Effect of Observed Biological and Non-biological Movement on Human Action and Coordination

Dynamical systems researchers have understood the stable patterning of interpersonal and non-biological environmental rhythmic limb coordination to be constrained by the selforganized entrainment processes’ coupled oscillators. Recently, it has been demonstrated that an individual’s rhythmic limb movements exhibit greater variability when viewing spatially incongruent biological limb movements, but not when viewing spatially incongruent non-biological movements. Some researchers have concluded that a ‘mirrorneuron system’ might mediate the intrinsic bidirectional link between perception and action underlying interpersonal, but not environmental, coordination (e.g., Kilner et al., 2003; Tognoli et al., 2007). The current study aimed to: 1) contest this recent finding; and 2) demonstrate that the self-organized entrainment processes of coupled oscillators can explain the differing influences of biological and non-biological movements. In the first experiment, participants intentionally coordinated arm movements with spatially congruent and incongruent arm movements of a confederate, a robotic image with computer generated movement, and a robotic image producing pre-recorded human movement. Results revealed more stable coordination for congruent and biological movement than incongruent and robotic movement, respectively. The second experiment investigated the influence of biological and non-biological movement on unintentional coordination. Consistent with dynamical systems theory, coordination was found in both biological and non-biological conditions. Effect of Observed Biological and Non-biological Movement 3 Effect of observed biological and non-biological movement on human action and coordination Movement coordination is present at all levels of daily life. From the natural, unpracticed coordination that occurs between two friends walking and talking together to the skillful, trained coordination of lanky basketball players or quick-footed dancers, coordination plays an undeniable role in human social interaction. Despite not sharing a common cognitive or neural mechanism and common limbs, multiple individuals still come together in dynamic cooperative relationships binding them in a way that resembles a single organism’s functioning (Asch, 1952; Newtson, Hairfield, Bloomingdale, & Cutino, 1987). For such coordination to occur, there must be a perceptual link between these two separate acting systems. Further there must be a strong link between the perception and action processes of each interactant. Without such a link between perception and action, coordination among cognitively and psychically separate systems can not occur. The perception of another’s actions, and the influence of this perception on one’s actions, is what causes multiple systems to come together in synchrony. The dynamic systems theory and the mirror systems theory are two popular explanations for the relationship between perception and action, and how this relationship can result in interpersonal coordination. According to mirror systems theory there are neurons in our brain that represent certain movements. These neurons fire when we observe or perform an action (Blakemore & Frith, 2005). The perception of an action causes firing in the same part of the brain as the performance of that action, and thus perception and action directly influence each other. On the other hand, dynamic systems theory states that all systems have a tendency to become coordinated with their Effect of Observed Biological and Non-biological Movement 4 environment. Such coordination leads to a more stable state of existence—a systemenvironment equilibrium. Dynamic systems posits as long as there is sufficient perception of environmental rhythms, a system’s actions will naturally come into self-organized coordination with what is perceived. Thus the basis for both theories is that when we perceive another person’s movement, we are likely to coordinate with this movement. Kilner, Paulignan, and Blakemore (2003) conducted a study on perceiving and making arm movements, and explained their results through the mirror systems theory. The present study aims to replicate their findings, and show that they can be better explained using the dynamic systems theory. Mirror neurons and mirror systems theory Mirror neurons refer to cells that were discovered in area F5 of the pre-motor cortex of Macaque monkeys that fire when a monkey both executes and observes a particular action (Gallese, Fadiga, Fogassi, & Rizzolatti, 1996). Since the discovery of mirror neurons in monkeys, a number of functional magnetic resonance imaging (fMRI), positron emission tomography (PET), electroencephalography (EEG), and transcranial magnetic stimulation (TMS) studies have found evidence to suggest that a similar “mirror neuron” system might also exist in humans (Grezes, et al., 1998; Cochin, et al., 1999; Grezes, et al., 1999; Strafella & Paus, 2000; Buccino, et al., 2001; Iacobani et al., 2001; Aziz-Zadeh, et al., 2002; Muthukumaraswamy & Johnson, 2004). Theoretically, the mirror neuron system is argued to account for why observing an action causes an automatic activation of brain areas normally associated with action execution (Press et al., 2005), whereby this neural activation represents itself behaviorally Effect of Observed Biological and Non-biological Movement 5 through action imitation (Brass & Heyes, 