- Email: [email protected]
Generalizability of perturbation-evoked cortical potentials: Independence from sensory, motor and overall postural state
Neuroscience Letters 451 (2009) 40–44
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Generalizability of perturbation-evoked cortical potentials: Independence from sensory, motor and overall postural state George Mochizuki a,b,∗ , Kathryn M. Sibley a,c , Hannah J. Cheung a , Joanne M. Camilleri a,d , William E. McIlroy a,b,c,d,e a
Mobility Team, Toronto Rehabilitation Institute, University Centre, Toronto, Canada Heart and Stroke Foundation Centre for Stroke Recovery, Sunnybrook Health Sciences Centre, Toronto, Canada Institute of Medical Science, University of Toronto, Toronto, Canada d Graduate Department of Rehabilitation Science, University of Toronto, Toronto, Canada e Department of Kinesiology, University of Waterloo, Waterloo, Canada b c
a r t i c l e
i n f o
Article history: Received 19 August 2008 Received in revised form 10 November 2008 Accepted 11 December 2008 Keywords: Balance Posture Evoked potentials Human
a b s t r a c t Following disturbances to postural stability, balance recovery reactions are evoked by numerous sensory inputs and characterized by motor reactions involving different patterns of activity, depending on postural task conditions. It remains unknown whether well-documented cortical responses to instability share common spatio-temporal characteristics, despite variations in the sensory, motor, and postural components of the reactions. The objective was to explore the spatio-temporal profile of cortical potentials evoked by instability requiring either upper- or lower-limb compensatory responses. The hypothesis that upper- and lower-limb balance-correcting reactions are associated with evoked cortical potentials (N1, P2) featuring similar spatio-temporal characteristics was tested by inducing postural perturbations in seated (SIT) or standing (STAND) positions. For both conditions, N1 amplitude was greatest at FCz, with no significant differences in the timing of N1 peak (SIT: 142.4 ± 7.95 ms; STAND: 148.4 ± 4.10 ms) or N1 amplitude (SIT: 37.16 ± 6.99 V; STAND: 39.08 ± 4.51 V). The amplitude of the P2 potential (measured at CPz) was significantly larger in the STAND condition (37.87 ± 6.14 V) than in the SIT (23.66 ± 6.21 V) condition. Significant differences in P2 peak time between tasks were absent (SIT: 319.9 ± 11.45 ms; STAND: 322.7 ± 7.61 ms). Though differences in the amplitude of components of evoked potentials may reflect the extent of cortical involvement in different aspects of postural control, similarities in the spatio-temporal components of cortical potentials between tasks reflects generalizable cortical processing of unexpected stimuli independent of the sensory, motor, or postural aspects of the response. © 2008 Elsevier Ireland Ltd. All rights reserved.
There is emerging evidence of the potential role of cortical contributions to the control of upright stability [10] and compensatory balance reactions [1,15,20]. The putative involvement of cortical processes to balance control have been highlighted by the measurement of evoked cortical potentials that are of greatest amplitude at fronto-central electrode sites and are time-locked to the onset of the perturbation [1,6,20]. Within the context of balance control, the amplitude of the N1 potential has been shown to be modified by predictability [1], attention [18], and alterations in proprioception [23] even when the characteristics of the perturbations themselves remain constant. The amplitude of the P2 potential has been thought to be related to sensorimotor processing involved in the balance response [19]. These studies have emphasized the link between cortically evoked potentials and spe-
∗ Corresponding author at: Toronto Rehabilitation Institute, 550 University Ave, Rm 11019, Toronto, Ontario M5G 2A2, Canada. Tel.: +1 416 597 3422x7831. E-mail address: [email protected] (G. Mochizuki). 0304-3940/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2008.12.020
cific characteristics of the sensory inputs. In contrast, one could also make the argument that the evoked cortical response reflects more generalizable aspects of control (rather than a link to specific sensory or motor aspects of control). For example, there is a suggestion that the N1 reflects an error signal [1], depicting the contrast between anticipated (stability) and actual (instability) physiological states, while the P2 may be linked to events related to resource allocation [20]. The purpose of this study was to provide additional insight into the specificity of the relationship between the cortical potentials and specific aspects of the balance control response. Arguably, if the evoked potentials are uniquely associated with the sensory or motor characteristics of the movement, then the spatio-temporal characteristics of the evoked potential would vary based on different types of balance reactions. Alternatively if the evoked potentials represent a more generalizable aspect of control, then the spatio-temporal properties would not be influenced by changes in the specific characteristics of the sensory inputs or motor responses during the balance reactions.
G. Mochizuki et al. / Neuroscience Letters 451 (2009) 40–44
Compensatory balance reactions can be evoked under a variety of postural states that require a range of possible reactions. The common aspects of these responses are rapid onset latencies, scalability in the amplitude of responses, and coordinated activity appropriate to counter the direction of perturbation. These evoked responses are specific to postural control, elicited only when a period of instability exists [2] and are of shorter latency than voluntary activation of the same muscle [7]. To supplement traditional studies using standing models of postural control we have developed a seated balance model to examine the control of compensatory reactions in the upper limb [7,21]. In this seated model, the primary sensory inputs (when the hand is in prior contact with the support surface) and the primary motor responses of the involved upper limb are linked. This contrasts with the sensory and motor components of standing balance responses where the sensory inputs originate from sources that are not directly involved in the motor response [11]. While the temporal characteristics of compensatory upper- and lower-limb muscle responses are comparable in both change-in-support [12] and fixed-support paradigms [4], it is unknown whether the cortical responses evoked by whole-body instability are uniquely influenced by the position of the body when the compensatory reaction is made or by the limb used for balance recovery following postural instability. Differences in the spatial location of the largest amplitude cortical potential, especially if it is lateralized, might suggest that the evoked cortical activity is limb-specific. Conversely, centrally located cortical potentials evoked in association with a unilateral upper-limb response would indicate that centrally located potentials evoked during lower-limb responses are not a manifestation of lower-limb motor events. Similarities in the timing and spatial characteristics of the cortical potentials evoked by upper- and lower-limb compensatory responses would support the argument that such potentials are associated with more generalizable aspects of control. The objective of this study was to determine whether wholebody postural instability requiring either upper- or lower-limb compensatory responses evoked cortical potentials that were spatially and temporally linked. As noted, independence of the characteristics of the spatio-temporal aspects of the evoked potentials from the specific sensory and motor elements of the reaction reinforces the idea that the cortical responses may reflect more generalizable aspects of the underlying control. It was hypothesized that perturbation-evoked cortical potentials, specifically the N1 and P2, are common regardless of the sensory, motor, or postural components of the compensatory response and therefore would have similar timing and spatial location when comparing between seated and standing perturbations. Portions of this manuscript have been presented in abstract form [14]. Eight participants (22–44 years, 172.3 ± 7.5 cm, 72.0 ± 14.9 kg) were recruited to take part in this study. All participants were free of neuromuscular disorders and informed consent was obtained from all participants. This study was conducted with approval from the Research Ethics Board at the Toronto Rehabilitation Institute and in accordance with the Declaration of Helsinki. Temporally unpredictable perturbations to standing balance (STAND) were elicited using the same paradigm as in previous studies [15]. Briefly, postural instability in the posterior direction was initiated by the release of a 2.8 kg load that was electromagnetically coupled to a cable affixed to a belt worn at the level of the sternum while participants stood on a force plate. Participants were required to stand on the force plate with the lateral border of the heels 24 cm apart with the load attached prior to the start of the first task. The participant’s gaze was fixated on a target approximately 3 m ahead. Online monitoring of the centre of pressure (COP) was used to keep the participant in the same position at the beginning of each trial. Adjustments in the initial COP were made if participants displayed anticipatory adjustments in their starting position.
