UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Sensorimotor loop delays in the control of human stance Rasman, Brandon Gerald 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


24-ubc_2017_february_rasman_brandon.pdf [ 2.22MB ]
JSON: 24-1.0340025.json
JSON-LD: 24-1.0340025-ld.json
RDF/XML (Pretty): 24-1.0340025-rdf.xml
RDF/JSON: 24-1.0340025-rdf.json
Turtle: 24-1.0340025-turtle.txt
N-Triples: 24-1.0340025-rdf-ntriples.txt
Original Record: 24-1.0340025-source.json
Full Text

Full Text

 SENSORIMOTOR LOOP DELAYS IN THE CONTROL OF HUMAN STANCE  by  Brandon Gerald Rasman  B.KIN., The University of British Columbia, 2014   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF   MASTER OF SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Kinesiology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  November 2016   © Brandon Gerald Rasman, 2016   ii  Abstract Maintaining upright stance involves a time-critical process in which the central nervous system monitors postural orientation and modulates muscle activity accordingly. Visual, vestibular and somatosensory systems detect body motion that the balance controller utilizes to update standing control. The time delays between motor output and the resulting sensory feedback are expected and likely accommodated for. Consequently, we perceive whole-body movement as being a consequence of our own actions. Balance control, however, also involves processes that do not rely on conscious perception, allowing us to maintain standing balance almost effortlessly. Recent studies have demonstrated that both the perception and vestibular control of balance are modulated when sensory signals of whole-body movement do not match self-generated ankle torques. The aim of this thesis was to explore the temporal properties of the sensorimotor loops driving the perception and vestibular control of standing balance.  Using a robotic balance simulator, experimentally-induced time delays were introduced between human participant’s ankle-produced torques and body movement. The first experiment used a psychophysical design to determine what delay is needed for humans to perceive a change in balance control. All participants were able to perceive a 300 ms delay with 100% success, with an average 69% correct threshold of 155 ms. In the second experiment, participants were exposed to a virtual vestibular perturbation while they balanced their body at different induced delays. Vestibular-evoked muscle responses attenuated with increasing loop delays, falling to amplitudes 84% smaller than baseline when a 500 ms delay was introduced between the produced torques and body movement.  This is the first study to explore the time domain relationship between sensory and motor signals in standing, and the results reveal and describe temporal constraints of the sensorimotor iii  control of balance. The present findings will act as springboard for studying postural control mechanisms in the future, encouraging the use of this robotic simulator to alter sensorimotor relationships during ongoing balance control. Using interventions like induced delays, we can decipher the natural processes that govern posture, and explore the adaptability and plasticity of these systems.                        iv  Preface  The protocols used in these studies were reviewed by The University of British Columbia Clinical Research Ethics Board (UBC CREB# H13-01951). All participants provided written informed consent prior to participation in these studies and every effort has been made to ensure that the subjects are not identified in this thesis.  I was the lead investigator on the project, responsible for concept development, data collection and analysis, and document composition. Dr. Ryan M. Peters and Dr. Jean-Sébastien Blouin were involved in concept development. Dr. Jean-Sébastien Blouin was the supervisory author on the project and was involved in the concept formation and thesis revisions. Dr. J. Timothy Inglis, Dr. Romeo Chua, and Dr. Ryan M. Peters were involved in thesis revisions. The experiments contained in this thesis have not been submitted for publication at the time of thesis submission.              v  Table of Contents  Abstract .......................................................................................................................................... ii Preface ........................................................................................................................................... iv Table of Contents .......................................................................................................................... v List of Tables ............................................................................................................................... vii List of Figures ............................................................................................................................. viii Acknowledgements ...................................................................................................................... ix Dedication ...................................................................................................................................... x Introduction ................................................................................................................................... 1 Chapter 1: Sensorimotor control of standing balance............................................................... 4 1.1 – Human standing.................................................................................................................. 4 1. 2 – Internal models for sensorimotor control .......................................................................... 5 1.2.1 – Perception vs action: evidence of separate internal models? ...................................... 8 1.2.2 – Principles governing the internal model of standing control ....................................... 9 1.3 – Sense of agency ................................................................................................................ 11 1.4 – Temporal windows in sensorimotor internal models ....................................................... 12 1.5 – Conscious sense of control in standing ............................................................................ 15 First aim of this thesis: Examining the perceptual sensitivity to sensorimotor loop delays in standing control ..................................................................................................................... 17 1.6 – Vestibular control of standing balance ............................................................................. 18 1.6.1 – Unique integration for the vestibular control of standing balance ............................ 19 Second aim of this thesis: Examining the sensitivity to sensorimotor loop delays in standing for the vestibular control of standing balance....................................................................... 21 Chapter 2: Experiments, analyses, and results ........................................................................ 22 2.1 – General experimental set-up............................................................................................. 22 2.1.1 – Backboard motion control ......................................................................................... 23 2.1.2 – Visual display and control ......................................................................................... 23 2.1.3 – Adjustable time delay ................................................................................................ 26 2.1.4 – Data acquisition ......................................................................................................... 26 2.1.5 – Participants ................................................................................................................ 26 Experiment 1: Perceiving sensorimotor loop delays in standing ........................................ 28 2.2 – Introduction ...................................................................................................................... 28 vi  2.3 – Methods ............................................................................................................................ 29 2.3.1 – Induced delay detection ............................................................................................. 29 2.3.2 – Analysis ..................................................................................................................... 32 2.3.3 – Statistical analysis...................................................................................................... 34 2.4 – Results .............................................................................................................................. 34 2.4.1 – Perceptual thresholds ................................................................................................. 34 2.4.2 – Perceptual response times .......................................................................................... 38 Experiment 2: Sensorimotor delays in stance and the vestibular control of balance ....... 39 2.5 – Introduction ...................................................................................................................... 39 2.5 – Methods ............................................................................................................................ 40 2.5.1 – Participant familiarization with EVS ......................................................................... 40 2.5.2 – Vestibular stimulation................................................................................................ 40 2.5.3 – Data collection ........................................................................................................... 41 2.5.4 – Experimental protocol ............................................................................................... 42 2.5.5 – Signal processing & data reduction ........................................................................... 45 2.5.6 – Dependent measures & statistical analysis ................................................................ 47 2.6 – Results .............................................................................................................................. 48 2.6.1 – Backboard angular position RMS .............................................................................. 48 2.6.2 – Vestibular-evoked balance responses ........................................................................ 50 Chapter 3 – Discussions, interpretations and future directions ............................................. 54 3.1 – Sensorimotor loop delays in standing .............................................................................. 54 3.2 – Perceptual sensitivity to induced sensorimotor loop delays............................................. 55 3.3 – Processes responsible for facilitating and attenuating vestibular control ........................ 58 3.3.1 – An internal model for the vestibular control of balance ............................................ 58 3.3.2 – Subcortical and cortical influences on balance control ............................................. 60 3.4 – Differences between perception and vestibular control of standing balance ................... 62 3.5 – Broader considerations and future directions ................................................................... 65 Conclusion ................................................................................................................................... 68 Bibliography ................................................................................................................................ 69 Appendices ................................................................................................................................... 81 Appendix A: Vestibular-evoked muscle responses ................................................................... 81 Appendix B: Gain and phase of vestibular-evoked muscle responses ...................................... 82 vii   List of Tables  Table 1: Participant perceptual thresholds .............................................................................. 36 Table 2: Mean participant perceptual response times ............................................................ 38                       viii  List of Figures  Figure 1: Robotic balance simulator ......................................................................................... 25 Figure 2: Example of delays presented in an experimental block of trials............................ 31 Figure 3: Raw data of a single subject for experiment 1 ......................................................... 32 Figure 4: Psychometric function for a single participant ........................................................ 35 Figure 5: Psychometric functions for all participants ............................................................. 37 Figure 6: Sample of stimuli and raw data for experiment 2 ................................................... 44 Figure 7: Mean sway RMS ......................................................................................................... 49 Figure 8: Vestibular-evoked muscle responses with induced delays in standing.................. 51 Figure 9: Vestibular-evoked muscle responses for all participants ....................................... 52 Figure 10: Vestibular-evoked muscle response amplitudes .................................................... 53 Figure 11: An internal model for standing balance control .................................................... 55            ix  Acknowledgements  Firstly, I would like thank my supervisory committee. To my mentor and friend, Jean-Sébastien Blouin, thank you for giving me the opportunity to work alongside you. It has been a pleasure and honour to have you as a supervisor. From my first day in the lab as an undergraduate, you have challenged and motivated me. I admire your meticulous attention to detail and creativity. In my eyes, you are the model for what a scientist should be. Beyond academia, I have truly valued our discussions about life and the world. You continue to inspire me to be a better man. To Tim Inglis and Romeo Chua, thank you for instilling in me a love and passion for physiology and neuroscience. Working with and learning from you has been a precious experience. To Ryan Peters, thank you for being a constructive mentor and supportive friend. There is no doubt in my mind that you will have great success for years to come.  Secondly, I’d like to thank my lab mates for all the help and companionship over the years. You all helped in providing an amazing work environment. I’d also like to thank all of my participants for volunteering their time for my experiments. To my true friends, thank you for being generous and supportive, and for being a rowdy crew to have adventures with.  Finally, I’d like to express my gratitude for my family. You all know how much I love you, but I don’t express my appreciation enough. I whole-heartedly say that I could not ask for a better family. Your love and consideration is limitless, and I experience that from you all every day. You mean more to me than anything in the world, and I’m forever grateful and blessed to have you in my life. The time I spend with you is what I cherish the most. I love you.  What a privilege it has been to have relationships with you all. Thank you! x  Dedication It is in human nature to explore the unknown…1  Introduction  The timing of events and how we interpret them is imperative to our ability to perceive and control movement. Our sensory receptors detect stimuli, informing the central nervous system (CNS) of the current state within and around the body. The motor system utilizes this sensory information to update control over our body segments and limbs. This important relationship between sensory and motor signals defines movement control as a sensorimotor process. Sensorimotor loops bear inherent delays which arise from sensory transduction, conduction time, central processing, and motor output (Miall & Wolpert, 1996). The CNS accommodates for these delays to allow for fine control of movement. During movement, motor actions are coupled with sensory consequences that arise in a given time frame. The CNS attributes this sensorimotor relationship as causal, allowing for fine adjustments and smooth ongoing control. For example, when a soccer player swings his or her leg to strike a ball, the CNS sends signals for leg muscle contractions to produce the desired action and expects leg motion to occur in a timely manner. This temporal link allows the CNS to associate the movement as self-generated, giving rise to our conscious sense of agency and precise control of movement. In this way, our ability to move and coordinate our actions in a meaningful manner is limited by the time-sensitive relationship between sensations and motor outputs.  The present research focuses on the temporal relationships of sensorimotor signals in human stance. The ability to stand is often taken for granted, but is nonetheless a complex process. Standing upright requires the CNS to monitor postural position while activating numerous muscles to counteract the effect of gravity on the body. The contraction of muscles ultimately results in forces and torques that are delivered through the feet and cause the whole-body to move. This motion is conveyed to the CNS by the sensory cues arising from multiple 2  systems. In healthy humans, the time that passes between motor output and resulting sensory input is expected and accommodated for. We perceive whole-body movement as being a consequence of our muscle actions. Furthermore, we can maintain standing balance relatively effortlessly. How would these processes be affected if the temporal relationship between sensory and motor cues of standing was altered? If a delay was introduced, could we perceive it or even maintain upright balance? How long of a delay is required to be consciously detected or disrupt balance control? My research focuses on the sensitivity to sensorimotor loop delays for human standing balance. Specifically, I examine how time delays between motor outputs and postural sway affect one’s sense of standing control as well as vestibular balance control during standing.  My research addressed these questions by assessing human participants in standing balance experiments. All experiments, which will be described in detail later, involved the use of a robotic balance simulator which allows the manipulation of human standing dynamics. Through use of this simulator, delays between participant’s ankle-produced torque and the resulting body movement were artificially introduced. Through a behavioural-based analysis, I investigated the temporal parameters of the sensorimotor processes responsible for conscious perception of standing and vestibular balance control. Psychophysical techniques and analyses were used to determine the timing principles for perception of standing control, or sense of standing agency. Correspondingly, a virtual vestibular perturbation signal was provided to participants to assess the presence and features of a postural muscle response, providing insight into the sensorimotor temporal properties driving the vestibular control of standing balance. Therefore, the underlying theme of this thesis was to characterize how humans respond to additional sensorimotor loop delays during upright stance. 