An eye-tracking study of how coach's nonverbal communication affects memorization of basketball tactical scenes

Abstract

Numerous studies have been conducted to examine the effects of an instructor's pointing gestures on learning performance across a wide range of academic domains. There are, however, few clear instructional guidelines for enhancing learning from this visual guidance. This study tested the combined effects of the coach's pointing gesture and gaze guidance on attention and recall performance in field-based sports. This study used a 3 (control vs. gesture vs. gesture/gaze) × 2 (novice vs. expert) between-subjects factorial design. Results showed a significant interaction effect between expertise level and experimental conditions on visual attention and memorization performance. Novice players scored higher after studying tactical instructions with pointing gestures and gaze guidance than players who studied the same instructions with pointing gestures, who in turn performed better than players who received instructions without cues. Expert players showed the same level of performance when they received the instructions without cues and with pointing gestures. However, they showed lower performance (notably in the visual attention and the mental effort measures) when they received the video with pointing gestures and gaze guidance compared to the instructions without gestures and with pointing gestures. These results suggest that the effectiveness of nonverbal visual guidance changes depending on the players’ level of expertise.

Keywords

Cognitive load expertise reversal effect gaze guidance mental effort pointing gestures visual attention

Introduction

In the context of team performance, contact between coaches and players is mainly established through verbal (i.e., speech) and nonverbal communication (i.e., hand gestures, facial expressions, eye gaze, human-like movements).^1,2 Therefore, effective communication is important to help players to learn the strategies and tactics presented in practice.^3–6 During tactical training sessions, coaches spend most of their time giving information, instructions, or commands to athletes.⁷ They also use static diagrams drawn on a tactical board to convey essential information about the game through verbal description.⁸ This tool can be used to explain to players the different interactive elements and the transitions necessary to build a hierarchically organized mental representation of the game situation.^9–11 However, the physical presence of the coaches could lead to a splitting of attention between them and the game elements they refer to, such that the players’ attention is automatically drawn to the coaches’ face, creating an additional cognitive load (i.e., split-attention effect).^12,13 These attentional difficulties are supposed to be compensated by the presence of the coach's pointing gestures that draw players’ attention toward the relevant information of the game.³ However, as yet, there is little empirical evidence about whether the incorporation of other types of visual guidance (such as gaze cues) affects the effectiveness of pointing gestures, and thereby influences players’ memorization of game elements.

Pointing gestures are defined as extending the hand and index finger to direct attention to a specific object in the physical environment.^14,15 Research conducted in the non-sporting domain has shown that speech instruction accompanied by pointing gestures can convey information that cannot be conveyed by speech only and leads to improved learning performance.^16,17 In an influential study, Koumoutsakis et al.¹⁸ compared the role of the instructor's pointing gestures to no gesture in learning a mathematics lesson. They found that the instructor's pointing gestures improved students’ learning and transfer performance compared to students in the no-gesture condition. Likewise, Valenzeno et al.¹⁶ investigated the role of pointing gestures in learning a symmetrical lesson. The results showed that children who received instructions with pointing gestures performed much better than children who received instructions without pointing gestures.

In addition to the positive effect of pointing gestures on learning, studies of eye-tracking have shown that these cues are effective in influencing learners’ visual attention.^19–21 A potential explanation is that an instructor's pointing gestures could help avoid the negative effects of split attention by directing students’ attention to the appropriate relevant elements in the learning material.¹⁹ This can reduce the need for visual search and prevent interference of irrelevant elements, freeing up resources that can be devoted to learning. In a study by Pi et al.,²⁰ eye-tracking was used to examine how much time they spent watching the task. Results showed that students who observed the instructor's gesture paid more attention to the relevant task area she was referring to than those who did not receive any cues.

Although previous studies have shown that pointing gestures positively affect learning and visual attention, others have reported that these cues did not improve learners’ performance.^22–24 For example, Yeo et al.²⁵ studied the role of pointing gestures in learning mathematics instructions. They found that the use of pointing gestures does not always facilitate learners’ understanding and that these cues can be unhelpful and even harmful. Similar results were obtained by,²⁶ in which the condition with pointing gestures led to similar learning outcomes as the condition without cues for adult and child learners. A plausible explanation for the negative results could be that individuals have a tendency to maintain their focus on the instructor, even in situations where a pointing gesture is being utilized. In an eye-tracking study, Gullberg and Holmqvist²⁷ found that attention was mainly focused on the speaker's face (over 90% of viewing time), with only a minority of attention given to the gestures produced during a face-to-face communication. Despite these contradictory observations, it is, therefore, crucial to investigate the possible effects of pointing gestures as well as the moderating variables that may alter these effects.

Research on nonverbal communication has shown that gestures accompanying speech are generally processed in parallel with gaze cues.^28–29 In the literature, researchers distinguish three types of instructor's gaze: direct gaze, gaze guidance, and fixed gaze.^30–32 Several studies have shown that gaze guidance is the most powerful attentional gaze cue and has a positive effect on learning (gaze guidance principle³³;), as people tend to automatically follow other people's gaze to look at what they are looking at.^28,34 However, despite the positive effects of pointing gestures and gaze guidance, little attention has been paid to the influence of combining these cues in real instructional settings and whether they are superior to pointing gestures alone. In a recent study of a biology lesson by Pi et al.,²⁰ researchers found that learners in the pointing gestures condition and in the pointing gestures and gaze guidance condition showed higher learning performance and a better visual search pattern than in the no cues condition.