41
In the SIT condition, participants sat in a custom-built chair mounted on a hinge that permitted movement of the chair in the sagittal plane [21]. An electromagnet held the chair in an upright position while it was preloaded with 10 mm diameter elastic cord. Release of the magnet caused the chair to rapidly tilt backwards. A vertical pole was mounted approximately 0.5 m in front and slightly to the right of the participant, and was adjusted for arm length. With the shoulder flexed to approximately 90◦ and the elbow slightly flexed, the participant’s right hand was secured to the hand-hold with an elastic compression wrap. Release of the magnet and tilting of the chair induced a compensatory arm reaction on the handhold. For both conditions, participants received 30 perturbations. Participants wore disposable ear plugs so that they could not receive auditory cues about the upcoming perturbation except for a synchronizing verbal cue. The release of the electromagnet occurred at random intervals and the timing of the release was controlled by the experimenter. Fig. 1 depicts the paradigms used to induce postural instability. Electroencephalographic (EEG) signals were recorded on a 64 channel electrode cap (Quick-Cap, Neuroscan, El Paso, TX) based on the International 10-20 System. Recordings were made from Fz, FCz, Cz, CPz, C1, C2, C3 and C4. All channels were referenced to linked mastoids and the impedance for all channels was maintained at a level below 5 k. EEG signals were amplified (gain: 2010), filtered online (band pass: DC-200 Hz) using a SynAmps2 amplifier (Neuroscan, El Paso, TX), sampled at 1000 Hz and stored for offline analysis. For the STAND task, bipolar Ag–AgCl electrodes (10 mm diameter) were placed 25 mm apart on the tibialis anterior (TA) muscles, bilaterally. For the SIT task, bipolar electrodes were placed on the anterior deltoid (AD) and biceps brachii (BB) muscles, bilaterally. The skin was shaved, cleaned and abraded prior to application of the electrodes to attempt to achieve impedances at or below 5 k. The EMG signals were amplified (2000×) and filtered online (10–300 Hz) (Noraxon, Scottsdale, AZ), sampled at 1000 Hz and stored for offline analysis. Analysis of the EEG data was performed using Scan v4.3 (Neuroscan, El Paso, TX). Averages were assembled over 500 ms preceding and 1500 ms after the onset of perturbation. Points within this epoch were averaged across participants for both tasks. Epochs were filtered offline (high pass: DC; low pass 30 Hz) and artefacts caused by eye movement were removed from the recordings. To quantify the task-dependent differences in EEG, the amplitude and latency of the evoked potentials were measured. The latency of the N1 response was taken as the time of the peak relative to perturbation onset (time = 0). N1 amplitude was quantified as the difference in voltage between the N1 peak and perturbation onset (time = 0). The latency of the P2 was taken as the time of the peak relative to perturbation onset (time = 0), while the amplitude of the P2 was quantified as the voltage difference from perturbation onset to peak voltage within a 200 ms epoch ranging from 200 to 400 ms after perturbation onset. Prior to analysis, EMG signals were full-wave rectified. Onset of the EMG response burst was defined as the time at which the EMG amplitude reached +5 SD greater than the mean baseline value taken over a 100 ms block, 500 ms prior to perturbation onset. The magnitude of the EMG response for both TA and AD activity was quantified by measuring the integrated EMG activity for 200 ms following EMG onset (iEMGpost ). Within-subject comparisons of the effect of task (SIT vs. STAND) on EEG (N1 and P2 amplitude and latency) and EMG measures (primary muscle burst onset time) were performed using paired t-tests. Site-by-site comparison of N1 and P2 amplitude was assessed using one-way repeated measures ANOVA for each task, with electrode site (Fz, FCz, Cz, CPz, C1, C2, C3, C4) as the factor. Post hoc analysis was performed using the Bonferroni test. Values are reported as
42
G. Mochizuki et al. / Neuroscience Letters 451 (2009) 40–44
Fig. 1. Illustration of the paradigms used to generate whole-body instability in the STAND (A) and SIT (B) conditions.
mean ± standard error unless otherwise stated. Significance level was set at p ≤ 0.05. On average, the N1 potential recorded at FCz reached its peak amplitude 145.4 ± 4.39 ms after perturbation onset. There were no significant differences in the timing or amplitude of the N1 at FCz between conditions. N1 amplitude peaked at 142.4 ± 7.95 and 148.4 ± 4.10 ms after perturbation onset for the SIT and STAND conditions, respectively. N1 amplitude was 37.16 ± 6.99 V in the SIT condition and 39.08 ± 4.51 V in the STAND condition. A large
P2 was observed in the STAND condition and based on visual inspection, was largest at the CPz electrode site (Fig. 2). While there were no differences in P2 peak time between tasks (STAND: 322.7 ± 7.61 ms; SIT: 319.9 ± 11.45 ms), P2 peak amplitude was significantly higher in the STAND (37.87 ± 6.14 V) than in the SIT (23.66 ± 6.21 V) condition (t[7] = −2.96, p = 0.021). Site-by-site comparison of N1 amplitudes for both the SIT and STAND conditions revealed a significant effect of site (SIT: F(1,7) = 24.61, p < 0.001; STAND: F(1,7) = 19.35, p < 0.001). Post hoc
Fig. 2. Grand average (n = 8) perturbation-evoked cortical potentials at each electrode site for the SIT (black) and STAND (gray) conditions. The amplitude of the N1 is largest at FCz and the amplitude of the P2 is largest at CPz (highlighted panels). Note the similarities in the features of the evoked potentials between tasks prior to the occurrence of the P2.