3   In the first chapter, I provide an introduction and relevant background on human stance. I then discuss current theories of sensorimotor processing that are believed to be responsible for movement control. Here, I introduce the internal model theory, a concept which has widespread applications and provides the foundation for the questions addressed in this thesis. I then go on to define and discuss the perceptual processes and vestibular balance control that are investigated. In Chapters 2, I present the two experiments conducted in this study and their analyzed results. Chapter 3 discusses the meaning of the results and associates their relevance to sensorimotor control of posture and movement. This thesis concludes with suggested steps to progress investigations in sensorimotor integration and control.               4  Chapter 1: Sensorimotor control of standing balance  1.1 – Human standing  Bipedal stance is mechanically unstable. When the human body is upright, its centre of mass (COM; averaged location of total body mass) is located at roughly 55% of body height, placing the brunt of the load on the legs and feet (Cotton 1931; Woodhull et al. 1985). This body mass and its consequent pull from gravity requires an ongoing modulation of leg muscle activity to keep the COM within the base of support and prevent the body from toppling over (Joseph & Nightingale, 1952; Portnoy & Morin, 1956). Passive contributions from muscles, ligaments, bones and other tissues are not adequate to keep the body upright (Loram & Lakie, 2002; Morasso & Sanguineti, 2002). Therefore, standing is an active process involving control from the CNS (Fitzpatrick et al. 1996; Winter 1995; Peterka 2002). This is primarily achieved through the orchestrating of plantar flexor and dorsiflexor moments, which act to stabilize the COM in the anterio-posterior (AP) direction or sagittal plane. Different combinations of trunk and leg muscles are also involved in controlling the body position in the medio-lateral (ML) direction or frontal plane. Quiet standing, however, is not a static process. The body’s COM is in regular motion as the timing and magnitude of ankle-produced torques change. This continuous movement is referred to as postural sway. For the CNS to control postural sway, it requires a time-sensitive detection of the body state to precisely modulate neural drive to muscles involved in standing upright.  It is understood that the CNS relies on visual, vestibular, somatosensory, and to some extent auditory signals to govern the standing body. Early animal works by Sherrington (1948) demonstrated that these sensory inputs were responsible for automatic reflexes which could 5  collectively function to maintain upright posture. However, most of these observations were made with decerebrate animals, which removes the role of central processing at neural levels above the brainstem. Years after those early investigations, numerous studies showed that providing different combinations of visual (Berthoz et al. 1979; Lee and Lishman, 1975; Lestienne et al. 1977), vestibular (Nashner & Wolfson, 1974; Lund & Broberg, 1983; Day et al. 1997) and somatosensory (Allum et al. 1983; Horak & Nashner, 1986; Johansson et al. 1988; Kavounidias et al. 1999) inflow changed postural sway in the upright human. These findings, as well as many others, revealed the nervous system’s use for these inputs to be far more sophisticated than simple reflexes. As each of these sensory inputs provides information regarding the position of the body, they can be collectively used to represent postural orientation and motion. Indeed, standing has been described as a process that relies on multisensory integration (Gurfinkel & Levick, 1991; Fitzpatrick et al. 1996; Horak & Macpherson 1996; Massion 1992; 1998; Mergner and Rosemeier 1998; van der Kooji 1999; Peterka 2002; Mergner et al. 2005; Héroux et al. 2015). While the details of these theoretical models vary, they share a common feature of the CNS consolidating different sensory cues for standing motor control. Regardless of the precise mechanism, the CNS must overcome the delays associated with sensory signals conveying body position to modulate motor outputs in a timely manner and keep the body balanced.   1. 2 – Internal models for sensorimotor control When dealing with any controller, it is important to consider its functioning within the time dimension. The sensorimotor system contains various delays in sensory transduction (stimulus detection, sensory conduction), central processing (neural computations), and motor output (motor conduction, electromechanical delay). In such, online feedback of movement 6  (negative feedback loop) is delayed and is not useful for the control of rapid movements (Hollerbach 1982; Gerdes & Happee 1994). This problem can be solved through use of a feedforward controller, where motor commands are executed without needing feedback for correction (Keele 1981; Arbib 1981). In this method, however, the accuracy and precision of movements are entirely dependent on the motor commands. Hybrid models of the feedforward and feedback control of movement have been proposed to account for the limitations of each process (Meyer et al. 1988; Milner 1992; Plamondon & Alimi, 1997), but these theories cannot explain how a deafferented patient corrects for errors in movement (Bard et al. 1999). An increasingly accepted concept that has dominated many fields of sensorimotor control is the internal model theory.  The basic premise of the internal model is that the CNS simulates the behaviour of a natural system (eg: movement control of hand) to facilitate finer control and overcome inherent delays in sensory feedback (Wolpert et al. 1995; Miall & Wolpert, 1996; Desmurget & Grafton, 2000). The birth of this theory can be traced back to original proposals that the CNS uses information from the motor command to predict the sensory consequences of self-generated movements and determine error present in actual feedback (von Holst and Middaelstadt, 1950; Sperry 1950; Schmidt 1975). Through this process, the brain compares expected and actual afference to determine whether the sensory inputs are the consequence of self-generated actions (re-afference) or external interventions (ex-afference).  Over the years, evidence for an internal model and its functions in sensorimotor integration have been presented in a variety of fields. Neuronal recordings in animals have provided compelling evidence for the internal model. In the electric fish, a predicted re-afference signal is integrated with actual sensory inflow at the first level of central processing, allowing for 7  an attenuation of self-generated sensory inputs (Bell et al. 1999; 2001). In the non-human primate, neurons in the vestibular nuclei demonstrate a selective encoding to passive head rotations, with a large attenuation to those actively generated (Boyle et al. 1996; McCrea et al. 1999; Roy and Cullen, 2001). Cullen and her colleagues have investigated the nature of this selective processing, using the vestibular system as a framework to explore the CNS mechanisms of sensory attenuation. Importantly, they demonstrate that sensory attenuation does not simply occur in the presence of multiple sensory and motor signals (vestibular, proprioceptive, efference copy of motor command). Rather, the spatial and temporal parameters of these signals must be linked such that actual sensory re-afference matches with the predicted re-afference (Roy and Cullen, 2001; 2004; Brooks and Cullen, 2013; Brooks and Cullen, 2014; see Cullen 2011 for review). For example, when the spatiotemporal dynamics of head movement are artificially altered, vestibular signals associated with self-generated head motion are not attenuated (Brooks and Cullen, 2014). This is presumably a result of an inaccurate prediction of re-afference for the changed sensorimotor environment (Brooks and Cullen, 2014). However, if the non-human primate is exposed to this altered control system over successive head movements, the vestibular signal is suppressed, suggesting that the CNS adapted to the changed dynamics and an accurate prediction becomes available (Brooks and Cullen, 2015). Collectively, these studies provide strong support of an internal model for sensorimotor integration.   In humans, the opportunities to record neurons involved in central processing are sparse. However, numerous behavioural and perceptual studies have reported evidence of the internal model in the human CNS. Humans perceive externally-generated tactile stimuli and physical force to be more ticklish or intense then a self-generated counterpart (Blakemore et al. 1999; Shergill et al. 2003). This may be reflective of the attenuation of self-generated sensory signals. 8  Furthermore, hand and arm reaching movements, which often require rapid adjustments, are believed to rely on feedforward processing (Prablanc and Martin, 1992; Wolpert et al. 1995). The internal model has also been proposed as the basis for human motor learning, in which the CNS adapts to sensory errors by adjusting its predictions and movement patterns (Shadmehr & Mussa-Ivaldi, 1994; Kording & Wolpert, 2004). Indeed, it is generally accepted that the sensorimotor control of movement relies on an internal model (Miall & Wolpert 1996; Wolpert et al. 2011; Franklin & Wolpert 2011).  1.2.1 – Perception vs action: evidence of separate internal models?  The human body is constantly in motion, and the CNS is challenged to determine which motion is self-generated. Many movements are executed volitionally, and are therefore cortically controlled and consciously perceived. That being said, we are clearly not aware of all of our actions during movement. For example, when catching a football, the receiver is conscious of the reaching and hand adjustment that is necessary. However, it is doubtful that a player will focus on the muscles they are activating in the arms and hands, or the compensatory adjustments that are occurring throughout the body to prepare for the catch (e.g.: trunk & leg muscle activity). From the perspective of movement control, it is vital that the CNS generates motor commands for desired whole-body postural states, identifies types of afferent inflow and predicts upcoming states. But what proportion of that information is available to consciousness?   Over the years, many reports have demonstrated a disparity between awareness of movement and movement control. Humans are poor at accurately perceiving the localization of a visual target that changes position during saccadic eye movements, a phenomenon known as saccadic suppression (Bridgeman et al. 1975; 1979). Despite this reduced ability to consciously 9  detect a change in target location during eye saccades, hand pointing or reaching movements to the target are accurate, suggesting the motor system has access to information of the target shift and adjusts accordingly (Bridgeman et al. 1979; Castiello et al. 1991; Goodale et al. 1994). Similarly, participants are poor at realizing that their self-generated movements are deviating when artificially provided visual feedback of the movement is accurate (Fourneret & Jeannerod, 1998). The division between perception and action is further portrayed through the dual process theory of visuomotor control (Milner & Goodale, 1995). In this theory, there is a ventral and dorsal stream of visual processing that are responsible for visuomotor perception (object recognition, function, and understanding) and action (physical interactions with external space), respectively (Milner & Goodale, 1995). This has been experimentally supported in clinical patients. Humans with visual agnosia struggle to comprehend what action is needed to interact with a given object, yet are successful when physically intervening (Goodale et al. 1991; Milner et al. 1991). Comparatively, optic ataxia patients can identify and verbalize the appropriate action and movements required, but fail when attempting (Perenin & Vighetto, 1988).   1.2.2 – Principles governing the internal model of standing control  Internal models are presumed to be at play for standing control, and similar disparities between perception and action have also been observed in human standing. Standing is believed to rely on both cortical and subcortical processes, implying that conscious centres in the brain may not have access to all the sensorimotor information associated with upright posture. Therefore, the cortical mechanisms contributing to perceiving and controlling stance may rely on a set of principles that differ from subcortical processes (Luu 2010; Luu et al. 2012). Regarding 10  the interplay between the motor control of standing and the perception of body posture, two hypotheses have been postulated.  The first suggests that both the perception of stance and balance control are governed by a common internal representation of postural state. This is a reflection of the body schema theory proposed by Victor Gurfinkel. Here, the sensory signals of postural orientation are integrated into an internal framework that is used for a cognitive representation of posture and motor control (Gurfinkel et al. 1988; Gurfinkel & Levick 1991). Therefore, one’s conscious perception arises in parallel with sensorimotor control, and the two processes should reflect one another. In regards to standing, Gurfinkel and his colleagues presented several experimental reports to support this model (Popov et al. 1986; Gurfinkel et al. 1988; Gurfinkel et al. 1989). Notably, spatial transformations of the vestibular-evoked balance response were observed to align with perceived head-on-body position, rather than actual orientation (Popov et al. 1986; Gurfinkel et al. 1989). Others have incorporated this understanding into models of standing control (Massion 1998; Merger & Rosemeier 1998).  The second hypothesis refutes this theory, arguing that perceptual processes and the control of standing balance are not derived from the same internal representation (Dalton et al. 2016). This is supported by reports of discrepancies between conscious postural awareness and balance control. Luu (2010) performed a series of experiments to investigate human perception of force (and torque) exerted through the feet during stance. When participants were asked to voluntarily contract lower limb muscles to replicate the effort perceived during free standing, the torque generated was one-third of that produced in quiet stance (Luu 2010). This is not due to passive torque contributions of standing (Luu 2010). In addition, the corticomuscular coherence (cortex-soleus muscles) in standing was one-third of that exhibited in a torque-equivalent 11  voluntary contraction task (Luu 2010). Taking these findings together, Luu (2010) proposed that a large proportion of the descending motor drive to muscles for standing balance arises subcortically, and is not accessible to conscious perception. Another study assessed perception and vestibular-evoked balance responses when humans participated in robotic balance simulations (Luu et al. 2012). The details of this experiment will be described further in later sections (see 1.5 and 1.6). For the purpose of this discussion, it is important to emphasize that the engagement of the vestibular control of balance was modulated independent of whether humans perceived that they were controlling the motion of their standing bodies (Luu et al. 2012). More recently, Dalton and colleagues (2016) examined the spatial tuning of the vestibular-evoked whole-body balance response under prolonged static head-turned postures. In some conditions, participants exhibited vestibular-evoked balance responses which were not spatially aligned with perceived head orientation (Dalton et al. 2016). It was suggested that the conscious awareness of posture and the spatial transformation of vestibular signals for standing balance do not rely on the same internal representation.   1.3 – Sense of agency Our interactions with the surrounding world are heavily influenced by how actions are perceived. There are two possible origins for a given body movement: 1) self-generated or 2) caused by an external source. When the brain attributes body motion with the former origin, this gives rise to a sense of agency – that I am the source and controller of my actions.  The sense of agency arises from a sensorimotor integration that consolidates actions with outcomes (Frith et al. 2000; Blakemore et al. 2002; Bays et al. 2005; Moore & Fletcher, 2012). Actual sensory inflow (re-afferent and ex-afferent) is relayed to a comparator, where it is determined whether the predicted sensory consequences of the movement match the actual 12  sensory afference (Frith et al. 2000; Blakemore et al. 2002; Frith 2005). This conceptual model suggests that if there is discrepancy between predicted and actual sensory consequences, the movement is classified as externally-driven. In contrast, a close match between signals results in a sense of agency over movement. Therefore, the CNS relies on this comparator model to determine whether a sensory event is self-generated, and thus attributed to one’s motor actions. In reality, the brain’s implementation of this model is not perfect. Because the comparison is dependent on actual and predicted re-afference, the model’s performance is limited by the sensitivity of sensory receptors and accuracy of predictions (Frith 2000). For example, the proprioceptors in the hand have given thresholds for joint movement (Burke et al. 1988; Hall & McCloskey, 1983) such that minute motion may not modulate sensory afferents. In such a case, the comparator would not have access to accurate actual re-afference signals when considering the prediction of re-afference. Similarly, the sensory prediction is limited by the accuracy of the forward model which is challenged to model a plant with nonstationary and nonlinear properties (Franklin & Wolpert, 2011). In either case, the CNS may fail to attribute movement to its appropriate origin (self- or externally-generated) or fail to detect the movement all together. Schizophrenia patients provide a compelling example of the former, in which the CNS has a so-called “misattribution” of movement origin and resulting perceptual delusion of influence (Frith 1994; Daprati et al. 1997; Frank et al. 2001; Frith 2005). This disorder has been experimentally demonstrated to be a result of imprecise predictions of re-afference (Synofizik et al. 2010).   