1.    Introduction
Performance in middle-distance running is unique and characterized by intermediate factors of biomechanics (1, 2) and physiology (3, 4), with the challenge being to run at high velocities while still maintaining economical movement (5). Understanding the biomechanical principles, including stride length, stride frequency, flight time, and ground contact time, can make a significant difference in middle-distance running performance (6, 7). Running involves a cycle of flight and support phases, with efficient running typically consisting of about 65% flight and 35% support (8). Key kinematic factors such as step length, contact time, flight time, and step frequency regulate this cycle (9, 10). Optimal stride length and frequency help minimize energy use while maximizing speed. Efficient biomechanics, including shorter ground contact and longer flight times, enhance running economy and performance (11, 12).
Swinnen et al. (13) indicated that runners tend to adopt a stride frequency that minimizes energy consumption, with higher frequencies leading to reduced ground contact time and increased hip power during leg swing. Moreover, a systematic review by Van Hooren, Jukic (14) indicated that a higher cadence correlates with lower oxygen costs, suggesting that optimizing stride frequency can enhance running economy and performance. Experienced runners naturally select stride lengths that minimize oxygen uptake, demonstrating a capacity for self-optimization (15). In addition, differences in ground contact time significantly affect running economy, with shorter contact times linked to better performance, as seen in comparisons between North African and European runners (16).
Study reported that high-intensity interval training (HIIT) is a powerful method for enhancing muscle mass, strength, and aerobic power (17), all of which contribute to improvements in the biomechanical aspects of running performance. Uphill running serves as an effective form of resistance training because it closely mimics running biomechanics while offering sport-specific strength benefits. This aligns with findings from a study that highlighted how progressive resistance training can enhance performance and reduce injury risk across various sports disciplines (18).
Several studies have examined the acute effects of uphill running on the kinematic characteristics of a runner (12, 19-23). Forinstance, a study found a 4% increase in step frequency and a 4% decrease in stride length as incline increased from 0% to 7% (24). A similar study reported that running speed and step length decreased by 5.2% during uphill running compared to downhill running and were 3.0% slower compared to running on level surface (21). Another research also reported that, in comparison with running on flat ground, uphill running is characterized by a higher step frequency, more internal mechanical work, shorter swing/aerial phase duration, and a greater duty factor/contact time (25). This is because, to increase the body's potential energy during uphill running, lower limb muscles must perform a greater total mechanical effort than they do in level running. Uphill running also significantly raises metabolic costs, with a reported 53% increase in energy expenditure at steeper inclines (24). Moreover, shorter flight time and longer groung contact time during the stride cycle were characteristics of the uphill running (26), while higher step frequency (SF) is associated with steeper slopes during uphill running (27, 28). Steeper slopes (e.g., +20%) require greater muscle strength and energy expenditure, influencing running economy differently than level or shallow slopes (29). Transitioning from level to uphill running alters biomechanics, enhancing joint angles and range of motion, which can improve overall running mechanics.
Incorporating uphill running into training regimens can enhance performance by improving metabolic efficiency and kinematic adaptations (30). To enhance level running performance, coaches often incorporate uphill running into their training regimens to strengthen lower limb muscles. Research indicates that running on incline surface significantly increases muscle activation, particularly in the rectus femoris, vastus medialis, biceps femoris, and gastrocnemius, compared to level running (31). He found that uphill running leads to a 16.54% increase in lower limb muscle activation compared to level running. This muscle engagement is crucial for developing strength and endurance, which are essential for improved running efficiency. Previous study reported that MRS, step rate, contact time, and step time were improved significantly higher in the combined uphill-downhill group performed at ±3o incline after 8weeks of training by 4.3%, 4.3%, −5.1%, and −3.9% respectively (32) than in horizontal running relatively smaller improvements 1.7%, 1.2%, 1.7%, and 1.2% respectively. However, a similar study reported a 4.8% improved running speed from a 6-week combined uphill-downhill running on ±3o slope compared to horizontal running. However, no significant change observed in step rate and ground contact time (33). An other study reported that combined uphill-downhill training group increased their MRS by 3.7%, stride rate by 3.1% at (p < 0.05), after training 6-weeks of training while no change observed in contact time and flight time as it compared to the horizontal training group (34).