G. Mochizuki et al. / Neuroscience Letters 451 (2009) 40–44
Fig. 3. Average absolute N1 (filled) and P2 (open) peak amplitude in the SIT (A) and STAND (B) conditions.
analysis confirmed that the largest amplitude N1 was located at the FCz electrode and was significantly larger than all other sites except Cz and Fz in the SIT condition and Cz and C1 in the STAND condition (although there was a tendency towards significance for C1 (p = 0.057)). Site-dependent differences in P2 amplitude were less prominent. ANOVA revealed only a tendency towards site-by-site differences in the SIT (p = 0.075) and STAND (p = 0.099) conditions. There were no measurable asymmetries (i.e. N1 amplitudes that were lateralized) in the SIT (unilateral upperlimb response) condition as compared to the STAND (bilateral lower-limb response) task. Fig. 3 displays the absolute value of the average amplitude of the N1 and P2 potentials for both tasks. EMG data from two participants were excluded from analysis because of technical issues. For the participants included in the analysis, reactive EMG bursts were observed for all trials. In the STAND condition, the magnitude of the TA EMG response was 16.49 ± 5.72 mV ms. The magnitude of the AD response was 30.76 ± 6.55 mV ms in the SIT condition. Because the onset of AD was earlier than BB, comparisons of EMG onset between conditions were made between the TA and AD muscles. The onset of reactive muscle activity differed significantly between conditions. On average, the timing of AD onset was 139.1 ± 10.83 ms after perturbation onset in the SIT condition. This response was significantly slower [t(1,5) = 2.86, p = 0.035] than TA onset, which took place 113.2 ± 4.15 ms after perturbation onset in the STAND condition. The main findings of this study were that (1) the amplitude of the N1 potential in both tasks was largest at fronto-central sites and was not lateralized, (2) neither the amplitude, timing nor location of the largest amplitude N1 potential differed between upper- and lower-limb responses, despite differences in muscle onset latencies between responses, and (3) the later P2 potential was larger in standing than in sitting. The similarities in the timing, amplitude, and spatial representation, in spite of the profound differences in sensory drive, motor reactions, and postural state, support the view that the N1 response represents a more generalizable event linked to the control of upright stability. Comparing the N1 vs. P2 across tasks may indicate that the early and later components of perturbation-evoked potentials represent different aspects of cortical involvement in postural control. Similarities in the spatio-temporal characteristics and amplitude of the N1 potential across tasks reinforce the view that the response reflects a more generalizable event rather than being
43
specifically linked to the details of the afferent input or motor output. The timing of the N1 is consistent with previous studies showing that the amplitude of the peak N1 occurs approximately 150 ms after perturbation onset in both sitting [22] and standing [15] conditions. The absence of spatial asymmetry (i.e. a lateralized potential) in both tasks also supports previous reports [3] and confirms that the evoked potential was driven by the perturbation, and not by the mode of sensory information or motor response. This is of specific importance given that in the SIT condition, the sensory drive most likely originated from somatosensory sources with the palmar surface of the hand already in contact with the support post at the time of perturbation onset, in contrast to the proprioceptive inputs from the lower limb suggested to trigger balance reactions involving the lower limbs in the STAND condition. Note that vestibular and/or visual inputs are not considered essential inputs to the initiation of the rapid onset compensatory reactions; rather, postural state-specific proprioceptive inputs are proposed to be the important inputs for evoking balance reactions [5]. The between-task similarities in N1 amplitude and the absence of pre-perturbation cortical activity in the STAND condition also rule out potential postural set differences between conditions, where muscles involved in the compensatory responses may already be engaged in maintaining upright stance. Application of the error signal hypothesis [1] to the current findings suggests that the perception of an error between expected and actual physiological states was comparable. Despite this, there was a difference in the onset of the reactive muscle response (∼139 ms for anterior deltoid, ∼113 ms for tibialis anterior). Although the EMG onset times are consistent with those reported in previous studies involving whole-body instability in sitting [7,21] and standing [15], it is conceivable that this difference indicates that the relative velocity of the perturbation was slower in the SIT condition. However, an onset latency of 139 ms is still within the range to be considered an automatic postural response for the anterior deltoid and is faster than voluntary activation of the shoulder musculature [7,for review see 9]. It cannot be known with certainty whether the perturbation characteristics between conditions were similar; yet, in spite of the potential relative differences in perturbation characteristics, the spatio-temporal characteristics and amplitude of the N1 cortical potentials were remarkably consistent, reinforcing the view that the N1 represents an event that is somewhat independent of the specific characteristics of the perturbation characteristics. The presence of the P2 potential in the present study is consistent with previous reports of late, positive cortical potentials evoked by postural instability [19,20]. It is not presently clear why the amplitude of this later potential was different between the two tasks (SIT and STAND). It is possible that differences in P2 reflected specific sensory or motor differences between the tasks (i.e. sensory activity, muscles recruited). However, such specific associations to sensory and motor differences would not be supported by the similarities in the spatial properties of the P2 between the two tasks. It is also possible that the differences were associated with later phases of control that differ between the seated and standing conditions. Quant et al. [19] reported that the magnitude of the P2 was larger when individuals passively experienced postural perturbations as opposed to actively countering postural instability. While both of the current tasks involved active corrections, the latter phases of the reactions were likely different. For example, in the standing condition there was a persisting challenge to stability since individuals still needed to reposition the centre of mass over the small base of support (feet). In contrast, in the seated condition, the individuals may have been more easily able to restore equilibrium due to the large moment arm possible when gripping the handhold. In addition, in the sitting condition there may have been an addi-
44
G. Mochizuki et al. / Neuroscience Letters 451 (2009) 40–44
tional sense of security because of the trunk support and the safety stops. The consequence of not appropriately attending to postural stability in a seated position may have less meaning as the threat of physical harm due to a fall is recognized as being smaller and may be influenced by the actual threat of injury; that is, a perception of less threatening consequences during a more ‘natural’ seated balance task. That being said, the majority of the differences in the challenge to re-stabilization and control occur after the time of the peak P2 (300 ms). While it is not clear what task-related differences accounted for the differences in P2 amplitude, it is interesting that the timing of the P2 (300 ms) coincided with events that seem linked to attention related processing. For example, previous studies have revealed disruptions of dual task performance during standing within 325 ms [13] and 480 ms [16] following the onset of a postural perturbation. While interpreting the characteristics of the perturbation-evoked potentials in relation to traditional P300 research may be tenuous, there is an abundance of evidence suggesting that the amplitude of later evoked potentials are related to stimulus context, attention, and meaningfulness (salience) [for reviews see 8,17]. Within the context of balance recovery, these cognitive parameters may be more important when an individual is required to actively maintain standing balance following postural stabilization as the consequence of not attending to standing postural stability may lead to a subsequent fall. In summary, we have demonstrated that the cortical potentials evoked by bouts of whole-body instability in a seated or standing position requiring upper- and lower-limb compensatory responses, respectively, share in their spatio-temporal characteristics. This supports the view that cortically evoked potentials during the control of reactions to preserve stability are most likely associated with generalizable aspects of control rather than the specific features of the stimulus (i.e. specific sensory inputs or motor outputs). While the amplitude of the early N1 potential was also consistent between perturbation types, the amplitude of the later P2 potential was larger for perturbations experienced in the standing position. These differences may reflect the cortical processing of the distinction in functional requirements during the stabilization period immediately following the initial compensatory response. Acknowledgements The authors wish to thank G. Sung and N. Amani for assistance with data collection and analysis. This work is supported by funding from the Heart and Stroke Foundation of Ontario, the Heart and Stroke Foundation of Canada (GM), and the Natural Science and Engineering Research Council (WEM, KMS). We acknowledge the support of the Toronto Rehabilitation Institute who receives funding under the Provincial Rehabilitation Research Program from the Ministry of Health and Long Term Care in Ontario.