1.4 – Temporal windows in sensorimotor internal models According to the internal model, a movement leads to predicted re-afference and actual re-afference. Because actual re-afference has associated delays, those temporal features must be 13  accounted for within the model (ie: prediction, comparator). These signals of movement are compared for congruency by using the spatial and temporal parameters of the signals. So how does the integration handle signals which are spatially congruent, but temporally incongruent? The important thing to note in this scenario is that the executed movement (and resulting re-afference) is appropriately scaled for spatial parameters (kinematics and kinetics), but the CNS receives this confirmation earlier or later than expected. In this case, the CNS may attribute the delayed sensory feedback as ex-afferent, to some extent.  The brain, however, demonstrates flexibility when processing different sensorimotor delays. Humans can perceive time differences of 20-30 ms between sensory cues (Hirsh & Sherrick, 1961). This sensitivity, however, appears to decrease when sensory signals are self-generated (Haggard et al. 2002; Eagleman & Holcombe, 2002). For example, when a participant attempts to simultaneously move the index finger and ankle joint, the movements are perceived to occur at the same time despite the onset of ankle movement leading the finger by ~20 ms (Paillard 1948). It was proposed that the CNS accommodates for the longer motor pathway for the foot movement to temporally align re-afferent signals (Paillard 1990; Blouin et al. 2004). This preceding movement of the ankle is not observed in a de-afferented patient (Bard et al. 1992), indicating that the absence of actual re-afference limits the sensorimotor system from adapting to temporal differences between sensory consequences of movement. Furthermore, when an auditory tone occurs 250 ms after a voluntary key button press, humans perceive the press to occur 15 ms later and tone 46 ms earlier (Haggard et al. 2002). Thus, it appears that the perceptual system attempts to link the sensory auditory signal closer in time with the causal key press motor command. This “perceptual” or “temporal binding”, such that there is a broadening of the temporal window between expected and actual sensory consequences of movement, has 14  been reported when humans repeatedly experience delayed sensory feedback (Haggard et al. 2002; 2003; Park et al. 2003; Stetson et al. 2006; Heron et al. 2009). It is possible that the CNS adjusts the temporal components of its predictions in order to establish a sense of agency (Haggard et al. 2002; Eagleman & Holcombe, 2002). Correspondingly, there appears to be a broad time window between self-generated movements and actual sensory feedback in which a sensory attenuation for perception can occur. This has been demonstrated when humans report perceived tactile sensations (Blakemore et al. 1999; Bays et al. 2005). When the time span between movement and sensory feedback is artificially increased, humans are more sensitive to the tactile inputs (Blakemore et al. 1999; Bays et al. 2005). An unaltered temporal relationship results in the greatest sensory attenuation, but substantial suppression still occurs up to ~300 ms (Blakemore et al. 1999; Bays et al. 2005). Bays and colleagues (2005) further showed that sensory attenuation can occur when the actual sensory inflow leads or lags the movement, so long as the sensory input is linked to the event and corresponding prediction.  Similar processes appear to be at work for the motor control system. During reaching movements towards a target, accuracy degrades when visual feedback of the movement is artificially displaced (e.g.: through prism goggles), but improves after the participant interacts with the altered feedback through self-generated actions (Welch et al. 1974; Welch 1978). However, when provided with visual feedback regarding the end position of the hand when reaching to a target, humans do not demonstrate the same adaptation to prism-displaced visuomotor feedback if it is delayed beyond 50 ms (Kitazawa et al. 1995; Tanaka et al. 2011). This suggests that the visual re-afference needs to be tightly coupled temporally to the reaching movements for it to be used for motor correction (Kitazawa et al. 1995). Recent evidence, however, has shown that delayed visuomotor feedback can be used for the motor adaptation if 15  the sensorimotor system has the opportunity to calibrate to the delayed signals (Honda et al. 2012; Botzer & Karniel, 2013). Importantly, these studies provided ongoing visuomotor feedback, rather than only at the cessation of the movement as in the previous investigations (Kitazawa et al. 1995; Tanaka et al. 2011). Both studies demonstrated that experiencing delayed feedback over successive movements (> 150 trials) allowed the CNS to adapt to the new temporal relationship between motor output and sensory input, and therefore use the lagging visual feedback to adapt reaching movements to a shifted target. These changes are likely to occur through the updating of the prediction arising from forward sensory models (Honda et al. 2012; Botzer & Karniel, 2013). Important to note, these effects in motor control can occur without comparable changes in perception, and vice versa (Tanaka et al. 2011; Botzer & Karniel, 2013).  Collectively, it appears that both perceptual and motor control systems have thresholds with some flexibility regarding the temporal discrepancy between actual and predicted re-afference. The focus of our discussion will now turn towards the sensorimotor coupling for perception of standing and balance control. I will then propose methods to explore the temporal properties of these relationships.  1.5 – Conscious sense of control in standing Several different theories have been proposed on how a conscious representation of postural state is formed and used by the CNS (Paillard 1991; Gurfinkel & Levick, 1991; Massion 1998; Mergner & Rosemeier 1998). However, there are few descriptions on how a conscious sense of control (or agency) is attributed with bipedal posture. To control efficiently the whole-body in space, a representation of self- and externally-generated movements is required.  16  The study by Luu et al. (2012), which was mentioned earlier (see section 1.2), shed some understanding on this conscious perception. Through a robotic apparatus, participants were tasked to actively balance their body by modulating lower leg muscle activity. In some trials, however, the robot was programmed to move based on a trajectory independent to the participants’ current ankle torques. This motion profile was a replica of previous active balance trials. The participants were kept naïve to the changed condition, and therefore continued to balance their body by modulating ankle torques. Despite the shift of whole-body motion control from human to computer, participants rarely perceived or reported a change in how postural sway was generated (self or external). When informed of the possible shift in control, participants only demonstrated chance level success at perceiving the change. The experimental conditions employed by Luu et al. (2012) have important implications. The human controlled simulations provided sensory signals of whole-body motion (postural sway) that were spatially and temporally (aside from a fixed ~42 ms delay between the computer and robot) congruent with motor commands to lower leg muscles. From an internal model perspective, the sensory inflow can be classified as re-afferent – and a sense of standing control should be present to consciousness. This was the case as participants perceived that they were in control of their balance state. In contrast, the computer-controlled simulations resulted in postural sway afferent cues that were incongruent with leg muscle motor commands. Therefore, this sensory information should be classified ex-afferent to the CNS – and the sense of standing control should be absent, or at least altered. However, this was often not the case, as humans demonstrated a poor ability to perceive the shift in body control. This raises questions about the sensitivity of the sensorimotor integration process that grants conscious sense of standing control. What are the parameters needed for a human to perceive that they are not in control of 17  their upright body? As demonstrated by Luu et al. (2012), simply having an incongruent relationship between motor commands and body motion is not adequate. The brain can be fooled if the motion is somewhat similar to regular standing.  The question takes a different route when we consider the temporal relationship between signals. What if a greater delay was introduced between leg muscle motor commands and postural sway? In this scenario, the sensory and motor signals are still spatially congruent, as body movement is coupled and appropriately scaled to ankle torques. However, the temporal incongruence may challenge the CNS to classify delayed postural sway signals as self-generated. The first experiment of this thesis examines what time delay is tolerated before a human consciously detects a change of balance control.  First aim of this thesis: Examining the perceptual sensitivity to sensorimotor loop delays in standing control Here, I investigate the temporal congruency between motor commands and whole-body motion (re-afference) that is required for a human to consciously perceive that they are controlling their upright body. This question was addressed through a psychophysical experiment. Human participants performed balancing trials on a robotic simulator that moves the whole-body in response to ankle torques. The time delay between ankle torques and postural sway was varied (up to 300 ms) and participants indicated when they perceive a change in the control of body movement. Using this psychophysical approach, I extracted the thresholds of time-coupling needed for a human to consciously perceive a control of standing.  18  1.6 – Vestibular control of standing balance Standing balance can be described as the endeavour and processes involved to keep the body upright and control movement of the COM. To balance the body, the CNS needs to monitor and control the body’s limbs and segments. While this process involves a complex coordination between the different sensorimotor systems, standing feels almost effortless to the healthy human. This is due to the fact that the balance system operates largely at an unconscious and subcortical level (Luu 2010). In particular, assessing the vestibular system’s influence over balance reveals this automatic process of the CNS.  Historically, electrical vestibular stimulation (EVS) has been used to probe and assess vestibular contributions to standing balance (Nashner & Wolfson, 1974; Lund & Broberg 1983; Day et al. 1997; Dakin et al. 2007; for reviews, see Fitzpatrick & Day, 2004; Forbes et al. 2015). By applying electrical current to the mastoid processes, either mon- or binaurally, an artificial vestibular error signal can be produced (Fitzpatrick & Day, 2004). The CNS interprets this signal as an unexpected head movement and produces whole-body motor responses to counteract the ‘virtual’ perturbation (Lund and Broberg 1983; Day et al. 1997; Fitzpatrick and Day, 2004). As its origin is from vestibular signals, postural responses to EVS can be described as a reflection of the vestibular control of standing balance (Guerraz & Day, 2005; Luu et al. 2012).  EVS has been used to investigate fundamental properties of the balance system, and accumulating evidence implies the functioning of a sensorimotor process that considers postural state. Changes in sensory feedback from visual (Smetanin et al. 1990; Fitzpatrick et al. 1994; Day et al. 1997; Welgampola & Colebatch, 2001; Day and Guerraz, 2007), proprioceptive (Lund & Broberg 1983; Horak & Hlavacka, 2001; Day & Cole, 2002; Mian & Day 2009), and foot sole cutaneous (Magnusson et al. 1990; Muise et al. 2012) channels have been shown to modulate the amplitude and spatial transformation of these balance responses. In addition, the balance control 19  system appears to take whole-body stability into its integration, where anisotropic plane stability can modulate the direction of the response to vestibular stimulation (Mian & Day, 2014). Furthermore, vestibular-evoked responses are only present in appendicular muscles that are actively contributing to the motion of the upright body, whereas they disappear if posture is externally supported (Britton et al. 1993; Fitzpatrick et al. 1994; Luu et al. 2012; Forbes et al. 2015). Such intricate modulations emphasize that vestibular-evoked balance responses are not reflective a simple reflex response to vestibular inputs, but rather a system that considers the context of the task at hand. Importantly for the present research, is understanding how the balance system determines if a muscle is influencing the body’s position.  1.6.1 – Unique integration for the vestibular control of standing balance  The investigation of Luu et al. (2012), which was discussed in the previous section on perception, also used their design to study the vestibular control of balance. While the participants performed balancing trials atop the robotic simulator, EVS was delivered to evoke vestibular responses in the soleus muscles. The soleus is a major contributor of plantar flexor torque, and therefore has a strong balance role in normal standing. When the motion of the whole-body was congruent with the participant’s ankle torques (through robotic platform motion), vestibular-evoked responses were observed in the soleus muscle. In contrast, these responses attenuated when the robotic simulator moved the body irrespective of human ankle actions. As mentioned in section 1.2, this modulation of vestibular control was not affected by whether or not the participant was aware they had control over their postural sway. Luu et al. (2012) interpreted these observations by suggesting that the vestibular control of balance 20  operated under its own re-afference principle to determine which muscles actions are congruent with body motion, and transmits vestibular error signals to those muscles accordingly. More recently, Héroux et al. (2015) used a head-velocity coupled EVS signal to alter vestibular afference during human standing. This artificially altered vestibular signal was first interpreted as ex-afferent head motion, causing participants to sway with a greater variability and reduced vestibular muscle responses to an independent ex-afferent EVS signal. Participants were then exposed to a “calibration period” in which they experienced the coupled EVS while standing with their eyes open (or on a stable support surface). This provided the CNS with re-afferent visual (or proprioceptive) cues of motion in addition to the artificial head-velocity coupled vestibular signal. After this exposure, participant’s postural sway reduced and the vestibular-evoked muscle responses to an (ex-afferent) EVS signal increased in amplitude. The authors proposed that a recalibration of a forward vestibular sensory model occurred and permitted the balance controller to identify the artificial head velocity-coupled EVS signal as re-afference.  Both of these studies provide evidence indicating that an internal model is used to modulate the vestibular control of standing balance. The presence and scaling of vestibular-evoked balance responses are dependent on whether the CNS identifies a muscle’s activity as a contributor to upright body motion and balance control (Luu et al. 2012; Héroux et al. 2015). In considering this internal model, we can question the temporal windows in which it operates. The natural question then arises, what temporal relationship is required between muscle activity and postural sway for the balance system to deem a match between actual and predicted re-afference?  21   Second aim of this thesis: Examining the sensitivity to sensorimotor loop delays in standing for the vestibular control of standing balance  The present research investigates the temporal congruency between motor commands and whole-body motion (re-afference) that is required to engage the vestibular control of standing balance. This is achieved by having human participants perform standing balance trials in a robotic balance simulator. While balancing, EVS was delivered to participants in an attempt to evoke vestibular responses in the lower leg muscles. In specific trials, body motion was delayed (up to 500 ms) with respect to ankle-produced torques. In this manner, I examined the level of temporal incongruence between lower limb output and body movement the balance system can tolerate before reducing, and eventually abolishing, the vestibular-evoked muscle response.              22  Chapter 2: Experiments, analyses, and results  2.1 – General experimental set-up A custom-designed robotic balancing simulator (Figure 1) was used for both experiments in this study. The robotic device consists of a foot platform and backboard which are independently controlled by two rotary motors (SCMCS-2ZN3A-YA21, Yaskawa, Japan). Both motors have a resolution of 1048576 encoder readings per revolution (i.e. 0.00034°). Fixed atop the foot platform is a force plate (OR6-7-1000; AMTI, MA, USA) capable of reading forces and torques in all three spatial dimensions. The footplate and backboard can be held stationary or rotated in the anterior-posterior plane (via motor control). Located to the left of the robotic device is a visual projection screen. The screen provides visual cues of motion regarding the balance simulation. The simulator is designed for human interaction as a means to replicate and/or alter the dynamics of standing. The apparatus is programmed with a model of quiet standing balance, specifically, an inverted pendulum with a distributed mass (Luu et al. 2011; Shepherd 2014). The program uses the participant’s mass and height of the centre of mass (COM) to simulate a load experienced during regular standing. Therefore, each participant’s mass and COM height is measured prior to simulations. To determine mass (kg), participants stood relaxed on the force plate and the vertical force (Fz) was recorded. Participants’ COM height was approximated by laying the participant supine on a large board resting on a dowelling (acting as a fulcrum). The board was positioned over the dowelling such that it rested in a balanced state. The participant was then shifted atop the board so that it again became balanced. The measured distance from the dowelling’s centre to the ankle joints was then used for COM 23  location. After a participant’s parameters were entered into the computer program, the balance simulations were run.  2.1.1 – Backboard motion control In all experiments, the foot platform was held stationary while the backboard moved the upright body in response to ankle-produced torques - replicating whole-body standing sway. The balance simulator was controlled by a real-time system (PXI-8119; National Instruments, TX, USA) running at 2000 Hz. Backboard motion was controlled by a participant’s ankle torque (measured by the force plate). The backboard consisted of a metal frame with a lining of medium density foam. The board was adjusted relative to the frame to account for the subject’s natural standing angle. The backboard was rotated in the AP plane about the participant’s ankles with angular limits of 6° anterior and 3° posterior from vertical. If participants crossed these limits, it was considered as a “virtual fall” and they were required to modulate their ankle torques to return to an upright standing position. Soft motion limits of 50°/s and 1000°/s2 were programmed to encompass the physical limits of sway during standing balance (Pospisil et al. 2012). The delay between a motion command and the motor responses was approximately 20 ms (Shepherd 2014). While the simulation ran, the robotic apparatus updated the angle of the inverted pendulum presented to the subject through motion of the backboard and visual cues via the projection screen.  