Experimental Brain Research

, 214:567

Effects of the model’s handedness and observer’s viewpoint on observational learning

Authors

  • Hassan Rohbanfard
    • Département de kinésiologieUniversité de Montréal
    • Département de kinésiologieUniversité de Montréal
Research Article

DOI: 10.1007/s00221-011-2856-z

Cite this article as:
Rohbanfard, H. & Proteau, L. Exp Brain Res (2011) 214: 567. doi:10.1007/s00221-011-2856-z

Abstract

Observation promotes motor skill learning. However, little is known about the type of model and conditions of observation that can optimize learning. In this study, we investigated the effects of the model’s handedness and the observer’s viewpoint on the learning of a complex spatiotemporal task. Four groups of right-handed participants observed, from either a first- or third-person viewpoint, right- or left-handed models performing the task. Observation resulted in significant learning. More importantly, observation of same-handed models resulted in improved learning as compared with observation of opposite-handed models, regardless of the observer’s viewpoint. This suggests that the action observation network (AON) is more sensitive to the model’s handedness than to the observer’s viewpoint. Our results are consistent with recent studies that suggest that the AON is linked to or involves sensorimotor regions of the brain that simulate motor programming as if the observed movement was performed with one’s own dominant hand.

Keywords

Allocentric observationEgocentric observationRelative timingObservation perspectiveModel handednessAction observation networkMotor skill learning

Introduction

Observation facilitates learning of a motor skill by permitting the observer to determine the key spatial and/or temporal features of the task, which spares him or her the need to create a cognitive representation of the action pattern through trial and error (Blandin et al. 1994; Carroll and Bandura 1982; Newell 1981; Pollock and Lee 1992; Schmidt and Lee 2005; Scully and Newell 1985; for reviews on observational learning, see McCullagh and Weiss 2001; Wulf and Mornell 2008; Wulf et al. 2010). These findings are supported by recent brain imaging studies, which indicated that an “action observation network” (AON; including premotor cortex, inferior parietal lobule, superior temporal sulcus, and supplementary motor area) engages the observer in processes similar to those that occur during physical practice (Brown et al. 2009; Buccino et al. 2001; Cisek and Kalaska 2004; Cross et al. 2009; Dushanova and Donoghue 2010; Fogassi et al. 2005; Frey and Gerry 2006; Gallese et al. 2002; Grafton et al. 1997; Shmuelof and Zohary 2006).

However, little is known about the attributes of a good model. In real-life situations, one may observe a model performing a task with the opposite hand (hereafter called an opposite-handed model), such as when a left-handed trainer shows a right-handed pupil how to putt or drive a golf ball. Moreover, the model may be observed from different viewpoints. The pupil may face the trainer (third-person observation) or be located such that he or she has the same perspective as the trainer (first-person observation). From a third-person perspective, when a left-handed trainer faces a right-handed pupil, they both swing the club in the same direction, as if the pupil was facing a mirror. A right-handed trainer (hereafter called a same-handed model) facing a right-handed pupil will swing the club in the opposite direction. However, when a same-handed trainer is observed from a first-person perspective, the pupil observes the trainer swinging the club in the same direction as the pupil. Does the pupil learn more when observing an opposite-handed trainer or is the skill learned more easily when the pupil observes a same-handed trainer who is either (a) facing him or her (i.e., third-person observation) or (b) placed in the same perspective (i.e., first-person observation)?1