References [1] A.L. Adkin, S. Quant, B.E. Maki, W.E. McIlroy, Cortical responses associated with predictable and unpredictable compensatory balance reactions, Exp. Brain Res. 172 (2006) 85–93. [2] T.C. Britton, B.L. Day, P. Brown, J.C. Rothwell, P.D. Thompson, C.D. Marsden, Postural electromyographic responses in the arm and leg following galvanic vestibular stimulation in man, Exp. Brain Res. 94 (1993) 143–151. [3] J.M. Camilleri, K.M. Sibley, K.F. Zabjek, W.E. McIlroy, Investigating the role of the cortex following unexpected whole body perturbations, Program No. 353.10, 2006 Neuroscience Meeting Planner, Society for Neuroscience, Atlanta, GA, 2006 (Online). [4] P.J. Cordo, L.M. Nashner, Properties of postural adjustments associated with rapid arm movements, J. Neurophysiol. 47 (1982) 287–302. [5] M.C. Do, Y. Breniere, S. Bouisset, Compensatory reactions in forward fall: are they initiated by stretch receptors? Electroencephalogr. Clin. Neurophysiol. 69 (1988) 448–452. [6] R.B. Duckrow, K. Abu-Hasaballah, R. Whipple, L. Wolfson, Stance perturbationevoked potentials in old people with poor gait and balance, Clin. Neurophysiol. 110 (1999) 2026–2032. [7] W.H. Gage, K.F. Zabjek, S.W. Hill, W.E. McIlroy, Parallels in control of voluntary and perturbation-evoked reach-to-grasp movements: EMG and kinematics, Exp. Brain Res. 181 (2007) 627–637. [8] A. Kok, On the utility of P3 amplitude as a measure of processing capacity, Psychophysiology 38 (2001) 557–577. [9] B.E. Maki, W.E. McIlroy, The role of limb movements in maintaining upright stance: the “change-in-support” strategy, Phys. Ther. 77 (1997) 488–507. [10] B.E. Maki, W.E. McIlroy, Cognitive demands and cortical control of human balance-recovery reactions, J. Neural Transm. 114 (2007) 1279–1296. [11] C. Maurer, T. Mergner, B. Bolha, F. Hlavacka, Vestibular, visual, and somatosensory contributions to human control of upright stance, Neurosci. Lett. 281 (2000) 99–102. [12] W.E. McIlroy, B.E. Maki, Early activation of arm muscles follows external perturbation of upright stance, Neurosci. Lett. 184 (1995) 177–180. [13] W.E. McIlroy, R.G. Norrie, J.D. Brooke, D.C. Bishop, A.J. Nelson, B.E. Maki, Temporal properties of attention sharing consequent to disturbed balance, Neuroreport 10 (1999) 2895–2899. [14] G. Mochizuki, H.J. Cheung, K.M. Sibley, J.M. Camilleri, W.E. McIlroy, Late but not early components of perturbation-evoked cortical potentials differ between sitting and standing postural tasks, Arch. Phys. Med. Rehabil. 89 (2008) e40. [15] G. Mochizuki, K.M. Sibley, J.G. Esposito, J.M. Camilleri, W.E. McIlroy, Cortical responses associated with the preparation and reaction to full-body perturbations to upright stability, Clin. Neurophysiol. 119 (2008) 1626–1637. [16] R.G. Norrie, B.E. Maki, W.R. Staines, W.E. McIlroy, The time course of attention shifts following perturbation of upright stance, Exp. Brain Res. 146 (2002) 315–321. [17] J. Polich, Updating P300: an integrative theory of P3a and P3b, Clin. Neurophysiol. 118 (2007) 2128–2148. [18] S. Quant, A.L. Adkin, W.R. Staines, B.E. Maki, W.E. McIlroy, The effect of a concurrent cognitive task on cortical potentials evoked by unpredictable balance perturbations, BMC Neurosci. 5 (2004) 18. [19] S. Quant, A.L. Adkin, W.R. Staines, W.E. McIlroy, Cortical activation following a balance disturbance, Exp. Brain Res. 155 (2004) 393–400. [20] S. Quant, B.E. Maki, W.E. McIlroy, The association between later cortical potentials and later phases of postural reactions evoked by perturbations to upright stance, Neurosci. Lett. 381 (2005) 269–274. [21] K.M. Sibley, B. Lakhani, G. Mochizuki, W.E. McIlroy, Influences of perturbation amplitude and task difficulty on electrodermal responses to whole-body instability, Parkinsonism Relat. Disord. Supplement 1 (2008) S38. [22] K.M. Sibley, K.F. Zabjek, J.M. Camilleri, W.E. McIlroy, Effects of perturbation amplitude on cortical potentials evoked by whole-body perturbations, in: International Congress on Gait and Mental Function Book of Abstracts, 2006, p. 50. [23] W.R. Staines, W.E. McIlroy, J.D. Brooke, Cortical representation of whole-body movement is modulated by proprioceptive discharge in humans, Exp. Brain Res. 138 (2001) 235–242.