2.1.2 – Visual display and control  The visual projection screen was located 0.77 m to the left of the participant. The screen rested 0.18 m above the ground, and was 2.4 m tall. It is a 55.13° arc of a 4.36 m radius 24  horizontal circle. A projector (W1080ST; BenQ, Taipei, Taiwan) was located 3 m behind the screen at ground level (resolution of 1920x1080p). The visual scene was controlled by a desktop computer with a dedicated graphics card (Quadro K4000, Nvidia, CA, USA), which received commands and angle updates from the real time system over a network connection. The selected visual projection produced the scene of a city courtyard, with a water fountain situated adjacent to the standing participant (Vizard 20137; WorldViz, CA, USA). Rendering and projection of the visual scene took approximately 70 ms, 50 ms longer than the response time of the motors. A linear least-squares predictor algorithm was used to predict the visual angle and synchronize the visual motion with the motors at a delay of 20 ms (Shepherd 2014). Participants wore active 3D glasses (DLP Link 3D Glasses; BenQ, Taipei, Taiwan), modified to block out peripheral vision and limit the participant’s field of view to approximately ± 45° horizontally and ± 30° vertically. Participants also wore earplugs and noise cancelling headphones (with audio of a water fountain) to remove noise produced by the motors as well as other extraneous sounds.    25   Figure 1: Robotic balance simulator In all experiments, participants stood barefoot atop of a force plate that was fixed to the footplate which was held stationary. The torso was secured to the backboard with two seatbelt straps – one across the chest and upper arms, and the other across the waist. As the torques measured at the force plate changed, the backboard motor rotated the backboard (and attached body) in the anterio-posterior plane. The visual display scene (not depicted here) simulated that of the participant standing in a courtyard next to a water fountain. Not depicted in the figure, all participants wore 3D goggles, ear plugs, and noise-cancelling headphones (with audio of a fountain). Image provided by Myles Shepherd and modified with permission.   Figure 1Figure 2: Robotic Balance Simulator  In all experiments, participants stand barefoot atop of a force plate that is fixed to the footplate which will be held stationary. The torso is secured to the backboard with two 26  2.1.3 – Adjustable time delay As previously stated, there was a 20 ms delay between motion commands and motor responses driving the robotic simulation, and this constant delay was unavoidable without the use of a predictive model. Based on the interests for the present research, the balance simulation was modified so that the delay could be adjusted through the computer program. Due to the computer-to-motor delay (20 ms), any delay entered into the program must be added to the constant delay to determine the total delay value (ie: 80 ms program delay + 20 ms constant delay = 100 ms total delay).  In both experiments, participants were informed that in some trials the robotic simulation may have a “change of control, in which the body’s movement may seem abnormal and standing balance may become more difficult.” Participants were not explicitly informed of the experimentally-induced delays.   2.1.4 – Data acquisition Different data were recorded depending on the experiment. The data that could be recorded included: backboard angular position, ground reaction forces and torques, electromyography (EMG), electrical vestibular stimuli (EVS), a button switch signal, and the time delay signal. Force plate, EMG, EVS, and button switch data were lowpass filtered at 40 kHz prior to being digitized at 2000 Hz (PXI-6289; National Instruments, TX, USA).  2.1.5 – Participants  Preliminary testing with the robotic simulation suggested that participants could adapt and improve their balance performance (reduced sway and sway variability) with additional 27  delays in the balance simulation. This became noticeable when the user was exposed to the delays for longer durations (> 60 s), especially at delays less than or equal to 200 ms. To minimize potential adaptation between testing experiments 1 & 2, different participants were recruited for the experiments in this study.  Twenty healthy participants (Experiment 1: 8 men, 2 women, age: 25.4 (SD: 3.5), weight: 70.6 kg (SD: 11.9), height: 174.2 cm (SD: 7.4); Experiment 2: 8 men, 2 women, age: 25.2 (SD: 3.5), weight: 77.6 kg (SD: 10.6), height: 175.8 cm (SD: 6.3) with no known history of neurological or vestibular deficits participated in this study. The experimental protocol was verbally explained and written and informed consent was obtained. The experiments were approved by the University of British Columbia ethics committee (No. H13-01951) and were conducted in accordance with the ethical guidelines set forth by the Declaration of Helsinki.   All participants completed familiarization sessions on the robotic standing balance simulator prior to participating in the experiments. This familiarization involved performing the normal (baseline) standing condition for two sessions of 3 minutes.    All descriptive and statistics in this work will be presented as mean ± the standard deviation.      28  Experiment 1: Perceiving sensorimotor loop delays in standing 2.2 – Introduction The first experiment of this thesis aimed to describe human perceptual sensitivity to the sensorimotor loop delays associated with standing. Recognizing how our motor actions influence body position is fundamental to consciously controlling movement. While standing, muscle actions cause the body to move creating re-afferent postural sway signals. When an external source perturbs the upright body, the movement elicits ex-afferent sensory signals and should be corrected for. Similarly, whole-body motion that is spatially and temporally incongruent with self-generated muscle actions should be identified as ex-afferent. Therefore, in standing, humans should be able to perceive instances when body motion is not associated with regular standing control. Despite this fact, humans are poor at perceiving an external source (robotic) of control when the sway is similar to normal standing (Luu et al. 2012). How sensitive are humans at perceiving changes between ankle-produced torques and whole-body movement during standing when the only alteration is in the time domain? In this first experiment, I used a psychophysical approach to vary the temporal lag between the mechanical actions of the participant and the robot-induced whole-body movements (output of the simulation). Participant responses were used to determine the threshold of experimentally-induced delays in postural sway needed for humans to detect a change in standing control. To my knowledge, this is the first study to investigate ongoing perception of standing control while the sensorimotor delay is altered. Several studies, however, have artificially increased the delay of sensory feedback in different sensorimotor tasks, such as hand reaching and perception of tactile sensation (Kitazawa et al. 1995; Blakemore et al. 1999; Haggard et al. 2002; Bays et al. 2005; Tanaka et al. 2011; Honda et al. 2012). Importantly, a common finding from these experiments was that delays ≥ 100 ms 29  resulted in sensorimotor discrepancies that were consciously perceived and affected the task at hand. For example, induced 100 ms delays in sensory feedback reduced the attenuation of self-generated tactile sensations (Blakemore et al. 1999; Bays et al. 2005) and limited prism adaptation during pointing (Kitazawa et al. 1995; Tanaka et al. 2011). Taking these findings into account, I hypothesized that perceptual thresholds to delays during standing would be ≥ 100 ms.   2.3 – Methods 2.3.1 – Induced delay detection  In this experiment, participants stood quietly while being exposed to intervals of differing induced delays. Prior to the trials, participants were informed that the balance simulation “may change its control, in which the body’s movement may seem abnormal and standing balance may become more difficult.” Using the robotic apparatus, participants balanced themselves on the robotic inverted pendulum for blocks of trials lasting ~260 s. Participants indicated their perceived balance control through use of a button switch that they held in their hand for the duration of the simulation. Participants were instructed to hold down the button switch when they perceived a change of balance control. Induced delays were presented through a variation of the psychophysical method of constant stimuli. Bayesian-based adaptive procedures are often used to map out psychometric functions for the purpose of identifying thresholds, and hold the benefit of using information from previous trials to adjust stimulus level and efficiently estimate psychometric curves (Kontesevich & Tyler, 1999; Tong et al. 2013; Peters et al. 2015). I originally designed a paradigm with a Bayesian adaptive procedure to identify induced delay detection thresholds. However, preliminary testing revealed a significant caveat with this method in regard to the questions of this thesis. Participants demonstrated very high success at detecting 30  changes of control at delays above 200 ms, which resulted in the majority of trials delivering delay values in the 100-150 ms range (via adaptive procedure’s algorithm). I was interested in comparing the perceptual performance of participants with the vestibular responses at different induced delays. The second (vestibular) experiment included delays up to 500 ms (see below), and a somewhat similar range of values was desired for the first (perceptual) experiment. In such, the Bayesian method was limiting for this study as very few trials are delivered at delays above 200 ms due to participants’ prowess in the Bayesian task. Therefore, I opted for the method of constant stimuli – where 20 trials of each delay could be delivered while still providing data used to estimate psychometric functions and extract accurate thresholds (Gescheider 1997). A maximum delay stimulus of 300 ms was selected because preliminary results demonstrated that participants always perceived the change of control at this induced delay. Six experimentally-induced delays were tested: 50, 100, 150, 200, 250 and 300 ms. In the testing blocks, participants attempted to maintain standing balance on the balance robot. The baseline simulation delay (20 ms) lasted 7-10 s before changing to one of the six experimental delays (50-300 ms) lasting for 8 s (trial presentation), after which it returned to baseline for another 7-10 s. Catch delays of 20 ms were also delivered for 8 s periods within each block of trial. In this manner, 14 delays (2 of each; 6 experimental, 1 catch) were delivered in a random order for each block of trials. An example of the delays delivered in a block of trials is depicted in Figure 2. A total of 10 blocks were performed, providing 20 instances of trial presentation for each experimentally-induced delay.    31       Figure 2: Example of delays presented in an experimental block of trials Six experimentally-induced delays (50, 100, 150, 200, 250, 300 ms) and catch trials (20 ms delay) were presented twice in each block of trials. Participants began the block standing at the baseline simulation (20 ms delay). Further delays were presented to the subject through a change from baseline to the experimental delay value. The presented delay lasted 8 s (represented by red lines) prior to returning to baseline. Delays were presented in a random order across an experimental block of trials. Each participant completed ten experimental blocks, resulting in a total of twenty presentations (or trials) of each delay.  32    2.3.2 – Analysis All data were imported into MATLAB for analysis. First, the button switch recordings from all participant blocks (10 blocks per participant) were used to determine response performance (Figure 3). The following criteria was used to classify a participant’s response. A hit (true positive) was identified if the button switch was on at any time during an 8 s Figure 3: Raw data of a single subject for experiment 1 Participants completed 10 experimental blocks of trials lasting ~260 seconds each. Shown above is 110 seconds of raw data from one block of a single subject. The top plot displays backboard angular position during the balancing task. The bottom plot shows the delay values used for the balance simulation (black line) and participant button switch responses (dashed gray line).  33  experimental delay presentation. A false alarm (false positive) was identified if the button switch was pressed during the baseline delay simulation. False alarms were not assigned to periods in which the button switch was left on for a brief period after an experimental delay presentation. Mean detection rates for each delay were computed by dividing the number of detected control changes (delay trial presented with button switch on) by the total number of delay trials (20 for each delay value), providing a resolution of 5% for each delay level.   Psychometric functions for detecting the induced delays were generated for all participants and for grouped data. The proportion of “correct” responses (ie: when the participant perceived a change of standing control) was plotted with respect to the stimulus intensity (delay value). Using a MATLAB glmfit function, a sigmoid function based on a normal (Gaussian) distribution was fitted to the response rates to obtain a psychometric curve. The curve was interpolated to find the delay amplitude corresponding to 69% correct detection performance (corresponding to a d’ value of 1 in a single-interval detection task). The 69% correct delay amplitude was taken as the participant’s threshold estimate. All participant thresholds were then used for statistical analysis. For illustrative purposes, a psychometric function was similarly generated by averaging all participants’ responses. Response times were also calculated from participant data. Delay detection times were defined as the time between a delay and button switch onsets. Return to baseline balance control detection times were defined as the time between delay offset and the release of the button. To determine premature perceptions of balance control returning to normal, I also counted cases where participants indicated a return to baseline control prior to the induced delay returning to 20 ms.  34  2.3.3 – Statistical analysis   For statistical purposes, data were analyzed using STATISTICA 6.0 (StatSoft) to determine if participant thresholds were ≥ 100 ms. All participant 69% thresholds were used as the dependent variables. These were compared to an expected value of 100 ms using a one-sample t-test (α = 0.05).   2.4 – Results  2.4.1 – Perceptual thresholds  All participants were able to maintain standing balance on the robot with a delay of 20 ms. When the delay value increased, participants balanced with greater sway variability and range, which was observed to be more exaggerated at greater delay values. For example, at 300 ms, participants often incurred a virtual fall (reaching anterior or posterior limits of the balance simulator). Participants indicated a change of balance control by holding a button switch. Every participant was able to perform the task. Detection responses and a psychometric function from a representative participant is presented in Figure 4. The proportion of correct responses increased with induced delays, such that the participant had 100% detection of the 300 ms delay. A similar trend was observed for group data as every participant could detect the 300 ms delay as a change of balance control. Psychometric functions were generated for each participant (Figure 5) and the 69% correct thresholds were extracted. I hypothesized that participants would exhibit perceptual thresholds greater than 100 ms (expected value tested). A one-sample t-test was run with 69% correct thresholds as the dependent variable. Group data demonstrated a mean perceptual threshold of 155 ms (± 31; see Table 1) and the t-test returned a significant effect (t9 = 5.551, p < 0.001), confirming the hypothesis.  35        Figure 4: Psychometric function for a single participant The proportion of correct response detections (% out of 20 delay trials) for a single participant is plotted with respect to the experimentally-induced delay (blue circles). A curve was fitted (black line) to the plotted points using a MATLAB glmfit function. The dashed lines illustrate the interpolation of a 69% correct threshold for induced delays. Note that the delay axis ranges from 0 – 500 ms. This is for illustrative purposes (comparing with Experiment 2), as the maximum delay presented in Experiment 1 was 300 ms. 36  Table 1: Participant perceptual thresholds Participant ID Perceptual detection thresholds (ms) S01 203 S02 170 S03 128 S04 120 S05 154 S06 161 S07 138 S08 128 S09 137 S10 208 Mean 155    37                     Figure 5: Psychometric functions for all participants Individual participant (thin) and mean (thick) fitted psychometric functions. The delay axis ranges from 0 – 500 ms for illustrative purposes. Mean 69% correct threshold extracted from the psychometric functions was 155 ms (SD: 31) 38  2.4.2 – Perceptual response times  When participants were tasked to report each time they perceived a change of standing balance control, the rate of detection varied between the six induced delay levels (demonstrated in psychometric functions, see Figure 5), and the detection times varied as well. Detection times were excluded from the average when participants pressed the button prior to the delay onset, or released the button prior to the delay offset (resulting in a negative value for detection time). Mean detection times ranged from 2.6 to 4.8 seconds for the 300 and 100 ms conditions, respectively (see Table 2). When the induced delay returned to baseline (20 ms) participants required at least ~2 seconds to perceive that balance control had returned to normal (Table 2). For 9.5% of the trials, participants perceived that balance control had returned to normal prior to the simulation returning to a baseline delay. Finally, on average, participants also had false alarms 4% (± 2.8%) of the time, where they indicated that a change of balance control had occurred despite no change in the induced delay. Table 2: Mean participant perceptual response times  Induced delay (ms) 50 100 150 200 250 300 Delay detection time (s) 4.1  ± 1.8 4.8 ± 2.0 4.0 ± 1.9 3.7 ± 1.8 2.8 ± 1.5 2.6 ± 1.3 n 7 60 128 157 189 194 Return to baseline detection time (s) 3.5 ± 2.5 2 ± 1.2 2.2 ± 1.3 2.3 ± 1.2 2.6 ± 1.5 2.4 ± 1.2 n 2 47 108 144 175 182  Data are reported as mean ± SD As responses detection rates varied with stimulus level (see Figures 4 and 5), this resulted in different sample sizes (n) of detection times for different delay trials.   