Brain imaging studies have revealed that observing a left- or right-hand reach-and-grasp movement from a first-person perspective (also called egocentric) resulted in larger blood oxygenation level-dependent (BOLD) responses in the contralateral anterior intraparietal cortex of right-handed observers, which is similar to when one performs the observed task (Shmuelof and Zohary 2006, 2008). However, when the observation was from a third-person perspective (i.e., the model faces the observer, also called allocentric), BOLD data revealed a larger activation of the ipsilateral anterior superior parietal lobule (aSPL; Shmuelof and Zohary 2008). Observing a right-handed action increased BOLD activity in the right hemisphere. In this vein, using transcranial magnetic stimulation (TMS), Alaerts et al. (2009) had participants watch videos showing left- and right-hand extension movements from a first- or third-person perspective. Electromyography data demonstrated an increase in the left primary motor cortex excitability when participants observed the right hand from a first-person perspective or when participants observed the left hand from a third-person perspective (see also Hesse et al. 2009). Therefore, learning a motor skill might be facilitated when observing a same-handed trainer from a first-person perspective or an opposite-handed trainer from a third-person perspective. However, the observation of increased activation of different brain regions does not necessarily mean that behaviorally significant differences in the learning of a complex motor task will accompany these activation changes. Therefore, our goals were to determine whether same-handed or opposite-handed models would better promote learning and whether this effect would be mediated by the observer’s perspective.

Method

Participants

Seventy-two self-declared right-handed students (46 females) from the Département de kinésiologie at the Université de Montréal participated in this experiment. All participants reported normal or corrected-to-normal vision, had no prior experience with the experimental task, were unaware of the goals of the study, and signed an informed consent form. The participants were each paid $20 CDN for their participation. Two additional participants served as right- and left-handed novice models. One author (H. R.) served as the expert model with both right and left hands. The Health Sciences Ethics Committee of the Université de Montréal approved this study.

Apparatus and task

Two apparatuses were used in this study. The first apparatus (hereafter called the right apparatus) is illustrated in Fig. 1 (top left panel). This apparatus consisted of a wooden base with three barriers (height: 11 cm, width: 8 cm) and a start button embedded in a final target (11 × 8 cm). The barriers were placed perpendicular to the wooden base at the beginning of each trial, closing a micro switch circuit. The micro switches were connected to a computer via the I/O port of an A-D converter (National Instruments), and a millisecond timer was used to record the total movement time (TMT) and the time required to complete each segment of the task (intermediate times; ITs). The frontal (a negative value indicates that the barrier is located to the left of the starting base) and the sagittal Cartesian coordinates of the first, second, and third barriers relative to the start button were −12.5 and 9 cm, −13.5 and 41.5 cm, and 0 and 29 cm, respectively. While sitting in front of the apparatus, a participant’s task was to initiate his or her movement from the starting position, trip the first, second, and third barriers successively, and then end their movement on the target. The participants were asked to perform each segment of the task in an IT of 300 ms, leading to a TMT of 1,200 ms. The second apparatus (hereafter called the left apparatus; see Fig. 1, top right panel) was the mirror image of the first one.
https://static-content.springer.com/image/art%3A10.1007%2Fs00221-011-2856-z/MediaObjects/221_2011_2856_Fig1_HTML.gif
Fig. 1

Top. View of the observer and of the model when using the left and right apparatuses. The participants’ task was to initiate their movement from the starting position, trip the first, second, and third barriers successively, and then end their movement on the target in a clockwise (left panel) or a counterclockwise (right panel) motion. Bottom. In all groups, the participants knew that they would perform the task using their right hand and right apparatus. During the acquisition phase, the KR was provided on a computer screen illustrating the time spent on each segment of the task (IT) and the TMT

In a pilot study, we determined the natural relative timing used by individuals who physically practiced the experimental task (Collier and Wright 1995) with their right, dominant hand. Specifically, three participants who did not take part in the present study were asked to complete the task in a movement time of 1,200 ms for 60 trials. Although we recorded both TMTs and ITs, the participants received feedback only on the TMT following each trial. Twenty practice trials allowed the participants to approach the goal TMT of 1,200 ms. Data from the remaining forty trials revealed stable relative timing both within and across participants. On average, the participants used a relative timing of 17.2, 29.0, 23.2, and 30.6% to complete the first, second, third, and fourth segments of the experimental task, respectively (within- and between-participant variability fluctuated between 1.0 and 2.4%; see also Blandin et al. 1999 for a similar observation). Therefore, the task required that participants modify the naturally emergent relative timing pattern for this task.