2005). It is important to note, however, that this response is to general, not specific, aspects of the observed movement (Blakemore & Frith, 2005), whereby the mirror neuron system activates a motor prototype in the observer and not an “instant and accurate imitation” (Vogt & Thomaschke, 2006). In addition, because the amount and the effects of activation are directly related to the observer’s intention, mirror systems theory holds that activation of these neurons does not always elicit behavioral imitation. Indeed, activation in the premotor and parietal cortices is greatest when observers have the intent to imitate (Grèzes, Costes & Decety, 1999). Thus, when the intent to imitate is absent, witnessing movement either facilitates the execution of the witnessed movement or interferes with performance of a different movement (Press et al., 2005). This latter point defines a key prediction of mirror systems theory, namely, that observed movements that are qualitatively different from performed movements cause an interference effect. Specifically, because observed and performed movements activate the same neurons, an incompatibility between them is argued to impair movement. For instance, individuals instructed to either tap or lift their fingers, with a photo of either the movement they were to produce or the alternate movement as their signal to begin, exhibited significantly slower movement onset times when the photograph was of the alternate movement compared to the to-be-produced movement (Brass, Bekkering, & Prinz, 2001). Similar interference effects have been demonstrated in studies investigating bar grasping, opening and closing of hands, and arm swinging (Craighero, et al., 2002; Kilner, Paulignan, & Blakemore, 2003; Vogt, Taylor, & Hopkins, 2003; Blakemore & Frith, 2005). Effect of Observed Biological and Non-biological Movement 6 Researchers have suggested that such interference effects should only occur when the observed movement is made by a human, or more specifically a biological conspecific (Blakemore & Frith, 2005; Kilner, Paulignan, & Blakemore, 2003). Drawing on research which demonstrates how infants, as early as 3 months old, are able to distinguish between dots moving in a human-like (biological) manner and those moving in a robotic (non-biological) manner (Bertenthal, 1993), these authors have claimed that the mirror system more effectively represents human movement, because the brain processes biological and non-biological motion differently. The distinction between biological and non-biological motion is partially explained by biological motion primarily being mentally processed in the STS (Grèzes, et al., 2001; Grossman & Blake, 2001; Grossman et al., 2000; Allison, Puce & McCarthy, 2000; Frith & Frith, 1999), which is also activated during motion performance. Although the area of the brain in which nonbiological motion is processed has yet to be discovered, it does not appear to be processed in the STS, thus no interference effect would occur while witnessing robotic motion (Blakemore & Frith, 2005). To verify this claim, Kilner, Paulignan, & Blakemore (2003) further examined the dependence of the interference effect on witnessing human motion. Participants made horizontal or vertical arm movements in the presence of a confederate (biological motion) or a robotic arm (non-biological motion), with four conditions for each. In two of the conditions, the movements of the participant and confederate/robot were congruent— both moved either horizontally or vertically. In the other two conditions, the movements of the participant and confederate/robot were incongruent—the participant moved Effect of Observed Biological and Non-biological Movement 7 vertically while the confederate/robotic arm moved horizontally, or the participant moved horizontally while the confederate/robotic arm moved vertically. The subjects coordinated in time with the confederate or robotic arm’s movements, regardless of congruency or direction, and the authors measured the end-point variability of the participant’s movements in the plane opposite to the intended plane of movement. The results revealed an increase in end-point variability for the incongruent condition compared to the congruent condition, but only when the participant coordinated with the confederate—no difference was found between the congruent and incongruent conditions when the participant coordinated with the robot. Thus, an interference effect was only observed for incongruent human movement. The subjects perceptually represented a type of movement that was in opposition to the one they intended to make. Because this representation and action occur in the same neural location, the participant’s action was hindered by the incongruent perception. One the other hand, robotic movement is represented in a different neural location than human movement, and thus perception of robotic motion had no effect on the participant’s action. The interference effect was only present in the confederate condition, with almost no effect for the robotic condition. Although the study by Kilner et al., (2003) suggests that there are separate neural representations for biological motion and robotic motion, more recent research has shown that the more non-biological motion resembles human movement, the more likely imitation (and thereby interference) is to occur (Press, Bird, Flach &Heyes, 2005). Participants in this study were required to open their hand while looking at a picture