39  Experiment 2: Sensorimotor delays in stance and the vestibular control of balance  2.5 – Introduction The second experiment of this thesis assessed the vestibular control of balance under conditions of induced delay. Previous findings have demonstrated that in order for vestibular signals to evoke postural responses, muscles supporting the body need to be engaged in a balancing task (Britton et al. 1993; Fitzpatrick et al. 1994; Luu et al. 2012). Current theory suggests that the human balance system can attribute muscle actions to balance control when the motor commands (and associated efference copies) are congruent with sensory signals of self-motion (Luu et al. 2012; Héroux et al. 2015). Therefore, based on current understanding of the vestibular control of standing balance, it appears that this vestibular-driven control can be facilitated or attenuated depending on the extent to which the balance system classifies postural motion cues as re-afferent or ex-afferent, respectively. What remains unclear is the temporal decoupling between motor commands and whole-body sway the CNS can tolerate before attenuating, and eventually abolishing, vestibular responses. In this experiment, participants were probed with electrical vestibular stimuli while performing standing trials with varying levels of induced delay. If the engagement of vestibular-evoked muscle responses is dependent on a tight temporal coupling of sensory and motor signals, the balance system should categorize the delayed motion as ex-afferent and consequently restrict vestibular error signals from modulating muscle activity. However, because the delayed motion is still proportional to ankle-produced torques (spatially congruent), the sensory signals which have shorter delays may still be considered congruent and partially re-afferent. Therefore, I hypothesized that the vestibular-evoked muscle responses would attenuate with increasing induced delays, until completely abolishing. Previous studies have found sensory attenuation of self-generated stimuli to be 40  minimal with induced delays ≥ 300 ms (Blakemore et al. 1999; Bays et al. 2005). In light of these findings, I expected that the similar delays would result in the greatest suppression of vestibular responses.  2.5 – Methods  2.5.1 – Participant familiarization with EVS For some, first time exposure to EVS can be startling, nauseating, or simply uncomfortable – leading to a distraction from the task at hand. Therefore, all participants in this experiment had previous exposure to the stimulus being used by completing familiarization session prior to experimental trials. This familiarization involved receiving the stimulus while standing on the robot at its normal (baseline) condition. Minimum familiarization involved a 20 s vestibular stimulus (one trial, see below).  2.5.2 – Vestibular stimulation Binaural bipolar electrical vestibular stimulation was delivered through carbon rubber electrodes (9 cm2) coated with electrode gel (Spectra 360, Parker Laboratories, NJ, USA) and secured over the mastoid processes of each ear. The vestibular stimuli were generated offline with custom-designed computer code (LabVIEW National Instruments, TX, USA). The stimuli were sent as analog signals via a data acquisition board (PXI-6289, National Instruments, Austin, TX, USA) to an isolated constant current unit (DS5, Digitimer, Hertfordshire, England). For this experiment, a short exposure to vestibular stimulus was required to minimize the possible effects of adapting to the induced delays of postural sway. Participants were exposed to a stochastic electrical vestibular stimulus (SVS), a white noise signal that has been filtered to contain a set bandwidth of frequencies. SVS provides several advantages to traditional EVS 41  (square, ramp, sinusoid) counterparts, eliciting muscle and whole-body responses with shorter testing times and increased subject comfort. The lower limb demonstrates vestibular-evoked muscle responses with frequencies up to 25 Hz (Dakin et al. 2007; Dakin et al. 2011). Participants were therefore exposed to 20-second SVS signals with flat power spectrums from 0-25 Hz and peak amplitudes of ± 4.0 mA. Each experimental balance condition was performed four times (see below) resulting in 80 seconds of SVS delivery for each condition. To create an overall 80s signal interval of SVS with well distributed power between 0-25 Hz, four different signals were used with rms ranging from 1.47 to 1.61 mA.  2.5.3 – Data collection  Surface electromyography (EMG) was recorded from the right soleus (SOL) muscle. The surface of the skin was cleaned with an alcohol swab and abraded with gel (Nu-Prep, Weaver and Company, Aurora, CO, United States). Self-adhesive Ag-AgCl surface electrodes (Blue Sensor M, Ambu A/S, Ballerup, Denmark) were positioned over the belly of the muscle in a bipolar configuration, in line with the muscle fibres with an inter-electrode distance of 2 cm centre-to-centre. To reduce electrical noise from the motors, two ground electrodes were used: a nickel plated disc electrode was coated with Spectra 360 gel and secured to the medial malleolus of the tibia, and a Velcro strap electrode was secured around the right lower leg – overlaying the tibial tuberosity and head of the fibula. Surface EMG signals were amplified (×5000, Neurolog, Digitimer Ltd., Hertfordshire, United Kingdom) and band-pass filtered prior to digitization (10-1000 Hz). Force plate signals were amplified ×4000 (MSA-6; AMTI, Watertown, MA, USA) prior to being digitized. EMG, force plate, vestibular stimuli, and backboard angular position 42  signals were digitized and recorded at 2000 Hz via a digital acquisition board (PXI-6289; National Instruments, Austin, TX, USA).  2.5.4 – Experimental protocol  The experiment consisted of a block of trials lasting ~25 s each. All trials were performed with the eyes open (focusing on visual projection screen) and involved SVS delivery. The head was turned left and oriented such that Reid’s plane (from the inferior orbital margin to the external acoustic meatus) was pitched ~19º up from horizontal. This position aligns the virtual head rotation signal evoked by transmastoid electrical stimulation such that it is akin to motion around the roll axis (Fitzpatrick & Day 2004; Day & Fitzpatrick 2005) and oriented in the anterior-posterior plane of the body, which consequently maximizes the soleus muscle response to stimulation (Cathers et al. 2005; Dakin et al. 2007). To monitor head position, the participant wore a headband affixed with a laser pointer directed behind the participant. Prior to the trials, the participant’s head was guided and oriented to the appropriate position and a target marker for the laser was placed on a wall behind and out of sight to the participant. Throughout the trials, head position was monitored and verbal feedback provided to ensure that the desired head position was maintained. All trials involved participants attempting to maintain quiet standing at fixed experimentally-induced delays while being exposed to SVS. Participants were informed that the robotic simulator’s parameters would be changed for some trials and at times feel less stable. They were instructed to do their best to maintain a relaxed upright stance, without focusing on changing their strategy. Six delays were tested: 20 (baseline), 100, 200, 300, 400, and 500 ms. The trials began with a short period of data collection (~3-5 s) while the participant stood quietly 43  at the delay condition. Soon after, the participant was warned of the upcoming stimulus, and probed with SVS (20 s; see Figure 6). Following the completion of the SVS, the robot was halted in preparation for the next trial. Each of the 6 experimental delays were provided 4 times, resulting in a total of 24 trials.  44   Figure 6: Sample of stimuli and raw data for experiment 2 A) Sample of 20-s vestibular stimuli provided in a trial and its power spectrum. B) Participants performed standing balance trials at 6 different delay conditions (20, 100, 200, 300, 400, 500 ms). Presented above are raw data from one subject (all conditions) showing soleus electromyography (EMG) and backboard angular position (°) during vestibular stimulation. 45  2.5.5 – Signal processing & data reduction  After acquisition, all data were imported into MATLAB for processing. EMG data was bidirectionally highpass filtered at 30 Hz (sixth order Butterworth) and rectified (Dakin et al. 2007; Dakin et al. 2014). All four trials for a given delay period were cut to include only data in which SVS was delivered. Trials were cut into segments of 2048 data points and were concatenated to create 76 segments equating to 77.824 s data records for each subject. The RMS of backboard angular position (º) was also calculated for each subject at each condition. For example, for the 100 ms delay condition for one subject, the RMS of backboard angular position was calculated in the period of SVS delivery for each trial (4 trials, 20 s each). The average RMS of those four trials was then used as the RMS value for analysis. The data for individual subjects were then used to evaluate the presence of vestibular-evoked muscle responses and extract dependent variables (vestibular response magnitudes, RMS of backboard position) for statistical analyses. Vestibular-evoked muscle responses were quantified by computing coherence and cross-covariance (described below) estimates between the input vestibular stimulus and output muscle activity. These functions provide an assessment of the relationship between two signals in the frequency and time domains (Halliday et al. 1995; Dakin et al. 2007). These measures were calculated individually for each condition and subject. Data were sectioned into 1.024 s segments prior to calculating the autospectra for the vestibular stimuli and muscle activity as well as the cross-spectra of the vestibular stimuli with muscle activity. This provided 76 segments for a given delay condition for an individual subject. In order to represent group data for descriptive purposes and simplify graphical presentation, data from all subjects were concatenated to create a single pooled data set, providing 760 segments for each delay condition. The resulting spectra were averaged and all calculated spectra had a frequency resolution of 0.97656 Hz. 46  Coherence is a measure of the linear relationship between two signals in the frequency domain. Here, it was calculated to estimate how much of the muscle activity power (EMG) was related to the input vestibular stimulus power (SVS), as well as providing information relating to the timing between the two signals through phase coherence estimates. Similar to a coefficient of determination, coherence ranges from 0 to 1, where a value of 1 indicates a perfect linear relationship between SVS-EMG frequencies and a value of 0 indicates an absence of a relationship (Halliday et al. 1995; Amjad et al. 1997). Coherence was calculated by squaring the cross spectrum (SVS-EMG) and dividing it by the product of the two auto spectra (SVS; EMG). Confidence limits were calculated from the number of 1.024 s segments (single subject = 76; pooled = 760) and coherence values were deemed significantly different from zero when they surpassed the 95% confidence limit (Halliday et al. 1995). Coherence was only used for descriptive purposes in the analysis, providing an indication of the presence and strength of the vestibular stimulus-soleus muscle relationship in the frequency domain. The cross-covariance function was used to evaluate the time domain correlation between the vestibular input stimulus (SVS) and muscle activity (EMG). Cross-covariance functions were computed by taking the inverse Fourier transform of the cross-spectrum (Halliday et al. 1995). The cross-covariance was normalized by the product of vector norms of the vestibular input and muscle activity output. This procedure transforms values into coefficients of correlation (r) ranging from -1 to 1, providing meaningful units of magnitude and allowing comparison between participant responses (Dakin et al. 2010). Normalized 95% confidence intervals were calculated for each condition for each participant’s and pooled cross-covariance to determine where responses were significant (Halliday et al. 1995). Electrical vestibular stimuli-lower limb EMG cross-covariance responses are represented by a biphasic pattern with opposite peaks, defined as 47  short (~60 ms) and medium latency (~100 ms) responses (Dakin et al. 2007; Dakin et al. 2010; Luu et al. 2012). By convention, anode right/cathode left currents are represented as a positive vestibular signal. Therefore, a positive cross-covariance function indicates that anode right currents induced muscle excitation or that anode left currents induced muscle inhibition. For comparison across delay conditions, the peak-to-peak amplitude of the cross-covariance functions were extracted from each participant’s response and used as a measure of response magnitude. When either of the short or medium latency peaks did not surpass the 95% confidence limit, the value of that peak was set to zero.   To further describe the SVS-EMG transfer functions across delay conditions, gain and phase frequency estimates were calculated from the pooled data sets. Gain and phase were calculated only at frequencies exhibiting significant coherence, since both estimates are only meaningful when a relationship between two signals exists. These results are presented in Appendix B.   2.5.6 – Dependent measures & statistical analysis After the coherence and cross-covariance functions were calculated for each participant and condition, the dependent measures were extracted for statistical analysis with (Statistica 6.0, (StatSoft). To determine the effect of the induced delay on the vestibular-evoked muscle response magnitude, a one-way repeated measures ANOVA was performed with peak-to-peak cross-covariance amplitudes as the dependent variable and time delay as the independent variable (20, 100, 200, 300, 400, 500 ms). To evaluate how postural sway changed with condition, a one-way repeated measures ANOVA was also performed with mean RMS of backboard angle (°) as the dependent variable and delay as the independent variable (20, 100, 200, 300, 400, 500 ms). Mauchly’s tests were conducted to determine whether the data violated 48  the assumption of sphericity. If a main effect was detected (p<0.05), its decomposition was performed using a Tukey HSD test to determine which conditions exhibited a significant difference from the baseline. Effect sizes (ES) were calculated using the partial eta-squared method.  2.6 – Results This experiment was designed to determine whether the vestibular control of balance attenuates or supresses its influence over muscle activity with temporally incongruent sensorimotor signals. Ten participants performed a series of ~25 second trials in which they stood on a balance simulator with an induced loop delay (20, 100, 200, 300, 400, 500 ms) between ankle-produced moments and body motion and were probed with SVS. Coherence between vestibular stimulation and soleus muscle activity was used on a descriptive basis, providing a measure of the relationship between the two signals in the frequency domain. The dependent variables analyzed for statistical analysis were: RMS of backboard angular position and peak-to-peak amplitude of the cross-covariance vestibular-muscle response.   2.6.1 – Backboard angular position RMS  A stable upright posture was easily maintained by all subjects at 20 and 100 ms delay conditions. As the delay increased to levels ≥ 200 ms, participants swayed with more variability and would at times incur a virtual fall (hitting the limits of 6° anterior or 3° posterior). Raw data traces of backboard angular position from a representative participant are shown in Figure 6B. The pattern of increasing sway range and variability with increasing delays was demonstrated for grouped data. The RMS of backboard position at delays of 20, 100, 200, 300, 400 and 500 ms 49  yielded means of 1.23, 1.39, 2.60, 3.46, 3.85 and 4.21 degs, respectively. A one-way repeated measures ANOVA revealed that backboard position RMS increased with increasing delays (F5,45 = 102.127, p < 0.00001, ηp2 = 0.910). A Tukey HSD test revealed that backboard RMS did not change between baseline (20 ms) and 100 ms conditions (Tukey HSD: p = 0.943). Backboard angular position RMS values were significantly larger than baseline values at 200, 300, 400, and 500 ms (Tukey HSD: p < 0.001; see Figure 7) conditions, such that the 500 ms delay condition had RMS values 242% greater than baseline.    Figure 7: Mean sway RMS  Mean (n=10) backboard angle RMS values are plotted with respect to the delay condition. Backboard angle RMS at delays of 200 ms and above were significantly different from baseline (20 ms). Error bars are SDs. *p < 0.001.  50  2.6.2 – Vestibular-evoked balance responses Participants were able to balance the robot and maintain an upright position during the baseline (20 ms) condition with ease. For trials with greater delays, participants demonstrated an increased difficulty to stabilize their body, such that they could not maintain upright stance in delay conditions of 400 and 500 ms. Data from a representative participant when balancing at different delays are presented in Figure 8A. At the baseline (20 ms) delay, coherence was significant at frequencies up to 20 Hz. The biphasic vestibular-evoked muscle response (cross-covariance) was largest at 20 and 100 ms delay conditions, with short (64 ms) and medium (103 ms) latency peaks exceeding the 95% confidence interval. As the induced delay reached levels ≥ 200 ms, both the coherence and cross-covariance amplitude attenuated. This trend was also observed in the all participants (Figure 9) and pooled data (Figure8B).       51   Figure 8: Vestibular-evoked muscle responses with induced delays in standing Coherence and cross-covariance measures are shown for one participant (A) and pooled estimates (n = 10) (B) for each delay condition. The horizontal dashed line for coherence plots represents the 95% confidence limit. While 95% confidence intervals differ for each condition, the most conservative confidence interval (from the 500 ms condition) is presented as a dashed line for illustrative purposes. Significant coherence bandwidths up to ~25 Hz were observed for 20, 100, and 200 ms conditions.    