While the focus of most of the previous studies was on the acute effects of uphill running on different physiological and biomechanical characteristics of athletes (10, 12, 21, 22, 24, 28, 29, 35-38), the chronic effects of training on sloped surface on the kinematics of characterstics of athletes in horizontal running performance are unclear (21). Although some recent long-term studies have explored the combined effects of uphill and downhill training compared to level or downhill running, there is still a notable gap in the literature regarding the isolated effects of uphill training and the relative effects of different hill gradients on running kinematics. This is important because uphill running imposes distinct biomechanical and muscular demands compared to flat or downhill running. Understanding how various uphill gradients influence long-term kinematic adaptations could provide valuable insights for coaches and athletes, enabling more targeted training strategies and performance improvements.Therefore, the present study aims to investigate the chronic effects of three distinct uphill gradients on the kinematic characteristics of middle-distance runners and to identify which gradient is most effective in enhancing these characteristics.
Recent existing long-term studies were focused on comparing the effects of combined uphill-downhill training to horizontal and downhill running on different performance parameters. Still a gap in literature covering the relative effects of different uphill gradients and the separate effects of uphill training on the kinematic characteristics of athletes as uphill running demands a different effort and biomechanical behavior than running on a flat and downhills. Thus, understanding the long-term effects on kinematic parameters at several gradients may still important and might assist coaches and athletes in the training process and performance improvement by providing detailed feedback. Therefore, the present study aimed to determine the chronic effects of three different uphill gradients on the kinematic characteristics of middle-distance runners and to identify the most effective uphill gradient in improving those characteristics. 
2.    Methods
2.1. Participants
A total of forty moderately trained middle-distance runners (24 males and 16 females), with age (M-18.54±1.02, F-17.44 ± 1.09 years [mean ± SD], with 2.11 ± 0.74 years of structured training experience took part in the study. Written informed consent was obtained before the intervention, and the study was approved by the Ethics Committee of Bahir Dar University, Bahir Dar, Ethiopia (Protocol No. IRERC 05/2024). All procedures were conducted in accordance with the Declaration of Helsinki on the treatment of human subjects. Participants were told about the study’s procedures, methods, benefits, and risks. After that, they read and signed the consent form to participate in the study. Participants were recruited based on predefined inclusion criteria. The inclusion criteria were healthy middle-distance runners from Bahirdar University Sports Academy who had no muscular, neurological, or tendon injuries and weren't taking any medication, athletes having at least 6-months of trainin experiance at regional or natonal level. The exclusion criteria were athletes who had less than 6 months of training, those with a lower-body injury in the last 3 months, and anyone taking medication. Eligible athletes were informed about the study through announcements at training sessions and were invited to participate voluntarily. Written informed consent was obtained from all participants prior to enrollment. Participants were randomly assigned to one of three training groups or the control group. Each group consisted of 10 participants (steep gradient group, 7.6%; moderate gradient group, 5.1%; shallow hill group, 2.5%; and control group). The inclusion of specific gradient levels is based on the premise that different gradients have different levels of intensity depending on the categories specified (39, 40). This helps to analyze how each intensity level affects kinematic characteristics, such as step rate (SR), step length (SL), contact time (CT), and flight time (FT). This variety allows researchers to determine which gradient provides the best balance between training load and performance improvement.
2.2. Study design
The study adopted a pre-post parallel group experimental design, with measurements conducted before and after the 8-week intervention period. The participants were randomly assigned to one of the three training groups or the control group. Each group included 10 participants (steeper hill group, 7.6%; intermediate hill group, 5.1%; shallow hill group, 2.5%; and control group). The reason behind including specific gradient levels is because different hill gradients provide varying levels of intensity based on the given categories (41), which helps in examining how each intensity level affects kinematic measures. This variation allows researchers to determine which gradient offers the optimal balance between training load and kinematic improvements.