Experimental groups and procedure

The participants were randomly assigned to one of the six following groups: one control group, one physical practice group (PP), and four observation groups (left-handed model/first-person [L-1st]; left-handed model/third-person [L-3rd]; right-handed model/first-person [R-1st]; right-handed model/third-person [R-3rd]; for details see below and Table 1). After having been informed of the movement sequence to be performed, the participants completed the following four experimental phases: pre-test (PRT), acquisition (ACQ), and 10-min and 24-h retention tests (RET10 and RET24). The movement pattern, ITs and TMT, were illustrated on a poster located directly in front of the apparatus during all experimental phases.
Table 1

Groups and experimental phases

Groups

Phase

Pre-test

Acquisition

10-min Retention

24-h Retention

Physical practice

Perform 10 trials—No KR

Perform 40 physical practice trials + KR

Perform 10 trials—No KR

Perform 10 trials—No KR

Observers L-3rd

Perform 10 trials—No KR

Observe left-handed model 3rd person 40 trials + KR

Perform 10 trials—No KR

Perform 10 trials—No KR

Observers L-1st

Perform 10 trials—No KR

Observe left-handed model 1st person 40 trials + KR

Perform 10 trials—No KR

Perform 10 trials—No KR

Observers R-3rd

Perform 10 trials—No KR

Observe right-handed model 3rd person 40 trials + KR

Perform 10 trials—No KR

Perform 10 trials—No KR

Observers R-1st

Perform 10 trials—No KR

Observe right-handed model 1st person 40 trials + KR

Perform 10 trials—No KR

Perform 10 trials—No KR

Control

Perform 10 trials—No KR

No practice

Perform 10 trials—No KR

Perform 10 trials—No KR

In the PRT, a participant used his or her right hand to perform 10 trials using the right apparatus. No knowledge of the results (KR) was provided during this phase.

During ACQ, a participant in the physical practice group physically practiced the task with his or her right hand on the right apparatus for 40 trials. After each trial, the participant received the KR in milliseconds on both the TMT and ITs. The KR was presented on a computer screen and remained visible for 7 s (see bottom of Fig. 1). The participants in all observation groups watched a film of a model performing the experimental task for 40 trials. The video was presented on a 37-in. monitor (Sony Bravia KDL-37M3000) located directly in front of the participant. Following each trial, the KR concerning the performance of the model on both the TMT and ITs was presented on the monitor for 7 s. The observation groups differed by the type of model the participants watched (right-handed or left-handed model) and whether the film was presented using a first- or third-person perspective. The right-handed models were shown using the right apparatus, whereas the left-handed models were shown using the left apparatus. For all groups, the videos were closed captioned so that the observers could clearly see the apparatus and the model’s motion throughout the duration of a trial. The performance of the expert model was nearly perfect and was similar for both the right and left hands. On average (SD), when using his right hand, the expert spent 301 (6.6), 303 (7.2), 300 (6.7), and 299 (5.8) ms on the first, second, third, and fourth segment of the task, respectively (TMT = 1,202, SD = 10.1). When using his left hand, the expert spent 304 (7.5), 301 (5.9), 303 (7.1), and 298 (5.5) ms on segments 1–4, respectively (TMT = 1,205, SD = 11.7). The right- and left-handed novice models showed gradual and similar improvements during practice (see Fig. 2). We opted to use a mixed schedule of observation. Specifically, the participants watched a film showing 20 trials performed by a novice model who gradually improved his performance through practice and 20 trials performed by an expert model. Thus, observation provided a template of what needed to be done (expert model) and on how to correct one’s movement to be successful (novice observation). The model was alternated every five trials (e.g., novice: trials 1–5 and expert: trials 6–10). The participants were informed that they would observe both an expert and a novice model. Prior to each set of 5 trials, they were also reminded that they would observe the expert or novice model. All participants in the observation groups were informed at the beginning of the acquisition phase that they would have to perform the task using their right hand and the same apparatus as in the pre-test. The participants in the control group did not practice during this phase. Instead, they read a newspaper or magazine provided to them for the same duration as the observation or physical practice for the other groups (approximately 10 min).
https://static-content.springer.com/image/art%3A10.1007%2Fs00221-011-2856-z/MediaObjects/221_2011_2856_Fig2_HTML.gif
Fig. 2

