LEARNING A COMPLEX MOTOR SKILL FROM VIDEO AND POINT-LIGHT DEMONSTRATIONS 1, 2

Size: px

Start display at page:

Download "LEARNING A COMPLEX MOTOR SKILL FROM VIDEO AND POINT-LIGHT DEMONSTRATIONS 1, 2"

Linda Cummings
6 years ago
Views:

1 Perceptual and Motor Skills, 2010, 111, 1, Perceptual and Motor Skills 2010 LEARNING A COMPLEX MOTOR SKILL FROM VIDEO AND POINT-LIGHT DEMONSTRATIONS 1, 2 SÉRGIO T. RODRIGUES, MARCELA Laboratory of Information, Vision, and Action (LIVIA) UNESP State University of São Paulo, Bauru, SP, Brazil Summary. The aim of this study was to compare the learning process of a highly complex ballet skill following demonstrations of point-light and video models. 16 participants divided into point-light and video groups (n demonstration and practice, with a retention test a day later. Measures of head and trunk oscillation, coordination disparity from the model, and movement time difference showed similarities between video and point-light groups; ballet experts evaluations indicated superiority of performance in the video over the point-light higher relevance of information contained in biological motion models applied to learning of complex motor skills. Perceiving other persons actions, emotions, and intentions is a very human movement, one of the aspects necessary to observational learning - serving others, into motor commands to reproduce the observed behavior body segments in a new and functionally integrated manner. Important visual information available to the learner when observing others activity cations of information regarding this relative motion are stronger in the initial stage of motor learning than in the more advanced ones (Newell, 1 Address correspondence to Sérgio Tosi Rodrigues, Laboratory of Information, Vision, and Luis 2 This study was supported by scholarships to M. C. Ferracioli (Programa Institucional de - - The authors would like to thank M.Sc. Rodrigo Carvalho for his technical support on constructing the video models, and Mrs. Marina Denardi for revising the manuscript. DOI / PMS ISSN

2 2 S. T. RODRIGUES, ET AL. - More recent developments in computer science and technology have al- manipulate many variables of interest (e.g., Vanrie & Verfaillie, 2004; Ma, The literature on visual analysis of human motion has shown that observers easily perceive a variety of aspects when watching a point-light walking animation, such as the action being performed (Norman, Pay- Comparing motor learning processes of groups who watched either point-light or regular video demonstrations is a way of studying how the visually perceived relative motion modulates the motor action (e.g., Al- et al. perceiving and reproducing the movement, point-light displays should be el itself facilitates the motor learning process, in comparison to no-model groups or verbal guidance and discovery methods (e.g., Al-Abood, et al., 2001; Horn, et al Removing structural information via point-light can facilitate the perception of motion in some cases because it leaves only salient relative mo- - that physical therapists made more reliable assessments of lumbar sta- sources.

3 LEARNING THE PIROUETTE FROM BIOLOGICAL MOTION 3 - body parts can contribute more to motion perception than the proximal ones, due to its higher spatial complexity and magnitude (Mather, Rad- suggest that observers are less sensitive to the point-light information - tance of information about the elbow motion for recognition, but showed - Children who observed a point-light model dribbling a basketball were less accurate in their movement outcomes and form when imitating the movements than a group that viewed a video model and a no- et al adult and child participants; results from one of their experiments showed but children s action reproduction was superior when they watched a vid- - tion of biological motion is already present in one-year-old babies (Fox & - a point-light model as compared to a video model, independently of (reg- et al plots. Similar results were found by Horn, et al point-light, and no-model groups in the same soccer chip-shot task. Participants who observed the models showed immediate and enduring changes to more closely imitate the model s relative body part motions. The researchers also measured participants visual search behavior, which was formance accuracy or learning (Horn, et al

