Neuroscience Letters 566 (2014) 6–10
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Explicit and implicit knowledge of environment states induce adaptation in postural control José A. Barela a,b,∗ , Matthias Weigelt c , Paula F. Polastri d , Daniela Godoi e , Stefane A. Aguiar a , John J. Jeka f a
Institute of Physical Activity and Sport Sciences, Cruzeiro do Sul University, São Paulo, SP, Brazil Institute of Biosciences, São Paulo State University, Rio Claro, SP, Brazil c Department of Sport and Health, Paderborn University, Paderborn, Germany d Department of Physical Education, São Paulo State University, Bauru, SP, Brazil e Department of Physical Education and Human Movement, Federal University of São Carlos, São Carlos, SP, Brazil f Department of Kinesiology, Temple University, Philadelphia, PA, United States b
h i g h l i g h t s • Increased velocity of visual motion decreases coupling to visual information. • Knowledge of visual surrounding manipulation reduces coupling to visual information. • Implicit and explicit knowledge of environment states reduce sway amplitude.
a r t i c l e
i n f o
Article history: Received 13 November 2013 Received in revised form 28 January 2014 Accepted 14 February 2014 Keywords: Sensory re-weighting Adaptive control Posture Vision
a b s t r a c t The aim of this study was to investigate the effects of explicit and implicit knowledge about visual surrounding manipulation on postural responses. Twenty participants divided into two groups, implicit and explicit, remained in upright stance inside a “moving room”. In the fourth trial participants in the explicit group were informed about the movement of the room while participants in the implicit group performed the trial with the room moving at a larger amplitude and higher velocity. Results showed that postural responses to visual manipulation decreased after participants were told that the room was moving as well as after increasing amplitude and velocity of the room, indicating decreased coupling (down-weighting) of the visual influences. Moreover, this decrease was even greater for the implicit group compared to the explicit group. The results demonstrated that conscious knowledge about environmental state changes the coupling to visual information, suggesting a cognitive component related to sensory re-weighting. Reweighting processes were also triggered without awareness of subjects and were even more pronounced compared to the first case. Adaptive re-weighting was shown when knowledge about environmental state was gathered explicitly and implicitly, but through different adaptive processes. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Our daily activities require the control of a stable and (at the same time) flexible upright stance. To achieve this, the nervous system must process a stimulus-rich and continuously changing environment, requiring the ongoing integration of multisensory
∗ Corresponding author at: Instituto de Ciências da Atividade Física e Esporte – ICAFE, Universidade Cruzeiro do Sul, São Paulo, SP 01506-000, Brazil. Tel.: +55 11 3385 3103. E-mail address: [email protected] (J.A. Barela). http://dx.doi.org/10.1016/j.neulet.2014.02.029 0304-3940/© 2014 Elsevier Ireland Ltd. All rights reserved.
information to update our estimate of self-motion. This integration process is related to the mechanism of sensory re-weighting, which is defined as the ability to select and decrease/increase the influence of a specific sensory stimulus on postural control (e.g., [3,13]). Recently, sensory re-weighting has been rigorously demonstrated and uncovered by manipulating the amplitude and velocity of visual stimuli [1,8,18], vibrating ankle muscles [19], and/or changing somatosensory and visual cues simultaneously [17]. In all cases, changes in postural control were observed, that relied on the down- or up-weighting of the sensory cue’s influence on balance (i.e. upright stance). This indicates that the central nervous system modulates the contribution of the available sensory cues
J.A. Barela et al. / Neuroscience Letters 566 (2014) 6–10
based upon the reliability that they would furnish. Such a mechanism has been characterized as nonlinear and much research has been conducted to understand the mechanistic underpinnings of sensory re-weighting [7,9,10,16,20]. Sensory re-weighting in postural control responses has been observed when participants were told about forthcoming changes in sensory cues. Such instruction-based influences of knowledge on sensory re-weighting were first observed by Nashner and colleagues [6,12] and denominated as central settings adjustments of postural responses. In a more recent study, participants were told that a vision manipulation was occurring and subsequently down-weighted the influence of such sensory cue manipulation [4]. Nevertheless, evidences demonstrate that such down-weighting is stimulus-dependent [1]. Adaptation in postural control system regarding sensory reweighting processes has also been observed after amplitude and velocity of visual motion were increased, leading to downweighting of visual stimulus as well [2,8]. Specifically, it has been demonstrated that when the amplitude of the visual scene motion changes from low amplitude to high amplitude, the postural control system responds relatively quick. However, when the direction of the jump is reversed (from high to low) the response is significantly slower [8,14]. Therefore, the postural control system adapts rapidly the coupling to visual information through sensory reweighting processes when the amplitude of visual motion is high because this represents a threat to stability. In the case of low visual motion amplitude, since stability is not threatened the system can adjust its functioning slowly. Importantly, adaptation due to stimulus properties changes does not involve conscious, since individuals were never informed about visual manipulation [2,14]. Explicit knowledge about visual surrounding manipulation, acquired when individuals were informed about visual motion and asked to resist its influence, as well as implicit knowledge, acquired when participants were exposed to higher velocity and larger amplitude of visual motion without conscious knowledge of the manipulation, have been demonstrated to decrease coupling to visual information (e.g., [1,8]). While several studies provided evidence to confirm these two adaptation mechanisms, no study has ever investigated differences between these two processes with the purpose to discuss how attention and cognitive aspects related to postural control can affect sensory re-weighting compared to behaviors that occur implicitly, that is, without conscious mechanisms involved. Therefore, the purpose of this study was to compare the effects of implicit and explicit information about environmental state in the coupling between visual information and postural control. We employed the moving room paradigm [5,11,15,16] in which participants’ task was to fixate a visual marker on the wall ahead of them and to maintain an upright stance, while the walls of the room moved sinusoidally on a stationary floor. Participants acquired knowledge of the room’s movement either explicitly, by simply being told that the room was moving, or implicitly, by drastically changing the motion parameters of the room (amplitude and velocity), without informing participants about such a change.