Root mean square error on the intermediate times as a function of the number of trials for the right-handed (RH) and the left-handed (LH) novice models used in the present study

Ten minutes and 24 hours after the end of the acquisition phase, all participants performed retention tests similar to the pre-test described above.2 No KR was provided during these phases.

Data analysis

To determine the accuracy and consistency of the participants’ movements, we computed the absolute constant error (|CE|) and variable error (VE) of the total movement time, respectively. We opted to use |CE| instead of CE because, within all groups, approximately half of the participants undershot the target TMT, whereas the other half overshot it. For ITs, we computed a root mean square error (RMSE), which presents in a single score how much each participant deviated from the prescribed relative timing pattern
$$ {\text{RMSE}} = \sqrt {\mathop \sum \limits_{\text{Segment 1}}^{\text{Segment 4}} \left( {\frac{{({\text{IT}}i - 300)^{2} }}{4}} \right)} , $$
where ITi is the intermediate time of segment i for each trial.

The data of the three dependent variables were individually submitted to two analyses. In the first analysis, we determined whether the observation led to significant learning of the TMT and ITs of the task. The data were submitted to an ANOVA contrasting 6 groups (physical practice, control, observers R-1st, observers R-3rd, observers L-1st, and observers L-3rd) × 3 experimental phases (pre-test, retention 10 min, retention 24 h) × 2 blocks (trials 1–5, 6–10) using repeated measurements on the last two factors. Next, to determine whether some observation conditions resulted in better learning of the task, the data from the observation groups were submitted to an ANOVA contrasting the two handedness of the models (right vs. left) × 2 perspectives of observation (first- vs. third-person) × 2 experimental phases (retention 10 min, retention 24 h) × 2 blocks (trials 1–5, 6–10) using repeated measurements on the last two factors.

Before computing the different ANOVAs, three specific assumptions of the ANOVA were tested. The z scores of the skewness and kurtosis values were calculated to test the normality of the distribution (Tabachnick and Fidell 2007). To verify the homogeneity of the variances, Hartley’s Fmax test was used. Finally, the degrees of freedom were adjusted as suggested by Greenhouse and Geisser (1959) when Mauchly’s test of sphericity was significant. However, the original degrees of freedom were presented when the effects were found to be significant following the Greenhouse-Geisser correction. All significant effects are reported at P < 0.05 and were corrected for the number of comparisons (Bonferroni adjustment; Cardinal and Aitken 2006).

Results

Total movement time

|CE|. The results of the first analysis revealed significant main effects of group, F (5, 66) = 3.64, and phase, F (2, 132) = 31.97, and a significant group × phase interaction, F (10, 132) = 2.29. As illustrated in Fig. 3 (upper panel), the breakdown of the interaction revealed no significant difference between groups in PRT, F < 1; however, in both the 10-min, F (5, 66) = 8.54, and the 24-h retention tests, F (5, 66) = 10.10, the control group had a significantly larger |CE| than the five remaining groups, which did not differ significantly from one another. Figure 3 also illustrates that all experimental groups, but not the control group, exhibited a significantly reduced |CE| when comparing the PRT to the other two experimental phases. In addition, when comparing the 10-min to the 24-h retention test, we found a significant increase in |CE| for the physical practice group (P = 0.019) but not for the four observation groups (P > 0.90 for all groups).
https://static-content.springer.com/image/art%3A10.1007%2Fs00221-011-2856-z/MediaObjects/221_2011_2856_Fig3_HTML.gif
Fig. 3