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Generalizability of perturbation-evoked cortical potentials: Independence from sensory, motor and overall postural state George Mochizuki a,b,∗ , Kathryn M. Sibley a,c , Hannah J. Cheung a , Joanne M. Camilleri a,d , William E. McIlroy a,b,c,d,e a
Mobility Team, Toronto Rehabilitation Institute, University Centre, Toronto, Canada Heart and Stroke Foundation Centre for Stroke Recovery, Sunnybrook Health Sciences Centre, Toronto, Canada Institute of Medical Science, University of Toronto, Toronto, Canada d Graduate Department of Rehabilitation Science, University of Toronto, Toronto, Canada e Department of Kinesiology, University of Waterloo, Waterloo, Canada b c
a r t i c l e
i n f o
Article history: Received 19 August 2008 Received in revised form 10 November 2008 Accepted 11 December 2008 Keywords: Balance Posture Evoked potentials Human
a b s t r a c t Following disturbances to postural stability, balance recovery reactions are evoked by numerous sensory inputs and characterized by motor reactions involving different patterns of activity, depending on postural task conditions. It remains unknown whether well-documented cortical responses to instability share common spatio-temporal characteristics, despite variations in the sensory, motor, and postural components of the reactions. The objective was to explore the spatio-temporal profile of cortical potentials evoked by instability requiring either upper- or lower-limb compensatory responses. The hypothesis that upper- and lower-limb balance-correcting reactions are associated with evoked cortical potentials (N1, P2) featuring similar spatio-temporal characteristics was tested by inducing postural perturbations in seated (SIT) or standing (STAND) positions. For both conditions, N1 amplitude was greatest at FCz, with no significant differences in the timing of N1 peak (SIT: 142.4 ± 7.95 ms; STAND: 148.4 ± 4.10 ms) or N1 amplitude (SIT: 37.16 ± 6.99 V; STAND: 39.08 ± 4.51 V). The amplitude of the P2 potential (measured at CPz) was significantly larger in the STAND condition (37.87 ± 6.14 V) than in the SIT (23.66 ± 6.21 V) condition. Significant differences in P2 peak time between tasks were absent (SIT: 319.9 ± 11.45 ms; STAND: 322.7 ± 7.61 ms). Though differences in the amplitude of components of evoked potentials may reflect the extent of cortical involvement in different aspects of postural control, similarities in the spatio-temporal components of cortical potentials between tasks reflects generalizable cortical processing of unexpected stimuli independent of the sensory, motor, or postural aspects of the response. © 2008 Elsevier Ireland Ltd. All rights reserved.
There is emerging evidence of the potential role of cortical contributions to the control of upright stability [10] and compensatory balance reactions [1,15,20]. The putative involvement of cortical processes to balance control have been highlighted by the measurement of evoked cortical potentials that are of greatest amplitude at fronto-central electrode sites and are time-locked to the onset of the perturbation [1,6,20]. Within the context of balance control, the amplitude of the N1 potential has been shown to be modified by predictability [1], attention [18], and alterations in proprioception [23] even when the characteristics of the perturbations themselves remain constant. The amplitude of the P2 potential has been thought to be related to sensorimotor processing involved in the balance response [19]. These studies have emphasized the link between cortically evoked potentials and spe-
∗ Corresponding author at: Toronto Rehabilitation Institute, 550 University Ave, Rm 11019, Toronto, Ontario M5G 2A2, Canada. Tel.: +1 416 597 3422x7831. E-mail address: [email protected] (G. Mochizuki). 0304-3940/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2008.12.020
cific characteristics of the sensory inputs. In contrast, one could also make the argument that the evoked cortical response reflects more generalizable aspects of control (rather than a link to specific sensory or motor aspects of control). For example, there is a suggestion that the N1 reflects an error signal [1], depicting the contrast between anticipated (stability) and actual (instability) physiological states, while the P2 may be linked to events related to resource allocation [20]. The purpose of this study was to provide additional insight into the specificity of the relationship between the cortical potentials and specific aspects of the balance control response. Arguably, if the evoked potentials are uniquely associated with the sensory or motor characteristics of the movement, then the spatio-temporal characteristics of the evoked potential would vary based on different types of balance reactions. Alternatively if the evoked potentials represent a more generalizable aspect of control, then the spatio-temporal properties would not be influenced by changes in the specific characteristics of the sensory inputs or motor responses during the balance reactions.
G. Mochizuki et al. / Neuroscience Letters 451 (2009) 40–44
Compensatory balance reactions can be evoked under a variety of postural states that require a range of possible reactions. The common aspects of these responses are rapid onset latencies, scalability in the amplitude of responses, and coordinated activity appropriate to counter the direction of perturbation. These evoked responses are specific to postural control, elicited only when a period of instability exists [2] and are of shorter latency than voluntary activation of the same muscle [7]. To supplement traditional studies using standing models of postural control we have developed a seated balance model to examine the control of compensatory reactions in the upper limb [7,21]. In this seated model, the primary sensory inputs (when the hand is in prior contact with the support surface) and the primary motor responses of the involved upper limb are linked. This contrasts with the sensory and motor components of standing balance responses where the sensory inputs originate from sources that are not directly involved in the motor response [11]. While the temporal characteristics of compensatory upper- and lower-limb muscle responses are comparable in both change-in-support [12] and fixed-support paradigms [4], it is unknown whether the cortical responses evoked by whole-body instability are uniquely influenced by the position of the body when the compensatory reaction is made or by the limb used for balance recovery following postural instability. Differences in the spatial location of the largest amplitude cortical potential, especially if it is lateralized, might suggest that the evoked cortical activity is limb-specific. Conversely, centrally located cortical potentials evoked in association with a unilateral upper-limb response would indicate that centrally located potentials evoked during lower-limb responses are not a manifestation of lower-limb motor events. Similarities in the timing and spatial characteristics of the cortical potentials evoked by upper- and lower-limb compensatory responses would support the argument that such potentials are associated with more generalizable aspects of control. The objective of this study was to determine whether wholebody postural instability requiring either upper- or lower-limb compensatory responses evoked cortical potentials that were spatially and temporally linked. As noted, independence of the characteristics of the spatio-temporal aspects of the evoked potentials from the specific sensory and motor elements of the reaction reinforces the idea that the cortical responses may reflect more generalizable aspects of the underlying control. It was hypothesized that perturbation-evoked cortical potentials, specifically the N1 and P2, are common regardless of the sensory, motor, or postural components of the compensatory response and therefore would have similar timing and spatial location when comparing between seated and standing perturbations. Portions of this manuscript have been presented in abstract form [14]. Eight participants (22–44 years, 172.3 ± 7.5 cm, 72.0 ± 14.9 kg) were recruited to take part in this study. All participants were free of neuromuscular disorders and informed consent was obtained from all participants. This study was conducted with approval from the Research Ethics Board at the Toronto Rehabilitation Institute and in accordance with the Declaration of Helsinki. Temporally unpredictable perturbations to standing balance (STAND) were elicited using the same paradigm as in previous studies [15]. Briefly, postural instability in the posterior direction was initiated by the release of a 2.8 kg load that was electromagnetically coupled to a cable affixed to a belt worn at the level of the sternum while participants stood on a force plate. Participants were required to stand on the force plate with the lateral border of the heels 24 cm apart with the load attached prior to the start of the first task. The participant’s gaze was fixated on a target approximately 3 m ahead. Online monitoring of the centre of pressure (COP) was used to keep the participant in the same position at the beginning of each trial. Adjustments in the initial COP were made if participants displayed anticipatory adjustments in their starting position.