The study adhered to the CONSORT guidelines for improving the quality and clarity of reporting of experimental studies (42), and registered with pactr.samrc.ac.za (Registration Number NCT01234567) on 29/11/2024. We determined the appropriate sample size for our study using G*Power (43). Based on the results of a previous study (34). We assumed a normal distribution for all responses within a group and estimated the required sample size to be 36 participants to determine an effect size of 0.62 between the pre- and post-experimental groups with a power of 80% at a significance level of 0.05. In order to account for possible drop-outs, an additional 10% allowance was made for sample recruitment, and the required sample size was increased to 40.
To minimize potential bias, trained personnel were utilized to create the random allocation, enroll participants, and assign them to interventions. Given the small sample size and the simplicity of execution, a straight-forward random lottery approach was employed to assign participants across four study groups (each consisting of ten participants): steeper hill group (STHG), intermediate hill group (IHG), shallow hill group (SHG), and control group (CG) in a 1:1:1:1 ratio as illustrated in Fig. 1. Allocation concealment was guaranteed through the use of sealed envelopes. To reduce bias, blinding was implemented when possible. The outcome assessors remained unaware of the group assignments.
Participants underwent baseline assessments of the kinematic parameters from a 30m maximal sprint test prior to the intervention. The same research team oversaw all training and testing activities.

2.3. Procedures
To assess the effects of uphill training on the spatial-temporal variables, a 30-m horizontal maximal sprint running test was used to measure stride rate (SR), step length (SL), contact time (CT), and flight time (FT). After completing a standard warm-up period of 10–15 min followed by dynamic stretching (44), a standard flying 30’s consisting of a 30 m acceleration zone to enable the runner to reach their maximum speed and a 30 m maximal velocity zone to reach maximal sprinting speed on a 400 m oval outdoor track at the university was used to test kinematic characteristics of runners, as in previous studies (45, 46). This is because, while middle-distance running primarily emphasizes aerobic capacity and endurance, short sprint tests such as the 30-meter sprint are commonly used in sports science to assess acceleration, neuromuscular power, and anaerobic performance, which are also relevant to middle-distance performance particularly during race starts, tactical surges, and final sprints. These elements can significantly influence race outcomes in competitive settings. Moreover, the 30-meter sprint test provides a standardized, reliable, and easily replicable measure of explosive speed and lower-limb power, which complements other physiological assessments. It was not intended to represent the entirety of middle-distance performance but rather to provide insight into specific physical qualities that contribute to overall athletic capability. The participants completed three sprint runs over a distance of 30 meters, beginning from a standing start and taking a recovery period of 6 to 8 minutes between each attempt. The fastest time was measured to two decimal places. The timing commenced when the athlete's torso crossed the starting cone at 30 meters and concluded at the 60-meter cone marker (47, 48).
Testing was conducted on the same day for all participants. The trials with the highest MRS values were chosen for further analysis of kinematic characteristics. A Canon SX70 HS camera, capable of recording in 1080p resolution at a sampling rate of 240 Hz, was utilized to capture a 30-meter sprint and the associated kinematic characteristics. The camera was positioned on a tripod at a height of 1.5 meters, located 7.5 meters from the side of the track, perpendicular to the direction of movement at the participant's sagittal plane. The recorded footage was analyzed later using Kinovea 0.9.5 motion analysis software. The kinematic characteristics examined included contact time (ms), flight time (ms), step length (m), and step frequency (Hz).
The step rate was determined by counting the number of steps taken in a 30m sprint run from video data using slow motion analysis and divided by the time used to complete the distance. In addition, step length was determined by dividing the running distance by the number of steps taken to complete the distance. The contact time (CT) and flight time (FT) were calculated by counting the frames in contact and flight on the video data, then dividing by the sampling rate, 240 (1 frame = 240 Hz ≈ 0.0042s). A full stride (2 consecutive steps of right and left foot) was used to analyze contact time. Although there was no significant difference in contact time between the right and left foot, the right foot was used throughout the analysis in contact time. The CT was defined and calculated as the time between initial contact with the ground and the last frame of contact before toe-off. The FT was defined and calculated as the time between toe-off and subsequent initial contact of the contralateral foot. Initial contact and toe-off were visually detected.