The effect of including pointing gestures and gaze cues in learning materials is explained by cognitive load theory (CLT^35,36;), which posits that these cues focus learners’ attention on the most relevant information, reduce visual search due to the split-attention effect, and require fewer visuo-spatial resources to perform eye movements.^37–40 In addition to CLT, research based on the cognitive theory of multimedia learning (CTML⁴¹;) has shown that the inclusion of attention-guiding cues (e.g., pointing gestures and gaze) can improve the effectiveness of instructional materials.^42–44 This phenomenon is referred to as the signaling principle,⁴⁵ which states that students learn better when instructional materials contain cues that highlight the relevant elements of the material.⁴⁶ This principle is also consistent with the “attention-guiding principle”, as most cues highlight information in order to attract learners’ attention and enhance learning.⁴⁷ Furthermore, according to social agency theory, a recent extension of CTML, cues promote deeper learning by providing a richer cognitive schema that encourages learners to activate their social response and engage in a deeper cognitive process of organizing and integrating the material with their prior knowledge.⁴⁵

Interestingly, research on the embodied cognition perspective states that cognitive processes are rooted in the interactions of the human body with the physical environment.^48–50 According to this perspective, observing the instructor's presentation of information in the form of speech, pointing gestures, and gaze guidance could lead to a higher quality of cognitive schema than in the form of speech alone, as gestures and gaze are more than simple visual cues as they are processed by the motor system. Moreover, research on social cognition has shown that gestures and gaze can elicit joint attention, cooperation, and shared intentions.⁵¹ Therefore, they can serve as social motivators for learners to focus their attention on the instructor's intended meaning. Although embodied cognition and social cognition are distinct areas of research, they are interconnected through the role of the mirror neuron system.^52,53 The mirror neuron system, which has been extensively studied, plays an important role in understanding the actions of others and is responsible for our ability to learn by observing and imitating others.^54,55 This suggests that the integration of sensorimotor processes is not only necessary for physical movements and social interaction, but also for higher-level cognitive processes such as understanding and predicting the thoughts and intentions of others.⁵⁶

Although previous studies have shown the potential benefit of the instructor's pointing gesture and gaze as instructional support for novice learners, they have not examined this effect with expert learners.^16,21,22,43 Several studies have shown that instructional techniques that are effective for novice learners may be ineffective or even harmful for expert learners.^10,57 This effect is known as the expertise reversal effect and states that additional tools hinder the learning process because they present redundant content that is already known.⁵⁸ In other words, presenting the same information in multiple modalities is redundant for expert learners and forces them to process irrelevant material incrementally.⁵⁹ In a sports-related field, Khacharem⁶⁰ studied the effect of adding a red circle to a static football diagram. They found that novice players performed better in the condition with the cues than in the condition without cues, whereas this strategy was ineffective for expert players who did not benefit from the visual cues. Consequently, visual cues and prior knowledge may overlap and cause the expertise reversal effect. The presentation of extraneous cues (e.g., pointing gestures and eye gaze) could be redundant for experienced learners, as their prior knowledge helps to select the appropriate information from the visual display.^61,62 According to the redundancy principle,⁶³ Richter et al.⁶⁴ have shown that experienced learners are not able to ignore redundant information. Therefore, this could lead to an additional cognitive load on working memory resources that might hinder expert learning.

The present study examined the effects of coaches’ pointing gestures on the memorization of basketball offensive tactics by measuring learners’ recall performance, mental effort, and visual attention patterns. Participants watched a video lecture in which a coach verbally explained the development of an offensive play scene in one of three conditions: the coach looked directly at the camera without pointing gestures or gaze cues (control condition), looked directly at the camera and made pointing gestures toward the relevant elements of the diagram (gesture condition), or made pointing gestures accompanied by gaze shifts toward the relevant elements of the play (gesture/gaze condition). Based on the above literature, the study was designed to test the following hypotheses:

H1: Regarding the recall score, we predicted an interaction between conditions and expertise. Specifically, novices in the control condition will recall fewer game elements than novices in the gesture condition (H1a) and the novices in the gesture/gaze condition (H1b); furthermore, novices in the gesture condition will recall fewer game elements than novices in the gesture/gaze condition (H1c). However, experts in the control condition will recall more elements of play than experts in the gesture condition (H1d) and the experts in the gesture/gaze condition (H1e); furthermore, experts in the gesture condition will recall more game elements than experts in the gesture/gaze condition (H1f).

H2: Regarding the recall time, we predicted an interaction between conditions and expertise. Specifically, novices in the control condition will reconstruct the game system less quickly than novices in the gesture condition (H2a) and novices in the gesture/gaze condition (H2b); furthermore, novices in the gesture condition will reconstruct the game system less quickly than novices in the gesture/gaze condition (H2c). However, experts in the control condition will reconstruct the game system faster than experts in the gesture condition (H2d) and the experts in the gesture/gaze condition (H2e); furthermore, experts in the gesture condition will reconstruct the game system faster than experts in the gesture/gaze condition (H2f).

H3: Regarding the mental effort, we predicted an interaction between conditions and expertise. Specifically, novices in the control condition will report higher mental effort than novices in the gesture condition (H3a) and novices in the gesture/gaze condition (H3b); furthermore, novices in the gesture condition will report higher mental effort than novices in the gesture/gaze condition (H3c). However, experts in the control condition will report lower mental effort than experts in the gesture condition (H3d) and experts in the gesture/gaze condition (H3e); furthermore, experts in the gesture condition will report lower mental effort than experts in the gesture/gaze condition (H3f).

H4: Regarding the visual attention, we predicted an interaction between conditions and expertise. Specifically, novices in the control condition will tend to fixate on the relevant diagrams for a shorter duration and with a smaller number of fixations, while they will show a longer duration and greater fixations on the coach and switch more frequently between the coach and the relevant diagrams than novices in the gesture condition (H4a) and novices in the gesture/gaze condition (H4b); furthermore, novices in the gesture condition will tend to fixate on the relevant diagrams for a shorter duration and with a smaller number of fixations, while they will show a longer duration and greater fixations on the coach and switch more frequently between the coach and the relevant diagrams than novices in the gesture/gaze condition (H4c). However, experts in the control condition will tend to fixate on the relevant diagrams for a longer duration and with a larger number of fixations, while they will show a shorter duration and fewer fixations on the coach and switch less frequently between the coach and the relevant diagrams than experts in the gesture condition (H4d) and experts in the gesture/gaze condition (H4e); furthermore, experts in the gesture condition will tend to fixate on the relevant diagrams for a longer duration and with a larger number of fixations, while they will show a shorter duration and fewer fixations on the coach and switch less frequently between the coach and the relevant diagrams than experts in the gesture/gaze condition (H4f).

Method

Participants

To estimate the sample size required for this experiment, an a priori power analysis was performed using G*Power 3.1.9.7.⁶⁵ Analysis of variance (ANOVA) was considered: Fixed effects, special effects, main effects, and interactions, an effect size of 0.40 (F = 3.09), an alpha of 0.05, a power of 0.95. This power analysis indicated that a total sample of 100 participants would be required. Therefore, a total of 144 participants with two different skill levels were included in this study. Seventy-four experienced basketball players who had been playing for 10 to 15 years (M_age = 24.89 years, SD = 2.05), and seventy-four novice student-athletes (M_age = 22.28 years, SD = 1.79) who had no specific basketball experience. They were physical education students and occasionally played basketball with friends or during physical education classes. The experts were recruited from the 2nd and 3rd French basketball leagues and had trained regularly over the past decade. They were selected based on factors such as the number of training sessions per week (at least 7 sessions), the average number of hours per week (M = 9.42, SD = 1.33), and the fact that they regularly competed at least three times per month. All participants were native French speakers and indicated that they had no visual problems. Approval for this experiment was obtained from the local ethics committee before testing began, and all rights of participants were respected.