Absolute constant error (upper panel) and variable error (middle panel) on the total movement time and root mean square error of the intermediate times (lower panel) as a function of experimental phases for the physical practice (PP), control, and four observation groups (L-3rd = Left-handed model/3rd person, L-1st = Left-handed model/1st person, R-3rd = Right-handed model/3rd person, R-1st = Right-handed model/1st person)

The results of the second analysis revealed no significant main effect or interaction. Thus, the four observation groups exhibited significant improvements in the accuracy of TMT, with no differences related to the conditions of observation.

VE. The results of the first analysis revealed significant main effects of phase, F (2, 132) = 41.48, and block, F (1, 66) = 5.06, and a significant phase × block interaction, F (2, 132) = 3.31. The breakdown of the interaction revealed a significantly larger VE in block 1 than in block 2 in the PRT (71 and 61 ms, respectively), whereas no significant between-block difference was found in both the 10-min (50 and 50 ms for blocks 1 and 2, respectively) and the 24-h retention phases (42 and 40 ms for blocks 1 and 2, respectively). Post hoc comparisons of the phase main effect revealed that VE significantly decreased from one experimental phase to the next (P < 0.01 for all phases).

The results of the second analysis revealed a significant main effect of phase, F (1, 44) = 12.21, and a significant phase × view interaction, F (1, 44) = 5.81. The breakdown of the interaction revealed a decrease in VE from the 10-min to the 24-h retention test when observing from the first-person perspective (49 and 36 ms, respectively) but not when observing from the third-person perspective (45 and 42 ms, respectively).

Intermediate times

The results of the first ANOVA computed on the RMSE revealed significant main effects of phase, F (2, 132) = 70.91, and group, F (1, 66) = 10.32, and significant phase × group, F (10, 132) = 4.26, and phase × block interactions, F (2, 132) = 4.61. The breakdown of the phase × group interaction revealed the following. As illustrated in Fig. 3 (bottom panel), there were no significant between-group differences in the PRT, F < 1. In both the 10-min and 24-h retention tests, the control group had a significantly larger RMSE than did the remaining five groups, which did not differ significantly from one another; F (5, 66) = 19.24 and 24.14, for the 10 min and 24 h tests, respectively. Figure 3 also illustrates that all experimental groups, but not the control group, exhibited significantly reduced RMSEs when comparing the PRT to the two retention tests. However, when comparing the 10-min to the 24-h retention test, we found a significant increase in the RMSE for the physical practice group (P = 0.005) but not for the observation groups (P > 0.14 for all groups). The breakdown of the phase × block interaction revealed that the participants had a significantly larger RMSE in block 1 than in block 2 of the PRT (123 and 117 ms, respectively), whereas no significant difference between blocks 1 and 2 was found in either the 10-min (81 and 82 ms, respectively) or the 24-h retention test (77 and 78 ms, respectively).

The results of the second analysis revealed significant main effects of phase, F (1, 44) = 9.26, and handedness, F (1, 44) = 13.10. Post hoc comparisons revealed that the participants showed decreased RMSEs when comparing the 10-min (77 ms) to the 24-h retention test (68 ms). More importantly, the participants had a significantly smaller RMSE when observing a right-handed model (63 ms) than they did when observing a left-handed model (82 ms; Fig. 4).
https://static-content.springer.com/image/art%3A10.1007%2Fs00221-011-2856-z/MediaObjects/221_2011_2856_Fig4_HTML.gif
Fig. 4

Effects of model’s handedness (Left-handed [LH] vs. Right-handed [RH]) and observation viewpoint (1st person vs. 3rd person)

Thus, all forms of observation used in the present study led to significant learning of intermediate times. However, participants learned more when observing a same-handed model than they did when observing an opposite-handed model.