2. Methods 2.1. Subjects Twenty healthy young adults participated in this study, equally divided into two groups, one with explicit and one with implicit knowledge. The explicit group included eight males and two females (M = 22.3; SD = 1.57 years), whereas the implicit group four males and six females (M = 22.0; SD = 2.40 years). The recruited participants were undergraduate or graduate students and all had normal or corrected-to-normal vision. In addition, all of them gave
7
their informed consent prior to participation according to procedures approved by the Institutional Review Board.
2.2. Procedures Participants were asked to maintain upright stance inside a moving room at 1 m away from the frontal wall and to look at a target attached at this frontal wall. The moving room consisted of three walls and a roof (2.1 m long × 2.1 m wide × 2.1 m height), mounted on wheels so that it could be moved back and forth by a servomotor mechanism, while the floor remained motionless. The walls and the roof were white with black stripes painted on the walls, creating a pattern of 42 cm wide vertical white and 22 cm wide vertical black stripes. A 20-Watt fluorescent lamp was attached to the ceiling and used to maintain consistent light throughout data collection. The servomotor mechanism consisted of a controller (Compumotor, Model APEX 6151), a controlled stepper motor (Compumotor, Model N0992GR0NMSN), and an electrical cylinder (Compumotor, Model EC3-X3xxN-10004a-Ms1-MT1M), which connected the servomotor to the moving room’s structure. Specialized software (Compumotor, Motion Architect for Windows) controlled the servomotor mechanism, moving the room continuously away from and toward the participant (anterior/posterior direction). A movement analysis system (OPTOTRAK 3020 – 3D Motion Measurement System) was placed behind the participants. One OPTOTRAK IRED marker was placed on the participant’s back (at the 8th thoracic vertebra level) and another one on the frontal wall of the moving room. These markers provided information about the participant’s trunk sway and the moving room’s displacement, respectively, in the anterior–posterior, medial–lateral, and vertical directions, with a sampling rate of 100 Hz. For each participant, 7 trials of 60 s were collected. The first three trials were named pre-change, and the room was oscillated with peak-to-peak amplitude of 0.5 cm, peak velocity of 0.6 cm/s, and frequency of 0.2 Hz. In the fourth trial, named change trial, different manipulations were applied to each group. In the explicit group, participants were verbally (explicitly) informed that the room was moving before the fourth trial started and were asked to resist to this movement (while the parameters of the room’s movement remained the same as used in previous trials during the fourth trial). In the implicit group, the parameters of the room’s movement were changed in the fourth trial to peak-to-peak amplitude of 3.5 cm and peak velocity of 3.5 cm/s. Importantly, no further verbal (explicit) information was given to the participants of the implicit group about the room’s movement. In the remaining trials (Trials 5–7), named post-change trials, the room was oscillated using the same parameter values of the pre-change trials (Trials 1–3), i.e., with peak-to-peak amplitude of 0.5 cm and a peak velocity of 0.6 cm/s. Thus, the difference between the two groups was related to the change trial (Trial 4), in which the explicit group was told about the room’s movement (with the oscillation parameters remaining the same), while the implicit group experienced a change in the oscillation parameter, but was not informed about the room’s movement. When participants in the explicit group were informed about room movement, before the fourth trial, they were explicitly told that the room had been moving in the previous trials and would continue to move in the next trials during the entire experiment. In order to assure participants would be unaware of the movement of the room at the beginning of the experiment, the wheels in which the room was mounted were covered in a way participants could not see them and a random sound (white noise) was used during the entire experiment to mask any possible sound produced by the motor that moved the room. None of the participants
8
J.A. Barela et al. / Neuroscience Letters 566 (2014) 6–10
in both groups reported to have noticed room movement during pre-change trials. 2.3. Data analysis Analyses of body sway magnitude and the relationship between body sway and the moving room was performed only in the anterior–posterior direction, since this was the direction of the manipulation applied. First, mean trunk sway amplitude was calculated by obtaining the standard deviation of trunk position after the average trunk position and a first order polynomial were subtracted from the trunk position time series. The coupling between visual information and trunk sway was examined through the variables gain, phase, stimulus frequency sway amplitude (SFSA) and coherence. To calculate gain and phase, the frequency response function (FRF) was derived from the trunk sway Fourier Transforms divided by the visual stimulus Fourier Transforms. Gain and phase were computed as the absolute value and the argument of the transfer function, respectively, at the driving frequency for each participant across all trials. Gain corresponded to the ratio between the body response amplitude and the visual stimulus amplitude, at the driving stimulus frequency (0.2 Hz). Gain values of one indicate a response amplitude similar to the stimulus amplitude and lower/higher values indicate that the response amplitude was lower/higher than the stimulus driving amplitude. Phase defines the temporal relationship between body sway and the visual stimulus. Positive (negative) phase values indicate sway was ahead (behind) the visual stimulus. In addition, SFSA was calculated as the spectral magnitude of body sway at the frequency of the room motion (0.2 Hz). Finally, coherence measured the strength of the relationship between room movement and body sway. Coherence values close to one indicate strong dependency between these two signals, and coherence values close to zero indicate weak or no dependency between the signals. 2.4. Statistical analysis
Fig. 1. Mean trunk sway amplitude values of young adults in the pre- and postchange trials for both explicit and implicit groups. Error bars represent standard deviation.