Participants were randomly assigned to one of three experimental conditions: (i) control condition (24 novices and 24 experts); (ii) gesture condition (24 novices and 24 experts); (iii) gesture/gaze condition (24 novices and 24 experts).

Apparatus

The apparatus consisted of a Lenovo laptop (L340 Gaming) placed at a distance of 50 cm from the participant. Stimuli were displayed on a 15.6-inch screen with a viewing angle of 45°. The video-based modeling examples were recorded using a Canon 5d Mark III camera and edited using Adobe Premium Pro 22.1.1 and Adobe Audition 2.0. To record eye movements, we used the Tobii Pro Glasses 2, a wireless eye-tracking system that employs a camera and near-infrared diodes that record eye movements at a sampling rate of 50 Hz. Participants completed the calibration task while looking at a calibration card for a few seconds. Finally, Tobii Pro Lab software version 1.171 was used to analyze the eye movement data.

Materials

Three conditions of a video lecture were created. All conditions were accompanied by the same auditory verbal explanation of the development of an offensive basketball system and lasted 82 s. The game system was initially developed by two professional basketball coaches who have sufficient experience in basketball (M_age = 36 years). The tactical combination consists of 5 players performing 9 actions, starting with a pass and ending with a shot. Each video presentation contained the same tactical combination, and the same coach gave the instructions while standing to the left of the game system. In the control condition, the coach stood still (without making hand gestures or gaze shifts) and looked at the camera while giving verbal descriptions. In the gesture condition, the coach pointed to the corresponding elements of the game system and looked directly at the camera without making gaze shifts. Finally, in the gesture/gaze condition, the coach pointed and looked at the same relevant element mentioned in the speech. Figure 1 shows a screenshot of each experimental condition.

Dependant variables

Visual attention

To investigate the effects of pointing gestures and gaze guidance on the player's visual attention, an area of interest (AOI) was the relevant diagrams of play according to which the coach was talk about and another AOI according to the coach's entire body. For each participant, eye-tracking was used to calculate the percentage of total fixation duration and the number of fixations on the relevant diagrams and the coach, and finally the saccades count between the two AOIs (Figure 2).

Recall performance

Participants were asked to recreate the tactical scene action by action by positioning the game elements (ball and/or player) in their correct locations. They were specifically instructed to use conventional arrows commonly used in basketball to present the evolution of the game system. A solid arrow for example, was used to represent player movement, while dashed arrows were used to represent a pass. An independent rater scored the total number of right and wrong responses given by each participant in their description of the system. Each right response received one point, while wrong answers were assigned zero points. Consequently, the score could range from 0 to 19.

Mental effort

The questionnaire developed by Paas⁶⁶ was used to measure participants’ perceived cognitive load during learning. The statement was “How much mental effort did you invest to learn the game system?”. Participants were required to evaluate their mental effort on a scale ranging from very, very low mental effort (1) to very, very high mental effort (9). This scale is recognized as an effective and noninvasive measure of cognitive load on working memory.^67,68

Procedure

The experience was conducted in the university laboratory (30 min). Participants were tested individually, they filled out the demographic questionnaire, the consent form, and the prior knowledge test. The procedure was conducted in two phases: a study phase and a test phase.

Study phase

After putting on the glasses, each participant was asked to look at the center of a calibration card to ensure that the eye-tracking was automatically calibrated. Then, one of the three video conditions was randomly projected once on the screen, and participants were asked to memorize the evolution of the game system as quickly and accurately as possible.

Test phase

Immediately after the visualization, the eye-tracker recording was stopped and the glasses were removed. The participant was asked to perform two tasks. (i) Rate the perceived mental effort while studying the game system. (ii) Reconstruct as quickly and as precisely as possible the evolution of the game system (position of the ball carrier, position of the players, and their movements) by drawing it on a paper containing three empty basketball courts (see Figure 3).

Data analyses

To ensure that the assumptions necessary for parametric tests were not violated, both Levene's test for homoscedasticity and Kolmogorov-Smirnov's test for normality of distribution were employed. To analyze the data, we performed a between-subjects analysis of variance (ANOVA) with a 3 (condition: control, gesture, and gesture/gaze) × 2 (expertise: novice vs. expert) factorial design. For each dependent variable, participants’ means and standard deviations were measured. The dependent variables were recall score (0–19), recall time (sec), mental effort invested in studying the game system (0–9), total fixation duration on the diagrams and the coach (%), number of fixations (count) on the diagrams and on the coach, and saccades count between the diagrams and the coach. Planned contrasts of ANOVA were conducted to test the specific effect of each condition. Planned contrasts were more appropriate than Post Hoc analyses (multiple comparisons) when specific hypotheses, which is in our case, are formulated.⁶⁹ Effect sizes are expressed as partial eta squared ( $η_{p}^{2}$ ), with values of 0.01, 0.06, and 0.14 and Cohen's d values of 0.20, 0.50, and 0.80 representing small, moderate, and large effect sizes, respectively. All planned contrasts were orthogonal and the effects were determined to be statistically significant at a significance level of 0.05 (α). Descriptive statistic for all dependant variables were presented in Table 1.

Table 1.

Means (and standard deviations) for the recall accuracy, recall time, mental effort, and visual attention measures of novice and expert players under different experimental conditions.

