Background and Purpose
Motor learning can occur through various practice modalities such as physical execution, action observation, and motor imagery. According to the simulation theory, these methods engage similar neural mechanisms and activate overlapping brain regions, suggesting a shared basis for skill acquisition. While several neurophysiological and behavioral studies have supported this claim, there is also evidence indicating distinct underlying mechanisms for these methods.
Despite extensive research on movement outcomes, few studies have examined the effects of these practice methods on coordination-related variables such as functional degrees of freedom (fDOF). fDOF reflects the flexibility and synergistic structure of neuromuscular coordination and can be quantified using techniques like principal component analysis.
This study investigates whether physical practice, action observation, and motor imagery similarly affect movement accuracy and fDOF during a golf putting task. Exploring these effects may clarify the extent to which these methods share underlying mechanisms and inform the development of optimized training protocols.
 
Methods
This semi-experimental study used a pretest–posttest design with four groups: physical practice, motor imagery, action observation, and control. Forty right-handed male university students (mean age = 25 ± 3.4 years) with normal or corrected-to-normal vision participated. Participants were randomly assigned to groups. The experimental task was golf putting on a 4×9 m artificial green using a standard club and ball. The target was a 4 cm diameter circle fixed to the green.
After signing consent forms, participants completed a revised Movement Imagery Questionnaire. Seventeen reflective markers were placed bilaterally on key anatomical points. Each participant performed 12 trials aiming at a target 2.44 meters away. Kinematic data were captured with a six-camera SIMI motion analysis system at 200 Hz sample rate. Participants wore tight black clothing to improve marker visibility.
During acquisition, each group completed 162 practice trials across nine blocks (18 trials per block, 2-minute rest between blocks). The physical practice group performed actual strokes without verbal or visual feedback. The action observation group watched videos of matched physical participants. The motor imagery group imagined each putt while holding the club at the start position and verbally confirmed each trial. The control group did not train but read an article on exercise benefits during the same period.
Twenty-four hours later, all participants completed a retention test identical to the pretest. Putting accuracy was assessed by measuring radial error—the distance between the ball’s edge and the target. Kinematic data were filtered with a 4th-order Butterworth filter (7 Hz cutoff frequency), then standardized (Z-scores). Principal component analysis (PCA) on 54 variables (X, Y, Z coordinates of 18 markers) was used to estimate functional degrees of freedom (fDOF).
A 4 (Group) × 2 (Test Phase) mixed-design ANOVA was used to analyze both accuracy and kinematic data. All statistical analyses were performed using SPSS version 16, with the significance level set at p < 0.05.
 
Results
The analysis of participants’ imagery ability (both visual and kinesthetic dimensions) revealed no significant differences between groups, indicating homogeneity in baseline mental imagery skills (all F < 1).
Regarding putting accuracy, results from the 4 (Group) × 2 (Test Phase) mixed ANOVA indicated significant main effects of Group (F(3,36) = 19.78, p <.001, η²p = 0.62), Test Phase (F(1,36) = 102.36, p < 0.001, η²p = 0.74), and their interaction (F(3,36) = 18.57, p < 0.001, η²p = 0.60). Post hoc comparisons showed no significant differences between groups at the pretest phase (p = 0.98), whereas all pairwise comparisons at the posttest were significant (all p < 0.05). In terms of accuracy, performance ranked from best to worst as follows: Physical (M = 20.13), Observation (M = 26.39), Imagery (M = 29.52), and Control (M = 36.23). Significant improvements from pretest to posttest were observed in all groups except the control group (p = 0.81), indicating the effectiveness of all three training methods, particularly physical and observational practice.
For functional degrees of freedom (fDOF), the ANOVA revealed significant main effects for Group (F(3,36) = 3.03, p = 0.05, η²p = 0.21), Test Phase (F (1,36) = 8.72, p = 0.005, η²p = 0.19), and their interaction (F(3,36) = 5.76, p = 0.003, η²p = 0.32). There were no significant group differences at pretest (p = 0.94), but during the posttest, the physical practice group showed significantly lower fDOF than all other groups (all p < 0.05), and the observation group showed significantly lower fDOF than both the imagery and control groups (all p < 0.05). No significant difference was found between the imagery and control groups (p = 0.08). Mean posttest fDOF values were: Physical = 4.66, Observation = 5.16, Control = 5.72, Imagery = 6.06. Furthermore, significant within-group reductions in fDOF from pretest to posttest were observed in the physical and observation groups (both p < 0.05), but not in the imagery or control groups. These findings suggest that physical and observational training were more effective in enhancing coordination and reducing movement redundancy.
 
Conclusion
The results demonstrated that while all experimental groups improved in accuracy from pretest to posttest, physical practice yielded the highest performance, followed by action observation and motor imagery. Functional degrees of freedom (fDOF) decreased significantly only in the physical and observation groups, indicating more efficient movement coordination. These findings suggest that the three practice methods may rely on distinct underlying mechanisms. Based on internal model theory (32), motor imagery may lack effective sensory feedback integration, leading to less precise motor tuning. In contrast, action observation allows visual feedback but still lacks proprioceptive engagement, resulting in intermediate effectiveness. Physical practice, combining both feedback modalities, produced the most efficient learning. The results align with models emphasizing error correction and feedback-based adaptation (30,35-37). Future research should manipulate feedback availability directly and include EEG or coordination metrics to further explore neural and behavioral mechanisms. Overall, combining these methods may optimize motor learning depending on task demands.
 
Article Message
This study highlights the differential effectiveness of physical execution, action observation, and motor imagery in motor learning. While all three methods improved performance, physical practice showed superior outcomes in both accuracy and movement coordination. Action observation ranked second, outperforming motor imagery. Functional degrees of freedom analysis revealed more efficient neuromuscular control in the physical and observational groups. These results suggest that feedback availability and sensory engagement are critical for optimizing internal model updating. The findings support using combined or complementary practice strategies based on task complexity, offering practical implications for motor learning and rehabilitation interventions.
Ethical Considerations
This study was reviewed and approved by the Ethics Committee of Shiraz University.
Authors’ Contributions

Conceptualization: Davoud Fazeli
Data Collection: Davoud Fazeli, Fatemeh Jabbari, Hossein Taghizadeh, Leila Ghohestani
Data Analysis: Davoud Fazeli, Fatemeh Jabbari, Hossein Taghizadeh, Leila Ghohestani
Manuscript Writing: Davoud Fazeli, Fatemeh Jabbari, Hossein Taghizadeh, Leila Ghohestani
Review and Editing: Davoud Fazeli, Fatemeh Jabbari, Hossein Taghizadeh, Leila Ghohestani
Responsible for funding: None declared
Literature Review: Davoud Fazeli, Fatemeh Jabbari, Hossein Taghizadeh, Leila Ghohestani
Project Manager: Davoud Fazeli
Any Other Contributions: None

Conflict of Interest
The authors declare no conflict of interest related to the publication of this article.
 
Acknowledgments
The authors express their sincere gratitude to the participants whose cooperation made this research possible.