52   Figure 9: Vestibular-evoked muscle responses for all participants Coherence and cross-covariance estimates are shown for all participants for each delay condition. The horizontal dashed line for coherence plots represents the 95% confidence limit. Confidence intervals are not presented for cross-covariance functions as the values change for different participants.  53  Peak-to-peak amplitudes of cross-covariance responses at delays of 20, 100, 200, 300, 400 and 500 ms yielded means of 0.117, 0.117, 0.064, 0.039, 0.026 and 0.018, respectively (see Appendix A). Four of the ten participants did not have a vestibular response at 500 ms. A one-way repeated measures ANOVA revealed that vestibular-evoked muscle response magnitudes attenuated with increasing delays (F5,45 = 36.375, p < 0.00001, ηp2 = 0.802). A Tukey HSD test revealed that vestibular response amplitude did not change between baseline (20 ms) and 100 ms conditions (Tukey HSD: p = 1.0). Vestibular responses were significantly smaller than baseline values at 200, 300, 400, and 500 ms (Tukey HSD: p < 0.001; see Figure 10), such that the 500 ms delay condition had responses 84% smaller than baseline.   Figure 10: Vestibular-evoked muscle response amplitudes Mean (n=10) peak-to-peak amplitudes for cross-covariance responses are plotted with respect to the delay condition. Responses at delays of 200 ms and above were significantly different from baseline (20 ms). Error bars are SDs. *p < 0.001. 54  Chapter 3 – Discussions, interpretations and future directions   The control of upright stance is dependent on a tight sensorimotor relationship between ankle torque motor commands and consequent sensations associated with postural sway. This association tasks the CNS with linking sensory and motor signals in the spatial and temporal domains. Focusing on this relationship, the experiments in this thesis investigated how the central sensorimotor integrations governing the control of human stance are affected by additional sensorimotor loop delays.    These experiments targeted two integratory processes of upright stance: 1) conscious perception of standing control and 2) vestibular control of balance; with the aim to reveal temporal properties of sensorimotor integration for these processes.   3.1 – Sensorimotor loop delays in standing The sensorimotor control of stance can be represented by an internal model, in which actual afference and predicted re-afference are aligned for comparison (see Figure 11). In the present experiments, the time between ankle-produced torques and consequent body movement was artificially increased. This occurs downstream from the efference copy which is relayed to the forward model that generates predicted re-afference. As a result, actual and predicted afferent inflow are temporally incongruent, leading to a state estimation error. Here, the error is only in the time domain, as spatial parameters of the signals remain appropriately scaled to the motor command and congruent with the prediction.      55              3.2 – Perceptual sensitivity to induced sensorimotor loop delays  The first experiment evaluated human ability to detect a change of standing balance control when additional delays were added to the sensorimotor loop. Confirming my hypothesis, Figure 11: An internal model for standing balance control Sensory systems detect whole-body movement (actual afference) which is compared with a forward prediction of sensory feedback to the motor command (predicted re-afference) to determine the presence of errors and whether a movement was self-generated or externally-produced. Through the robotic balance simulator, a delay (δ-t) was introduced between ankle-produced moments and resulting body sway. This added delay creates afferent signals which are temporally incongruent with predicted re-afference. Model adapted from Héroux et al. (2015).  56  all participants exhibited perceptual thresholds which were greater than delays of 100 ms. These results suggest that the brain can accommodate for different sensorimotor loop delays without detecting a shift in the control of stance. Seeing as we spend our adult lives with relatively constant sensorimotor neural latencies, it seems surprising that a delay greater than 100 ms is required to be consciously perceived. However, this result should not be viewed as a fault or decrement of CNS, but rather a reflection of flexible sensorimotor processing.  At the perceptual level, it has been proposed that the CNS utilizes the computations from an internal model (namely, a comparison between predicted and actual afference) to establish a conscious sense of agency and control (Frith et al. 2000; Blakemore et al. 2002; Frith 2005). Spatiotemporal congruency between expected and actual re-afference allows the brain to confirm that a movement was self-generated (Frith et al. 2000; Blakemore et al. 2002). Delayed sensory feedback results in a temporal discrepancy between signals and a state estimation error that may lead the CNS to conclude the movement was externally produced. Humans are capable of perceiving temporal differences of 20-30 ms between sensory cues (Hirsh & Sherrick, 1961). However, this perceptual sensitivity is decreased when sensory signals arise from self-generated movements (Haggard et al. 2002; Eagleman & Holcombe, 2002). In a phenomenon described as “perceptual” or “temporal binding”, the temporal window between predicted and actual sensory consequences of movement can be broadened when afference is delayed (Haggard et al. 2002; 2003; Park et al. 2003; Stetson et al. 2006; Heron et al. 2009). For example, when an auditory tone follows a self-produced keypress, the tone is perceived to occur sooner than it is actually presented, with shifts up to ~100 ms (Haggard et al. 2002).  Similar phenomena have been illustrated when subjects are tasked to perceive the intensity of self-generated tactile sensations (Blakemore et al. 1999; Bays et al. 2005). When a participant touches their hand, the tactile 57  stimulus is largely attenuated due to the sensorimotor congruence and cancellation of afferent feedback (Blakemore et al. 1999). As the time between self-movement and tactile feedback is artificially increased, the sensations become more prominent or intense – yet considerable sensory attenuation persists up to ~ 300 ms (Blakemore et al. 1999; Bays et al. 2005). Interestingly, these previous results align with the present findings – as participants had 100% success at detecting a 300 ms delay, suggesting a large incongruence in sensorimotor integration. Collectively, these findings can be explained through the processes of an internal model. When only disrupting signals in time domain, the spatial components of signals match for state estimation, and therefore, a large temporal discrepancy may be required for error detection.   In the current study, participants performed continuous standing balance tasks (~260 seconds) as the delay between ankle torques and postural sway was randomly altered (20-300 ms). They were kept naïve to the changing delays, and were only informed that at times the nature of balance control may change “such that standing balance feels abnormal”. Here, it is important to emphasize that participants were always in control of their body motion. That is, the robotic simulator moved the body in the anterior-posterior plane in proportion to the torques created by the participant. Therefore, whole-body movement was always coupled to human actions, resulting in a causal relationship and spatial congruency between motor commands (and predicted re-afference) and actual afference.  As observed previously, the CNS may have broadened its temporal threshold for sensory congruency in order to establish a sense of agency despite larger than expected sensorimotor loop delays (Haggard et al. 2002; Eagleman & Holcombe, 2002).   Another possibility is to consider the different types of balance control errors that are associated with different delays, and the possibility that humans use these errors to perceive 58  changes. Qualitatively analyzing the data through visual inspection, it was observed that subjects swayed with greater variability and amplitude when large delays were delivered (Figure 3). Indeed, this was observed for all participants. Therefore, it is important to appreciate the uncertainty regarding what information participants used to determine when “balance control had changed”. It is possible that factors relating to postural stability, such as body sway displacement or velocity, were used for perceptual detections. Further analysis and follow-up experiments are needed to describe how postural stability and sway variability is associated with perceptual responses.  3.3 – Processes responsible for facilitating and attenuating vestibular control  Experiment 2 demonstrated an attenuation of vestibular-evoked muscle responses with increasing sensorimotor delays in standing. What mechanism could be responsible for this modulation? There are two possible explanations, which are not mutually exclusive.  3.3.1 – An internal model for the vestibular control of balance One possibility is that the vestibular control of balance operates under the principles of an internal model, gating its control based on sensorimotor congruency. When the upright body is free to move in space, vestibular-evoked responses are found in appendicular muscles directly contributing to the balance state (Britton et al. 1993; Fitzpatrick et al. 1994; Luu et al. 2012). Vestibular error signals are not transmitted to such muscles when whole-body motion is decoupled from motor commands, leading Luu and colleagues (2012) to propose that an internal model relying on visual, vestibular, and somatosensory channels of re-afference is implemented by the CNS to modulate vestibular control. Héroux et al. (2015) expanded this proposed model 59  by using a head-velocity coupled EVS signal to alter the gain of vestibular afference during human standing. This artificial signal provided net vestibular afference that did not match predicted re-afference, resulting in increased postural sway and attenuated vestibular responses to an independent SVS signal. After participants experienced this new vestibular afference with unaltered visual or somatosensory inflow, the vestibular control of balance was facilitated and sway returned to normal levels. More recently, Forbes et al. (2016) used a robotic balance simulator to alter the mechanics of standing as a means to assess how the nervous system associates new relationships between vestibular error signals and balancing motor commands. In one experiment, the relationship between ankle-torque motor commands and body movement was reversed – leading to a consequent reversal of the direction and torque of the vestibular-evoked balance responses. These results provided compelling evidence that the nervous system can rapidly assess changes in standing balance dynamics and transform vestibular signals accordingly for balance control (Forbes et al. 2016). Both of these studies (Héroux et al. 2015; Forbes et al. 2016) supported the views of Luu et al. (2012), reaffirming that vestibular balance control relies on an internal model of sensorimotor integration.  Taking the above findings and proposals, it is possible that vestibular-evoked muscle responses are dependent on the state estimation process discussed in section 3.1 (sensorimotor loop delays): where larger errors result in greater attenuation. This may explain the current findings. As the sensorimotor delay (δ-t) between motor output and sensory detection of postural sway increased, greater state estimation errors between actual and predicted afferent inflow resulted. Consequently, the vestibular-evoked muscle response attenuated with increasing delays.   60  3.3.2 – Subcortical and cortical influences on balance control  Another possible explanation for the observed vestibular response attenuation relates to which sensorimotor processes of the brain are involved in the balancing task. In quiet standing, a large proportion of descending motor drive is believed to originate at subcortical levels, as evident from absent or low level corticomuscular coherence (EEG-lower leg muscles) during standing tasks (Masakado et al. 2008; Luu 2010; Murnaghan et al. 2014). It is possible that subcortical areas may take the reins for balance in typical standing control.  Although it remains unclear which specific brain regions are involved in the vestibular control of balance, it is probable that they include areas in the brainstem (vestibular nuclei, reticular formation) and cerebellum (flocculonodular lobe, vermis, fastigial nuclei) as these anatomical regions are prominent in vestibular processing, sensory integration and postural control (Pompeiano 1967; Batton et al. 1977; Carleton & Carpenter, 1983; Pompeiano et al. 1991; Blakemore et al. 1999; Roy & Cullen, 2001; 2004; Brooks & Cullen, 2013; Brooks et al. 2015). The vestibular nuclei, which represent the first stage of central processing for vestibular signals, are a strong candidate due to their sensory convergence and descending spinal tracts (Carleton & Carpenter, 1983; 1984; Barmack, 2003). Both vestibular and reticular formation nuclei share prominent connections with the cerebellum, a proposed location for sensorimotor integration and control (Raymond et al. 1996; Blakemore et al. 1999; Miall et al. 1993; Bastian, 2006). Importantly, neurons within the vestibular nuclei and cerebellum (rostral fastigial nucleus) have been observed to robustly encode ex-afferent head and body movement signals while cancelling re-afferent self-motion (Boyle et al. 1996; Roy & Cullen, 2001; 2004; Brooks & Cullen, 2013; Brooks et al. 2015) - computations which would be vital for the functioning of an internal model for balance control. Therefore, such subcortical networks would be sensitive to vestibular error inputs (ie: EVS) and consequent vestibular-evoked responses would be expected.  61  Alternatively, when the nature of the balancing task changes from the typical principles of normal standing, cognitive processes and thus cortical centres may spring into action. For example, modulation of cortical neural activity increases when quadrupeds perform a novel postural task, such as standing or walking on a tilted surface (Beloozerova et al. 2005; Karayannidou et al. 2009). Here, the artificially increased sensorimotor loop delay challenges the CNS to adjust to a new temporal relationship between muscle actions and whole-body movement. This atypical balance task may increase cortical demand and require increased motor drive from higher levels of sensorimotor control, such as the frontal and parietal areas (Dum & Strick, 1996; Geyer et al. 2000). I theorize that the vestibular control of balance may be partially suppressed if these centres are primarily controlling the body’s balance state. It is important to acknowledge, however, that there is indirect evidence that cortical areas may indirectly influence the vestibular control of balance through corticobulbar pathways (Marsden et al. 2005), indicating that cortico-vestibular interactions and connections can’t be overlooked (Keizer & Kuypers 1984; 1989; Akbarian et al. 1993; 1994; for review, see Fukushima 1997) when interpreting this balance mechanism.    In summary, a transition of balance control motor drive between subcortical to cortical processes may explain why vestibular responses are attenuated with increasing sensorimotor loop delays. This scenario would not negate the possibility of internal models being involved. On the contrary, we could presume that both subcortical and cortical mechanisms employ internal models for balance control. The uncertainty lies in understanding the features of those two models: sensory channels involved, efference copies accessed, and influence of conscious perception. Future investigations should focus on determining the relationship between subcortical and cortical mechanisms in balance control.  62  3.4 – Differences between perception and vestibular control of standing balance  A long-standing question in sensorimotor control is determining the extent in which conscious perception of body state influences different motor processes. This debate also stands in postural control. In particular, two hypotheses have been proposed to account for how the CNS represents and controls the body: 1) both the perception of postural state and balance control are driven by a common internal representation of postural state and 2) perceptual processes and the control of standing balance are not derived from the same internal representation. Evidence supporting each theory was discussed in Chapter 1, and I will refer the reader to section 1.2.2 to review this discussion.   However, I will revisit the findings of Luu et al. (2012) as the study design has important similarities and differences to the present experiments. In their study, humans balanced on a robotic simulator (of similar design to the one used here) by adjusting their ankle torques. During the balancing trials, control of body movement would randomly switch from human (responding to measured moments) to computer control (following a trajectory independent of current moments). Important to note, the computer trajectories were replicas of previous human control trials, and control transitions would only occur when the robotic platform was in a relatively quiet motion state (within 0.017 rad and 0.00085 rad s-1 of the computer trajectory). Additionally, computer controlled trajectories would only last 4 s prior to returning to human control (which lasted for 8 – 65 s). This design was to provide two sensorimotor conditions: whole-body movement with congruent (human control) and incongruent (computer control) signals. Luu and colleagues (2012) were interested in determining whether the task-dependent nature of the vestibular-evoked balance responses was related to the conscious perception of standing control. Therefore, a group of participants were informed about the study protocol and 63  were tasked to indicate when they perceived a change of balance control by pressing and holding a button switch. Vestibular stimulation was also provided throughout the balance trials.   The results of these experiments found the engagement of the vestibular control of balance to occur independent of perception (Luu et al. 2012). Vestibular-evoked muscle responses were found to attenuate or facilitate between control conditions. In contrast, participants demonstrated poor ability at detecting these changes, with only chance-level success. Furthermore, on average, attenuation and facilitation of vestibular control occurred in less than 1 s whereas perceptual response time to balance control changes took more than 2 s. These findings, along with others presented in Chapter 1 (Luu 2010, Héroux et al. 2015; Dalton et al. 2016) suggest that perception of body state and balance control do not rely on the same internal representation.   