Discussion

Observation results in learning of a motor skill (Blandin et al. 1999; Hayes et al. 2010; McCullagh and Weiss 2001; Shea et al. 2000; Wulf and Mornell 2008; Wulf et al. 2010). However, little is known concerning the type of model and conditions of observation that can optimize motor skill learning. In the present study, our goals were to determine whether a same-handed or an opposite-handed model would better promote learning of a complex spatiotemporal task and whether this effect would be mediated by the observer’s perspective.

The results of the present study are straightforward. Regardless of the models’ handedness or the observers’ perspective, results from the two retention tests revealed that all observation groups learned to complete their movements in the prescribed TMT at least as accurately as did the physical practice group and significantly better than the control group. These results were consistent with previous findings indicating that the TMT can be learned through observation (Blandin et al. 1999; Rohbanfard and Proteau 2011; Trempe et al. 2011). In addition, our results indicated that the observation permitted the participants to learn the TMT and ITs concomitantly (We return to this point below). Importantly, the physical practice group, unlike the observation groups, did not perform as well in the 24-h retention test as they did in the 10-min retention test. This is consistent with our previous study (Rohbanfard and Proteau 2011) and suggests that a 10-min retention test might reflect short-term performance effects more than learning effects (see Schmidt and Lee 2005) following the physical practice.

In addition, the results revealed that, for both same-handed and opposite-handed models, different observation perspectives (first- vs. third-person) resulted in no significant differences in the learning of the task (Fig. 4). This was consistent with the results of previous behavioral studies investigating the effects of the observation viewpoint in imitation (Ishikura and Inomata 1995; Sambrook 1998). For example, Ishikura and Inomata had participants observe a model performing a sequential movement task from a first-person or a third-person perspective. Their results showed that the first-person group significantly outperformed the third-person group on immediate recall tests. However, both groups performed equally well in a series of delayed retention tests completed 1 day, 1 week, or 5 months later. Ishikura’s and Inomata’s results suggest that although the first-person perspective resulted in better performance than did the third-person perspective, it resulted in similar long-term retention and, thus, learning of the task.

More notably, our results revealed that the observation of a same-handed model led to significantly better learning of the ITs than did observation of an opposite-handed model from both first- and third-person perspectives. For the first-person perspective, our results were consistent with behavioral studies (Blandin et al. 1999; Boutin et al. 2010; Gruetzmacher et al. 2011; Heyes and Foster 2002; Osman et al. 2005) and brain imaging studies, which demonstrated that observation results in contralateral activation of brain regions that were solicited during physical practice (Aziz-Zadeh et al. 2002, 2006; Maeda et al. 2002; Pilgramm et al. 2010; Shmuelof and Zohary 2006, 2008). This may explain why right-handed participants were more accurate when they observed a same-handed model than when they observed an opposite-handed model from a first-person perspective. Furthermore, our results supported Michel and Harkins’s study (1985), in which right- and left-handed participants observed a knot-tying task performed by a right-handed or left-handed model from a first-person perspective. Their results showed that the participants learned significantly faster from a same-handed model than from an opposite-handed model. Together, these results suggest that the observation of an opposite-handed model requires some additional processing of information (e.g., transformation of visual information), which could limit or slow the learning of a new motor skill.