4 4 S. T. RODRIGUES, ET AL. In general, the above research shows evidence supporting the claim ognition and movement replication, although some discrepancies are evident. In particular, results seem to depend heavily on task constraints; if - - et al., study, observation of the model s action is most important in the replication of the movement (Hodges, et al The complexity of the task is also a consideration in this debate. Motor learning studies based on point-light stimuli have used relatively sim- nature of the task. Horn, et al dition, as the complexity and novelty of the movement increases, a more sophisticated processing system operates that perhaps is more reliant on information about intra-limb and inter-limb coordination and relative motion, as suggested by Hodges, et al Although the debate about the relevance to motor learning of the in- 10 years (Hayes, et al such as the role of outcome goal and task complexity. The present study - - The aim of the present study was to compare the learning process of a highly complex ballet skill following demonstrations of point-light and - vant for movement recognition and reproduction than in the video model, body coordination of the learners was assessed by global performance compared to the model were used to describe the performance. The point-

5 LEARNING THE PIROUETTE FROM BIOLOGICAL MOTION 5 measured variables, indicating the use of salient information available in the point-light model during the motor learning process. M Participants Sixteen female undergraduate students (M SD with normal or corrected-to-normal visual acuity volunteered in the present study. They had no previous experience in ballet and were naive as to pate in the study as volunteers; the study was approved by the universi- point-light group, resulting in two groups of eight participants each. Apparatus screen dimensions were 56 by 42 cm. MATLAB ysis. Task and Video Construction cal ballet. It involves a complete turn around the body s longitudinal axis - power to generate body rotation, eye-head coordination to provide visu- - details involved and its markedly long learning process; demonstrations process, which is under investigation in the present study. Participants of both groups had to copy the model s action presented on a television ular video format while the point-light group watched an edited version of the same video clip showing only point lights on a black background. previous biological motion-recognition test using three video clips vary-

6 6 S. T. RODRIGUES, ET AL. F en dehors in the : it is a turn executed in the direction of the leg to be raised (e.g., half-tip with a preparation position third position preparation position before keeps a position named retiré passé ing the number of markers, color, and disappearance during the movement. Eighty-one undergraduate students (M SD domly divided into three groups, with normal or corrected-to-normal with higher recognition ratings (74.1% of participants were able to visu- - - to the model s head, shoulders, elbows, wrists, hips, knees, ankles, and extremities of hands and feet; these markers never disappeared throughout the presentation.

7 LEARNING THE PIROUETTE FROM BIOLOGICAL MOTION 7 Procedure and Design observe the model on the screen and then copy the model s movement as tached to participants nose end, top of the head (intersection of saggital blocks of 20 trials during two consecutive days of data collection. Prior to each practice block, participants watched their respective model video three times, seated 2.4 m from the television screen. The visual angle of the - - et al - trial of each block and the twentieth trial of the retention block were video recorded for each participant in top and front views for further analysis. Dependent Measures and Data Analysis Global performance score. - served performance to model performance; the respective mean values 10 was assumed to represent a movement identical to the model s demonstration. Each expert evaluated a randomly ordered data set of all trials selected from all trials of all participants and an additional 20% of the trials for each participant. Thus, during the assessment of each trial, the experts were not aware of the participant s group or trial number, or about which participant was in the next trial viewed. Pearson s correlations between were.59,.56, and.59, respectively. For the 20% of trials evaluated twice by Video-based analysis of movements of the model and participants was conducted; the re- as a criterion to compare groups. Changes in head-trunk coordination

8 S. T. RODRIGUES, ET AL across trials and with the model. et al Hayes, et al uses the disparity between each participant and model s data in each tri- - Head and trunk oscillations. Oscillations of trunk and head were ob- respectively, of the top of the head marker and a calculated midpoint between shoulders. These oscillations were measured in both anteriorposterior and medial-lateral axes. - - the model. Statistical Analysis Data for each dependent measure obtained throughout the practice tor. Data for each dependent measure obtained at the end of the practice period and 24 hr. later were compared by using a Group (video, point- the last factor. When violations of the assumption of sphericity for repeated measures ANOVA were observed, Greenhouse-Geisser adjustments were used; pairwise comparisons and Bonferroni adjustments for multi- R Trunk and head oscillations (anterior-posterior and medial-lateral