What implications do the present findings have on how the CNS represents and governs balance control? Here, induced delays between ankle-produced torques and postural sway were used to create sensorimotor signals which were temporally incongruent. In the first experiment, participants were found to have an average perceptual threshold of ~155 ms to correctly perceive a change in balance control. Perceptual response time varied with delay magnitudes, where the shortest response times were on average 2.6 seconds (300 ms delay). For the balance simulation, a 300 ms induced delay was a robust intervention, causing participants to sway with greater variability (often incurring virtual falls) and detect the change 100% of the time. Still, humans required at least a couple of seconds to detect the changes in balance control, aligning with previous findings of Luu et al. (2012). This may be due to the natural temporal relationship of sensory and motor signals in normal standing. In quiet stance, calf muscle adjustments occur 2.6 times per second (Loram et al. 2005) and the body sways with a mean frequency of 0.12 Hz 64  (Rougier & Farenc, 2000). Therefore, several adjustments in muscle activity – and resulting postural sway – may be needed in order for the conscious brain to associate alterations in the control of standing balance. In the second experiment, significant attenuation of the vestibular-evoked balance response occurred at 200 ms (46% smaller than baseline), but only one participant did not exhibit a significant response. At the extreme condition of 500 ms, pooled coherence was not significant and mean cross-covariance amplitude was 84% smaller than baseline (4 out of 10 participants had no response). Conceptually, it is difficult to directly compare 69% perceptual thresholds with vestibular-evoked muscle responses (coherence and cross-covariance). It is more prudent to extract different information from the psychometric functions. Notably, every participant demonstrated 100% success at detecting a change of balance control at an induced delay of 300 ms. Comparatively, mean vestibular response amplitude was 67% smaller at 300 ms when compared to baseline. Coherence was largely attenuated at 300 ms, with variable points of significant values across the 0-25 Hz bandwidth.   From my results, it appears that a delay of 300 ms between ankle-produced torques and postural sway results in a temporal discrepancy between predicted and actual sensory signals which is large enough to always be perceived. On the other hand, while the vestibular control of balance is attenuated at 300 ms, responses can still be present up to 500 ms. The sensorimotor integration governing vestibular balance control appears to have a greater temporal window for afferent comparisons. Therefore, compared to perception, greater temporal discrepancies can be tolerated for the vestibular control of balance to be engaged. This aligns with the second hypothesis presented for how the CNS represents the body in postural state: where perception and vestibular balance control do not rely on the same internal representation. Moving forward, a logical next step would be to compare the time constants associated with perceptual detection 65  and vestibular response attenuation for a given delay. This could be achieved by performing perceptual trials while vestibular stimuli are provided simultaneously. Perceptual detection time could then be determined, and time-frequency analysis could be used to measure the time constants for vestibular attenuation and facilitation (Blouin et al. 2011; Luu et al. 2012).  3.5 – Broader considerations and future directions   The experiments in this thesis were designed to investigate specific mechanisms of the CNS, specifically, perceptual and vestibular processes of balance control. That being said, I believe the findings have applicable meaning to wider audience.  Firstly, I must emphasize the novelty and advantage of the robotic balance simulator used in the present experiments. To my knowledge, the Sensorimotor Physiology Laboratory at UBC is the only facility which has an apparatus that can have participants controlling whole-body postural state (ie: standing balance) via closed-loop control. This allows investigators to alter the dynamics of postural control (ie: standing balance) in various ways (eg: delays, gain changes). Why is this so advantageous? Historically, many investigations of sensorimotor control have primarily used discrete motor tasks such as reaching, touching and grasping (Prablanc & Martin, 1992; Shadmehr & Mussa-Ivaldi, 1994; Bays et al. 2005; Franklin & Wolpert, 2011; Wolpert et al. 2011). These types of studies have provided foundational findings regarding how the CNS interprets the body’s environment and controls movement. Balance control, however, represents a unique sensorimotor task in that it requires ongoing use and updating of the control system. That is to say - standing balance is a never-ending endeavour as we never reach a “termination point” or a “period of complete steadiness” in free stance. In addition, balance control is known to rely on multiple sensory channels (visual, vestibular, somatosensory, and auditory). Therefore, 66  with the ability to manipulate the dynamics of human balance control, we have a window into investigating ongoing processes of sensorimotor integration and movement control. Regarding delay intervention in particular, further questions can be addressed. As mentioned several times in this thesis, a primary role of motor control systems (represented by internal models) is the ability to adapt to novel sensorimotor environments and principles. Here, I only investigated how the CNS responds to additional sensorimotor loop delays when first exposed to the temporal change. A sensible next step would be to investigate if and how the CNS can adapt to these sensorimotor loop delays, in both perceptual and vestibular balance domains. Indeed, an adaptation protocol would reveal important properties regarding the plasticity of the CNS to deal with altered temporal relationships between motor commands and sensory signals. An adaptation experiment is being conducted in our laboratory. Another consideration of this artificial delay design is its potential application to clinical populations. For example, multiple sclerosis (MS) is a prevalent neurological disease, especially in young adults. One of the debilitating effects of the disease is the impairment of balance control, leading to falls and reduced mobility (Zwibel 2009; Cameron & Lord, 2010; Peterson et al. 2013). While MS results in damage to various parts of the nervous system, a critical deficit from MS demyelination is the slowing of spinal somatosensory conduction (Cameron et al. 2008). In such, additional delays are added to the sensorimotor loop of postural control, resulting in delayed postural responses to perturbations (Cameron et al. 2008) and increased postural sway (Spain et al. 2012; Gera et al. 2016). Studies have sought to design exercise and strength training, as well as motor learning interventions to improve balance control in MS patients (Kasser at al. 2010; Gera et al. 2016). Seeing as MS patients have difficulty dealing with greater sensorimotor loop delays for postural 67  control, future studies should utilize the robotic balance simulator used in the present experiments to adapt MS patients to sensorimotor loop delays.                       68  Conclusion  The experiments in this thesis investigated how additional sensorimotor loop delays in humans standing affect perceptual and vestibular processes of balance control. The first experiment revealed that the perceptual threshold for detecting a change in standing control is 155 ms, while loop delays ≥ 300 ms are needed to be detected with certainty. The second experiment found that vestibular-evoked balance control (elicited through electrical vestibular stimulation) is attenuated as sensorimotor loop delays increase, falling to levels 84% smaller than baseline at a 500 ms delay.  The results from both of these experiments have been interpreted in the context of an internal model of sensorimotor integration. Through a robotic balance simulator, a delay was added between ankle-produced torques and body movement, resulting in a temporal incongruence between actual and predicted afference. For perception, this temporal incongruence is believed to be referred to as a state estimation error – leading to humans detecting that there was a change of standing balance control. Regarding vestibular responses, I suggest that the temporal delay led to a similar state estimation error for the systems that govern the vestibular control of balance. As a consequence, the lower-leg muscle response to vestibular stimulation was partially suppressed.  The results of this thesis provide important insight into balance control systems and sensorimotor integration processes. This is the first study to demonstrate how altering the time domain relationship of sensory and motor signals in standing affects the sensorimotor control of standing balance. Moving forward, investigations should consider designing experiments to test the balance system’s adaptability to modulated sensorimotor loop delays, in both healthy and clinical populations. 69  Bibliography Akbarian S, Grusser OJ, Guldin WO (1993). Corticofugal projections to the vestibular nuclei in squirrel monkeys - further evidence of multiple cortical vestibular fields. J Comp Neurol 332, 89–104  Akbarian S, Grusser OJ, Guldin WO (1994). Corticofugal connections between the cerebral cortex and brainstem vestibular nuclei in the macaque monkey. J Comp Neurol 339, 421–37  Allum JHJ (1983). Organization of stabilizing reflex responses in tibialis anterior muscles following ankle flexion perturbations of standing man. Brain Res 264, 297–301.  Amjad AM, Halliday DM, Rosenberg JR & Conway BA (1997). An extended difference of coherence test for comparing and combining several independent coherence estimates: Theory and application to the study of motor units and physiological tremor. J Neurosci Meth 73, 69–79.  Arbib MA (1981). Perceptual structures and distributed motor control. In Handbook of Physiology, Section 1: The Nervous System (Motor Control) (Vol. 2) (Brooks, W.B., ed.), pp. 1449–1480, Williams and Wilkins.  Bard C, Paillard J, Lajoie Y, Fleury M, Teasdale N, Forget R, Lamarre Y (1992) Role of afferent information in the timing motor commands: a comparative study with a deafferented patient. Neuropsychologia 30, 201–206.  Bard C, Turrell Y, Fleury M, Teasdale N, Lamarre Y, Martin O (1999). Deafferentation and pointing with visual doublestep perturbations. Exp. Brain Res. 125, 410–416.  Barmack NH (2003). Central vestibular system: vestibular nuclei and posterior cerebellum. Brain Res Bull 60, 511-541.  Bastian AJ (2006). Learning to predict the future: the cerebellum adapts feedforward movement control. Curr. Opin. Neurobiol. 16, 645–649.  Batton RR, 3rd, Jayaraman A, Ruggiero D, & Carpenter MB (1977). Fastigial efferent projections in the monkey: an autoradiographic study. J. Comp. Neurol. 174, 281–305  Bays PM, Wolpert DM & Flanagan JR (2005). Perception of the consequences of self-action is temporally tuned and event driven. Current Biology 15, 1125-1128.  Bell CC (2001). Memory-based expectations in electrosensory systems. Curr Opin Neurobiol 11, 481–487.  Bell CC, Han VZ, Sugawara Y & Grant K (1999) Synaptic plasticity in the mormyrid electrosensory lobe. J Exp Biol 202, 1339–1347. 70   Bellozerova IN, Sirota MG, Orlovsky GN & Deliagina TG (2005). Activity of pyramidal tract neurons in the cat during postural corrections. J Neurophysiol 93, 1831-1844.  Berthoz A, Lacour M, Soechting JF & Vidal PP. (1979). The role of vision in the control of posture during linear motion. Prog Brain Res 50, 197–210, 1979.  Blakemore SJ, Frith CD & Wolpert DM (1999). Spatio-temporal prediction modulates the perception of self-produced stimuli. J. Cogn. Neurosci 11, 551–559.  Blakemore SJ, Wolpert DM, & Frith CD (1999). The cerebellum contributes to somatosensory cortical activity during self-produced tactile stimulation. Neuroimage 10, 448–459.  Blakemore SJ, Wolpert DM & Frith CD (2002). Abnormalities in the awareness of action. Trends in Cognitive Sciences 6, 237-242.  Blouin JS, Bard C, Paillard J (2004). Contribution of the cerebellum to self-initiated synchronized movement: a PET study. Exp Brain Res 155, 63-68.  Blouin JS, Dakin CJ, van den Doel K, Chua R, McFadyen BJ & Inglis JT (2011). Extracting phase-dependent human vestibular reflexes during locomotion using both time and frequency correlation approaches. J Appl Physiol 111, 1484–1490.  Botzer L & Karniel A (2013). Feedback and feedforward adaptation to visuomotor delay during reaching and slicing movements. European Journal of Neuroscience 38, 2018-2123.  Boyle R, Belton T & McCrea RA (1996). Responses of identified vestibulospinal neurons to voluntary eye and head movements in the squirrel monkey. Ann NY Acad Sci 781, 244–263.  Britton TC, Day BL, Brown P, Rothwell JC, Thompson PD, & Marsden CD (1993). Postural electromyographic responses in the arm and leg following galvanic vestibular stimulation in man. Exp Brain Res 94, 143-51.  Bridgeman B, Hendry D, & Stark L (1975). Failure to detect displacement of the visual world during saccadic eye movements. Vision Res 15, 719–722.  Bridgeman B, Lewis S, Heit G, Nagle M (1979). Relation between cognitive and motor-oriented systems of visual position perception. Journal of Experimental Psychology: Human Perception and Performance 5, 692-700.  Brooks JX & Cullen KE (2013). The primate cerebellum selectively encodes unexpected self-motion. Current Biology 23, 947–955.   71  Brooks JX & Cullen KE (2014). Early vestibular processing does not discriminate active from passive self-motion if there is a discrepancy between predicted and actual proprioceptive feedback. J Neurophysiol 111, 2465–78.  Brooks JX, Carriot J & Cullen KE (2015). Learning to expect the unexpected: rapid updating in primate cerebellum during voluntary self-motion. Nat. Neurosci 18, 1310–1317.   Burke D, Gandevia SC & Macefield G (1988). Responses to passive movement of receptors in joint, skin and muscle of the human hand. J Physiol 402, 347–361.  Cameron MH, Horak FB, Herndon RR, Bourdette D (2008). Imbalance in multiple sclerosis: a result of slowed spinal somatosensory conduction. Somatosens Mot Res. 25, 113-122.  Cameron MH & Lord S (2010). Postural control in multiple sclerosis: implications for fall prevention. Curr Neurol Neurosci Rep. 10, 407-412.  Carleton SC & Carpenter MB (1983). Afferent and efferent connections of the medial, inferior and lateral vestibular nuclei in the cat and monkey. Brain Res. 278, 29–51.  Carleton SC & Carpenter MB (1984). Distribution of primary vestibular fibers in the brainstem and cerebellum of the monkey. Brain Research. 294, 281-298  Cathers I, Day BL & Fitzpatrick RC (2005). Otolith and canal reflexes in human standing. J Physiol 563, 229–234.  Castiello U, Paulignan Y & Jeannerod M (1991). Temporal dissociation of motor responses and subjective awareness. A study in normal subjects. Brain 114, 2639–2655.  Cotton FS (1931). Studies in the centre of gravity changes. A new method for finding the height of the centre of gravity in man, with some applications. Aust J Exp Biol Med Sci 8, 53-67.   Cullen KE, Brooks JX, Jamali M, Carriot J & Massot C (2011). Internal models of self-motion: computations that suppress vestibular reafference in early vestibular processing. Exp Brain Res 210, 377-388.  Dakin CJ, Lee Son GM, Inglis JT, & Blouin JS (2007). Frequency response of human vestibular reflexes characterized by stochastic stimuli. J. Physiol. 583, 1117–1127.  Dakin CJ, Luu BL, van den Doel K, Inglis JT, & Blouin JS (2010). Frequency-specific modulation of vestibular-evoked sway responses in humans.  J. Neurophysiol. 103, 1048–1056.  Dakin CJ, Inglis JT, & Blouin JS (2011). Short and medium latency muscle responses evoked by electrical vestibular stimulation are a composite of all stimulus frequencies. Exp. Brain Res. 209, 345–354.   72  Dakin CJ, Dalton BH, Luu BL, & Blouin JS (2014). Rectification is required to extract oscillatory envelope modulation from surface electromyographic signals. J. Neurophysiol. 112, 1685-1691.  Dalton BH, Rasman BG, Inglis JT, Blouin JS (2016). The internal representation for head orientation differs for conscious perception and balance control. Submitted. Daprati E, Franck N, Georgieff N, Proust J, Pacherie E, Dalery J & Jeannerod M (1997). Looking for the agent: an investigation into consciousness of action and self-consciousness in schizophrenic patients. Cognition 65, 71–86.  Day BL & Cole J (2002). Vestibular-evoked postural responses in the absence of somatosensory information. Brain 125, 2081–2088.  Day BL & Guerraz M (2007). Feedforward versus feedback modulation of human vestibular-evoked balance responses by visual self-motion information. J Physiol 582, 153–161.  Day BL, Séverac Cauquil A, Bartolomei L, Pastor MA & Lyon IN (1997). Human body-segment tilts induced by galvanic stimulation: a vestibularly driven balance protection mechanism. J Physiol 500, 661– 672.  Desmurget M & Grafton S (2000). Forward modeling allows feedback control for fast reaching movements. Trends Cogn. Sci. 4, 423–431.  Dum RP & Strick PL (1996). Spinal cord terminations of the medial wall motor areas in macaque monkeys. J Neurosci 16, 6513–6525.  Eagleman DM & Holcombe AO (2002). Causality and the perception of time. Trends Cogn Sci 6, 323–325.  Fitzpatrick R, Burke D & Gandevia SC (1994). Task-dependent reflex responses and movement illusions evoked by galvanic vestibular stimulation in standing humans. J Physiol 478, 363–372.  Fitzpatrick R, Burke D & Gandevia SC (1996). Loop gain of reflexes controlling human standing measured with the use of postural and vestibular disturbances. J Neurophysiol 76, 3994–4008.  Fitzpatrick RC & Day BL (2004). Probing the human vestibular system with galvanic stimulation. J Appl Physiol 96, 2301–2316. Fourneret P & Jeannerod M (1998). Limited conscious monitoring of motor performance in normal subjects. Neuropsychologia 36, 1133–1140.   Forbes PA, Dakin CJ, Vardy AN, Happee R, Siegmund GP, Schouten AC & Blouin JS (2013). Frequency response of vestibular reflexes in neck, back and lower limb muscles. J. Neurophysiol. 110, 1869–1881.  73  Forbes PA, Siegmund GP, Schouten AC & Blouin JS (2015). Task, muscle and frequency dependent vestibular control of posture. Front Integr Neurosci 8, 94.   Forbes PA, Luu BL, Van der Loos HFM, Croft E, Inglis JT, Blouin JS (2016). Transformation of vestibular signals for the control of standing in humans. J Neurosci (In press).  