of either a human or robotic hand in an open/compatible position or a closed/incompatible Effect of Observed Biological and Non-biological Movement 8 position. Although the human hand had a stronger effect on performance, making the robotic hand similar in size, color, and brightness to the human hand elicited both an imitation and interference effect. Thus, it is unclear if the Kilner et al (2003) findings are due to separate neural representations of a biological species and robotic creation. There is a growing body of literature on the self-organized emergence of visual rhythmic coordination, according to which coordination between the movements of two systems emerges spontaneously so long as there is an informational link between the systems (Schmidt, Bienvenu, Fitzpatrick, & Amazeen, 1998; Schmidt, Christianson, Carello, & Baron, 1994). Of particular relevance, is that this latter research has demonstrated how the human movements of one individual can become intentionally and unintentionally synchronized to the movements of another individual (biological) or environmental stimulus (non-biological) and that such coordination is constrained by the self-organized entrainment process of coupled oscillators. Under this understanding, increases in the variability of movement are not the result of ‘interference’ or different neuro-cognitive processes, but rather are the result of differences in the inherent and lawfully defined stabilities of the different movement patterns produced or the strength of the informational coupling that links the respective movements (Kelso, 1995; Kugler & Turvey, 1987). Thus the variability observed in the above mentioned studies for congruent and incongruent movements is a result of the inherent stabilities of these movement patterns, and the variability observed for biological and non-biological Effect of Observed Biological and Non-biological Movement 9 movement is caused by the different coupling links between two biological systems and between a biological system and an environmental stimulus. Dynamic systems theory Dynamical systems theory holds that the components of any system or systems mutually influence each other and that the organization of a system’s behavior is the result of natural laws and constraints, and emerges without an inside agent controlling it (Kelso, 1995). This self-organizing behavior often results in the synchrony or coordination of the system’s components. With respect to the coordination of behavioral rhythms, this dynamical systems approach has been effectively used to understand the self-organized synchronization of many different biological systems, from the flashing of fireflies (Hanson, 1978) and chirping of crickets (Walker, 1969) to intrapersonal human interlimb coordination (i.e. bimanual coordination of the finger, leg or arm movements of a single individual) (e.g., Kelso, 1995; Turvey, Rosenblum, Schmidt, & Kugler, 1986). Pertinent to this study, this behavioral coordination has been well demonstrated in the interpersonal movement coordination that occurs both intentionally and unintentionally between two individuals or an individual and an environmental stimulus (Richardson, Marsh, & Schmidt, 2005; Schmidt, Bienvenu, Fitzpatrick, & Amazeen, 1998; Schmidt, Christianson, Carello, & Baron, 1994; Schmidt & O’Brien, 1997; Schmidt, Richardson, Arsenault, & Galantucci, 2007). For all the above systems (as well as many more) the patterning of the coordination that occurs is known to be governed by the self-organized entrainment processes of coupled oscillators. More specifically, the coordinated behavior of two Effect of Observed Biological and Non-biological Movement 10 rhythmic limb or body movements is consistent with the dynamics of coupled limit-cycle oscillators. Thus, for 1:1 frequency locked coordination, the dynamic stabilities of such coordination can be captured using a motion equation for the collective variable of relative phase ) ( R L θ θ φ − = , the difference in the phase angles of the left and right movements/oscillators. Typically, this motion equation, know as the Haken-Kelso-Bunz (HKB) equation, takes the following form: t Q b a ζ φ φ ω φ + − − Δ = 2 sin 2 sin & , (1) Here, coordination is measured using the change in relative phase ( ) as the order parameter—the unit of measure that specifies the spatial temporal order of the rhythmic units. The control parameters—the variables influencing the stability of coordination— are captured by the difference in eigenfrequency between the two rhythmic units (Δω), the coordination strength expressed in the coupling function ( φ& φ φ 2 sin 2 sin b a − ), and a Gaussian white noise process (ζ ) dictating a stochastic force of strength (Q ). Essentially, the smaller the differences in the eigenfrequencies of the two movements, the larger coupling strength, and the smaller the level of noise present in the system, the stronger and more stable the resulting coordination (Kelso, 1995; Schmidt & Richardson, 2008). The states of coordination that occur for human interlimb and interpersonal coordination and that are modeled by Equation 1 are refereed to as inphase and antiphase. Inphase coordination occurs at a relative phase angle between the coordinating Effect of Observed Biological and Non-biological Movement 11 components of 0o. In this symmetric or congruent phase mode, components are at the same point in their cycles at the same time. Antiphase coordination occurs at a relative phase angle between