Training Protocol: Prior to the study, all participants consistently engaged in running training at varying intensities, averaging four times a week and covering about 21.748 miles weekly. During the intervention, researchers added a hill training program to their existing routine and replaced their strength training with uphill workouts on a Cybex 530T pro plus USA motorized treadmill. The intervention group completed two uphill training sessions each week for eight weeks while still maintaining their regular running training during the rest days from uphill sessions. In contrast, the control group adhered to their usual training regimen. Although the incline varied across groups, training intensity was standardized by controlling for running speed relative to each athlete’s maximal sprinting capacity and by monitoring heart rate responses. This approach ensured that all participants trained at a comparable relative intensity, regardless of incline. Additionally, rest intervals were standardized to allow for consistent recovery across groups. The number of sets or and intervals were gradually increased over the course of the training program to systematically increase training load while allowing for adaptation and recovery. The steep uphill group involved 3-4 sets of 6-10 intervals lasting 30-90 seconds on a treadmill inclined at 7.6%, with participants running at 85-100% of their maximum heart rate and resting for 4-6 minutes between sets and 2-4 minutes between intervals. The intermediate uphill sessions included 2 sets of 6-10 intervals lasting 1-2 minutes on a treadmill set to a 5.1% incline, also at 85-100% HR max, with similar rest durations. The shallow uphill sessions consisted of 6-10 intervals lasting 2-3 minutes on a treadmill at a 2.5% incline, again at 85-100% HR max, with 2-4 minutes of rest between intervals, as detailed in Table 1. Participants in the control group continued with their standard weekly training programs.

2.4. Statistical analysis
Statistical analysis was conducted using IBM SPSS Statistics 27 (IBM Corporation). The results for the outcome variables are presented as mean ± standard deviation (SD). A two-way mixed (time [pre-post] X group) repeated-measures analysis of variance (ANOVA) was employed to determine if there were significant differences between pre-training and post-training tests among the training groups and any interaction effects for each variable. If significant main effects or interactions were found, post hoc Tukey’s HSD test with Bonferroni correction was applied for all pairwise comparisons. The normality of the data was assessed using the Shapiro-Wilk W test, and homogeneity of variance was evaluated with Levene’s test. All dependent variables (MRS, SR, SL, FT, and CT) were analyzed at a significance level of p < 0.05. The interpretation of data included the magnitude of the mean ± SD of the mean difference (MΔ), 95% confidence interval (95% CI), and p-values (50).
3.    Results
Table 2 presents a summary of the baseline descriptive data and the adjusted absolute changes in the kinematic characteristics of participants throughout the study. The statistical analysis of the pre-training assessments revealed no statistically significant differences between groups across all variables. This demonstrates that groups that were similar at baseline were created by the randomization process, which provides a solid foundation for comparing the three uphill groups to the control group.

A two-way mixed repeated measure ANOVA revealed a significant main effect of time on Vmax (F (1, 36) = 173.68, p < 0.001, η_p^2 = 0.82), indicating that Vmax significantly changed over the training period, as indicated in Table 3.  Average Vmax were significantly higher on post-test result (M = 7.87 ± 0.86 m·s-1) than pre-test result (M = 7.32 ± 0.77 m·s-1). Additionally, there is a significant main effect of hill gradient level on Vmax (F (3, 36) = 2.87, p = 0.049, η_p^2= 0.20), suggesting differences in Vmax between the different hill gradients as indicated in Table 3. 