strength between the visual stimulus and body sway were found when results between these participants were compared. 3.1. Mean sway amplitude Fig. 1 depicts mean sway amplitude in pre- and post-change trials for both explicit and implicit groups. ANOVA revealed a trial effect on mean sway amplitude, F(1,18) = 9.47, p < 0.01, 2 = 0.345, indicating that trunk sway was reduced in the post- (0.45 cm) compared to the pre-change trials (0.51 cm). No main effect of group, F(1,18) = 0.32, p > 0.05, 2 = 0.018, and no group and trial interaction, F(1,18) = 1.16, p > 0.05, 2 = 0.061, were found. 3.2. Gain and phase Fig. 2 depicts gain and phase between the visual stimulus and trunk sway in pre- and post-change trials for both groups. MANOVA revealed a trial effect, Wilks’ Lambda = 0.207, F(2,17) = 32.51, p < 0.001, 2 = 0.793. The MANOVA also showed a
Three 2 × 2 (Group × Trial) ANOVAs, with repeated measures on the last factor, were carried out to evaluate the effects of group (implicit and explicit) and trial (pre-change and post-change) on the dependent variables mean sway amplitude, SFSA, and coherence. The F-statistic from the averaged FRF values was tested for all groups of trials, revealing that the absolute value and the argument of these values were different from zero (p < 0.02). Following, a 2 × 2 (Group × Trial) MANOVA, with repeated measures on the last factor, was performed to verify the effects of group (implicit and explicit) and trial (pre-change and post-change) on the dependent variables gain and phase. For all variables, the average from the three first trials was used for the pre-change estimate, and the average from the three last trials for the post-change estimate. Post hoc analyses with Tukey’s correction were used to test the main effects. The ˛-level for all analyses was 0.05. 3. Results Participants in the explicit group reported that they did not discriminate the room oscillation, even after the information about the room movement had been provided. Participants in the implicit group were asked, at the end of the experimental session, whether or not they had noticed anything different or uncommon during the experiment. Six participants in the implicit group reported that the room had moved in the fourth trial and remained stationary in all other trials. The other four participants in this group did not report any awareness of the room’s oscillation. Despite these reports, data for all participants were pooled together in the implicit group, because no differences in body sway and in coupling
Fig. 2. Mean gain and phase values of young adults in the pre- and post-change trials for both explicit and implicit groups. Error bars represent standard deviation.
J.A. Barela et al. / Neuroscience Letters 566 (2014) 6–10
9
2 = 0.555, but no group effect, F(1,18) = 3.61, p > 0.001, 2 = 0.167, nor group and trial interaction, F(1,18) = 3.15, p > 0.05, 2 = 0.149, for coherence. Coherence was lower in post-change (0.75) compared to pre-change (0.91) trials.
4. Discussion
Fig. 3. Mean stimulus frequency sway amplitude (SFSA) values of young adults in pre- and post-change trials for both explicit and implicit groups. Error bars represent standard deviation.
group and trial interaction, Wilks’ Lambda = 0.663, F(2,17) = 4.32, p < 0.05, 2 = 0.337, but no group effect, Wilks’ Lambda = 0.946, F(2,17) = 0.48, p > 0.05, 2 = 0.054. Univariate tests indicated a trial effect on gain, F(1,18) = 51.27, p < 0.001, 2 = 0.740. The group and trial interaction for gain was also significant, F(1,18) = 7.53, p < 0.05, 2 = 0.295. Post hoc tests involving this interaction showed that gain values decreased in the post-change trials compared to the prechange trials in both groups. In addition, the post hoc tests revealed that gain values in the post-change trials were lower for the implicit group (0.41) compared to the explicit group (0.60). Univariate analysis for phase values showed no trial effect, F(1,18) = 1.18, p > 0.05, 2 = 0.062, and no group and trial interaction, F(1,18) = 0.05, p > 0.05, 2 = 0.001. Phase values close to zero for all trials indicated that trunk sway was temporally close to the room’s motion. 3.3. SFSA Fig. 3 depicts SFSA in pre- and post-change trials for both groups. ANOVA revealed a trial effect, F(1,18) = 52.56, p < 0.001, 2 = 0.745, and a group and trial interaction, F(1,18) = 7.83, p < 0.02, 2 = 0.303, but no group effect, F(1,18) = 0.69, p > 0.05, 2 = 0.037, for SFSA. Post hoc tests involving this interaction showed that SFSA values decreased in post-change trials compared to pre-change trials, for both groups. In addition, post hoc tests showed that SFSA values in the post-change trials were lower for the implicit group (0.23) compared to the explicit group (0.33). 3.4. Coherence Fig. 4 depicts coherence values in pre- and post-change trials for both groups. ANOVA revealed a trial effect, F(1,18) = 22.40, p < 0.001,
Fig. 4. Mean coherence values of young adults in pre- and post-change trials for both explicit and implicit groups. Error bars represent standard deviation.
