promoting effect on muscle recovery or even exerts an adverse impact on it. Our study also shows that running is detrimental to muscle recovery. However, the running speeds and intensities reported in the literature are inconsistent and lack a unified standard. Moreover, it is difficult for recreational runners to determine the appropriate intensity on their own; they need to seek help from professionals and rely on equipment to achieve the desired intensity. In our study, we used the heart rate reserve (HRR) displayed on sports watches—commonly used by recreational runners—as the standard for classifying running intensity. This aligns with the practical needs of the general public and is more easily accepted and adopted by recreational runners.

Our research results demonstrate that after the half-marathon: The passive rest group showed the earliest onset of thigh muscle recovery (starting at 24 h post-race) and returned to the pre-race level within 72 h; The moderate-intensity running group began muscle recovery later than the rest group (starting at 48 h post-race) and did not return to the pre-race level by 72 h; The high-intensity running group showed no signs of thigh muscle recovery even by 72 h post-race. Compared with the rest group, engaging in 30 min of moderate-intensity or high-intensity running for three consecutive days starting from 24 h after a half-marathon is adverse to the recovery of muscle micro-injuries, with high-intensity running being particularly harmful. Therefore, it is not recommended to use high-intensity or moderate-intensity running for recovery within 3 days after a half-marathon.

Our study still has some limitations: (1) Subjects had mixed FM/HM experience, and although all completed the same HM, prior experience may have influenced muscle damage susceptibility. Larger cohorts with uniform running backgrounds are needed; (2) DTI sequences and ROI delineation were manual, leading to potential inter-institutional variability. Automated analysis tools and standardized scanning protocols are required for widespread adoption; (3) Sleep quality, dietary intake, and hydration status were not standardized, which may have affected muscle recovery. Future studies will control these variables; (4) The small sample size (n = 24) limits the generalizability of results. Larger multi-center studies are needed to validate findings; (5) The accessibility of MRI scans and quality healthcare services is limited for many recreational runners, thus restricting the widespread adoption of the quantitative imaging evaluation methods used in this study and the practical application of our findings in the broader recreational runner population. We acknowledge this as a key limitation that may impede the translation of our results to real-world settings.

Conclusions

By using DTI and T2 mapping techniques, this study systematically revealed, for the first time, the differential effects of running interventions with different intensities on the recovery of thigh muscle microstructure after a half-marathon. This provides important imaging evidence for the selection of post-marathon recovery strategies.

The results show that, compared with the rest group, starting moderate-intensity running or high-intensity running at 24 h after a half-marathon is adverse to the recovery of muscle micro-injuries, with high-intensity running being particularly harmful. Therefore, it is not recommended to adopt high-intensity or moderate-intensity running for recovery within 3 days after a half-marathon.

Take home

1. Lower FA values on MRI reflect increased muscle fiber damage post-HM, while elevated T2 values indicate interstitial edema and inflammation—both are sensitive markers of muscle microdamage.

2. DTI combined with T2 mapping is a reliable non-invasive tool for evaluating post-marathon muscle recovery.

3. Contrary to some previous assumptions that running aids recovery, this study provides strong evidence that immediate passive rest post-HM promotes faster muscle repair. High-intensity and moderate-intensity running are not recommended for recovery within 3 days after a half-marathon.

4. Recreational runners should not self-determine post-marathon recovery intensity; professional guidance with heart monitoring is recommended, but recreational athletes minimally should monitor post-HM heart rates.

5. Standardization of DTI scanning sequences and ROI analysis methods is needed for broader clinical and sports science application.

Authors’ contributions

Jing Zhou and ChengYuan Xu drafted the first version of the manuscript.YaWen Wu, XinMiao Mao, ZhenQi Hu and ZhouHao Wei offered advice on the workflow and critically revised the manuscript.YuQing Li and Jie Huang provided insightful comments and revised the manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

This study was funded by the Medical and Health Science and Technology Project of Hangzhou (A20220387), the Medical and Health Science and Technology Program Project of Zhejiang Province (2023KY960, 2024KY1346).

Data availability

With permission from the Affiliated Hospital of Hangzhou Normal University, the dataset is available from the corresponding authors Dr. YuQing Li and Dr. Jie Huang. Restrictions do exist, the information isn’t available to the general public.

Declarations

Ethics approval and consent to participate

The protocol was approved by the Research Ethics Committee of the Affiliated Hospital of Hangzhou Normal University in accordance with the Declaration of Helsinki for human research [No.: 2024 (ES)-KS-011]. Written informed consents were obtained from all the included study participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.