Vmax were significantly higher on STHG (M = 8.74 ± 0.55 m·s-1), than IHG (M = 7.93 ± 0.79 m·s-1), SHG (M = 7.64 ± 0.54 m·s-1), and CG (M = 7.16 ± 0.76).  There was also a significant interaction effect between time and hill gradient level (F (3, 36) = 42.67, p < 0.001, η_p^2= 0.78), indicating that the effect of training time on Vmax varied depending on the hill gradient (Table 3). The post hoc Tukey’s Honestly Significant Difference (HSD) comparisons showed pre-post training differences at p < 0.05 for the STHG (MΔ = 1.28, 95% CI [1.11 – 1.44], p = 0.001), IHG (MΔ = 0.59, 95% CI [0.42 – 0.76], p = 0.001), and the SHG IHG (MΔ = 0.33, 95% CI [0.16 – 0.50], p = 0.001). However, no change was observed in the CG (Table 2). In addition, Vmax was significantly higher in the STHG compared to the CG (MΔ = 0.93, 95% CI [0.06 - 1.81], p = 0.032), while no statistically significant differences were detected in all other pairs of comparisons p > 0.05 (Table 4).

The analysis for SR revealed a statistically significant time effect of (F (1, 36) = 194.24, p < 0.001, η_p^2= 0.84), indicating that SR significantly changed over the training period.  The average SR was significantly higher in the post-test result (M = 4.52 ± 0.59 Hz) than in the pre-test result (M = 4.11 ± 0.41 Hz). Additionally, there is a significant main effect of hill gradient level in SR (F (3, 36) = 2.87, p = 0.049, η_p^2= 0.20), suggesting differences in SR between the different hill gradients. SR was significantly higher in STHG (M = 5.08 ± 0.51 Hz) than in IHG (M = 4.70 ± 0.39 Hz), SHG (M = 4.21 ± 0.38 Hz), and CG (M = 4.07 ± 0.45 Hz). There was also a significant interaction effect between time and hill gradient level (F (3, 36) = 59.37, p < 0.001, η_p^2= 0.83), indicating that the effect of training time on SR varied depending on the hill gradient. The post hoc Tukey’s Honestly Significant Difference (HSD) comparisons showed pre-post training differences at p < 0.05 for the STHG (MΔ = 0.95, 95% CI [0.83 – 1.07], p = 0.001), and IHG (MΔ = 0.61, 95% CI [0.50 – 0.73], p = 0.001). However, no significant change was observed in the SHG and CG (Table 3). In addition, SR was significantly higher in the STHG compared to the CG (MΔ = 0.52, 95% CI [0.02 - 1.03], p = 0.04), while no statistically significant differences were detected in all other possible pairs of comparisons p > 0.05 (Table 4).
The analysis for SL also showed a significant main effect of time (F (1, 36) = 76.033, p < 0.001, η_p^2= 0.679), indicating that SL significantly changed over the training period.  The average SL was significantly higher on the post-test result (M = 1.74 ± 0.09 m) than the pre-test result (M = 1.72 ± 0.09 m). However, there is no significant hill gradient effect on Vmax (F (3,36) = 2.080, p = 0.120, η_p^2 = 0.148). There was also a significant interaction effect between time and hill gradient level (F (3, 36) = 16.611, p < 0.001, η_p^2= 0.581), indicating that the effect of training time on SL varied across the hill gradients. The post hoc Tukey’s Honestly Significant Difference (HSD) comparisons showed pre-post training differences at the 0.05 level for the STHG and the IHG (MΔ = 0.061, 95% CI [0.049 – 0.073], d = 0.65), (MΔ = 0.028, 95% CI [0.016 – 0.040], d = 0.33) respectively. All other pairwise comparisons between group showed no differences, p > 0.05. Lastly, contact time and flight time didn’t show significant change within group and between group, as indicated in Table 3.
 4.   Discussion
The purpose of this study was to examine the chronic effects of uphill training on kinematic characteristics of middle-distance runners, with specific attention to identify the hill gradient that can best modify kinematic characteristics of middle-distance runners. The major findings of the present investigation suggest that; 8 weeks of high intensity uphill training can positively impact key kinematic characteristics of middle-distance runners.