coordinating components of 180o, and in this alternate or incongruent phase mode, the components are at opposite points in their cycles at the same time (Kelso et al., 1986; Stafford & Barnswell, 1985; Von Holst, 1939/1973). Although both inphase and antiphase are stable states of coordination, the stability of the two phase modes is not equal. A non-linear phase transition from antiphase to inphase coordination occurs when the frequency or tempo of movement (captured by the ratio of the parameters a and b in Equation 1) is increased (Kelso & Scholz, 1985; Hoyt & Taylor, 1981; Gilmore, 1981; Haken, 1977/1983). No such transition occurs from inphase to antiphase. Accordingly, antiphase coordination is known to be intrinsically less stable than inphase coordination. In addition to the transition from antiphase to inphase at increased frequencies of movement, the difference in the inherent stability of inphase and antiphase coordination is evidenced by the variability of antiphase coordination (as measured by the standard deviation of relative phase (SDφ)) being much greater than the variability of inphase coordination (Kelso et al., 1986) This variability is a result of the noise that is inherent to all movement systems (and all natural systems) and which continuously perturbs movement systems away from the stable states of coordination. Because antiphase is a weaker mode of coordination than inphase, the magnitude of these perturbations and thus the variability of the coordination, is much greater for antiphase compared to inphase. From a dynamical systems perceptive movement noise or variability is not necessarily Effect of Observed Biological and Non-biological Movement 12 negative or detrimental to movement coordination and action, and can even have an amplifying or resonating effect on coordination signals (Kelso, 1995; Riley & Turvey, 2002; Schmidt & Richardson, 2008). The present study The present study aimed to demonstrate how the differing influences of biological and non-biological movement of human movement variability and coordination are better explained by dynamic systems theory (and the self-organizing entrainment processes of coupled oscillators) than by the mirror systems theory. Two experiments were conducted and employed a similar methodology to that developed by Kilner et al (2003). In both experiments, participants observed a confederate or robotic computer stimulus producing either horizontal or vertical arms movements while producing congruent or incongruent horizontal or vertical arms movements. In Experiment 1, participants were instructed to intentionally coordinate their movements with the confederate or computer stimulus. In Experiment 2, the influence of observing biological and non-biological movements on the participant’s movement variability during unintentional coordination was examined by not instructing participants to coordinate with the confederate or biological stimulus. In addition to measuring the end-point variability of the movements (as done by Kilner et al., 2003), the stability and variability of the coordination that occurred between the participant and confederate or between the participant and computer stimulus was also measure. Furthermore, whereas Kilner et al. (2003) only examined the influence of observed human and robotic arm movements, both experiments included a biologicalEffect of Observed Biological and Non-biological Movement 13 computer stimulus condition, in which the motion of the computer stimulus was the prerecorded motion of the confederate, as well as the confederate and computer (robotic) stimulus conditions. The congruent and incongruent modes are analogous to the inphase and antiphase modes of coordination defined above. Accordingly, incongruent coordination is expected to be weaker than congruent coordination (Kelso et al., 1987, Stafford & Barnswell, 1985; Von Holst, 1939/1973), with greater movement and coordination variability in the incongruent conditions compared to congruent conditions. Additionally, the increased variability witnessed by Kilner during the purely biological trials and not the robotic trials could be due to biological motion’s natural variability. This variability in the incongruent confederate trials is not a result of interference in mirror neuron representation, but an attraction towards the more stable, congruent form of coordination. A difference in coordination is expected for the biological and nonbiological conditions, because of the greater strength of bidirectional coupling between biological systems compared to the weaker unidirectional coupling between a biological system and an environmental stimulus. Finally, a similar pattern of coordination across the biological and computer conditions is expected, especially for the unintentional coordination in Experiment 2. Experiment 1 Experiment 1 attempted to a) replicate the variability findings of Kilner et al. (2003) and b) investigate whether the differences in the variability between biological congruent and incongruent movements were byproducts of the stabilities of the Effect of Observed Biological and Non-biological Movement 14 coordination. A derivative of the method devised by Kilner’s et al. was employed, with participants intentionally coordinating horizontal and vertical congruent and incongruent arm movements with the arm movements of a confederate, a robotic image with computer generated movement, and a robotic image reproducing a pre-recorded human movement.