41
In the SIT condition, participants sat in a custom-built chair mounted on a hinge that permitted movement of the chair in the sagittal plane [21]. An electromagnet held the chair in an upright position while it was preloaded with 10 mm diameter elastic cord. Release of the magnet caused the chair to rapidly tilt backwards. A vertical pole was mounted approximately 0.5 m in front and slightly to the right of the participant, and was adjusted for arm length. With the shoulder flexed to approximately 90◦ and the elbow slightly flexed, the participant’s right hand was secured to the hand-hold with an elastic compression wrap. Release of the magnet and tilting of the chair induced a compensatory arm reaction on the handhold. For both conditions, participants received 30 perturbations. Participants wore disposable ear plugs so that they could not receive auditory cues about the upcoming perturbation except for a synchronizing verbal cue. The release of the electromagnet occurred at random intervals and the timing of the release was controlled by the experimenter. Fig. 1 depicts the paradigms used to induce postural instability. Electroencephalographic (EEG) signals were recorded on a 64 channel electrode cap (Quick-Cap, Neuroscan, El Paso, TX) based on the International 10-20 System. Recordings were made from Fz, FCz, Cz, CPz, C1, C2, C3 and C4. All channels were referenced to linked mastoids and the impedance for all channels was maintained at a level below 5 k. EEG signals were amplified (gain: 2010), filtered online (band pass: DC-200 Hz) using a SynAmps2 amplifier (Neuroscan, El Paso, TX), sampled at 1000 Hz and stored for offline analysis. For the STAND task, bipolar Ag–AgCl electrodes (10 mm diameter) were placed 25 mm apart on the tibialis anterior (TA) muscles, bilaterally. For the SIT task, bipolar electrodes were placed on the anterior deltoid (AD) and biceps brachii (BB) muscles, bilaterally. The skin was shaved, cleaned and abraded prior to application of the electrodes to attempt to achieve impedances at or below 5 k. The EMG signals were amplified (2000×) and filtered online (10–300 Hz) (Noraxon, Scottsdale, AZ), sampled at 1000 Hz and stored for offline analysis. Analysis of the EEG data was performed using Scan v4.3 (Neuroscan, El Paso, TX). Averages were assembled over 500 ms preceding and 1500 ms after the onset of perturbation. Points within this epoch were averaged across participants for both tasks. Epochs were filtered offline (high pass: DC; low pass 30 Hz) and artefacts caused by eye movement were removed from the recordings. To quantify the task-dependent differences in EEG, the amplitude and latency of the evoked potentials were measured. The latency of the N1 response was taken as the time of the peak relative to perturbation onset (time = 0). N1 amplitude was quantified as the difference in voltage between the N1 peak and perturbation onset (time = 0). The latency of the P2 was taken as the time of the peak relative to perturbation onset (time = 0), while the amplitude of the P2 was quantified as the voltage difference from perturbation onset to peak voltage within a 200 ms epoch ranging from 200 to 400 ms after perturbation onset. Prior to analysis, EMG signals were full-wave rectified. Onset of the EMG response burst was defined as the time at which the EMG amplitude reached +5 SD greater than the mean baseline value taken over a 100 ms block, 500 ms prior to perturbation onset. The magnitude of the EMG response for both TA and AD activity was quantified by measuring the integrated EMG activity for 200 ms following EMG onset (iEMGpost ). Within-subject comparisons of the effect of task (SIT vs. STAND) on EEG (N1 and P2 amplitude and latency) and EMG measures (primary muscle burst onset time) were performed using paired t-tests. Site-by-site comparison of N1 and P2 amplitude was assessed using one-way repeated measures ANOVA for each task, with electrode site (Fz, FCz, Cz, CPz, C1, C2, C3, C4) as the factor. Post hoc analysis was performed using the Bonferroni test. Values are reported as
42
G. Mochizuki et al. / Neuroscience Letters 451 (2009) 40–44
Fig. 1. Illustration of the paradigms used to generate whole-body instability in the STAND (A) and SIT (B) conditions.
mean ± standard error unless otherwise stated. Significance level was set at p ≤ 0.05. On average, the N1 potential recorded at FCz reached its peak amplitude 145.4 ± 4.39 ms after perturbation onset. There were no significant differences in the timing or amplitude of the N1 at FCz between conditions. N1 amplitude peaked at 142.4 ± 7.95 and 148.4 ± 4.10 ms after perturbation onset for the SIT and STAND conditions, respectively. N1 amplitude was 37.16 ± 6.99 V in the SIT condition and 39.08 ± 4.51 V in the STAND condition. A large
P2 was observed in the STAND condition and based on visual inspection, was largest at the CPz electrode site (Fig. 2). While there were no differences in P2 peak time between tasks (STAND: 322.7 ± 7.61 ms; SIT: 319.9 ± 11.45 ms), P2 peak amplitude was significantly higher in the STAND (37.87 ± 6.14 V) than in the SIT (23.66 ± 6.21 V) condition (t[7] = −2.96, p = 0.021). Site-by-site comparison of N1 amplitudes for both the SIT and STAND conditions revealed a significant effect of site (SIT: F(1,7) = 24.61, p < 0.001; STAND: F(1,7) = 19.35, p < 0.001). Post hoc
Fig. 2. Grand average (n = 8) perturbation-evoked cortical potentials at each electrode site for the SIT (black) and STAND (gray) conditions. The amplitude of the N1 is largest at FCz and the amplitude of the P2 is largest at CPz (highlighted panels). Note the similarities in the features of the evoked potentials between tasks prior to the occurrence of the P2.