In particular, Vmax was improved across all hill gradients, with substantial improvement observed at the steeper and intermediate hill gradients. While studies directly comparing the long-term effects of different uphill gradients on Vmax are limited, our findings align with earlier research. For instance, Tziortzis (51) showed a 3.3% improvement after 12 weeks of training on an uphill training slope of 8° on maximal running speed. This supports the notion that steeper gradients may elicit greater neuromuscular and biomechanical adaptations conducive to speed development. The key difference between the present study and the earlier work by Tziortzis is the shorter intervention duration eight weeks compared to twelve and the lower the gradient level (7.6%) compared to (14%). Interestingly, this reduced timeframe and hill gradient was still sufficient to elicit meaningful performance adaptations, suggesting that significant improvements can occur within a relatively brief training period and less steep gradients. Similar studies have reported improvements in Vmax of 3.7% and 3.5% using a 5.24% hill gradient compared to flat-surface training (34, 52). A key distinction between the present study and these previous investigations lies in the training design: while earlier studies employed a combined uphill-downhill training approach, the current study focused on the isolated effects of different uphill gradients. This distinction allows for a more precise understanding of how specific incline levels independently influence performance adaptations.
As the uphill training increases resistance, which helps in building strength and power (53), which are crucial for maximal running speed. The resistance provided by running uphill requires greater force production from the lower limb muscles, particularly the quadriceps, hamstrings, and calves. The rationale for the superiority of the steeper hill gradient against intermediate and shallow hill gradients is due to the increased muscular demand leading to hypertrophy and neuromuscular adaptations, improving the athlete's ability to generate force quickly (54). Uphill running necessitates more force application, which can increase muscle strength and power, contributing to improved acceleration and maximal velocity (55). 
The efficiency of uphill training is further demonstrated by the improvements observed in step rate (SR) and step length (SL), particularly at steeper and intermediate hill gradients. These enhancements were most pronounced at the steeper inclines. Similar findings were reported by (32, 56), who observed improvements in SR but no significant changes in SL. However, a key difference lies in the study design: the previous studies utilized a combined uphill and downhill training protocol on a 3° gradient, with comparisons made against horizontal and control groups. In contrast, the present study isolated the effects of different uphill gradients, allowing for a more targeted analysis of their specific impact on running kinematics. Notably, there is a lack of experimental data in the existing literature examining long-term changes in SR, SL, CT, and FT across varying uphill gradients. This highlights the novelty and contribution of the present study in addressing this gap and providing new insights into the gradient-specific adaptations in running mechanics.
It is well known that incline running modifies the length and frequency of strides, improving sprinting mechanics and allowing athletes to accelerate more effectively (57, 58). Sprinting performance requires the legs to traverse through the stride cycle at quicker rates and the muscles to shorten and lengthen more quickly, which is dependent on stride rate (59). Acute uphill running leads to increased stride rate and a shorter stride length as runners adjust to the incline by lifting their knees more and keeping their center of gravity over their lead foot (60, 61). This, in the long run, improves stride rate and stride length in horizontal running performance. Compared to intermediate and shallow hill training, steep hill training increases the body's potential energy by having lower limb muscles execute a higher net mechanical work. All joints, but especially the hip, produce more power to meet the increased needs of work as the running inclination rises. This suggests that compared to intermediate and shallow hill training, steep hill training necessitates higher levels of muscle activation. Therefore, adaptations to uphill training directly impact neuromuscular activation, which ensures the overall efficiency of movement, including step rate and step length (22). The study also demonstrated significant difference among training groups in Vmax and SR, while no significant difference observed in the rest of the variables (SL, CT, and FT). Vmax and SR was significantly different in the STHG compared to the control group. This suggests that uphill training at these gradients is particularly effective in enhancing maximal sprinting speed, likely due to increased neuromuscular demand and biomechanical adaptations. While the absence of significant changes in step rate (SR) and step length (SL) in the shallow hill gradient (SHG) group, as well as in contact time (CT) and flight time (FT) across all training groups, may initially seem unexpected, several plausible explanations can be considered. One key factor is the relatively short duration of the intervention—eight weeks—which, although sufficient to elicit improvements in certain performance variables such as Vmax and SR, may not provide enough time for more subtle neuromuscular and biomechanical adaptations to manifest in all kinematic parameters.