	Control	Gesture	Gesture/gaze
Novice
Recall accuracy	7.88 (2.33)	11.13 (2.72)	13.13 (2.17)
Recall time (s)	144.38 (62)	138.58 (49.32)	148.58 (48.53)
Mental effort	7.29 (1.40)	4.96 (1.68)	5.67 (1.31)
Total fixation duration on relevant diagrams (%)	54.77 (13.12)	70.10 (8.68)	79.57 (6.92)
Number of fixations on relevant diagrams	17.79 (3.78)	32.96 (7.95)	40.42 (5.68)
Total fixation duration on coach (%)	11.79 (5.15)	6.84 (2.19)	3.63 (2.56)
Number of fixations on coach	7.46 (3.88)	2.92 (2.86)	3.29 (2.61)
Saccades count between the relevant diagrams and the coach	17.79 (5.28)	12.71 (3.92)	14.17 (3.51)
Expert
Recall accuracy	15.58 (1.69)	15.38 (2.16)	14.58 (2.32)
Recall time (s)	86.46 (18.69)	89.17 (20.21)	88.75 (17.70)
Mental effort	3.21 (1.31)	3.79 (1.22)	5.38 (1.21)
Total fixation duration on relevant diagrams (%)	68.39 (6.43)	62.83 (6.23)	63.34 (16.28)
Number of fixations on relevant diagrams	17.04 (6.64)	15.25 (5.80)	14.88 (6.22)
Total fixation duration on coach (%)	5.11 (3.29)	5.09 (2.72)	8.66 (5.15)
Number of fixations on coach	3.25 (2.83)	4.29 (2.84)	4.38 (3.21)
Saccades count between the relevant diagrams and the coach	13.21 (3.97)	13.67 (2.63)	17.75 (2.91)

Results

Did the “gesture” or “gesture/gaze” conditions and the expertise level affect players’ recall accuracy during the test phase?

For the recall accuracy, ANOVA revealed a significant main effect for expertise, F(1, 138) = 141.712, p < .001, $η_{p}^{2}$ = 0.507, a significant main effect for condition, F(2, 138) = 11.326, p < .001, $η_{p}^{2}$ = 0.141 and a significant expertise × condition interaction, F(2, 138) = 23.152, p < .001, $η_{p}^{2}$ = 0.251. The results of planned contrasts showed that novices in the control condition achieved lower recall score than novices in the gesture condition (t ₍₁₃₈₎ = −5.015, p < .001, d = 1.35) and the gesture/gaze condition (t ₍₁₃₈₎ = −8.114, p < .001, d = 2.34). In addition, novice players in the gesture/gaze condition achieved higher recall score than novices in the gesture condition (t (₁₃₈) = 3.047, p = .003, d = 0.72). On the other side, the results of planned contrasts showed no significant differences for experts between the control condition and the gesture condition (t (₁₃₈) = 0.321, p = .748, d = 0.10) as well as between the control condition and the gesture/gaze condition (t (₁₃₈) = 1.355, p = .178, d = 0.49). Moreover, there were no significant differences between the gesture condition and the gesture/gaze condition (t (₁₃₈) = 1.037, p = .301, d = 0.36) (Figure 4).

Did the “gesture” or “gesture/gaze” conditions and the expertise level affect players’ recall time during the test phase?

For the recall time, ANOVA revealed a significant main effect for expertise, F(1, 138) = 69.13, p < .001, $η_{p}^{2}$ = 0.33, a non-significant main effect for condition, F(2, 138) = 0.178, p = .837, $η_{p}^{2}$ = 0.001 and a non-significant expertise × condition interaction, F(2, 138) = 0.228, p = 0.796, $η_{p}^{2}$ = 0.003. As shown in Figure 5, the analysis revealed that novices took more time to reconstruct the game system than experts (p < .001, d = 1.40). There were no significant differences in recall time between novices or experts across all conditions (p > .05).

Did the “gesture” or “gesture/gaze” conditions and the expertise level affect players’ mental effort during the study phase?

For the mental effort, the ANOVA revealed a significant main effect for expertise, F(1, 138) = 66.014, p < .001, $η_{p}^{2}$ = 0.324, a significant main effect for condition, F(2, 138) = 9.252, p < .001, $η_{p}^{2}$ = 0.118 and a significant expertise × condition interaction, F(2, 138) = 25.418, p < .001, $η_{p}^{2}$ = 0.269. The results of planned contrast showed that novices in the control condition invested more mental effort than novices in the gesture condition (t ₍₁₃₈₎ = −5.917, p < .001, d = 1.51) and novices in the gesture/gaze condition (t ₍₁₃₈₎ = −4.334, p < .001, d = 1.19). However, there were no significant differences for novices between the gesture condition and the gesture/gaze condition (t ₍₁₃₈₎ = 1.644, p = .104, d = 0.47). On the other side, the results of planned contrast showed no significant differences for experts between the control condition the gesture condition (t ₍₁₃₈₎ = 1.479, p = .141, d = 0.46). However, the results of planned contrast showed that experts in the gesture/gaze invested more mental effort than experts in the control condition (t ₍₁₃₈₎ = −5.586, p < .001, d = 1.72) and experts in the gesture condition (t ₍₁₃₈₎ = −4.122, p < .001, d = 1.31) (Figure 6).

Did the “gesture” or “gesture/gaze” conditions and the expertise level affect players’ visual attention during the study phase?

For the total fixation duration on the relevant diagrams, the analysis showed a non-significant main effect for expertise, F(1, 138) = 3.652, p = .058, $η_{p}^{2}$ = 0.026, a significant main effect for condition, F(2, 138) = 10.952, p < .001, $η_{p}^{2}$ = 0.137 and a significant expertise × condition interaction, F(2, 138) = 26.370, p < .001, $η_{p}^{2}$ = 0.277. The results of planned contrast showed that novices in the control condition spent less fixation time on the relevant diagrams than novices in the gesture condition (t ₍₁₃₈₎ = −5.088, p < .001, d = 1.38) and novices in the gesture/gaze condition (t ₍₁₃₈₎ = −8.074, p < .001, d = 2.36). Moreover, novices in the gesture/gaze condition spent more time fixating on the relevant diagrams than novices in the gesture condition (t ₍₁₃₈₎ = 2.934, p = .004, d = 1.21). On the other side, the results of planned contrast showed no significant differences for experts between the control condition and the gesture condition (t ₍₁₃₈₎ = 1.846, p = .067, d = 0.88), as well as between the control condition and the gesture/gaze condition (t ₍₁₃₈₎ = 1.636, p = .104, d = 0.41). Moreover, there were no significant differences between the gesture/gaze condition and the gesture condition (t ₍₁₃₈₎ = −0.190, p = .850, d = 0.04) (Table 1).