9 LEARNING THE PIROUETTE FROM BIOLOGICAL MOTION 9 during practice or in the retention test. Table 1 shows video and pointlight groups data collapsed across the trials. The models oscillations are remarkably smaller than those of participants, especially on the mediallateral axis. TABLE 1 Axis Model Video Point-light M SD M SD Trunk AP ML Head AP ML tion phase (F p 2 the point-light group (F p 2 TABLE 2 Group Retention Video M SE Point-light M SE 20 a,b,c 40 d 60 e a 100 f 120 d 140 b 160 c,e,f Note. A score of 10 represented a movement identical to the model s demonstration. Means p p SE - groups were more similar to the model s demonstration in their trunk pro-

10 10 S. T. RODRIGUES, ET AL. F ( F 4.0,56.1 p 2 scribes how video and point-light participants became faster as they prac- ference occurred between Trials 40 (M SE M sec., SE - F p 2 mately 0.1 sec. while the video group s decreased by approximately 0.4 their interaction. Interestingly, the groups performance ranks switched on the retention test with the point-light group s movement time decreasing from.21 (SE SE - SE SE tween the models and the point-light and video groups mean movement - D would occur for the point-light group, because biological motion models

11 LEARNING THE PIROUETTE FROM BIOLOGICAL MOTION 11 F represent one standard deviation. presumably preserve only the necessary visual information. The point- of the practice that occurred during the experiment. point-light group s learning process. Measures of head and trunk oscilla- - movement recognition and reproduction, but not more relevant than that

12 12 S. T. RODRIGUES, ET AL. (sec.) F in a regular videotape. The motor learning literature has shown similar et al. (2002, light groups during learning of a soccer chip shot; only a perceptual as- for a point-light than a video group (Horn, et al ing process of a bowling action, adult point-light and video groups re- performance when viewing the point-light compared to the video model (Hayes, et al task complexity and the absence of an explicit outcome goal (i.e., explicit A perceptual-motor improvement due to practice was expected for both groups. The present results partially support this expectation. Al- - reduced, reaching a magnitude similar to the model s movement time after 160 trials of practice, and maintained in the retention test. For instance, experienced ballet dancers did not change their movement time during of trunk rotation is not dependent on sensory updates (Denardi, Ferra-

13 LEARNING THE PIROUETTE FROM BIOLOGICAL MOTION 13 improvements only in the temporal structure of the learner s movement; spatial aspects such as the similarity with the model s head-trunk coordi- study to show performance improvement. Accordingly, Horn, et al - study, participants observing the video model presented a clear reduction in variability of the temporal phasing throughout practice trials, whereas type of video model watched by participants. The video group performed - performance improvement of both groups during practice as expected. Superiority of the video over the point-light group was similarly observed who watched point-light video models were less accurate in movement reproduction than those in the regular video group. This sort of evidence bowling (Hayes, et al. et al An additional aspect of biological motion perception may be related light groups did not become more similar to the model s movement pattern possibly because to set the head is a technical aspect of the pirou- - clearly distinct in time. Denardi, et al with mean expertise of 12 yr. use a particular strategy to maintain visual

14 14 S. T. RODRIGUES, ET AL. ence, and head and trunk anterior-posterior oscillation, perhaps indicat- et al - limited; the head was represented only with one point, which would potentially become a problem when learning to dissociate head and trunk movements. On the other hand, the video group did not take advantage of having more complete information on the associated movements of head due to visual aspects contained in video displays other than head/trunk dissociation. head-trunk coordination and body oscillations between groups and there the point-light group. The analysis based upon ballet experts evaluations, account more detailed aspects of the participant s movements. Presum- experts when deciding on the movement similarity between participant and model; this implies that the contribution of head-trunk coordination - - phases. Aspects other than head-trunk coordination, which would have sible for explaining the improvement of both groups throughout the trials regular video available for the video group contained information on body shape, environmental structure, head orientation, facial expressions, motor learning process. These data seem to suggest that, at least for the amount of practice provided, relative motion information (except regard- learned by both groups, but context-related information contained only in the video model improved the motor learning. With criteria used rela- -