The purpose of this study was to evaluate the effects of explicit and implicit knowledge about visual surrounding manipulation on postural responses. Participants who were told about the room’s movement (explicit knowledge) as well as participants who were exposed to higher room movement amplitude and velocity (implicit knowledge) decreased the influence of the visual manipulation on body sway. Moreover, the decreasing was more pronounced in the implicit compared to the explicit group, as evidenced by lower gain and SFSA values in post-change trials. The finding that explicit knowledge about visual surrounding manipulation can influence coupling between visual information and body sway has already been demonstrated [1,5]. Such results support the notion that cognitive components play a role in sensory re-weighting processes, increasing/decreasing the influence of one sensory channel on body oscillation through the conscious effort of a person to achieve a specific goal. This is reminiscent of top-down influences on balance control. At the same time, this study provides evidence for bottom-up influences on postural control. This was indicated by the fact that increasing the amplitude and velocity of the visual stimulus also led to down-weighting the influence of stimulus manipulation on trunk sway, consistent with previous studies [8,14,15]. Such observation indicates that the dynamics of the postural control system can trigger re-weighting processes without awareness of the subject or any ‘conscious’ involvement. In addition, we observed that the down-weighting of visual stimulus’ influence due to implicit knowledge, triggered by stimulus properties manipulation, was more pronounced than the one observed by the explicit knowledge, after the request to resist room motion. Another observation was that some participants in the implicit group, who reported to be aware that the room had been moved, did not show different values in all variables from those participants who were unaware of stimulus change. One would have expected that once the visual stimulus is large enough to be distinguished from the visual movement created by self-motion, it would essentially become explicit. Thus, the down-weighting of visual stimulus after the change trial for some participants in the implicit group could have been due to the same process that led to down-weighting in the explicit group, illustrating an indirect “explicit” effect. However, not only there were no differences in the data of participants who were aware and unaware about room’s movement in the implicit group, but our results also demonstrated that the down-weighting of visual stimulus was stronger in the implicit compared to the explicit group. Therefore, indirect explicit effects or an implicit intention to resist visual motion were not the cause of the down-weighting of visual stimulus in the implicit group. It is clear that the explicit and implicit manipulations are affecting re-weighting to visual information, but through different adaptive processes. Interestingly, adaptation caused by explicit and implicit knowledge of visual manipulation did not affect the temporal relationship between visual stimulus and body sway, since phase values remained unchanged across groups and trials. Although our results demonstrated that bottom-up manipulations are capable of reducing coupling to visual information even more strongly compared to cognitive manipulations, both adaptive processes did not reduce visual influence to a minimum. This result, which corroborates previous findings (e.g., [1]), demonstrates that even when if the visual sensory channel provides inaccurate
10
J.A. Barela et al. / Neuroscience Letters 566 (2014) 6–10
information, subjects are still not able to completely avoid the influence of visual stimulus on postural responses. Our results revealed important aspects of the coupling between visual information and body sway but also changes in the overall performance of the postural control system under the manipulations applied. Mean sway amplitude decreased equally when participants were informed about the room’s movement and when the velocity and amplitude of this movement increased, suggesting a conservative strategy of the postural control system in controlling body oscillation after a perturbation occurred, either implicitly or explicitly. In sum, our results revealed that postural control shows adaptive re-weighting, when knowledge about the visual information is gathered explicitly and implicitly, but through different adaptive processes. The functional role of re-weighting processes is an issue that deserves attention in future studies, since daily activity performed by humans are affected by an environment that constantly changes, challenging postural control and many other motor actions. References [1] A.M. Barela, J.A. Barela, N.M. Rinaldi, D.R. de Toledo, Influence of imposed optic flow characteristics and intention on postural responses, Motor Control 13 (2009) 119–129. [2] J.A. Barela, G.M. Focks, T. Hilgeholt, A.M. Barela, R.de P. Carvalho, G.J. Savelsbergh, Perception-action and adaptation in postural control of children and adolescents with cerebral palsy, Research in Developmental Disabilities 32 (2011) 2075–2083. [3] H. Forssberg, L.M. Nashner, Ontogenetic development of postural control in man: adaptation to altered support and visual conditions during stance, Journal of Neuroscience 2 (1982) 545–552. [4] P.B. Freitas Junior, J.A. Barela, Postural control as a function of self- and objectmotion perception, Neuroscience Letters 369 (2004) 64–68.
[5] P.B. Freitas Júnior, J.A. Barela, Postural control as a function of self-and objectmotion perception, Neuroscience Letters 369 (2004) 64–68. [6] F.B. Horak, H.C. Diener, L.M. Nashner, Influence of central set on human postural responses, Journal of Neurophysiology 62 (1989) 841–853. [7] J.J. Jeka, L.K. Allison, T. Kiemel, The dynamics of visual reweighting in healthy and fall-prone older adults, Journal of Motor Behavior 42 (2010) 197–208. [8] J.J. Jeka, K.S. Oie, T. Kiemel, Asymmetric adaptation with functional advantage in human sensorimotor control, Experimental Brain Research 191 (2008) 453–463. [9] T. Kiemel, K.S. Oie, J.J. Jeka, Multisensory fusion and the stochastic structure of postural sway, Biological Cybernetics 262 (2002) 262–277. [10] T.J. Klein, J. Jeka, T. Kiemel, M.A. Lewis, Navigating sensory conflict in dynamic environments using adaptive state estimation, Biological Cybernetics 105 (2012) 291–304. [11] J.R. Lishman, D.N. Lee, The autonomy of visual kinaesthesis, Perception 2 (1973) 287–294. [12] L.M. Nashner, Adapting reflexes controlling the human posture, Experimental Brain Research 26 (1976) 59–72. [13] L.M. Nashner, F.O. Black, C. Wall, Adaptation to altered support and visual conditions during stance: patients with vestibular deficits, Journal of Neuroscience 2 (1982) 536–544. [14] K.S. Oie, T. Kiemel, J.A. Barela, J.J. Jeka, The dynamics of sensory reweighting: a temporal asymmetry, Gait & Posture 21 (2005) S29. [15] K.S. Oie, T. Kiemel, J.J. Jeka, Multisensory fusion: simultaneous re-weighting of vision and touch for the control of human posture, Cognitive Brain Research 14 (2002) 164–176. [16] R.J. Peterka, Sensorimotor integration in human postural control, Journal of Neurophysiology 88 (2002) 1097–1118. [17] P.F. Polastri, J.A. Barela, T. Kiemel, J.J. Jeka, Dynamics of inter-modality reweighting during human postural control, Experimental Brain Research 223 (2012) 99–108. [18] N.M. Rinaldi, P.F. Polastri, J.A. Barela, Age-related changes in postural control sensory reweighting, Neuroscience Letters 467 (2009) 225–229. [19] F. Tjernstrom, P.A. Fransson, M. Patel, M. Magnusson, Postural control and adaptation are influenced by preceding postural challenges, Experimental Brain Research 202 (2010) 613–621. [20] H. van der Kooij, R. Jacobs, B. Koopman, F. van der Helm, An adaptive model of sensory integration in a dynamic environment applied to human stance control, Biological Cybernetics 84 (2001) 103–115.