G. Mochizuki et al. / Neuroscience Letters 451 (2009) 40–44
Fig. 3. Average absolute N1 (filled) and P2 (open) peak amplitude in the SIT (A) and STAND (B) conditions.
analysis confirmed that the largest amplitude N1 was located at the FCz electrode and was significantly larger than all other sites except Cz and Fz in the SIT condition and Cz and C1 in the STAND condition (although there was a tendency towards significance for C1 (p = 0.057)). Site-dependent differences in P2 amplitude were less prominent. ANOVA revealed only a tendency towards site-by-site differences in the SIT (p = 0.075) and STAND (p = 0.099) conditions. There were no measurable asymmetries (i.e. N1 amplitudes that were lateralized) in the SIT (unilateral upperlimb response) condition as compared to the STAND (bilateral lower-limb response) task. Fig. 3 displays the absolute value of the average amplitude of the N1 and P2 potentials for both tasks. EMG data from two participants were excluded from analysis because of technical issues. For the participants included in the analysis, reactive EMG bursts were observed for all trials. In the STAND condition, the magnitude of the TA EMG response was 16.49 ± 5.72 mV ms. The magnitude of the AD response was 30.76 ± 6.55 mV ms in the SIT condition. Because the onset of AD was earlier than BB, comparisons of EMG onset between conditions were made between the TA and AD muscles. The onset of reactive muscle activity differed significantly between conditions. On average, the timing of AD onset was 139.1 ± 10.83 ms after perturbation onset in the SIT condition. This response was significantly slower [t(1,5) = 2.86, p = 0.035] than TA onset, which took place 113.2 ± 4.15 ms after perturbation onset in the STAND condition. The main findings of this study were that (1) the amplitude of the N1 potential in both tasks was largest at fronto-central sites and was not lateralized, (2) neither the amplitude, timing nor location of the largest amplitude N1 potential differed between upper- and lower-limb responses, despite differences in muscle onset latencies between responses, and (3) the later P2 potential was larger in standing than in sitting. The similarities in the timing, amplitude, and spatial representation, in spite of the profound differences in sensory drive, motor reactions, and postural state, support the view that the N1 response represents a more generalizable event linked to the control of upright stability. Comparing the N1 vs. P2 across tasks may indicate that the early and later components of perturbation-evoked potentials represent different aspects of cortical involvement in postural control. Similarities in the spatio-temporal characteristics and amplitude of the N1 potential across tasks reinforce the view that the response reflects a more generalizable event rather than being
43
specifically linked to the details of the afferent input or motor output. The timing of the N1 is consistent with previous studies showing that the amplitude of the peak N1 occurs approximately 150 ms after perturbation onset in both sitting [22] and standing [15] conditions. The absence of spatial asymmetry (i.e. a lateralized potential) in both tasks also supports previous reports [3] and confirms that the evoked potential was driven by the perturbation, and not by the mode of sensory information or motor response. This is of specific importance given that in the SIT condition, the sensory drive most likely originated from somatosensory sources with the palmar surface of the hand already in contact with the support post at the time of perturbation onset, in contrast to the proprioceptive inputs from the lower limb suggested to trigger balance reactions involving the lower limbs in the STAND condition. Note that vestibular and/or visual inputs are not considered essential inputs to the initiation of the rapid onset compensatory reactions; rather, postural state-specific proprioceptive inputs are proposed to be the important inputs for evoking balance reactions [5]. The between-task similarities in N1 amplitude and the absence of pre-perturbation cortical activity in the STAND condition also rule out potential postural set differences between conditions, where muscles involved in the compensatory responses may already be engaged in maintaining upright stance. Application of the error signal hypothesis [1] to the current findings suggests that the perception of an error between expected and actual physiological states was comparable. Despite this, there was a difference in the onset of the reactive muscle response (∼139 ms for anterior deltoid, ∼113 ms for tibialis anterior). Although the EMG onset times are consistent with those reported in previous studies involving whole-body instability in sitting [7,21] and standing [15], it is conceivable that this difference indicates that the relative velocity of the perturbation was slower in the SIT condition. However, an onset latency of 139 ms is still within the range to be considered an automatic postural response for the anterior deltoid and is faster than voluntary activation of the shoulder musculature [7,for review see 9]. It cannot be known with certainty whether the perturbation characteristics between conditions were similar; yet, in spite of the potential relative differences in perturbation characteristics, the spatio-temporal characteristics and amplitude of the N1 cortical potentials were remarkably consistent, reinforcing the view that the N1 represents an event that is somewhat independent of the specific characteristics of the perturbation characteristics. The presence of the P2 potential in the present study is consistent with previous reports of late, positive cortical potentials evoked by postural instability [19,20]. It is not presently clear why the amplitude of this later potential was different between the two tasks (SIT and STAND). It is possible that differences in P2 reflected specific sensory or motor differences between the tasks (i.e. sensory activity, muscles recruited). However, such specific associations to sensory and motor differences would not be supported by the similarities in the spatial properties of the P2 between the two tasks. It is also possible that the differences were associated with later phases of control that differ between the seated and standing conditions. Quant et al. [19] reported that the magnitude of the P2 was larger when individuals passively experienced postural perturbations as opposed to actively countering postural instability. While both of the current tasks involved active corrections, the latter phases of the reactions were likely different. For example, in the standing condition there was a persisting challenge to stability since individuals still needed to reposition the centre of mass over the small base of support (feet). In contrast, in the seated condition, the individuals may have been more easily able to restore equilibrium due to the large moment arm possible when gripping the handhold. In addition, in the sitting condition there may have been an addi-
44
G. Mochizuki et al. / Neuroscience Letters 451 (2009) 40–44
tional sense of security because of the trunk support and the safety stops. The consequence of not appropriately attending to postural stability in a seated position may have less meaning as the threat of physical harm due to a fall is recognized as being smaller and may be influenced by the actual threat of injury; that is, a perception of less threatening consequences during a more ‘natural’ seated balance task. That being said, the majority of the differences in the challenge to re-stabilization and control occur after the time of the peak P2 (300 ms). While it is not clear what task-related differences accounted for the differences in P2 amplitude, it is interesting that the timing of the P2 (300 ms) coincided with events that seem linked to attention related processing. For example, previous studies have revealed disruptions of dual task performance during standing within 325 ms [13] and 480 ms [16] following the onset of a postural perturbation. While interpreting the characteristics of the perturbation-evoked potentials in relation to traditional P300 research may be tenuous, there is an abundance of evidence suggesting that the amplitude of later evoked potentials are related to stimulus context, attention, and meaningfulness (salience) [for reviews see 8,17]. Within the context of balance recovery, these cognitive parameters may be more important when an individual is required to actively maintain standing balance following postural stabilization as the consequence of not attending to standing postural stability may lead to a subsequent fall. In summary, we have demonstrated that the cortical potentials evoked by bouts of whole-body instability in a seated or standing position requiring upper- and lower-limb compensatory responses, respectively, share in their spatio-temporal characteristics. This supports the view that cortically evoked potentials during the control of reactions to preserve stability are most likely associated with generalizable aspects of control rather than the specific features of the stimulus (i.e. specific sensory inputs or motor outputs). While the amplitude of the early N1 potential was also consistent between perturbation types, the amplitude of the later P2 potential was larger for perturbations experienced in the standing position. These differences may reflect the cortical processing of the distinction in functional requirements during the stabilization period immediately following the initial compensatory response. Acknowledgements The authors wish to thank G. Sung and N. Amani for assistance with data collection and analysis. This work is supported by funding from the Heart and Stroke Foundation of Ontario, the Heart and Stroke Foundation of Canada (GM), and the Natural Science and Engineering Research Council (WEM, KMS). We acknowledge the support of the Toronto Rehabilitation Institute who receives funding under the Provincial Rehabilitation Research Program from the Ministry of Health and Long Term Care in Ontario.