For the number of fixations on the relevant diagrams, the ANOVA showed a significant main effect for expertise, F(1, 138) = 205.481, p < .001, $η_{p}^{2}$ = 0.598, a significant main effect for condition, F(2, 138) = 34.368, p < .001, $η_{p}^{2}$ = 0.332 and a significant expertise × condition interaction, F(2, 138) = 51.135, p < .001, $η_{p}^{2}$ = 0.426. The results of planned contrast showed that novices in the control condition paid less attention to the relevant diagrams than novices in the gesture condition (t ₍₁₃₈₎ = −8.342, p < .001, d = 2.44) and novices in the gesture/gaze condition (t ₍₁₃₈₎ = −12.162, p < .001, d = 4.69). Moreover, novices in the gesture/gaze condition paid greater attention to the relevant diagrams than novices in the gesture condition (t ₍₁₃₈₎ = 3.735, p < .001, d = 1.08). However, the results of planned contrast showed no significant differences for experts between the control condition and the gesture condition (t ₍₁₃₈₎ = 0.985, p = .326, d = 0.29), as well as between the control condition and the gesture/gaze condition (t ₍₁₃₈₎ = 1.348, p = .180, d = 0.33). Moreover, there were no significant differences between the gesture/gaze condition and the gesture condition (t ₍₁₃₈₎ = −0.373, p = .710, d = 0.06) (Table 1).

For the total fixation duration on the coach, the ANOVA revealed a non-significant main effect for expertise, F(1, 138) = 3.358, p = .069, $η_{p}^{2}$ = 0.024, a significant main effect for condition, F(2, 138) = 6.700, p = .002, $η_{p}^{2}$ = 0.089 and a significant expertise × condition interaction, F(2, 138) = 30.127, p < .001, $η_{p}^{2}$ = 0.304. The results of planned contrast showed that novices in the control condition spent more time fixating on the coach than novices in the gesture condition (t ₍₁₃₈₎ = −4.513, p < .001, d = 1.25) and novices in the gesture/gaze condition (t ₍₁₃₈₎ = −7.060, p < .001, d = 2.01). Moreover, novice players in the gesture/gaze condition spent less time fixating on the coach than novices in the gesture condition (t ₍₁₃₈₎ = 2.502, p = .014, d = 1.35). On the other hand, the results of planned contrast showed no significant differences for experts between the gesture/gaze condition and the control condition (t ₍₁₃₈₎ = −2.915, p = .004, d = 0.82), as well as between the gesture/gaze condition and the gesture condition (t ₍₁₃₈₎ = 2.940, p < .001, d = 0.87). However, there were no significant differences between the control condition and the gesture condition (t ₍₁₃₈₎ = 0.25, p = .980, d = 0.01) (Table 1).

For the number of fixations on the coach, the ANOVA showed a non-significant main effect for expertise, F(1, 138) = 0.013, p = .910, $η_{p}^{2}$ = 0.001, a significant main effect for condition, F(2, 138) = 6.963, p = .001, $η_{p}^{2}$ = 0.092 and a significant expertise × condition interaction, F(2, 138) = 11.210, p < .001, $η_{p}^{2}$ = 0.140. The results of planned contrast showed that novices in the control condition paid greater attention to the coach than novices in the gesture condition (t ₍₁₃₈₎ = −5.344, p < .001, d = 1.33) and novices in the gesture/gaze condition (t ₍₁₃₈₎ = −5.014, p < .001, d = 0.86). However, there were no significant differences between the gesture/gaze condition and the gesture condition (t ₍₁₃₈₎ = 0.384, p < .701, d = 0.48). On the other side, the results of planned contrast showed no significant differences for expert between the control condition and the gesture condition (t ₍₁₃₈₎ = −1.029, p = .305, d = 0.37), as well as between the control condition and the gesture/gaze condition (t ₍₁₃₈₎ = −0.217, p = .828, d = 0.37). Moreover, there were no significant differences between the gesture/gaze condition and the gesture condition (t ₍₁₃₈₎ = 0.801, p = .424, d = 0.03) (Table 1).

For the saccades count between the diagrams and coach, the ANOVA showed a non-significant main effect for expertise, F(1, 138) < 0.001, p = .983, $η_{p}^{2}$ = 0.001, a significant effect of condition F(2, 138) = 7.328, p < .001, $η_{p}^{2}$ = 0.096, and a significant expertise × condition interaction, F(2, 138) = 14.436, p < .001, $η_{p}^{2}$ = 0.173. The results of planned contrast showed that novices in the control condition switched between the two AOIs more frequently compared to novices in the gesture condition (t ₍₁₃₈₎ = −4.594, p < .001, d = 1.09) and novices in the gesture/gaze condition (t ₍₁₃₈₎ = −3.060, p = .003, d = 0.81). However, there were no significant differences for novices between the gesture condition and the gesture/gaze condition (t ₍₁₃₈₎ = 1.581, p = .116, d = 0.39). On the other side, the results of planned contrast showed that experts in the gesture/gaze condition switched between the two AOIs more frequently than experts in the control condition (t ₍₁₃₈₎ = −3.934, p < .001, d = 1.30) and experts in the gesture condition (t ₍₁₃₈₎ = 3.524, p = .001, d = 1.47). However, there were no significant differences between the control condition and the gesture condition (t ₍₁₃₈₎ = −0.414, p = .679, d = 0.14) (Table 1).

Discussion

The present study aimed to investigate the effect of pointing gestures and gaze guidance on recall performance, mental effort and distribution of visual attention in players with different levels of expertise.

The results indicated a positive effect of both conditions with pointing gestures (gesture and gesture/gaze) in comparison with the control condition for novice players. These players achieved higher recall performance (but not for recall time) and lower mental effort investment when learning from conditions with gestures, which supported H1a, H1b, H3a and H3b, but not H2a and H2b. In addition, eye-tracking data showed that in the conditions that the coach used pointing gestures, novice players paid greater attention to the relevant elements on the diagrams, less attention to the coach, and switched less frequently between the coach and the relevant diagrams as indicated by saccades count, which supported H4a and H4b. When novice players are presented with a system based-narration, they may find it difficult to extract the relevant information mentioned orally, thus reducing the cognitive resources available to organize and integrate tactical instructions. Observing the coach's pointing gestures allows these players to recognize the location of each game action as well as its starting and ending point, facilitating therefore the construction of a coherent mental model of the whole system and enhance learning. According to the embodied cognition perspective, when learners saw the instructor's gestures, they engaged with his motor and perceptual systems, which may help them understand his explanations and thus improve their learning performance.^49,50,70 Moreover, pointing gestures served to highlight the relevant information mentioned in the speech, limiting inefficient visual search and thus reducing extraneous cognitive load.^13,43,46 This explanation is consistent with the signaling principle, which states that adding cues such as pointing gestures to learning material can improve learners’ visual search and learning performance.^46,64