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Explicit and implicit knowledge of environment states induce adaptation in postural control José A. Barela a,b,∗ , Matthias Weigelt c , Paula F. Polastri d , Daniela Godoi e , Stefane A. Aguiar a , John J. Jeka f a
Institute of Physical Activity and Sport Sciences, Cruzeiro do Sul University, São Paulo, SP, Brazil Institute of Biosciences, São Paulo State University, Rio Claro, SP, Brazil c Department of Sport and Health, Paderborn University, Paderborn, Germany d Department of Physical Education, São Paulo State University, Bauru, SP, Brazil e Department of Physical Education and Human Movement, Federal University of São Carlos, São Carlos, SP, Brazil f Department of Kinesiology, Temple University, Philadelphia, PA, United States b
h i g h l i g h t s • Increased velocity of visual motion decreases coupling to visual information. • Knowledge of visual surrounding manipulation reduces coupling to visual information. • Implicit and explicit knowledge of environment states reduce sway amplitude.
a r t i c l e
i n f o
Article history: Received 13 November 2013 Received in revised form 28 January 2014 Accepted 14 February 2014 Keywords: Sensory re-weighting Adaptive control Posture Vision
a b s t r a c t The aim of this study was to investigate the effects of explicit and implicit knowledge about visual surrounding manipulation on postural responses. Twenty participants divided into two groups, implicit and explicit, remained in upright stance inside a “moving room”. In the fourth trial participants in the explicit group were informed about the movement of the room while participants in the implicit group performed the trial with the room moving at a larger amplitude and higher velocity. Results showed that postural responses to visual manipulation decreased after participants were told that the room was moving as well as after increasing amplitude and velocity of the room, indicating decreased coupling (down-weighting) of the visual influences. Moreover, this decrease was even greater for the implicit group compared to the explicit group. The results demonstrated that conscious knowledge about environmental state changes the coupling to visual information, suggesting a cognitive component related to sensory re-weighting. Reweighting processes were also triggered without awareness of subjects and were even more pronounced compared to the first case. Adaptive re-weighting was shown when knowledge about environmental state was gathered explicitly and implicitly, but through different adaptive processes. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Our daily activities require the control of a stable and (at the same time) flexible upright stance. To achieve this, the nervous system must process a stimulus-rich and continuously changing environment, requiring the ongoing integration of multisensory
∗ Corresponding author at: Instituto de Ciências da Atividade Física e Esporte – ICAFE, Universidade Cruzeiro do Sul, São Paulo, SP 01506-000, Brazil. Tel.: +55 11 3385 3103. E-mail address: [email protected] (J.A. Barela). http://dx.doi.org/10.1016/j.neulet.2014.02.029 0304-3940/© 2014 Elsevier Ireland Ltd. All rights reserved.
information to update our estimate of self-motion. This integration process is related to the mechanism of sensory re-weighting, which is defined as the ability to select and decrease/increase the influence of a specific sensory stimulus on postural control (e.g., [3,13]). Recently, sensory re-weighting has been rigorously demonstrated and uncovered by manipulating the amplitude and velocity of visual stimuli [1,8,18], vibrating ankle muscles [19], and/or changing somatosensory and visual cues simultaneously [17]. In all cases, changes in postural control were observed, that relied on the down- or up-weighting of the sensory cue’s influence on balance (i.e. upright stance). This indicates that the central nervous system modulates the contribution of the available sensory cues
J.A. Barela et al. / Neuroscience Letters 566 (2014) 6–10
based upon the reliability that they would furnish. Such a mechanism has been characterized as nonlinear and much research has been conducted to understand the mechanistic underpinnings of sensory re-weighting [7,9,10,16,20]. Sensory re-weighting in postural control responses has been observed when participants were told about forthcoming changes in sensory cues. Such instruction-based influences of knowledge on sensory re-weighting were first observed by Nashner and colleagues [6,12] and denominated as central settings adjustments of postural responses. In a more recent study, participants were told that a vision manipulation was occurring and subsequently down-weighted the influence of such sensory cue manipulation [4]. Nevertheless, evidences demonstrate that such down-weighting is stimulus-dependent [1]. Adaptation in postural control system regarding sensory reweighting processes has also been observed after amplitude and velocity of visual motion were increased, leading to downweighting of visual stimulus as well [2,8]. Specifically, it has been demonstrated that when the amplitude of the visual scene motion changes from low amplitude to high amplitude, the postural control system responds relatively quick. However, when the direction of the jump is reversed (from high to low) the response is significantly slower [8,14]. Therefore, the postural control system adapts rapidly the coupling to visual information through sensory reweighting processes when the amplitude of visual motion is high because this represents a threat to stability. In the case of low visual motion amplitude, since stability is not threatened the system can adjust its functioning slowly. Importantly, adaptation due to stimulus properties changes does not involve conscious, since individuals were never informed about visual manipulation [2,14]. Explicit knowledge about visual surrounding manipulation, acquired when individuals were informed about visual motion and asked to resist its influence, as well as implicit knowledge, acquired when participants were exposed to higher velocity and larger amplitude of visual motion without conscious knowledge of the manipulation, have been demonstrated to decrease coupling to visual information (e.g., [1,8]). While several studies provided evidence to confirm these two adaptation mechanisms, no study has ever investigated differences between these two processes with the purpose to discuss how attention and cognitive aspects related to postural control can affect sensory re-weighting compared to behaviors that occur implicitly, that is, without conscious mechanisms involved. Therefore, the purpose of this study was to compare the effects of implicit and explicit information about environmental state in the coupling between visual information and postural control. We employed the moving room paradigm [5,11,15,16] in which participants’ task was to fixate a visual marker on the wall ahead of them and to maintain an upright stance, while the walls of the room moved sinusoidally on a stationary floor. Participants acquired knowledge of the room’s movement either explicitly, by simply being told that the room was moving, or implicitly, by drastically changing the motion parameters of the room (amplitude and velocity), without informing participants about such a change.