References [1] A.L. Adkin, S. Quant, B.E. Maki, W.E. McIlroy, Cortical responses associated with predictable and unpredictable compensatory balance reactions, Exp. Brain Res. 172 (2006) 85–93. [2] T.C. Britton, B.L. Day, P. Brown, J.C. Rothwell, P.D. Thompson, C.D. Marsden, Postural electromyographic responses in the arm and leg following galvanic vestibular stimulation in man, Exp. Brain Res. 94 (1993) 143–151. [3] J.M. Camilleri, K.M. Sibley, K.F. Zabjek, W.E. McIlroy, Investigating the role of the cortex following unexpected whole body perturbations, Program No. 353.10, 2006 Neuroscience Meeting Planner, Society for Neuroscience, Atlanta, GA, 2006 (Online). [4] P.J. Cordo, L.M. Nashner, Properties of postural adjustments associated with rapid arm movements, J. Neurophysiol. 47 (1982) 287–302. [5] M.C. Do, Y. Breniere, S. Bouisset, Compensatory reactions in forward fall: are they initiated by stretch receptors? Electroencephalogr. Clin. Neurophysiol. 69 (1988) 448–452. [6] R.B. Duckrow, K. Abu-Hasaballah, R. Whipple, L. Wolfson, Stance perturbationevoked potentials in old people with poor gait and balance, Clin. Neurophysiol. 110 (1999) 2026–2032. [7] W.H. Gage, K.F. Zabjek, S.W. Hill, W.E. McIlroy, Parallels in control of voluntary and perturbation-evoked reach-to-grasp movements: EMG and kinematics, Exp. Brain Res. 181 (2007) 627–637. [8] A. Kok, On the utility of P3 amplitude as a measure of processing capacity, Psychophysiology 38 (2001) 557–577. [9] B.E. Maki, W.E. McIlroy, The role of limb movements in maintaining upright stance: the “change-in-support” strategy, Phys. Ther. 77 (1997) 488–507. [10] B.E. Maki, W.E. McIlroy, Cognitive demands and cortical control of human balance-recovery reactions, J. Neural Transm. 114 (2007) 1279–1296. [11] C. Maurer, T. Mergner, B. Bolha, F. Hlavacka, Vestibular, visual, and somatosensory contributions to human control of upright stance, Neurosci. Lett. 281 (2000) 99–102. [12] W.E. McIlroy, B.E. Maki, Early activation of arm muscles follows external perturbation of upright stance, Neurosci. Lett. 184 (1995) 177–180. [13] W.E. McIlroy, R.G. Norrie, J.D. Brooke, D.C. Bishop, A.J. Nelson, B.E. Maki, Temporal properties of attention sharing consequent to disturbed balance, Neuroreport 10 (1999) 2895–2899. [14] G. Mochizuki, H.J. Cheung, K.M. Sibley, J.M. Camilleri, W.E. McIlroy, Late but not early components of perturbation-evoked cortical potentials differ between sitting and standing postural tasks, Arch. Phys. Med. Rehabil. 89 (2008) e40. [15] G. Mochizuki, K.M. Sibley, J.G. Esposito, J.M. Camilleri, W.E. McIlroy, Cortical responses associated with the preparation and reaction to full-body perturbations to upright stability, Clin. Neurophysiol. 119 (2008) 1626–1637. [16] R.G. Norrie, B.E. Maki, W.R. Staines, W.E. McIlroy, The time course of attention shifts following perturbation of upright stance, Exp. Brain Res. 146 (2002) 315–321. [17] J. Polich, Updating P300: an integrative theory of P3a and P3b, Clin. Neurophysiol. 118 (2007) 2128–2148. [18] S. Quant, A.L. Adkin, W.R. Staines, B.E. Maki, W.E. McIlroy, The effect of a concurrent cognitive task on cortical potentials evoked by unpredictable balance perturbations, BMC Neurosci. 5 (2004) 18. [19] S. Quant, A.L. Adkin, W.R. Staines, W.E. McIlroy, Cortical activation following a balance disturbance, Exp. Brain Res. 155 (2004) 393–400. [20] S. Quant, B.E. Maki, W.E. McIlroy, The association between later cortical potentials and later phases of postural reactions evoked by perturbations to upright stance, Neurosci. Lett. 381 (2005) 269–274. [21] K.M. Sibley, B. Lakhani, G. Mochizuki, W.E. McIlroy, Influences of perturbation amplitude and task difficulty on electrodermal responses to whole-body instability, Parkinsonism Relat. Disord. Supplement 1 (2008) S38. [22] K.M. Sibley, K.F. Zabjek, J.M. Camilleri, W.E. McIlroy, Effects of perturbation amplitude on cortical potentials evoked by whole-body perturbations, in: International Congress on Gait and Mental Function Book of Abstracts, 2006, p. 50. [23] W.R. Staines, W.E. McIlroy, J.D. Brooke, Cortical representation of whole-body movement is modulated by proprioceptive discharge in humans, Exp. Brain Res. 138 (2001) 235–242.