When observing from a third-person perspective, our results were inconsistent with the pattern of activation revealed in TMS (Alaerts et al. 2009; Hesse et al. 2009) and brain imaging studies (Hesse et al. 2009; Kilner et al. 2009; Shmuelof and Zohary 2008). Specifically, recent studies have indicated a larger activation of the ipsilateral aSPL (Shmuelof and Zohary 2008) and a larger excitability of the ipsilateral primary motor cortex (Alaerts et al. 2009; Hesse et al. 2009) for movements observed from a third-person perspective. In addition, several studies have reported that young participants had a tendency to imitate the actions of others in a mirror-imaged manner when observed from a third-person perspective (Avikainen et al. 1999; Chiavarino et al. 2007; Iacoboni et al. 2001). Together, these results suggest that the observation of a left-handed model from a third-person perspective should have facilitated learning in the observers. A possible cause for these divergent findings is that the participants in the above studies observed familiar gestures or imitated familiar upper/lower limb movements immediately after having observed them, with no specific spatial and/or temporal constraints. This is in stark contrast from learning a complex spatiotemporal pattern, as in the present study and most sport-like activities. It is possible that the intent of learning in conjunction with the stringent temporal constraints of our task might require the AON to transform the visual information depicted from a third-person perspective to fit with how the observer must perform the task. This transformation is likely easier to do or more natural when a right-handed individual observes a right-handed model.

This is consistent with the predictive coding framework of the simulation of observed actions advocated by Kilner et al. (2007a, b), and Neal and Kilner (2010). This model, like others (Miall 2003; Wolpert et al. 2003), proposes that observers simulate an action they see using a generative or forward model of how they would perform the same action. However, in two recent papers, evidence was provided that the observer simulates the observed action as if performed by his or her dominant hand. Specifically, Neal and Kilner (2010) had right-handed participants observe video clips showing a right- or left-hand reach-and-grasp movement from either a first- or third-person perspective. In the “natural” condition, the videos were not altered in any significant way, whereas in the “manipulated” condition, the videos were reflected about the vertical midpoint. Thus, in the latter condition, a right-hand movement filmed from a first-person perspective was shown as a left-handed movement from a third-person perspective, and so on. The participants were informed that half of the videos they were about to see had been altered using video software. Following each video presentation, they had to indicate whether the video had been manipulated. Although the participants reported that they were guessing, videos of natural right-hand movements that were viewed from a first-person or third-person perspective were reported as natural significantly more often than left-handed videos that had been manipulated to look as right-handed movements. No such difference was found between the natural and manipulated left-hand movements. This suggests that observers always simulate what they observed using a generative or forward model of how they would perform the same action using their dominant right-hand. This model is used to predict what should be observed. The differences between the simulated and the observed movements would be greater when observing a left- than a right-handed model, which would explain why participants considered left-hand movements as manipulated more often than right-hand movements. More recently, Press et al. (2011) used magnetoencephalography to record the cortical activity of right-handed participants performing sinusoidal up and down movements with their left or right arm or observing video clips of individuals performing the same movements from a third-person perspective. Their results revealed that the observation of right- and left-hand actions elicited changes in the left hemisphere sensorimotor activation across time (Broadman area 4), according to the phase of the observed movement. These changes would be expected if one was executing the observed movement, indicating that observation activated the motor program required for its execution with the observer’s dominant right hand. Our results add to these previous findings by showing that the more accurate simulations hypothesized when a right-hand observer watched a right-hand movement also leads to a more accurate movement programming and execution, which would save practice time.

Finally, the benefit of observing a same-handed model for learning a complex motor skill might not apply to left-handed participants. Because left-handed people represent approximately 10% of the population (Ida and Mandal 2003), they show a tendency to accommodate to the right-sided world (Coren and Halpern 1991). This accommodation might result in left-handed people learning as quickly, and perhaps even more quickly, from right-handed rather than left-handed models. Future research is needed to address this issue.