2. Methods 2.1. Subjects Twenty healthy young adults participated in this study, equally divided into two groups, one with explicit and one with implicit knowledge. The explicit group included eight males and two females (M = 22.3; SD = 1.57 years), whereas the implicit group four males and six females (M = 22.0; SD = 2.40 years). The recruited participants were undergraduate or graduate students and all had normal or corrected-to-normal vision. In addition, all of them gave
7
their informed consent prior to participation according to procedures approved by the Institutional Review Board.
2.2. Procedures Participants were asked to maintain upright stance inside a moving room at 1 m away from the frontal wall and to look at a target attached at this frontal wall. The moving room consisted of three walls and a roof (2.1 m long × 2.1 m wide × 2.1 m height), mounted on wheels so that it could be moved back and forth by a servomotor mechanism, while the floor remained motionless. The walls and the roof were white with black stripes painted on the walls, creating a pattern of 42 cm wide vertical white and 22 cm wide vertical black stripes. A 20-Watt fluorescent lamp was attached to the ceiling and used to maintain consistent light throughout data collection. The servomotor mechanism consisted of a controller (Compumotor, Model APEX 6151), a controlled stepper motor (Compumotor, Model N0992GR0NMSN), and an electrical cylinder (Compumotor, Model EC3-X3xxN-10004a-Ms1-MT1M), which connected the servomotor to the moving room’s structure. Specialized software (Compumotor, Motion Architect for Windows) controlled the servomotor mechanism, moving the room continuously away from and toward the participant (anterior/posterior direction). A movement analysis system (OPTOTRAK 3020 – 3D Motion Measurement System) was placed behind the participants. One OPTOTRAK IRED marker was placed on the participant’s back (at the 8th thoracic vertebra level) and another one on the frontal wall of the moving room. These markers provided information about the participant’s trunk sway and the moving room’s displacement, respectively, in the anterior–posterior, medial–lateral, and vertical directions, with a sampling rate of 100 Hz. For each participant, 7 trials of 60 s were collected. The first three trials were named pre-change, and the room was oscillated with peak-to-peak amplitude of 0.5 cm, peak velocity of 0.6 cm/s, and frequency of 0.2 Hz. In the fourth trial, named change trial, different manipulations were applied to each group. In the explicit group, participants were verbally (explicitly) informed that the room was moving before the fourth trial started and were asked to resist to this movement (while the parameters of the room’s movement remained the same as used in previous trials during the fourth trial). In the implicit group, the parameters of the room’s movement were changed in the fourth trial to peak-to-peak amplitude of 3.5 cm and peak velocity of 3.5 cm/s. Importantly, no further verbal (explicit) information was given to the participants of the implicit group about the room’s movement. In the remaining trials (Trials 5–7), named post-change trials, the room was oscillated using the same parameter values of the pre-change trials (Trials 1–3), i.e., with peak-to-peak amplitude of 0.5 cm and a peak velocity of 0.6 cm/s. Thus, the difference between the two groups was related to the change trial (Trial 4), in which the explicit group was told about the room’s movement (with the oscillation parameters remaining the same), while the implicit group experienced a change in the oscillation parameter, but was not informed about the room’s movement. When participants in the explicit group were informed about room movement, before the fourth trial, they were explicitly told that the room had been moving in the previous trials and would continue to move in the next trials during the entire experiment. In order to assure participants would be unaware of the movement of the room at the beginning of the experiment, the wheels in which the room was mounted were covered in a way participants could not see them and a random sound (white noise) was used during the entire experiment to mask any possible sound produced by the motor that moved the room. None of the participants
8
J.A. Barela et al. / Neuroscience Letters 566 (2014) 6–10
in both groups reported to have noticed room movement during pre-change trials. 2.3. Data analysis Analyses of body sway magnitude and the relationship between body sway and the moving room was performed only in the anterior–posterior direction, since this was the direction of the manipulation applied. First, mean trunk sway amplitude was calculated by obtaining the standard deviation of trunk position after the average trunk position and a first order polynomial were subtracted from the trunk position time series. The coupling between visual information and trunk sway was examined through the variables gain, phase, stimulus frequency sway amplitude (SFSA) and coherence. To calculate gain and phase, the frequency response function (FRF) was derived from the trunk sway Fourier Transforms divided by the visual stimulus Fourier Transforms. Gain and phase were computed as the absolute value and the argument of the transfer function, respectively, at the driving frequency for each participant across all trials. Gain corresponded to the ratio between the body response amplitude and the visual stimulus amplitude, at the driving stimulus frequency (0.2 Hz). Gain values of one indicate a response amplitude similar to the stimulus amplitude and lower/higher values indicate that the response amplitude was lower/higher than the stimulus driving amplitude. Phase defines the temporal relationship between body sway and the visual stimulus. Positive (negative) phase values indicate sway was ahead (behind) the visual stimulus. In addition, SFSA was calculated as the spectral magnitude of body sway at the frequency of the room motion (0.2 Hz). Finally, coherence measured the strength of the relationship between room movement and body sway. Coherence values close to one indicate strong dependency between these two signals, and coherence values close to zero indicate weak or no dependency between the signals. 2.4. Statistical analysis
Fig. 1. Mean trunk sway amplitude values of young adults in the pre- and postchange trials for both explicit and implicit groups. Error bars represent standard deviation.