15 LEARNING THE PIROUETTE FROM BIOLOGICAL MOTION 15 evaluations clearly showed learning of both groups and the video group s superiority. One possibility to explain the present data is that expert ob- mance than the objective measures used. On the other hand, these gen- the multiple aspects involved in this complex task. Further investigation should consider a variety of movement elements as Gray, Neisser, Shap- - video using a rating system composed of thirteen dimensions (e.g., arm position, arm movement, balance, hesitation, head position, head move- distinguish between head-trunk coordination and other aspects from the experts point of view. Previous studies have suggested that too-simple tasks (Horn, et al., et al - and video groups. In order to overcome these limitations, the present study used a highly complex task, which lacks an explicit outcome goal, during the initial phase of motor learning. Present evidences support the to the debate about point-light and video models, such as the head-trunk ing learning of complex motor skills and to clarify the results heavy dependency on task constraints observed so far. REFERENCES & Journal of Motor Behavior, 33, & ence in biomechanical motion. Journal of Experimental Child Psychology, 37, &chology, & of the perception of biological or non-biological motion on movement execution. Journal of Sports Sciences, 25, & tion without familiarity cues. Bulletin of the Psychonomic Society, 9, & - no ballet. cias do Desporto,.

16 16 S. T. RODRIGUES, ET AL. & movements. Journal of Experimental Psychology: Human Perception and Performance, 31, & Science, & movements. 4, & Ecological Psychology, 3, & Journal of Sports Sciences, 25, & during observational learning? Journal of Sports Sciences, 25, & the role of visual search during observational learning from video and point-light models. Journal of Sports Sciences, 20, & coordination changes in response to video and point-light demonstrations without KR. Journal of Motor Behavior, 37, sis. Perception & Psychophysics, 14, & of identity, gender and emotion perception from biological motion. search Methods, & motion. 249, & Designing experiments and analyzing data: a model comparison perspective & tics in biomechanics and motor control. Journal of Sports Sciences, 19, Amsterdam: Elsevier. Pp & - Journal of Human Movement Studies, 7, & ception of biological motion. Psychology and Aging, 19, & point-light biological motion displays by young children. Perception, 30, & Perceptual and Motor Skills, &tion of biological motion displays. Acta Psychologica,

17 LEARNING THE PIROUETTE FROM BIOLOGICAL MOTION 17 & point-light walkers. Journal of Experimental Psychology: Human Perception and Performance, 31, terns by children: a perception/action approach. Unpublished doctorial thesis, International Institute for Sport and Human Performance, Univer. of Oregon, Eugene. Dicionário de ballet plications of a distal versus proximal approach. In G. Jansson, S. Bergstom, & W. Perceiving events and objects biological motion. Human Movement Science & a look at demonstrations. Irish Journal of Psychology, & motor skills: towards a visual perception perspective. Journal of Human Movement Studies, & angle: angle plots. Journal of Human Movement Studies, & Perception & Psychophysics, 67, & human point-light actions. 36, Perceptual and Motor Skills, 67, Biomechanics and motor control of human movement Accepted August 31, 2010.

6.4 The Ankle. Body Divided into Planes. Health Services: Unit 6 Arms and Legs. Body Movement Vocabulary

6.4 The Ankle. Body Divided into Planes. Health Services: Unit 6 Arms and Legs. Body Movement Vocabulary 6.4 The Ankle Body Movement Vocabulary When fitness professionals refer to movement of the body, the pattern of movement is described from the anatomical position This position can best be described as