Regarding the combination of pointing gestures and gaze guidance, the results showed that novices in the gesture/gaze condition achieved higher recall scores and invested the same mental effort as novices in the gesture condition, which supported H1c but not H2c and H3c. They also paid more attention to the relevant diagrams and less attention to the coach, and switched between the two AOIs in the same way as novices in the gesture condition which supported H4c. In other words, the results showed that the coach's pointing gesture combined with the gaze guidance attracted the novices’ attention even more and led to a weaker divided attention effect than the coach's pointing gesture alone. There are two possible explanations for these results. First, the coach's face attracts players’ attention even when he makes a pointing gesture. However, when the coach shifts his gaze from the players to the relevant game elements on the board, these are signals for the players to pay attention to the relevant elements on the board and to make connections between what the coach is saying and what he is indicating on the board. This explanation is supported by previous research on the role of gaze guidance in teaching^19,32 and consistent with the gaze guidance principle,³³ which states that learners could potentially improve their learning outcomes if they follow the instructor's gaze guidance. Second, the possible effect of combining pointing gestures and gaze guidance is explained from a social perspective. According to social agency theory,⁴¹ pointing and looking at relevant information could elicit a stronger social response (i.e., a sense of partnership) in learners that contributes to deeper learning. In addition, research on social cognition suggests that gestures and gaze can promote joint attention and shared intentions, which can increase learners’ motivation to engage in learning tasks.⁵¹

With regard to the experts’ performance, the results showed that the experts in the control condition achieved the same recall performance, same mental effort investment, and same visual search patterns as the experts in the gesture condition, which did not support H1d, H2d, H3d, and H4d. Furthermore, the results showed that expert players benefited more from the control condition than the gesture/gaze condition by achieving the same recall performance, lower mental effort investment, and better visual search, which supported H3e, and H4e, but not H1e and H2e. These results suggest that although players are experts in the field, this additional information (i.e., the pointing gesture) is redundant and unnecessarily burdens their working memory, which is associated with an increase in extraneous cognitive load in terms of CLT.^61,62,64 Through deliberate practice and years of experience in their field, expert players may have developed an advanced knowledge base that allows them to effectively orient their attention to the relevant information represented by the auditory description so that they can create an effective comprehensive mental model of the dynamic system.^60,71 As a result, any information provided to expert players that they do not need to process is redundant and should be removed.^58,70,71

On the other hand, when comparing the conditions with the gestures in expert players, the results showed that they benefited more from the gesture condition than from the gesture/gaze condition. Despite performing at the same recall performance, they invested less mental effort and showed better visual search (they spent less time fixating on the coach and switched between the coach and the relevant diagrams less frequently) in the gesture condition compared to the gesture/gaze condition, which supported H3f, H4f, but not H1f and H2f. For these players, the coach's gaze cues are external cues that tend to attract attention. Apart from the mere attentional distraction provided by the coach's gaze, the frequent gaze switches might have repeatedly distracted the experts from the diagrams of play drawn on the board. Because switching gaze between the player and the board automatically attracts attention, the experts who saw the gesture/gaze condition paid more attention to the coach than their counterparts the gesture condition. Since all relevant information is encoded in the verbal description, the coach's gaze cues are considered “extraneous” because they could lead to increased cognitive load. Therefore, adding pointing gestures and gaze guidance becomes unnecessary without providing any added value that could enhance memorization and visual search. The results are consistent with the expertise reversal effect which suggests that novice players profit more from highlighting cues than expert players since novice players are hardly able to disregard irrelevant details and do not automatically focus on relevant information.^3,58,60,72 In the case of a high level of expertise, signaling techniques (i.e., pointing gestures and gaze guidance) may contradict experts’ existing mental representation of new learning material.⁴¹ This represents a conflict that needs to be resolved as it leads to additional extraneous cognitive load.

Despite the numerous contributions of the study, three important limitations should be considered when interpreting the results. First, there are other types of human cues that have not yet been explored in the sports domain when representing game systems on the board. According to the embodiment principle,⁴¹ people learn more deeply when teachers show human-like gestures, movements, eye gaze, and facial expressions. Therefore, other types of hand gestures (e.g., tracing, beat gestures, depictive gestures), human-like movements, and facial expressions could be added to pointing gestures and gaze guidance and need to be used in the next studies to optimize players’ performance. Second, research examining the effects of non-human cues on tactical learning has shown that adding a red circle to learning material has a positive effect on novices, but reverses and becomes ineffective for experts.⁶⁰ Future studies should compare these different types of cues and show which technique is more effective in improving learning. Finally, the recall task was limited to a relatively simple game system. Studies examining the effects of human cues (e.g., pointing gestures) showed that they improved performance when they accompanied a complex verbal description.⁷³ Future studies could therefore address the question of whether pointing gestures and gaze cues could improve players’ performance in memorizing more complex content.

In summary, the present study confirms that when a static basketball diagram is presented on the board, players’ attention can be directed to the task area pointed out by the coach using pointing gestures. Moreover, it has been shown that gaze cues redirect players’ attention to a large extent and attract it at the same time. However, the level of expertise plays a crucial role in the tactical memorization process and must be taken into account when incorporating pointing gestures and gaze cues into the material. For novice players, pointing and looking at the relevant elements mentioned in the speech rather than just pointing or not giving cues is a better way to effectively encode and integrate tactical instructions. When the learner becomes an expert in the field, learning from the coach's speech appears to be sufficient to create a coherent mental model of the tactical instructions presented. This suggestion is in accordance with the expertise reversal effect, which states that the most appropriate learning technique depends critically on the learners’ level of prior knowledge.^10,58,60

Footnotes

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Houssem Ben Chikha

References

Camerino

Castañer

Anguera

TM.

Mixed methods research in the movement sciences: Case studies in sport, physical education and dance. Camerino

Castaner

Anguera

(eds). London, England: Routledge, 2014.

Evans

Reynolds

. The organization of corrective demonstrations using embodied action in sports coaching feedback: corrective demonstrations in sports coaching. Symb Interact 2016; 39: 525–556.