An alternative interpretation of our findings could be that the observers coded the movement sequence in visual-spatial coordinates rather than in a motor coordinates. Because the apparatus used by the opposite-handed model was a mirror image of that used by the L-1st and L-3rd groups, it could explain why observation of a left-handed model did not favor learning as much as observation of right-handed model. Partial support for this interpretation comes from two recent studies (Boutin et al. 2010; Gruetzmacher et al. 2011). In Boutin et al., right-handed participants observed a right-handed model moving a one-degree-of-freedom lever back and forth to reach a series of target locations as fast as possible. The sequence of target presentations was predetermined and repeated 140 times. Following observation, the participants performed the same task as the models with their right arm. They also performed two transfer tests using their left arm. For these tests, the sequence of target presentation remained the same as during observation, and thus, the visuospatial codes were maintained in transfer. In the second test, the sequence of target presentation was the mirror image of what had been observed, which kept the motor coordinates (pattern of arm flexion and extension) unchanged between the observation and test. The results revealed that the observers performed the task significantly faster when the visuospatial rather than the motor codes were maintained in transfer. This finding was replicated by Gruetzmacher et al., who used a relatively simple spatialtemporal movement sequence. Thus, it could be that an observer learns more from a same-handed model than from an opposite-handed model because he or she codes the observed movement pattern in visuospatial coordinates, which are maintained when observing a same-handed model. However, in both Boutin et al. and Gruetzmacher et al., the observers needed to learn the location of the targets, which could have made the visuospatial information more important than in the present study in which the target locations were known and the challenge was to learn a new imposed relative timing pattern. Future studies should address this question.

Concomitant learning of TMT and IT

In the present study, we found that observation permitted the participants to learn the TMT and ITs concomitantly. This finding differs from previous observations from our laboratory (Blandin et al. 1999). Blandin et al. used a four-segment timing task similar to that used in the present study; these authors reported that observers first learned to complete their movements in the prescribed TMT and then learned the relative timing pattern. There are two procedural differences between Blandin et al. and the present study that could explain these divergent results. First, the imposed TMT was longer in the present study than in Blandin et al. (1,200 ms vs. 900 ms). It could be that a more stringent TMT encouraged participants to learn to fit their movements within the appropriate time frame and then to proceed to make adjustments to relative timing. Second, we used a combination of expert and novice models (i.e., mixed model), whereas Blandin et al. used either an expert or a novice model. Thus, it could be that the observation of a mixed model, which has been shown to be more efficient for learning than the observation of either a novice or an expert model (Rohbanfard and Proteau 2011), permitted the observers to learn TMT and IT concomitantly in the present study.

Conclusion

Observation, regardless of the model’s handedness or the observer’s perspective, promoted learning of a new motor skill. However, better learning of the temporal sequencing of the task occurred when the right-handed observer viewed a right-handed model from either a first- or a third-person perspective. Thus, the AON is more sensitive to the model’s handedness than to the observer’s viewpoint because the AON is linked to or involves sensorimotor regions of the brain that simulate motor programming. Our results were consistent with recent findings indicating that this putative simulation of the observed movements occurs in the left hemisphere of right-handed observers regardless of the model’s handedness (Press et al. 2011). Our results also suggest that the observation for immediately reproducing the observed actions (i.e., imitation), as compared with observation for learning a new motor skill, might be based on different processes.

Footnotes
1

It should be noted that our study differs from previous work addressing the question of whether observational learning is effector dependent or independent (Boutin, Fries, Panzer, Blandin & Shea, 2010; Gruetzmacher, Panzer, Blandin, & Shea, 2011, Osman, Bird, & Heyes, 2005). In the present study, we want to determine whether a right-handed observer performing the task with his or her right hand (which is usually the case in most real-life situations, such as learning how to bowl, play golf, and tennis) can learn better from a right-handed or left-handed model depending on the observer’s perspective and not whether this learning can be transferred so that the observers can perform the task with both right and left hands.

 
2

It could be argued that because participants in the L-1st and L-3rd groups observed a left-handed model but performed the task using their right hand, a more appropriate label for these tests would be “transfer” rather than “retention”. However, because a transfer test evaluates the performance of the participants at a task that was different from what they specifically wanted to learn, we opted to use the “retention” label.

 

Acknowledgments

This work was supported by a Discovery grant (L.P.) provided by the Natural Sciences and Engineering Research Council of Canada.

Copyright information

© Springer-Verlag 2011