Franklin DW & Wolpert DM (2011). Computational mechanisms of sensorimotor control. Neuron 72, 425-442. Frith C. Theory of mind in Schizophrenia. In: David AS, Cutting JC, editors. The neuropsychology of schizophrenia. Hillsdale, NJ: Erlbaum Press; 1994. p. 147–61.  Frith C (2005). The self in action: lessons from delusions of control. Conscious Cogn, 14, 752–70.  Frith CD, Blakemore SJ, & Wolpert DM (2000). Abnormalities in the awareness and control of action. Philosophical Transactions of the Royal Society of London B, 355, 1771–1788.  Fukushima K. (1997). Corticovestibular interactions: anatomy, electrophysiology, and functional considerations. Exp Brain Res 117, 1–16.  Gerdes VG & Happee R (1994). The use of internal representation in fast gold-directed movements: a modeling approach. Biol. Cybern. 70, 513–524.  Gescheider GA. Psychophysics: The Fundamentals (3rd ed.). Mahwah, NJ: Erlbaum, 1997.  Gera G, Fling BW, Van Ooteghem K, Cameron M, Franks JS & Horak FB (2016). Postural motor learning deficits in people with MS in spatial but not temporal control of center of mass. Neurorehabil Neural Repair 8, 722-730.  Geyer S, Matelli M, Luppino G & Zilles K (2000). Functional neuroanatomy of the primate isocortical motor system. Anat Embryol (Berl) 202, 443–474.  Goodale MA, Milner AD, Jakobson LS, & Carey DP (1991). A neurological dissociation between perceiving objects and grasping them. Nature 349, 154–156.  Goodale MA, Jakobson LS, Milner AD, Perrett DI, Benson PJ, & Hietanen JK (1994). The nature and limits of orientation and pattern processing visuomotor control in a visual form agnosic. J. Cogn. Neurosci. 6, 46–56.  Gurfinkel VS, Levik Yu S, Popov KE, & Smetanin BN (1988). Body scheme in the control of postural activity. In Stance and Motion: Facts and Concepts, ed. V. S. Gurfinkel, M. E. Ioffe´, J. Massion and J.-P. Roll. Plenum Press, New York, pp. 185–193.  74  Gurfinkel VS, Popov KE, Smetanin BN, & Shlykov VY (1989). Changes in the direction of vestibulomotor response in the course of adaptation to protracted static head turning in man. Neurophysiology 21, 159 –164.   Gurfinkel VS, & Levick YS (1991). Perceptual and automatic aspects of the postural body scheme. In: J. Paillard (Ed.), Brain and Space, chap. 10, Oxford University Press. Oxford, 147-162.  Haggard P, Clark S, & Kalogeras J (2002). Voluntary action and conscious awareness. Nat Neurosci 5, 382–385.  Hall LA & McCloskey DI (1983). Detections of movements imposed on finger, elbow and shoulder joints. J Physiol 335, 519–533.  Halliday DM, Rosenberg JR, Amjad AM, Breeze P, Conway BA, Farmer SF (1995).  A framework for the analysis of mixed time series/point process data—theory and application to the study of physiological tremor, single motor unit discharges and electromyograms. Prog Biophys Mol Biol 64, 237–278.  Haggard P & Clark S (2003). Intentional action: conscious experience and neural prediction. Conscious Cogn 12, 695–707.  Heron J, Hanson J & Whitaker D (2009). Effect before cause: supramodal recalibration of sensorimotor timing. Plos One 4: e7681.  Héroux ME, Law TCY, Fitzpatrick RC & Blouin JS (2015). Cross-modal calibration of vestibular afference for human balance.  PLoS ONE 10(4): e0124532.  Hirsh IJ & Sherrick CE (1961). Perceived order in different sense modalities. J. Exp. Psychol. 62, 423–432.  Hollerbach, J.M. (1982). Computers, brains and the control of movement. Trends Neurosci. 5, 189–192.  Honda T, Hirashima M, & Nozaki D (2012). Adaptation to visual feedback delay influences visuomotor learning. PLoS ONE 7(5):e37900.  Horak FB & Nashner LM (1986). Central programming of postural movements: adaptation to altered support-surface configurations. J Neurophysiol 55, 1369–1381.  Horak FB & Macpherson JM (1996). Postural orientation and equilibrium. In: LB Rowell and JT Shepherd (eds). Handbook of Physiology, Section 12 Exercise: Regulation and Intergration of Multiple Systems, pp. 255-292. New York: Oxford Univ. Press. Horak FB & Hlavacka F (2001). Somatosensory loss increases vestibulospinal sensitivity. J Neurophysiol 86, 575–585. 75   Johansson R, Magnusson M, & Akesson M (1988). Identification of human postural dynamics. IEEE Trans Biomed Eng 35, 858–869.  Joseph J & Nightingale A (1952). Electromyography of muscles of posture: leg muscles in males. J Physiol 117, 484-491.  Karayannidou A, Beloozerova IN, Zelenin PV, Stoud EE, Sirota MG, Orlovsky GN, Deliagina TG (2009). Activity of pyramidal tract neurons in the cat during standing and walking on an inclined plane. J Physiol 587, 3795-3811.  Kasser SL & Jacobs JV (2014). Understanding and treating balance impairments in multiple sclerosis. J Clin Outcomes Manag. 21, 495-501.  Kitazawa S, Kohno T, & Uka T (1995). Effects of delayed visual information on the rate and amount of prism adaptation in the human. J Neurosci 15, 7644–7652.  Keele SW (1981). Behavioral analysis of movement. In Handbook of Physiology, Section 1: The Nervous System (Motor Control) (Vol. 2) (Brooks, W.B., ed.), pp. 1391–1414, Williams and Wilkins.  Keizer K & Kuypers HGJM (1984). Distribution of corticospinal neurons with collaterals to lower brainstem reticular formation in cat. Exp Brain Res 54, 107–20.  Keizer K & Kuypers HGJM (1989). Distribution of corticospinal neurones with collaterals to the lower brain stem reticular formation in the monkey (Macaca fascicularis). Exp Brain Res 74, 311–18.  Kavounoudias A, Gilhodes JC, Roll R, & Roll JP (1999). From balance regulation to body orientation: two goals for muscle proprioceptive information processing? Exp Brain Res 124, 80–88.  Kontsevich L, Tyler C (1999). Bayesian adaptive estimation of psychometric slope and threshold. Vision Res 39: 2729–2737.  Kording KP & Wolpert DM (2004). Bayesian integration in sensorimotor learning. Nature 427, 244–247.  Lee DN & Lishman JR (1975). Visual proprioceptive control of stance. J Hum Move Studies 1, 87–95.  Lestienne F, Soechting J, & Berthoz A (1977). Postural readjustments induced by linear motion of visual scenes. Exp Brain Res 28, 363–384.  76  Loram ID & Lakie M (2002). Direct measurement of human ankle stiffness during quiet standing: the intrinsic mechanical stiffness is insufficient for stability. J Physiol 545(Pt 3):1041-53.  Loram ID, Maganaris CN & Lakie M (2005). Human postural sway results from frequent, ballistic bias impulses by soleus and gastrocnemius. J Physiol 564, 295–311  Lund S & Broberg C (1983). Effects of different head positions on postural sway in man induced by a reproducible vestibular error signal. Acta Physiol Scand 117, 307–309.  Luu BL, Huryn TP, Van der Loos HF, Croft EA & Blouin JS (2011). Validation of a robotic balance system for investigations in the control of human standing balance. IEEE Trans Neural Syst Rehabil Eng 19, 382–390.  Luu BL, Inglis JT, Huryn TP, Van der Loos HF, Croft EA & Blouin JS (2012). Human standing is modified by an unconscious integration of congruent sensory and motor signals. J Physiol 590, 5783–5794.  Luu BL. Perception, perfusion and posture (doctoral thesis). Sydney, New South Wales, Australia: Univ. of New South Wales, 2010.   Magnusson M, Enbom H, Johansson R, & Wiklund J. Significance of pressor input from the human feet in lateral postural control. The effect of hypothermia on galvanically induced body-sway. Acta Otolaryngol 110, 321–327, 1990.  Marsden JF, Playford ED & Day BL (2005). The vestibular control of balance after stroke. J Neurol Neurosurg Psychiatry 76, 670–678.  Masakado Y, Ushiba J, Tsutsumi N, Takahashi Y, Tomita Y, Kimura A & Liu M (2008). EEG-EMG coherence changes in postural tasks. Electromyogr Clin Neurophysiol 48, 27–33.  Massion J (1998). Postural control systems in developmental perspective. Neuroscience and Biobehavioral Reviews 22(4), 465–472.  McCrea RA, Gdowski GT, Boyle R, & Belton T (1999) Firing behavior of vestibular neurons during active and passive head movements: vestibulo-spinal and other non-eye-movement related neurons. J Neurophysiol 82, 416–428.  Meyer DE, Abrams RA, Kornblum S, Wright CE & Smith JEK (1988). Optimality in human motor performance: ideal control of rapid aimed movements. Psychol. Rev.95, 340−370.  Milner, T.E. (1992) A model for the generation of movements requiring endpoint precision. Neurosci. 49, 487–496.  77  Milner AD, Perrett DI, Johnston RS, Benson PJ, Jordan TR, Heeley DW, Bettucci D, Mortara F, Mutani R, Terazzi E, et al. (1991). Perception and action in visual form agnosia. Brain 114, 405–428.  Milner AD & Goodale MA (1995). The visual brain in action. Oxford: Oxford University Press.  Mergner T & Rosemeier T (1998). Interaction of vestibular, somatosensory and visual signals for postural control and motion perception under terrestrial land microgravity conditions – a conceptual model. Brain Res. Rev 28, 118–135. Maurer C, Mergner T & Peterka RJ (2006). Multisensory control of human upright stance. Exp Brain Res 171, 231–250.  Morasso PG & Sanguineti V (2002) Ankle muscle stiffness alone cannot stabilize balance during quiet standing. J Neurophysiol 88(4), 2157-2162.   Moore JW & Fletcher PC (2012). Sense of agency in health and disease: A review of cue integration approaches. Consciousness and Cognition 21, 59-68.  Miall RC, Weir D, Wolpert DM & Stein J (1993). Is the cerebellum a smith predictor? Journal of Motor Behaviour 25(3), 203-216.   Miall RC & Wolpert DM (1996). Forward models for physiological motor control. Neural Networks 9(8), 1265-1279.    Mian OS & Day BL (2009). Determining the direction of vestibular-evoked balance responses using stochastic vestibular stimulation. J Physiol 587, 2869–2873.  Mian OS & Day BL (2014). Violation of the craniocentricity principle for vestibularly evoked balance responses under conditions of anisotropic stability. J. Neurosci 34(22), 7696-7703.   Muise SB, Lam CK, Bent LR (2012). Reduced input from foot sole skin through cooling differentially modulates the short latency and medium latency vestibular reflex responses to galvanic vestibular stimulation. Exp Brain Res 218, 63-71.  Murnaghan CD, Squair JW, Chua R, Inglis JT, Carpenter MG (2014). Cortical contributions to control of posture during unrestricted and restricted stance. J Neurophysiol 111, 1920-1926.  Nashner LM & Wolfson P (1974). Influence of head position and proprioceptive cues on short latency postural reflexes evoked by galvanic stimulation of the human labyrinth. Brain Res 67, 255–268.  Paillard J (1948) Quelques données psychophysiologiques relatives au déclenchement de la commande motrice. Année Psychologique pp 28–47.  78  Paillard J (1990) Réactif et prédictif: deux modes de gestion du geste de la motricité. In: Nougier V, Blanchi JP (eds) Pratiques sportives et modélisation du geste. Université Joseph-Fourier, Grenoble, pp 13–56.  Paillard J (1991). Motor and representational framing of space. In Brain and space, ed. Paillard J, pp. 163–182. Oxford University Press, Oxford.  Park J, Schlag-Rey M & Schlag J (2003). Voluntary action expands perceived duration of its sensory consequence. Exp Brain Res 149, 527–529.  Perenin MT & Vighetto A (1988). Optic ataxia: a specific disruption in visuo-motor mechanisms. I. Different aspects of the deficit in reaching for objects. Brain 111, 643-674. Peterka RJ (2002). Sensorimotor integration in human postural control. J Neurophysiol 88: 1097–1118 Peters RM, Rasman BG, Inglis JT, Blouin JS (2015). Gain and phase of perceived virtual rotation evoked by electrical vestibular stimuli. J Neurophysiol. 114, 264–273.  Peterson EW, Ari EB, Asano M, Finlayson ML (2013). Fall attributions among middle-aged and older adults with multiple sclerosis. Arch Phys Med Rehabil 94, 890-895.  Pompeiano O (1967). Functional organization of the cerebellar projections to the spinal cord. Progr Brain Res. 25, 282–321.  Pompeiano O, Andre P, Manzoni D (1997). Spatiotemporal response properties of cerebellar Purkinje cells to animal displacement: a population analysis. Neuroscience 81(3), 609–626. Popov KE, Smetanin BN, Gurfinkel VS, Kudinova MP & Shylkov VY (1986). Spatial perception and vestibulomotor responses in man. Neurophysiol 18, 548-554.  Portnoy H & Morin F (1956). Electromyographic study of postural muscles in various positions and movements. Am J Physiol 186, 122-126.  Prablanc C & Martin O (1992). Automatic control during hand reaching at undetected two-dimensional target displacements. Journal of Neurophysiology 67(2), 455-469  Pospisil ER, Luu BL, Blouin JS, Van der Loos HM, & Croft EA (2012). Independent Ankle Motion Control Improves Robotic Balance Simulator. IEEE Engineering in Medicine and Biology Conference, 6487-6491.  Plamondon, R & Alimi, AM (1997). Speed/accuracy trade-offs in target-directed movements. Behav. Brain Sci. 20, 1–21. Raymond JL, Lisberger SG, & Mauk MD (1996). The cerebellum: a neuronal learning machine? Science 272, 1126–1131. 79  Rougier P & Farenc I (2000). Adaptative effects of loss of vision on upright undisturbed stance. Brain Res, 871, 165–174.  Roy JE & Cullen KE (2001). Selective processing of vestibular reafference during self-generated head motion. J Neurosci 21, 2131–2142.  Roy JE & Cullen KE (2004). Dissociating self-generated from passively applied head motion: neural mechanisms in the vestibular nuclei. J. Neurosci. 24, 2102–2111.  Schmidt RA (1975). A schema theory of discrete motor skill learning. Psychological Review 82, 225-260.  Shadmehr R & Mussa-Ivaldi FA (1994). Adaptive representation of dynamics during learning of a motor task. J. Neurosci 14, 3208–3224.  Shepherd Thesis. Intersensory vestibular control of standing balance (masters thesis). Vancouver, British Columbia, Canada: Univ. of British Columbia.  Sherrington CS (1948). The integrative action of the nervous system. Cambridge. The University Press.  Shergill SS, Bays PM, Frith CD, & Wolpert DM (2003). Two eyes for an eye: the neuroscience of force escalation. Science 301, 187.  Smetanin AB, Popov KE & Shlykov VYu (1990). Changes in vestibular postural response determined by information content of visual feedback. Neirofiziologiya 22, 80–87.  Spain RI, St. George RJ, Salarian A, Mancini M, Wagner JM, Horak FB, Bourdette D (2012). Body-worn motion sensors detect balance and gait deficits in people with multiple sclerosis who have normal walking speed. Gait Posture 35, 573-578.  Stetson C, Xu C, Montague P, & Eagleman D (2006). Motor-sensory recalibration leads to an illusory reversal of action and sensation. Neuron 51, 651–659.  Synofzik M, Thier P, Leube DT, Schlotterbeck P, & Lindner A (2010). Misattributions of agency in schizophrenia are based on imprecise predictions about the sensory consequences of one’s actions. Brain 133, 262–271.  Tanaka H, Homma K, & Imamizu H (2011). Physical delay but not subjective delay determines learning rate in prism adaptation. Exp Brain Res 208, 257–268.  Tong J, Mao O, Goldreich D (2013). Two-point orientation discrimination versus the traditional two-point test for tactile spatial acuity assessment. Front Hum Neurosci 7, 579.  Welch RB, Choe’CS, & Heinrich DR (1974). Evidence for a three component model of prism adaptation. J Exp Psycho 1103, 700-705. 80   Welch RB (1978). Perceptual modification: Perceptual modification: Adapting to altered sensory environments. Academic Press, New York.  Welgampola MS & Colebatch JG (2001). Vestibulospinal reflexes: quantitative effects of sensory feedback and postural task. Exp Brain Res 139, 345–353.  Winter DA, Patla AE, Prince F, Ishac M & Gielo-Perczak K (1998). Stiffness control of balance in quiet standing. J Neurophysiol 80, 1211–1221.  Wolpert DM, Ghahramani Z, & Jordan M (1995). An internal model for sensorimotor integration. Science 269, 1880–1882.  Wolpert DM, Diedrichsen J, Flanagan JR (2011). Principles of sensorimotor learning. Nat Rev Neurosci 12, 739-751.  Woodhull AM, Maltrud K & Mello BL. (1985). Alignment of the human body in standing. Eur J Appl Physiol Occup Physiol 54, 109-115.  Woollacott M & Shumway-Cook A (2002). Attention and the control of posture and gait: a review of an emerging area of research. Gait Posture 16, 1–14.  van der Kooij H, Jacobs R, Koopman B & Grootenboer H (1999). A multisensory integration model of human stance control. Biol Cybern 80, 299–308.  von Holst E & Mittelstaedt H (1950). Das reafferenzprinzip. Naturwissenschaften 37, 464–476.  Zwibel HL (2009). Contribution of impaired mobility and general symptoms to the burden of multiple sclerosis. Adv Ther 26, 1043-1057.        81  Appendices  Appendix A: Vestibular-evoked muscle responses   Shown below is a table of group data for vestibular-evoked muscle responses at different induced delay conditions. Presented are the magnitudes (short latency, medium latency, and peak-to-peak amplitudes) and latencies (short and medium latency responses) of the SVS-Soleus EMG cross-covariance measures.  Table A: Vestibular-evoked muscle response magnitudes and latencies Delay (ms) SL peak (r) ML peak (r) Peak-to-peak (r) SL latency (ms) ML latency (ms) 20 0.059 ± 0.019 -0.059 ± 0.021 0.117 ± 0.038  58.3 ± 4.7 104.7 ± 16.9 100 0.057 ± 0.017 -0.060 ± 0.020 0.117 ± 0.033 58.4 ± 4.3 107.7 ± 16.4 200 0.036 ± 0.020 -0.027 ± 0.017 0.064 ± 0.033 62.8 ± 4.0 96.3 ± 4.3 300 0.021 ± 0.017  -0.017 ± 0.016 0.039 ± 0.027 59.7 ± 3.9 94.8 ± 3.5 400 0.020 ± 0.018 -0.006 ± 0.016 0.026 ± 0.026 63.0 ± 6.0 93.6 ± 5.1 500 0.012 ± 0.013 -0.007 ± 0.011 0.018 ± 0.020 61.7 ± 3.1 102.2 ± 12.2 Data are presented as mean ± SD      82  Appendix B: Gain and phase of vestibular-evoked muscle responses  Interpreting gain and phase estimates is valid over frequency ranges where there is significant coherence. As depicted in the pooled coherence estimates (figure B), only 20, 100, and 200 ms delay conditions were associated with significant SVS-Soleus EMG coherence over the 0-25 Hz bandwidth. Therefore, gain and phase were calculated for these three conditions (figure B). Across the three conditions, there was no observed difference in the phase response. Like coherence, gain was similar for 20 and 100 ms conditions, and decreased for the 200 ms condition. Observing the frequency responses across the 0-25 Hz bandwidth demonstrated a typical low-pass behaviour for the soleus muscle (Dakin et al. 2007; Forbes et al. 2013), where phase decayed across the entire region and gain was relatively flat up to ~15 Hz. An important result to highlight from this experiment is that the timing of the vestibular-evoked muscle response (short and medium latency peaks) did not appear to change with different delays. This is reflected both in the phase (Figure B) and cross-covariance (Figure 8) estimates, demonstrating no change in response lag in the frequency or time domains. 83   Figure B: Coherence, phase, and gain estimates for the soleus muscle elicited by a 0-25 Hz vestibular stimulus Pooled (n=10) coherence, phase, and gain plots between vestibular stimuli and soleus muscle activity are presented for the different delay conditions. Significant coherence bandwidths between 0-25 Hz were observed for 20, 100, and 200 ms conditions, and therefore only those conditions are presented for phase and gain. The dashed line represents the 95% confidence limit for coherence.  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items