Ben Chikha

Zoudji

Khacharem

Coaches’ pointing gestures as means to convey tactical information in basketball: an eye-tracking study. Int J Sport Exerc Psychol. Epub ahead of print 31 October 2022. doi:10.1007/s00426-023-01806-y http://dx.doi.org/10.1080/1612197x.2022.2138498

Kao

S-F

Hsieh

M-H

Lee

P-L

. Coaching competency and trust in coach in sport teams. Int J Sports Sci Coach 2017; 12: 319–327.

Klingner

Huijgen

BCH

Den Hartigh

RJR

, et al. Technical–tactical skill assessments in small-sided soccer games: a scoping review. Int J Sports Sci Coach 2022; 17: 885–902.

Schütz A and Kaul C. Means of verbal and non–verbal communication in coaching. In: Siegfried Greif, Heidi Möller, Wolfgang Scholl, Jonathan Passmore, Felix Müller (eds) International handbook of evidence-based coaching. Cham: Springer International Publishing, 2022, pp. 565–575.

Berntsen

Kristiansen

. Successful coach learning: digital workbook informed by pedagogical principles. Int J Sports Sci Coach 2019; 14: 310–323.

Krause

Weigelt

. Mental rotation and performance in basketball: effects of self-controlled and externally controlled time constraints on the processing and execution of tactic board instructions with varied orientations. Ger J Exerc Sport Res 2020; 50: 354–365.

Khacharem

Zoudji

Kalyuga

, et al. The expertise reversal effect for sequential presentation in dynamic soccer visualizations. J Sport Exerc Psychol 2013; 35: 260–269.

10.

Khacharem

Zoudji

Kalyuga

. Perceiving versus inferring movements to understand dynamic events: the influence of content complexity. Psychol Sport Exerc 2015; 19: 70–75.

11.

Schnotz

Lowe

A unified view of learning from animated and static graphics. In: Lowe

(ed.) Learning with animation: research implications for design, New York, NY, US: Cambridge University Press, 2008, ix, p. 304–356.

12.

Wang

Antonenko

. Instructor presence in instructional video: effects on visual attention, recall, and perceived learning. Comput Human Behav 2017; 71: 79–89.

13.

Van Wermeskerken

Van Gog

. Seeing the instructor’s face and gaze in demonstration video examples affects attention allocation but not learning. Comput Educ 2017; 113: 98–107.

14.

Colonnesi

Stams

GJJM

Koster

, et al. The relation between pointing and language development: a meta-analysis. Dev Rev 2010; 30: 352–366.

15.

Tuddert-Kennedy

. Hand and mind: what gestures reveal about thought. Lang Speech 1994; 37: 203–209.

16.

Valenzeno

Alibali

Klatzky

. Teachers’ gestures facilitate students’ learning: a lesson in symmetry. Contemp Educ Psychol 2003; 28: 187–204.

17.

Singer

Goldin-Meadow

. Children learn when their teacher’s gestures and speech differ. Psychol Sci 2005; 16: 85–89.

18.

Koumoutsakis

Church

Alibali

, et al. Gesture in instruction: evidence from live and video lessons. J Nonverbal Behav 2016; 40: 301–315.

19.

Ouwehand

Van Gog

Paas

. Designing effective video-based modeling examples using gaze and gesture cues. J Educ Techno Soc 2015; 18: 78–88.

20.

Zhang

Zhu

, et al. Instructors’ pointing gestures improve learning regardless of their use of directed gaze in video lectures. Comput Educ 2019; 128: 345–352.

21.

Wakefield

Novack

Congdon

, et al. Gesture helps learners learn, but not merely by guiding their visual attention. Dev Sci 2018; 21: e12664.

22.

Congdon

Novack

Brooks

, et al. Better together: simultaneous presentation of speech and gesture in math instruction supports generalization and retention. Learn Instr 2017; 50: 65–74.

23.

Kelly

. Broadening the units of analysis in communication: speech and nonverbal behaviours in pragmatic comprehension. J Child Lang 2001; 28: 325–349.

24.

Sekine

Kita

. The listener automatically uses spatial story representations from the speaker’s cohesive gestures when processing subsequent sentences without gestures. Acta Psychol (Amst) 2017; 179: 89–95.

25.

Yeo

Ledesma

Nathan

, et al. Teachers’ gestures and students’ learning: sometimes “hands off” is better. Cogn Res Princ Implic 2017; 2: 41.

26.

Ouwehand

Van Gog

Paas

. Effects of gestures on older Adults’ learning from video-based models: gestures in instruction for older adults. Appl Cogn Psychol 2015; 29: 115–128.

27.

Gullberg

Holmqvist

. What speakers do and what addressees look at: visual attention to gestures in human interaction live and on video. Pragmat Cogn 2006; 14: 53–82.

28.

Langton

Watt

Bruce

. Do the eyes have it? Cues to the direction of social attention. Trends Cogn Sci 2000; 4: 50–59.

29.

Streeck

. Gesture as communication I: its coordination with gaze and speech. Commun Monogr 1993; 60: 275–299.

30.

Stull

Fiorella

Mayer

. The case for embodied instruction: the instructor as a source of attentional and social cues in video lectures. J Educ Psychol 2021; 113: 1441–1453.

31.

Stull

Fiorella

Mayer

. An eye-tracking analysis of instructor presence in video lectures. Comput Human Behav 2018; 88: 263–272.

32.

Wang

. The instructor’s gaze guidance in video lectures improves learning. J Comput Assist Learn 2019; 35: 42–50.

33.

Mayer

Fiorella

Stull

. Five ways to increase the effectiveness of instructional video. Educ Technol Res Dev 2020; 68: 837–852.

34.

Clifford

CWG

Palmer

. Adaptation to the direction of others’ gaze: a review. Front Psychol 2018; 9: 2165.

35.

Sweller

. Cognitive load during problem solving: effects on learning. Cogn Sci 1988; 12: 257–285.

36.

Sweller

. Cognitive load theory and educational technology. Educ Technol Res Dev 2020; 68: 1–16.

37.

Castro-Alonso JC, Ayres P and Sweller J. Instructional visua lizations, cognitive load theory, and visuospatial processing. In: Juan C. Castro-Alonso (eds) Visuospatial processing for education in health and natural sciences. Cham: Springer International Publishing, 2019, pp. 111–143.

38.

Risko

Gilbert

. Cognitive offloading. Trends Cogn Sci 2016; 20: 676–688.

39.