Previous research suggests that longer training periods are often required to elicit measurable changes in variables such as CT and FT, especially when the training stimulus is moderate (62). Additionally, the specificity of the training stimulus plays a crucial role. The training protocols may have been more effective in targeting velocity-related adaptations rather than those influencing temporal gait characteristics like CT and FT. For instance, improvements in Vmax and SR may result more directly from enhanced force production and stride mechanics, whereas changes in CT and FT might require more focused interventions, such as plyometric or technique-specific drills. Studies have shown that resisted sprint training (e.g., sleds, uphill running) tends to produce moderate effects on acceleration and step frequency, but not always on CT or FT when compared to traditional sprinting (63). 
Individual variability in response to training is another important consideration. Although, baseline characteristics of the participant was no significant difference, individual athletes may differ in their baseline kinematic profiles, neuromuscular efficiency, and adaptability, which can influence how they respond to a given training stimulus. This variability can dilute group-level statistical significance, even when individual improvements are present. Finally, it is also possible that certain kinematic parameters, particularly CT and FT, are more resistant to change and may require either a longer training duration or a higher training volume and intensity to produce measurable effects (64). Future studies with extended intervention periods, larger sample sizes, and more targeted training modalities may help clarify these relationships and further elucidate the mechanisms underlying kinematic adaptations to uphill training. 
While this study constitutes a first in the literature, in comparing the chronic effects of three different uphill gradients on kinematic characteristics of middle-distance runners, it has methodological limitations that need to be acknowledged and addressed in the future. Due to the small number of participants available in the field, the study included a range of training experiences (6 months to 4 years), and mixed gender was part of the study. These may affect the result of the study, as athletes at different stages of their training journey and due to gender may respond differently to the same training regimen. This diversity of the study population may be considered as a limitation. However, the intervention study on athletes having some training experience while using a well-established controlled training protocol can be viewed as an advantage in the current study and for the future development of an efficacious training program. Future studies should examine the impact of training load, frequency, duration, and extra steep hill gradients on physiological performance parameters.
5.    Conclusion
The present study investigated the effects of different uphill training gradients on kinematic characteristics of middle-distance runners. The results of the present study demonstrated the positive effects of uphill training on key kinematic characteristics of middle-distance runners that are related to maximal running speed. The gains in Vmax were supported by improvements in SR and SL in favor of steep hill gradients. Different uphill training gradients appear to elicit specific training adaptations, implying that uphill training should be carefully tailored to the athlete's strengths and limitations. Until further evidence is available, runners can presume that changes in a range of kinematic characteristics induced by modifying the uphill-running loadings will result improvements in key kinematic characteristics. This could be applicable in middle-distance runners, requiring large anaerobic capacity, providing the ability to generate a high power output through ground reaction forces for high velocities over moderate distances (65). In addition, intermediate hills could be applied to longer distances, requiring relatively less power output and efficient use of energy through improved running mechanics and running economy (66). Moreover, coaches and athletes need to consider and apply uphill training as an efficient alternative training strategy for improving the kinematic characteristics of middle-distance runners.
Ethical Considerations
Compliance with ethical guidelines
The study was carried out in strict accordance with the ethical guidelines and principles using human subjects. The trial was approved by the Ethics Review of BDU, Sport Academy (Protocol number: IRERC 05/2024). Both parents/legal guardians of minors and minors/participants themselves were informed about the intervention and possible adverse events before the commencement of the trial and signed an informed consent form.
Funding
The author(s) acknowledge that financial support for the study was provided by Bahir Dar University, which played a crucial role in facilitating the research. 
Authors' contributions
YA, TT, and ZB conceived and designed the study. YA and ZB conducted the experiments and collected the data. YA, TT, and ZB analyzed the data. YA, TT, and ZB wrote the manuscript. All the authors read and approved the manuscript.
Conflicts of interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.