strength between the visual stimulus and body sway were found when results between these participants were compared. 3.1. Mean sway amplitude Fig. 1 depicts mean sway amplitude in pre- and post-change trials for both explicit and implicit groups. ANOVA revealed a trial effect on mean sway amplitude, F(1,18) = 9.47, p < 0.01, 2 = 0.345, indicating that trunk sway was reduced in the post- (0.45 cm) compared to the pre-change trials (0.51 cm). No main effect of group, F(1,18) = 0.32, p > 0.05, 2 = 0.018, and no group and trial interaction, F(1,18) = 1.16, p > 0.05, 2 = 0.061, were found. 3.2. Gain and phase Fig. 2 depicts gain and phase between the visual stimulus and trunk sway in pre- and post-change trials for both groups. MANOVA revealed a trial effect, Wilks’ Lambda = 0.207, F(2,17) = 32.51, p < 0.001, 2 = 0.793. The MANOVA also showed a
Three 2 × 2 (Group × Trial) ANOVAs, with repeated measures on the last factor, were carried out to evaluate the effects of group (implicit and explicit) and trial (pre-change and post-change) on the dependent variables mean sway amplitude, SFSA, and coherence. The F-statistic from the averaged FRF values was tested for all groups of trials, revealing that the absolute value and the argument of these values were different from zero (p < 0.02). Following, a 2 × 2 (Group × Trial) MANOVA, with repeated measures on the last factor, was performed to verify the effects of group (implicit and explicit) and trial (pre-change and post-change) on the dependent variables gain and phase. For all variables, the average from the three first trials was used for the pre-change estimate, and the average from the three last trials for the post-change estimate. Post hoc analyses with Tukey’s correction were used to test the main effects. The ˛-level for all analyses was 0.05. 3. Results Participants in the explicit group reported that they did not discriminate the room oscillation, even after the information about the room movement had been provided. Participants in the implicit group were asked, at the end of the experimental session, whether or not they had noticed anything different or uncommon during the experiment. Six participants in the implicit group reported that the room had moved in the fourth trial and remained stationary in all other trials. The other four participants in this group did not report any awareness of the room’s oscillation. Despite these reports, data for all participants were pooled together in the implicit group, because no differences in body sway and in coupling
Fig. 2. Mean gain and phase values of young adults in the pre- and post-change trials for both explicit and implicit groups. Error bars represent standard deviation.
J.A. Barela et al. / Neuroscience Letters 566 (2014) 6–10
9
2 = 0.555, but no group effect, F(1,18) = 3.61, p > 0.001, 2 = 0.167, nor group and trial interaction, F(1,18) = 3.15, p > 0.05, 2 = 0.149, for coherence. Coherence was lower in post-change (0.75) compared to pre-change (0.91) trials.
4. Discussion
Fig. 3. Mean stimulus frequency sway amplitude (SFSA) values of young adults in pre- and post-change trials for both explicit and implicit groups. Error bars represent standard deviation.
group and trial interaction, Wilks’ Lambda = 0.663, F(2,17) = 4.32, p < 0.05, 2 = 0.337, but no group effect, Wilks’ Lambda = 0.946, F(2,17) = 0.48, p > 0.05, 2 = 0.054. Univariate tests indicated a trial effect on gain, F(1,18) = 51.27, p < 0.001, 2 = 0.740. The group and trial interaction for gain was also significant, F(1,18) = 7.53, p < 0.05, 2 = 0.295. Post hoc tests involving this interaction showed that gain values decreased in the post-change trials compared to the prechange trials in both groups. In addition, the post hoc tests revealed that gain values in the post-change trials were lower for the implicit group (0.41) compared to the explicit group (0.60). Univariate analysis for phase values showed no trial effect, F(1,18) = 1.18, p > 0.05, 2 = 0.062, and no group and trial interaction, F(1,18) = 0.05, p > 0.05, 2 = 0.001. Phase values close to zero for all trials indicated that trunk sway was temporally close to the room’s motion. 3.3. SFSA Fig. 3 depicts SFSA in pre- and post-change trials for both groups. ANOVA revealed a trial effect, F(1,18) = 52.56, p < 0.001, 2 = 0.745, and a group and trial interaction, F(1,18) = 7.83, p < 0.02, 2 = 0.303, but no group effect, F(1,18) = 0.69, p > 0.05, 2 = 0.037, for SFSA. Post hoc tests involving this interaction showed that SFSA values decreased in post-change trials compared to pre-change trials, for both groups. In addition, post hoc tests showed that SFSA values in the post-change trials were lower for the implicit group (0.23) compared to the explicit group (0.33). 3.4. Coherence Fig. 4 depicts coherence values in pre- and post-change trials for both groups. ANOVA revealed a trial effect, F(1,18) = 22.40, p < 0.001,
Fig. 4. Mean coherence values of young adults in pre- and post-change trials for both explicit and implicit groups. Error bars represent standard deviation.