Sweller

Van Merriënboer

JJG

Paas

. Cognitive architecture and instructional design: 20 years later. Educ Psychol Rev 2019; 31: 261–292.

40.

Ayres

Sweller

. The split-attention principle in multimedia learning. In: Mayer

(eds) The Cambridge handbook of multimedia learning. Cambridge: Cambridge University Press, 2014, pp.206–226.

41.

Mayer

. Principles based on social cues in multimedia learning: personalization, voice, image, and embodiment principles. In: Mayer

(eds) The Cambridge handbook of multimedia Learning. Cambridge: Cambridge University Press, 2014, pp.345–368.

42.

Atkinson R, Lin L and Harrison C. Comparing the efficacy of different signaling techniques. In: G Siemens and C Fulford (eds) Proceedings of ED-MEDIA 2009--World Conference on Educational Multimedia, Hypermedia & Telecommunications. Waynesville, NC. Honolulu, HI, USA: Association for the Advancement of Computing in Education (AACE), 2009, Jun 22, pp. 954–962.

43.

Hong

Yang

. Effects of the instructor’s pointing gestures on learning performance in video lectures: pointing gestures enhance learning in video lectures. Br J Educ Technol 2017; 48: 1020–1029.

44.

Castro-Alonso

Wong

Adesope

, et al. Effectiveness of multimedia pedagogical agents predicted by diverse theories: a meta-analysis. Educ Psychol Rev 2021; 33: 989–1015.

45.

Mayer

. The Cambridge handbook of multimedia learning. Cambridge, England: Cambridge University Press, 2014.

46.

De Koning

Tabbers

Rikers

RMJP

, et al. Towards a framework for attention cueing in instructional animations: guidelines for research and design. Educ Psychol Rev 2009; 21: 113–140.

47.

Betrancourt

. The animation and interactivity principles in multimedia learning. In: Mayer

(eds) The Cambridge handbook of multimedia Learning. Cambridge: Cambridge University Press, 2005, pp.287–296.

48.

Alibali

Nathan

Wolfgram

, et al. How teachers link ideas in mathematics instruction using speech and gesture: a corpus analysis. Cogn Instr 2014; 32: 65–100.

49.

Barsalou

. Grounded cognition. Annu Rev Psychol 2008; 59: 617–645.

50.

Glenberg

. Embodiment as a unifying perspective for psychology: embodiment as a unifying perspective. Wiley Interdiscip Rev Cogn Sci 2010; 1: 586–596.

51.

Frischen

Bayliss

Tipper

. Gaze cueing of attention: visual attention, social cognition, and individual differences. Psychol Bull 2007; 133: 694–724.

52.

Fox

Bakermans-Kranenburg

Yoo

, et al. Assessing human mirror activity with EEG mu rhythm: a meta-analysis. Psychol Bull 2016; 142: 291–313.

53.

Fogassi

Ferrari

Gesierich

, et al. Parietal lobe: from action organization to intention understanding. Science 2005; 308: 662–667.

54.

Rizzolatti

Craighero

. The mirror-neuron system. Annu Rev Neurosci 2004; 27: 169–192.

55.

Binder

Dovern

Hesse

, et al. Lesion evidence for a human mirror neuron system. Cortex 2017; 90: 125–137.

56.

Van Gog

Paas

Marcus

, et al. The mirror neuron system and observational learning: implications for the effectiveness of dynamic visualizations. Educ Psychol Rev 2009; 21: 21–30.

57.

Chen

Kalyuga

Sweller

. The expertise reversal effect is a variant of the more general element interactivity effect. Educ Psychol Rev 2017; 29: 393–405.

58.

Kalyuga

. The expertise reversal principle in multimedia learning. In: Mayer

(eds) The Cambridge handbook of multimedia Learning. Cambridge: Cambridge University Press, 2014, pp.576–597.

59.

Sweller J, Ayres P and Kalyuga S. The expertise reversal effect. In: John Sweller, Paul Ayres, and Slava Kalyuga (eds) Cognitive load theory. New York: Springer New York, 2011, pp. 155–170.

60.

Khacharem

. Top-down and bottom-up guidance in comprehension of schematic football diagrams. J Sports Sci 2017; 35: 1204–1210.

61.

Homer

Plass

. Expertise reversal for iconic representations in science visualizations. Instr Sci 2010; 38: 259–276.

62.

Khacharem

Zoudji

Spanjers

IAE

, et al. Improving learning from animated soccer scenes: evidence for the expertise reversal effect. Comput Human Behav 2014; 35: 339–349.

63.

Kalyuga

Sweller

. The redundancy principle in multimedia learning. In: Mayer

(eds) The Cambridge handbook of multimedia Learning. Cambridge: Cambridge University Press, 2014, pp.247–262.

64.

Richter

Scheiter

Eitel

. Signaling text–picture relations in multimedia learning: the influence of prior knowledge. J Educ Psychol 2018; 110: 544–560.

65.

Faul

Erdfelder

Buchner

, et al. Statistical power analyses using G*Power 3.1: tests for correlation and regression analyses. Behav Res Methods 2009; 41: 1149–1160.

66.

Paas

. Training strategies for attaining transfer of problem-solving skill in statistics: a cognitive-load approach. J Educ Psychol 1992; 84: 429–434.

67.

Paas

Renkl

Sweller

. Cognitive load theory and instructional design: recent developments. Educ Psychol 2003; 38: 1–4.

68.

De Jong

. Cognitive load theory, educational research, and instructional design: some food for thought. Instr Sci 2010; 38: 105–134.

69.

Schad

Vasishth

Hohenstein

, et al. How to capitalize on a priori contrasts in linear (mixed) models: a tutorial. J Mem Lang 2020; 110: 104038.

70.

Arslan-Ari

. Learning from instructional animations: how does prior knowledge mediate the effect of visual cues? J Comput Assist Learn 2018; 34: 140–149.

71.

Khacharem

Zoudji

Kalyuga

. Which representation is best for communicating dynamic information? Memory 2019; 27: 943–951.

72.

Ben Chikha

Zoudji

Khacharem

Using pointing gestures to convey tactical information: investigating the roles of expertise and complexity. Psychol Res. Epub ahead of print 22 February 2023. doi:10.1007/s00426-023-01806-y

73.

McNeil

Alibali

Evans

. Role of gesture in children's comprehension of spoken language: now they need it, now they don't. J Nonverbal Behav 2000; 24: 131–150.