The purpose of this study was to evaluate the effects of explicit and implicit knowledge about visual surrounding manipulation on postural responses. Participants who were told about the room’s movement (explicit knowledge) as well as participants who were exposed to higher room movement amplitude and velocity (implicit knowledge) decreased the influence of the visual manipulation on body sway. Moreover, the decreasing was more pronounced in the implicit compared to the explicit group, as evidenced by lower gain and SFSA values in post-change trials. The finding that explicit knowledge about visual surrounding manipulation can influence coupling between visual information and body sway has already been demonstrated [1,5]. Such results support the notion that cognitive components play a role in sensory re-weighting processes, increasing/decreasing the influence of one sensory channel on body oscillation through the conscious effort of a person to achieve a specific goal. This is reminiscent of top-down influences on balance control. At the same time, this study provides evidence for bottom-up influences on postural control. This was indicated by the fact that increasing the amplitude and velocity of the visual stimulus also led to down-weighting the influence of stimulus manipulation on trunk sway, consistent with previous studies [8,14,15]. Such observation indicates that the dynamics of the postural control system can trigger re-weighting processes without awareness of the subject or any ‘conscious’ involvement. In addition, we observed that the down-weighting of visual stimulus’ influence due to implicit knowledge, triggered by stimulus properties manipulation, was more pronounced than the one observed by the explicit knowledge, after the request to resist room motion. Another observation was that some participants in the implicit group, who reported to be aware that the room had been moved, did not show different values in all variables from those participants who were unaware of stimulus change. One would have expected that once the visual stimulus is large enough to be distinguished from the visual movement created by self-motion, it would essentially become explicit. Thus, the down-weighting of visual stimulus after the change trial for some participants in the implicit group could have been due to the same process that led to down-weighting in the explicit group, illustrating an indirect “explicit” effect. However, not only there were no differences in the data of participants who were aware and unaware about room’s movement in the implicit group, but our results also demonstrated that the down-weighting of visual stimulus was stronger in the implicit compared to the explicit group. Therefore, indirect explicit effects or an implicit intention to resist visual motion were not the cause of the down-weighting of visual stimulus in the implicit group. It is clear that the explicit and implicit manipulations are affecting re-weighting to visual information, but through different adaptive processes. Interestingly, adaptation caused by explicit and implicit knowledge of visual manipulation did not affect the temporal relationship between visual stimulus and body sway, since phase values remained unchanged across groups and trials. Although our results demonstrated that bottom-up manipulations are capable of reducing coupling to visual information even more strongly compared to cognitive manipulations, both adaptive processes did not reduce visual influence to a minimum. This result, which corroborates previous findings (e.g., [1]), demonstrates that even when if the visual sensory channel provides inaccurate
10
J.A. Barela et al. / Neuroscience Letters 566 (2014) 6–10
information, subjects are still not able to completely avoid the influence of visual stimulus on postural responses. Our results revealed important aspects of the coupling between visual information and body sway but also changes in the overall performance of the postural control system under the manipulations applied. Mean sway amplitude decreased equally when participants were informed about the room’s movement and when the velocity and amplitude of this movement increased, suggesting a conservative strategy of the postural control system in controlling body oscillation after a perturbation occurred, either implicitly or explicitly. In sum, our results revealed that postural control shows adaptive re-weighting, when knowledge about the visual information is gathered explicitly and implicitly, but through different adaptive processes. The functional role of re-weighting processes is an issue that deserves attention in future studies, since daily activity performed by humans are affected by an environment that constantly changes, challenging postural control and many other motor actions. References [1] A.M. Barela, J.A. Barela, N.M. Rinaldi, D.R. de Toledo, Influence of imposed optic flow characteristics and intention on postural responses, Motor Control 13 (2009) 119–129. [2] J.A. Barela, G.M. Focks, T. Hilgeholt, A.M. Barela, R.de P. Carvalho, G.J. Savelsbergh, Perception-action and adaptation in postural control of children and adolescents with cerebral palsy, Research in Developmental Disabilities 32 (2011) 2075–2083. [3] H. Forssberg, L.M. Nashner, Ontogenetic development of postural control in man: adaptation to altered support and visual conditions during stance, Journal of Neuroscience 2 (1982) 545–552. [4] P.B. Freitas Junior, J.A. Barela, Postural control as a function of self- and objectmotion perception, Neuroscience Letters 369 (2004) 64–68.
[5] P.B. Freitas Júnior, J.A. Barela, Postural control as a function of self-and objectmotion perception, Neuroscience Letters 369 (2004) 64–68. [6] F.B. Horak, H.C. Diener, L.M. Nashner, Influence of central set on human postural responses, Journal of Neurophysiology 62 (1989) 841–853. [7] J.J. Jeka, L.K. Allison, T. Kiemel, The dynamics of visual reweighting in healthy and fall-prone older adults, Journal of Motor Behavior 42 (2010) 197–208. [8] J.J. Jeka, K.S. Oie, T. Kiemel, Asymmetric adaptation with functional advantage in human sensorimotor control, Experimental Brain Research 191 (2008) 453–463. [9] T. Kiemel, K.S. Oie, J.J. Jeka, Multisensory fusion and the stochastic structure of postural sway, Biological Cybernetics 262 (2002) 262–277. [10] T.J. Klein, J. Jeka, T. Kiemel, M.A. Lewis, Navigating sensory conflict in dynamic environments using adaptive state estimation, Biological Cybernetics 105 (2012) 291–304. [11] J.R. Lishman, D.N. Lee, The autonomy of visual kinaesthesis, Perception 2 (1973) 287–294. [12] L.M. Nashner, Adapting reflexes controlling the human posture, Experimental Brain Research 26 (1976) 59–72. [13] L.M. Nashner, F.O. Black, C. Wall, Adaptation to altered support and visual conditions during stance: patients with vestibular deficits, Journal of Neuroscience 2 (1982) 536–544. [14] K.S. Oie, T. Kiemel, J.A. Barela, J.J. Jeka, The dynamics of sensory reweighting: a temporal asymmetry, Gait & Posture 21 (2005) S29. [15] K.S. Oie, T. Kiemel, J.J. Jeka, Multisensory fusion: simultaneous re-weighting of vision and touch for the control of human posture, Cognitive Brain Research 14 (2002) 164–176. [16] R.J. Peterka, Sensorimotor integration in human postural control, Journal of Neurophysiology 88 (2002) 1097–1118. [17] P.F. Polastri, J.A. Barela, T. Kiemel, J.J. Jeka, Dynamics of inter-modality reweighting during human postural control, Experimental Brain Research 223 (2012) 99–108. [18] N.M. Rinaldi, P.F. Polastri, J.A. Barela, Age-related changes in postural control sensory reweighting, Neuroscience Letters 467 (2009) 225–229. [19] F. Tjernstrom, P.A. Fransson, M. Patel, M. Magnusson, Postural control and adaptation are influenced by preceding postural challenges, Experimental Brain Research 202 (2010) 613–621. [20] H. van der Kooij, R. Jacobs, B. Koopman, F. van der Helm, An adaptive model of sensory integration in a dynamic environment applied to human stance control, Biological Cybernetics 84 (2001) 103–115.
